Nanostructured Materials and Coatings for

Biomedical and Sensor Applications

NATO Science Series
A Series presenting the results of scientific meetings supported under the NATO Science
Programme.

The Series is published by lOS Press, Amsterdam, and Kluwer Academic Publishers in conjunction
with the NATO Scientific Affairs Division

Sub-Series

I. Life and Behavioural Sciences lOS Press

II.Mathematics, Physics and Chemistry Kluwer Academic Publishers
III.
Computer and Systems Science lOS Press
IV.Earth and Environmental Sciences Kluwer Academic Publishers
V. Science and Technology Policy lOS Press

The NATO Science Series continues the series of books published formerly as the NATO ASI Series.

The NATO Science Programme offers support for collaboration in civil science between scientists of
countries of the Euro-Atlantic Partnership Council. The types of scientific meeting generally supported
are "Advanced Study Institutes" and "Advanced Research Workshops", although other types of
meeting are supported from time to time. The NATO Science Series collects together the results of
these meetings. The meetings are co-organized bij scientists from NATO countries and scientists from
NATO's Partner countries - countries of the CIS and Central and Eastern Europe.

Advanced Study Institutes are high-level tutorial courses offering in-depth study of latest advances
in a field.
Advanced Research Workshops are expert meetings aimed at critical assessment of a field, and
identification of directions for future action.

As a consequence of the restructuring of the NATO Science Programme in 1999, the NATO Science
Series has been re-organised and there are currently Five Sub-series as noted above. Please consult
the following web sites for information on previous volumes published in the Series, as well as details of
earlier Sub-series.

http://www.nato-inVscience
http://www.wkap.nl
http://www.iospress.nl
http://www.wtv-books.de/nato-pco.htm

_~.l\_
.~
I

Series II: Mathematics, Physics and Chemistry - Vol. 102

Nanostructured Materials and
Coatings for 8iomedical and
Sensor Applications
edited by

Y.G. Gogotsi
Drexel University,
Department of Materials Engineering,
Philadelphia, U.SA

and

Iri na V. Uvarova
Institute for Problems of Materials Science,
National Academy of Science,
Kiev, Ukraine

....
"
Springer Science+Business Media, B.v.
Proceedings of the NATO Advanced Research Workshop on
Nanostructured Materials and Coatings for Biomedical and Sensor Applications
Kyiv, Ukraine
4-8 August 2002

A C.I.P. Catalogue record for this book is available from the Library of Congress.

ISBN 978-1-4020-1321-8 ISBN 978-94-010-0157-1 (eBook)

DOI 10.1007/978-94-010-0157-1

Printed on acid-free paper

AII Rights Reserved

© 2003 Springer Science+Business Media Dordrecht
Originally published by Kluwer Academic Publishers in 2003
Softcover reprint of the hardcover 1st edition 2003

No part of this work may be repraduced, stored in a retrieval system, or transmitted

in any form or by any means, electronic, mechanical, photocopying, microfilming,
recording or otherwise, without written permission fram the Publisher, with the exception
of any material supplied specifically for the purpose of being entered
and executed on a computer system, for exclusive use by the purchaser of the work.
Contents

Preface ......................................................................................................................... ix

Group photo ................................................................................................................. xi

Chapter 1. Materials for Biomedical Applications

"Nanostructured Materials in Medicine. State of the Art in Ukraine"

I. Uvarova ................................................................................................................ 1

"Current Trends in Biomaterial Coatings"

S.V. Mikhalovsky, M. Santin, L.I. Mikhalovska, A.W. Lloyd, S.P. Denyer ........ 15

"Injectable Nanoparticle Technology for In Vivo Remediation of Overdosed

Toxins"
R. Partch, E. Powell, Y.-H. Lee, M. Varshney, S. Kim, N. Barnard, D. Shah,
D. Dennis, T. Morey ............................................................................................. 27

"Nanostructured MeSi0 2 (Me=Ag, Cu) Coatings with Antibacterial Activity"

C.C. Trapalis, T. Vaimakis, A. Kharlamov, M. Kokkoris, G. Kordas ................... 41

"Low Friction Wear Resistant Sputtered Carbon Coatings for Biomedical

Applications"
S.K. Taylor, A.H.S. Jones, D. Meres, D.G. Teer, M. Elloy .................................... 51

"Mechanism of Interaction of Silicon Nitride Powders with Biochemical

Media and Their Toxic Effect"
V.A. Lavrenko, N.V. Boshitskaya, G.N. Makarenko ............................................ 63

"Microstructure and Strength Properties of Germanium Microwires for

Biomedical Devices"
D. Grabco, M. Dyntu, D. Meglei, O. Shikimaka ................................................... 71
vi

"The Researches of Properties of Biomaterials Based on Biological

Hydroxyapatite in Synthetic and Natural Physiological Mediums"
L.A. Ivanchenko, N.D. Pinchuk, V.S. Sulyma ..................................................... 77

"Preparation of Hydroxyapatite NanoParticles Using a Modified Precipitation

Method"
G. C. Koumoulidis, C. C. Trapalis, A. T. Sdoukos, and T. C. Vaimakis .............. 83

"Biomineralization of Hydroxyapatite/Titania Composites Prepared by

Explosive Shock Waves"
H. Beheri, Wafa 1. Abdel-Fattah, R. Pruemmer. .................................................... 91

"Prospects of Usage of Materials with Fullerene Coatings for Endoprosthesis

of Joints"
V.V. Lashneva, Yu.G. Tkachenko, V.A. Dubok, D.V. Schur, V.V. Sychev,
L.A. Matveeva ..................................................................................................... 103

"Nanodispersed Diamond Adsorbents for Biological Solution Cleaning"

G.P. Bogatyreva, M.A. Marinich, N.A. Oleynik, G.A. Bazaliy .......................... lll

Chapter 2. Thin Films and Coatings

"Tribological Characterization of Carbide-Derived Carbon (CDC) Films in

Dry and Humid Environments"
B. Carroll, Y. Gogotsi, A. Kovalchenko, A. Erdemir, M.J. McNallan ................ 119

"Nanostructured Ceramic Coatings Produced By Magnetron Sputtering"

V. Teixeira ........................................................................................................... 131

"Analysis of the Microhardness and the Porosity in Graded Thermal Barrier

Coatings"
A. Portinha, V. Teixeira, J. Martins, M.F. Costa, R. Vassen, D. Stoever. ........... 149

"Nanostructured Mesoporous Silica Films"

M.C. Gonyalves, G.S. Attard ............................................................................... 159

"Thin Film Crystallisation Behaviours for Oxides of Rare-Earth Metals on

Amorphous Silica"
A.F. Andreeva, A.M. Kasumov ........................................................................... 169

"Structure and Mechanical Properties of Nanostructured Metal-Carbon Films"

Y. Pauleau, F. Thiery, J. Pelletier, V.V. Uglov, A.K. Kuleshov, S.N. Dub,
M.P. Samtsov ....................................................................................................... 175
 Vll

"Thin Film Metal Matrixes for Biocorrosion Studies"

O.K. Dvoynenko, LA. Kozlova, V.M. Statsenko ................................................ 181

Chapter 3. Nanoparticles and Nanoceramics

"Synthesis and Properties ofNanostructured Carbon Materials: Nanodiamond,

Onion-Like Carbon and Carbon Nanotubes"
V.L. Kuznetsov, Yu.V. Butenko ........................................................................... 187

"Synthesis of Spherical, Non-aggregated Silica Nanoparticles"

A. Vital, U. Klotz, F. Graule, R. Mueller, H.K. Kammler, S.E. Pratsinis ........... 203

"Formation of Composite SiC-SiB 6 Nanopowders"

M. Lasova, M. Kakazey, 1.G. Gonzales-Rodriguez, G. Dominguez,
T. Tomila, L. Isaeva, L. Domasevich, A. Myratova ............................................ 211

"A Model of Electrical Conduction in Xerographic Toners as Polymer-

Conductive Filler Particulate Composites"
V. Skorokhod ....................................................................................................... 219

"Computerized Complex for Sintering Nanoceramics at High Pressures"

V.S. Urbanovich, G.G. Shkatulo .......................................................................... 229

"C 60 Fullerene Polymers"

R.A. Wood, M.H. Lewis, M.R. Lees, S.M. Bennington, M.G. Cain,
N. Kitamura .......................................................................................................... 239

"A Tight-Binding Molecular-Dynamics Approach to Structural and Electronic

Properties of a-SiC"
V.1. lvashchenko, P.E.A. Turchi, V.I. Shevchenko ............................................. 249

"Theory of Nanotube Nanodevices"

S.V. Rotkin .......................................................................................................... 257

Chapter 4. Sensor Materials

"Sol-Gel Processing of Lithia Aluminosilicates (LAS) Solid Solution for

Future Low Temperature Ceramic Sensor"
Wafa I. Abdel-Fattah, M.Sh. Fayed, Sh.R. Gouda, W.F.F. Mekky ..................... 279

"Structural and Optical Characterization of InN Thin Films - Novel Photonic

Materials for Photovoltaic and Sensor Applications"
V. Ya. Malakhov .................................................................................................. 29I
Ylll

"An Overview to Magnetic Beads Used in Electrochemical DNA Biosensors"

J. Wang, A. Erdem ............................................................................................... 297

"Mechanical Properties of Oxide Coatings Having Sensor Application"

D. Grabco, N. Palistrant, R. Jitaru, E. Rusu ......................................................... 305

"Approaches for Structured Immobilisation of Recognising Elements on the

Transducer Surface of Biosensors"
V.M. Starodub, A.V. Nabok, N.F. Starodub, L.I. Pirogova, W. Torbicz,
A.K. Ray .............................................................................................................. 311

"Performance of Amperometric Alcohol Electrodes Prepared by Plasma

Polymerization Technique"
D. <;6keliler and M. Mutlu ................................................................................... 327

"Chemical Active Centers at Surfaces ofSi-Based Materials"

A.E. Kiv, V.G. Litovchenko, D. Fuks, V.V. Golovanov, I.P. Lisovskyy,
T.I. Maximova ..................................................................................................... 333

"Deposition of Sn02-Based Thin Films by Reactive DC Magnetron

Sputtering for Gas Sensing Applications"
1. Miguel Correia-Pires, V. Teixeira, J.B. Almeida ............................................. 343

"Rehybridization at (110) Faces of Sn02 based gas sensor"

V.V. Golovanov, T.T. Rantala. T.S. Rantala, A. Kiv .......................................... 357

"Behaviour of Micro-Fabricated Composite Membrane as Amperometric

Glucose Biosensor"
eM. Pereira, J.M. Oliveira, F.Silva ................................................................... 365

"EQCM Study on the Polytyramine Modified Electrodes for the Preparation

of Biosensors"
eM. Cordas, A. Tenreiro, L.M. Abrantes ........................................................... 371

"Nanocoating of Particulate Surface in Colloidal Processing for Piezoelectric

Sensors Applications"
Wie-Heng Shih, Wna Y. Shih, Chia-Yi Yang, Huiming Gu, Jeong W. Yi ......... 377

Subject Index ............................................................................................................ 395

Preface

This volume contains papers that were presented at the NATO Advanced Research
Workshop on Nanostructured Materials and Coatings for Biomedical and Sensor
Applications held in Kyiv, Ukraine, 4-8 August, 2002. A total of 104 scientists from 14
countries participated in our ARW, making it a really international event. Participants
ranged from graduate students to senior researchers. They presented 16 tutorial lectures,
20 short talks and more than 70 posters. Invited speakers, from NATO and Partner
countries, presented some of the most recent developments in physics, chemistry and
technology of nanosized materials. A broad range of speakers having international
standing and representing NATO and partner countries, as well as university, industrial
and govemment research laboratories participated in this meeting and wrote papers for
this volume.
Foregoing ARW gathered together the scientists working in the area of nanosized
materials and coatings and their applications in biomedicine and sensors. The first
objective of this AR W was to discuss the current research covering a wide range of
physical and chemical properties of biomaterials and their use. Active discussion of oral
presentations and posters, and the round table discussion gave a good opportunity to
researchers from academia and industry to discuss the achievements in this field and
outline future directions in terms of technological developments and product
commercialisation in the fields of biomedicine and sensors. Particularly, advanced
ceramics and nanostructured carbons were covered in many presentations.
This ARW emphasized the connections between the physical, chemical and processing
data for nanosized materials, coatings and films. This AR W brought together scientists
from basic and applied research areas to initiate interactions aimed at translating basic
research achievements into engineering applications.
The workshop demonstrated that achievements in biomedical tt:chnology require
multidisciplinary research efforts. Those range from chemistry of complex oxide systems,
to physics and chemistry of different methods for their preparation and characterization,
mechanics of such systems and finally investigation of their interaction with, release of or
binding with biologically active (pharmaceuticals or toxins) molecules. Controlled-
release concepts for delivery of pharmaceuticals are mature and employed commercially.
Along with the different methods for selective and controlled delivery of medicine to
traumatized parts of human tissue, the removal of overdosed drugs in vivo is a very
important problem which only recently received attention. One of the most revolutionary
trends of material evolution for the past four decades is the innovative use of special
ceramics designed to reproduce different functions of living organisms, to deliver and
regulate medications and to investigate the processes of formatting and functioning of
bioplast. New interesting information in these areas was reported by the participants
from Egypt, USA, Israel, Ukraine, Portugal, Turkey and other countries. Particular
attention was dedicated to carbon nanotubes, fullerenes, nano-to-microscale carbon films,
which can be extremely useful in meeting the increasingly multifunctional needs of
advanced mechanical devices and have excellent biocompatibility.
ix
x

This volume is complementary to various specialized books (for example, on

nanopowders, sensors, or biomaterials) or more general books on nanomaterials or
nanotechnology. It aims to present an overview of research activities in nanostructured
materials for biomedical and sensor applications. All papers included in this book have
been peer-reviewed prior to publication. Some of the contributed papers will be
published in the special issues of Journal of Materials Processing and Manufacturing
Science and Poroshkovaya Metallurgiya.

This volume will be of interest for researchers and graduate students working in the area
of materials science and engineering, nanotechnology, biomaterials and sensors. It can
also be useful for engineers interested in production and use of relevant materials.

We acknowledge the contribution of the Director of Institute for Problems of Materials

Science Prof. V.V. Skorokhod and the members of the local Organizing committee Dr. L.
Kolomiets and Dr. T. Babutina for making this ARW a successful and smoothly running
meeting, and Mr. C. Avila for his help with editing this volume. Finally, we express our
sincere gratitude to the NATO Science Committee for granting us an award to organize
this meeting. Some of the administrative costs were borne by the Science and Technology
Center Ukraine, Drexel University and the Institute for Problems of Materials Science of
the National Academy of Science, Ukraine.

Yury Gogotsi Irina V. Uvarova

~
Department of Materials Engineering
Drexel University
1Jf/I
Institute for Problems of Materials Science
National Academy of Science
Philadelphia, PA 19104, USA Kyiv 03142, Ukraine
2002 NATO Advanced Research Workshop participants
 NANOSTRUCTURED MATERIALS IN MEDICINE. STATE OF THE ART IN
UKRAINE

1. UVAROVA
I.N.Frantsevich Institute for Problems of Materials Science,
Ukraininan Academy of Science,
3, Krzhizhanovsky Str.
03680, Kyiv, Ukraine

Most biological materials are composites at all levels from the organization of individual macromolecules to the
whole organism and are distinguished from any synthetic composites by their structural and organization
complexity. Bones. teeth and shells have a polymeric matrix reinforced by a mineral, hydroxyapatite or calcium
carbonate, which forms within the matrix. In bone, for example, a glycosaminoglycan matrIX is reinforced with
polymer fibres and with ribbon- or plate-like crystallites of hydroxyapatite, a calcium phosphate. The extent of
mineralization varies depending on the exact jUnction of the living organs. A typical level is about 38 vol. %
mineral with a crystalline thickness of 4 nm and lateral dimensions of 35 nm or more. The structure of
mammalian tooth is similar but the mineral level IS higher, about 86 vol %, and the individual crystallites are
larger. Shells of invertabrates have a comparable range of structures but the mineral is normally calcium
carbonate and the matrix is reinforced with chitin [I].
The Institutes of the National Academy of Science of Ukraine and Institutes of Ukrainian Academy
of MedIcal Science carry out many interesting investigations connected with using of hydroxyapatite. Among
them such Institutes should be mentioned as the Institute for Problems of Materials Science, Kharkiv Institute
of Monocrystals, Kharkiv National Umversity and Ukrainian Research Institute of Refractories, Institute of
Colloidal Chemistry and Chemistry of Water, Chemistry of Suljace, Institute of Gerontology AMSU,
Ukrainian SRI of Traumatology and Orthopedy, Institute of Microbiology and Virology of Ukrainian
Academy of Science (KyivJ and Institute of Experimental Pathology, Oncology and Radiology of Ukrainian
Ministry of Health Protection, Sytenko Institute of Spine and Joint Pathology and Kyiv Medical Academy of
Post graduation Education and so on.

1. Classification of different materials by their biomedical action

One of the most revolutionary trends of ceramics evolution for the past four decades
should be named the innovative use of special ceramics designed to reproduce different
functions of living organism, to deliver and regulate medications and to investigate the
processes of formatting and functioning of bioplast being the most perfect example of
nanostructured materials [2-8]. Ceramics used for these purposes are termed
"bioceramics". Bioceramics can be polycrystalline (alumina or hydroxyapatite),
bioactive glass, bioactive glass-ceramics, or bioactive composite (polyethylene-
hydroxyapatite). Many special ceramics and glasses have been developed during this
century for the health care industry; however, only a few have achieved human
clinical application. Clinical success requires the simultaneous achievement of a stable
implant - tissue interface and the match of the mechanical behavior of the implant and
the tissue to be replaced. It is important to take into account the progressive

Y.G. Gogotsi and I. V Uvarova (eds.),

Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 1-13.
© 2003 Kluwer Academic Publishers.
2

deterioration of tissue with age. The bone is especially defenceless to fracture in the
elderly because of the loss of bone density and strength with age.
Biomedical technologies to a great extent demand the attraction of the
multidisciplinary research effort. There is chemistry of complex oxides systems,
physicochemistry of different methods for their preparation and investigation,
mechanics and strength of such systems and finally medicine as a final and major stage
in biomedical technologies. Today Ukrainian biomedicine focused on the development
of novel medical techniques bases on the functionality of nanoparticles. There are
methods and preparative processes for manufacturing biomedical compatible ceramic
nanomaterials, nanostructured films and coatings on high strength substrates for
mechanical artificial valves, implants, and magnetic materials and for investigation of
nanoparticle behavior in intact organs of a living body. The investigations of
fundamental properties of bioactive ceramics are also very important for the
development of biosensors as materials and devices for control both different living
systems and medical techniques.
The different types of materials having the relationship to medicine can be
classified as biotoxic, bioinert, bioactive, or bioresorbable (Tabl. I) [2].

TABLE I. Biomaterial types determined in accordance with their reaction to

surrounding living tissue [2]

Medical action Reaction of organism Material

Biotoxic Atrophy, pathological change or Alloys containing cadmium,
rejection of living tissue near the vanadium, and other toxic elements,
material as a result of chemical, galvanic, carbon steels, carbides.
or other processes methylmethacrylate
Bioinert (to Coexistence with the material without Tantalum, titanium, alumina and
different degrees) noticeable change, separation from the zirconia ceramics
material by a layer of fibrous tissue
Bioactive Formation of direct biochemical bonds Bone autografts, high density
with the surface of the implanted hydroxyapatite and trica1cium
material, new bone free growth phosphate, some bioglasses
Bioresorbable Gradual dissolution of the material by Porous tricalcium phosphate, High
the biosystems of the organism, porous hydroxyapatite, calcium
replacement without toxicity and phosphate salts, some bioglasses,
rejection polyurethan
Biotoxicity is measured by local, system or general violations of organism metabolism
processes resulted from the action of biomaterial.
Bioinertness is determined by minimization of chemical action of a biomaterial on the
living organism.
Biocompatibility is an evidence of absolute regulation of living organism to the
chemical action ofbiomaterials and the absence of negative reactions between the living
organism and biomaterial.
Bioactivity determines the ability of a biomaterial to take part in metabolism processes
in a living body.
Bioresorbtion is solution of biomaterial resulted from metabolism processes in a living
body.
Biodegradation is the change of biomaterial properties due to the action of different
processes taking place in living organism.
 3

Biodestruction is a gradual destruction of biomaterial resulted from the action of

different processes in living organism.
Biofunctionality is an ability of biomaterial to promote some metabolism and other
functional processes in living body
The maximum bioinertness was found for the most refractory oxide ceramics
having the strongest chemical bonds and not being destroyed even by complex action of
different cells and ferment systems of living body. The material implanted in living
organism cannot be completely inert; all materials extract a response from living tissue.
The thickness of the interfacial zone or layer between the material and tissue influences
the relative level of the implant reactivity.
Metals cannot be combined into one group due to the principally different
nature of their chemical bonds, electrochemical processes occurring on their surfaces,
corrosion and loss of metal ions in the tissues. They can cause a variety of medical
problems. The reaction of living organism to a material substantially depends on the
physiological state of the letter. Thus, for example, in some cases splinters of carbon
steel may be encapsulated in a layer of fibrous tissue and retained in a human body
for tens of years (as a bioinert material), while in another organism at the same
situation toxicity leads to lethal results. These problems can be solved by using a
bioactive ceramic material, such as hydroxyapatite (HAP), as a coating on the porous
metal. The HAP coating also speeds up the rate of bone formation in the pores.

2. Advantages and disadvantages of bioceramics

Advantages of bioceramic materials are their bioinertness, biocompatibility, bioactivity,

bioresorption, and formation of direct biochemical bonds with the surface of the
implanted material.
Disadvantages of all bioceramic materials are their low crack resistance compared to
that of metals, practically complete absence of ductility, and difficulties of manipulation
and adjustment during a surgical operation (particularly high strength ceramics). One of
the methods to overcome these difficulties is using metallic implants with bioceramic
coatings, which makes it possible to combine the ductility and crack resistance of metals
with the advantages of ceramics. This possibility has attracted a great interest to the use
of such devices in modem medicine [9-16].
The use of monocrystals promises to improve the quality of the working
surface, thanks to the complete elimination of porosity, but for the same reason the
implant fixation is impaired. Kharkiv's scientists from the Institute of Monocrystals
carry out very interesting investigations in this direction and c:reate new sapphire
implants (Fig. 1) [17]. Functional advantages of sapphire implants comparing with more
known titanium analogs can be seen from data of Tabl. 2. The stability of sapphire in all
the acids and alkalis is incommensurably higher than titanium and other metals and
even polycrystalline alumina. It can be the main reason for wholt~ absence of patient
immune status changes. Authors established the increase of flexural strength in 1,5
times after heat treatment sapphire implants in hydrogen.
However, the use of single crystals does not eliminate the basic disadvantage
of corundum ceramics - low crack resistance.
4

Figure 1 Sapphire medical implants [l7]

TABLE 2. The comparison of sapphire and titanium implant functional characteristics

Implant Transfer of implant Decrease of bone Deflection of The formation of dense

material material to other issue at implant patient immune cuff round implant neck
body parts neck in 2 years status
Titanium + 0,5 +
Sapphire <0,2 +

2.1. CORUNDUM CERAMICS

Since 1977 the Sytenko Institute of Spine and Joint Pathology together with the
Ukrainian Research Institute of Refractories has been carrying out scientific work
connected with the use of corundum ceramics for orthopedy and traumatology [18-21].
During this period the vast experimental work aimed at the study of physico-
mechanical, toxic, biocompatible properties of ceramics has been accomplished.
Different clinical, radiological, morphological, mathematical and biomechanical
methods were used. The results of these investigations were treated by means of
variation statistics. Now the main attention of this work aims at investigation of
different ceramic compositions and the comparison of the possibilities of so called ultra
dense (poreless) Ah03 ceramics and porous corundum ceramics. Different types of
corundum ceramics are known to have ecological compatibility, adhesion ability with
biological tissues, compressive strength. Ultra dense and pure corundum ceramics are
produced by the Ukrainian Research Institute of Refractories from bauxites (natural raw
material) with the addition of MgO (0,15 0,20 %) from water soluble salt MgCIz 6H 2 0
before thermal treatment. This structure ensures high strength characteristics. The
properties of corundum ceramics after heating are given in Table 3. To produce porous
corundum ceramics a slip casting process with the addition of a pore former is used.
 5

Such ceramics have an apparent density from 1,0 to 1,5 glcm3 and a cold crushing
strength from 13 to 20 MPa. The peculiarities of the interrelation between an osseous
tissue and different types of ceramics (ultra dense and porous), determination of the
durability for the osseous-ceramic joining and the nature of biological reaction of soft
tissues surrounding the ceramic implant, the choice of most rational shapes and sizes of
ceramic implants have been studied in these cooperative work of the investigators-
ceramists and medical scientists. At dense and fixed contact the adhesion of implant
based on ultra dense ceramics to osseous tissue occurs at the expense of adhering
connective tissue to the finest roughness on the surface of ceramic implant. At the same
time an adhesion of porous implants takes place at the expense of growing of osseous
tissue in an implant pores with a subsequent development of blood vessels in contact
sites. The optimal pore size for the process of growing of osseous tissue into ceramic
material has been established by the authors to be 50 - 250 !lm .

TABLE 3 Properties of corundum ceramics

Ah01 content, %, not less 99,S
Prevailing size of crystals, ~m 16--18
Open porosity, % o
Apparent density, g/cm J 3,85
Bending strength, MPa 250
Crack resistance KI c, MPa mIl' 3-4
Modulus of elasticity, E 10', MPa 3,8

As a result of these experimental researches the most promising areas for

utilization of these materials have been proposed
- surgery of the spinal column.
- substitution of an osseous defect for any localization.
- reconstructive surgery of joints.
Ceramics based on tetragonal zirconium dioxide stabillized by yttrium or
cerium oxide are close to corundum ceramics in biocompatibilty, and have a higher
bend strength (up to 1200 MPa) and crack resistance (up to IS MPam I/2 ), but lower
values of Young modulus, compressive strength, and microhardness [22]. Other
advantages of zirconia ceramics are a lower coefficient of friction and less wear. The
disadvantages are manufacturing difficulties in attaining the n~quired porosity and
higher cost. It should be noted that certain ZrOz ceramic compositions (stabilized by
calcium or magnesium oxides) are not convenient for use as biomaterials, since their
surface roughness increases in a biomedium owing to grain boundary etching. Its
strength decreases in the biomedium due to aging the nature of wh ich is identical to that
of the phase composition change of ZrOz (Y 203) ceramics in the presence of a chemical
solvent and water. In spite of these drawbacks the replacement of corundum ceramics
with zirconium dioxide is a strong trend in the use ofbioinert materials.
6

2.2. NATURAL AND SYNTHETIC HYDROXYAPATITE

One of the main methods for nanocrystalline powder synthesis is the thermal
decomposition of different compounds resulting in creation of simple and complex
oxide systems with the best structure and level of mixture. Nanostructured materials
with the required particle sizes and the minimum of agglomerates can be created by
realizing the non-equilibrium state and control of the homogenization, coalescence and
coagulation processes at heat treatment. Hydroxyapatite (CalO(P04MOH)2 or HAP is
well known as an excellent bioactive material, capable of binding with living tissue.
Natural hydroxyapatite consists from composite organic matrix and non-organic mineral
of bone. The bones may be spongy and lamellar [23].
At cortical bone remodeling osteoclasts locate at the tip of the cone erode and a
canal forms within a bone. Osteoblasts presented on the lateral walls refill it and form
an osteon. At cancellous bone remodeling the process is begun with erosion of the
surface by osteoclasts and accompanied by formation of the cavity
Straight bones are harder and more structured. The bone consists of nano-
particles having the high hardness and organic amorphous matrix increasing its
durability. It is very interesting to create new nanostructured materials using bones after
their thermal treatment when organic matrix is moved off and the mineral of bones
remains. Adding the different glass phases and changing their quantity and the ratio
between bones mineral and amorphous phase the different composite materials can be
produced. In the context of synthetic composites, the biological materials have two
striking characteristics. The shape, size orientation and organization of the mineral in
the matrix show a high degree of sophistication when compared to the random
dispersion of particles in a filled polymer composite.
Hydroxyapatite is well known in surgery, orthopedy, traumatology and
oncology [24]. Hydroxyapatite ceramics having wide pore diapason up to ceramic foam
(higher than 80 %) have been created by the group of scientists led by prof. V.Dubok
from the Institute for Problems of Materials Science [25-28].
Different methods are known for modification of physical parameters of
inorganic bioceramics for functioning the components in living organism, such as the
use of plastic materials for filling up formation cavity, in other words, implants having
high bioactivity on the base of synthetic hydroxyapatite (SHAP), biological
hydroxyapatite (BHAP), glass-ceramics so called "bioglass of Hencha", containing
variable amounts of Ca, Si, P and Na oxides [29-31]. Today the most important task is
the development of new modifications of these ceramics to achieve the combination of
high bioactivity and mechanical strength. The advantage of glass-ceramics is the
possibility of wide variation of physico-chemical properties by changing the
composition, technology parameters, structure, size of particles and amount and size of
pores. Taking into account the process of growing of bone into implant the later must
have enough porous volume and a defined pore size.

2.3. BIOACTIVE GLASS AND GLASS-CERAMIC.

Bioactive glass and glass-ceramics implants have been used for more than 10
years to replace the small bones of the middle ear (ossicles) damaged by chronic
 7

infection. Bioactive glasses are also used in dentistry practice. The same bioactive glass
material has been used in a wide variety of clinical applications where bone grafting is
needed to fill spaces or augment the natural repair process. An especially important clinical
application ofbioactive implants is the use ofbioactive glass-ceramics in the repair of the
spine.
Cooperative researches of scientists from Institute for Problems of Materials
Science led by prof. L.Ivanchenko and the Institute of Gerontology AMS of Ukraine
and Ukrainian Scientific Institute of Traumatology and Orthopedy devoted to study of
composite materials based on silicate-boron-natrium oxide glasses with different
amounts of synthetic and biological hydroxyapatite established that such composite
materials have physico-mechanical parameters typical of silicate glass ceramics and
biological properties typical of "bioglasses of Xench" [32-39]. Biology activity is
usually determined by the rate of solution in physbiological mediums. The authors
established enhanced activity for glass ceramics with biological hydroxyapatite. By
varying the content of glass-ceramics and biological hydroxyapatite, and also
microstructures and particle sizes the authors aspired to approach the structure and
properties of different types of bone tissue. Mechanical compression strength equivalent
to that of enamel and other types of bone issue (140-300 MPa) was achieved. In these
cases the porous volume was equal 12-60 % vol. at the pore sizes in the limits of 30-
27511. For all the compositions the tearing away of implant from the cannon bone of
rabbit and inflammatory process for 1-25 weeks were not observed. The authors drew
the conclusion about the possibility of creating the material with a controllable rate of
biodegradation. Such ceramics are economical due to the use of cheap raw material.
Some parameters of samples made by the group of scientists from the Institute for
Problems of Materials Science led by prof. L. Ivanchenko [36] are shown in Tabl.4.
They established that porosity within created composite materials corresponds to
porosity of spongy bones (higher than 30 %), and liquid phase concentrate on surface
due to its surface tension. The mechanical strength of such compact implant increases
and they can be used as the assemblies. At preparation of granule from these materials
porosity can be higher and amount of glass phase can be smaller. Such granule variant
has been used for substitution of small defects in spongy bone.
The comparison of mechanical properties for sintered HAP with a tissue of
vertebrate is shown in Table 5 [39].
Ukrainian scientists from Kharkiv [30, 31] established that the best result of
growing for bone tissue of animal was in the case of using implants from synthetic
hydroxyapatite with a pore size of 80-15011, the volume of implants being under 3 % of
the total bone volume. The pores amount should be not smaller than 35 % vol. The
authors also established that the addition of a small amount (about 5 % mass) of silicate-
phosphate glass to SHAP powders not only increases the strength of such ceramics, but
also increases the bioactivity due to the presence of natrium and silicon ions.
Cooperative work connected with the preparation of high-quality
hydroxyapatite granules combining the high porosity and acceptable strength
characteristics on the base of nanosized (20-30 nm) initial powders is being carried out
by Kharkiv National University and Institute for Pathology of Spine and Joints. A foam
stirol was used as porous generator. So-called "dense" granules had only micropores
(porosity up to 32 %) and higher value of compressive strength (up to 48 MPa). More
8

porous granules (porosity up to 70 %) had the system of interconnected micro- and

macropores with the structure close to the mineral skeleton of spongy bone. In this case
the compressive strength was about 25 MPa. Both types of these granules were applied
in orthopedic surgery. In this case it is very important to choose the corrected rates of
heating and cooling and the temperatures determining the presence and amount of
amorphous phases. At low heating rates the porosity may be very big and the strength -
very small, at high heating rates cracks appear owing to the process of crystallization of
the amorphous phase of dehydroxylated calcium phosphate [40].

TABLE 4. Some parameters of glass ceramic specimens based on biological

hydroxyapatite

Amount ofBHAp Density, g/cm 3 Total pore Compressive Solubility, mass

Mol. % Mas. % volume, % strength, MPa %/day

3.0 32.95 2.588 .. 2.675 12 .. 14 70 .. 272 0.19

4.0 39.86 2.580 .. 2.715 18 .. 25 135 . .210 0.06
6.0 50.43 2.728 .. 2.750 18. .20 140 .. 217 0.09
12.0 68.70 2.855 ... 2.871 22 .. 25 91 .. 182 030
54 46.00 2.670 40. .42 20 .. 27 +0.4 ... +1.02**
6.0 50.43 2.503 20 .. 30 20 .. 28 035
12.9 73.92 2.850 20 .. 48 22. .25 0.15 ... 0.55
3.0' 32.95 2.010 ... 2.285 >55 - 0.05 ... 0.10
5.0* 45.59 2.432 ... 2.663 37 .. .40 8 .. 14 0.10 ... 1.00
5.0' 45.59 2.755 58 .. 72 II .. 14 0.03 .. 0.08
3.0' 32.95 2.492 ... 2.592 36 ... 47 35 .. 46 0.07 ... 0.09
5.0' 45.59 2.875 ... 3319 49 ... 60 17. .19 0.20 ... +0.12
* indicates that native bone was added. ** The + sign indicates that the mass of a
specimen increased during a time spent in physiological solution.

TABLE 5. The comparison of mechanical properties for sintered HAP with a

tissue of vertebrate
Properties Enamel (92- Bone (60-70% Sintered !-lAP
97%) Dense Porous
Density (g/cm 1 ) 2,9-3,0 1,5-2,2 3,156
Relative density 95-99,8 50-70
Orain size (nm) 7-30 200-2500
Mechanical strength (MPa):
Compression 250-400 140-300 270-900 30-100
Bending 100-200 80-250
Tension 20-114 90-120
Cross 35-95
Yaung modulus (OPa) 40-84 10-22 35-120
Fracture toughness (MPa.m Il2 ) 2,2-4,6 0,7-1,2
Coefficient of thermal conductivity 0,009 0,006 0,013
(W/cm.K)

The bioactivity of natural and synthetic hydroxyapatite and bioglass ceramics

containing different amounts of Ca and P oxides is connected with the capacity to form
chemical links with bone tissue. Hydroxyapatite is also used as a coating on orthopedic
and dental implants to enhance fixation of the implant in the surrounding bone.
Hydroxyapatite coatings that are prepared via plasma spraying typically consist of a
 9

mixture of crystalline and amorphous phases. The amorphous phase located on the outer
surface of the coating promotes the growth of osseous tissue compared to crystalline
sintered hydroxyapatite. However, the high dissolution of the amorphous phase can
influence implant stability. One way of controlling the amount of amorphous phase is a
heat treatment after the spraying procedure [41]. The formation of crystalline
hydroxyapatite at the coating/substrate interface via heat treatment has several
advantages. In this case the dissolution decreases and the substrate-coating bond
strength improves.

3. Ferromagnetic nanoparticIes with noble and platinum additions

Ferromagnetic nanoparticles are widely used in the medical institutions of Ukraine as

carriers not having toxic and allergic actions. The Institute of Colloidal Chemistry and
Chemistry of Water, Institute of Microbiology and Virology of Ukrainian Academy of
Science (Kyiv) and Institute of Experimental Pathology, Oncology and Radiology of
Ukrainian Ministry of Health Protection and Kyiv's Medical Academy of
Aftergraduation Education are carrying out multidisciplinary studies ultrafine
compositions of ferromagnetic nanoparticles based on a-Fe stabilized by oxide and
carbide iron with noble and platinum metals [42-48]. They are having potential for use
at medical treatment of festering and infective wounds, pathology of thyroid gland in
oncology for prolonged delivery of hormones into living body. Nanosized magnetic
materials based on iron with the particle sizes of 1-10 nm are known to be the carrier of
medicinal preparations for their directional delivery and decrease the loading on living
organism and increase the selectivity of action on pathologic tissues. The magnetic
particles can be used for advanced imaging techniques in medical applications.
An interesting fact studied in these works connects the action of magnetic fields
creating by nanosized ferromagnetic on the growth of plant cells. Such influence of
external magnetic field on the systems containing the ferromagnetic nanoparticles can
be interesting for the use of biotechnology in industry, for example at algae growing.
The action of magnetic field on ferromagnetic nanoparticles promotes the approach
ferromagnetic particles to the object of influence.
An interesting problem concerned with the technology of water cleaning is the
holding of considerable biomass of microorganism in clearing building. In this case
ultrafine ferromagnetic materials, especially with the addition of silver or other noble
metals, have a high bactericidal action for water disinfecting. The simplicity of moving
off such powders from water is their advantage.
Institute of Colloidal Chemistry and Chemistry of Water and the Institute of
Experimental Pathology, Oncology and Radiology are carrying out cooperative
investigations of action of magnetic field and ferromagnetic materials based on iron
nanoparticles and iron with small additions (from 0,1 to 10 % mass Pd or Pt) on normal
and transformed cells [47]. The authors of this work showed the possibility to delay the
cancerous growth in the cells coagulating with nanosized ferromagnetic. The best
results have been observed at the presence of 0,1 % mass Pd and Pt.. The decrease of
amount of free radicals in cancerous tissue after influence by nanostructured
10

ferromagnetic in author's opinion can be result in binding of free radicals by iron or

palladium.
The interesting works connecting with the deposition of coating based on
carbon containing compounds have begun in cooperative works of different Ukrainian
Academic Institutes. It is known that pyrolytic carbon coatings are used for prosthetic
heart valves. For example, 40 000 patients per year are in the United States that require
heart valve replacement, the availability of high-performance carbon bioceramics can
mean the difference between life and death.
Institute for Problems of Materials Science have been successfully working on the
problem offulerene deposition on the surfaces of titanium and steel.
 11
References

1. P.Calvert, S.Mann (1988) Review Synthetic and biological composites formed by in situ precipitation, J.
of Mater. Sci, 23, pp.3801-381 S.
2. V.A. Dubok (2000) Bioceramics - yesterday, today, tomorrow, Powder Metallurgy and Metal
Ceramics, N 7-8, pp.69-87.
3. L. Smith (1963) Ceramoplastic material as a bone substitute, Arch. Surg, 87, pp.653-661
4. SP Hulbert, S.J. Morrison, lJ. Klawitter (1972) Tissue reaction to three ceramics of porous and non-
porous structures, J. Biomed. Mater. Res., 6, pp. 347-374.
VADubok, O.N.Atamanenko, AY.Shinkaruk, EAShevchenko (2001) Study of hydroxyapatite
physical properties as the foundation for directional control of its bioactivity, 7th meeting and seminar
on Bioceramics cells and tissues, Biomimetic Engineering: a new role for ceramics, Faenza, Italy, June
13-15.
6. V.A.Dubok, A.Y.Shevchenko (2001) Structure and properties of hydroxyapatite nanopowders
synthesized by different methods, NATO Advanced Research Workshop Autumn School: Frontiers in
Molecular-Scale Science, Biopolymer Multifunctional Nanosystems, Kiev, 9-12 September 2001, SB-
R2.
7. VADubok, O.N.Atamanenko, EAShevchenko, A.V.Shinkaruk (2002) Study of bioceramics
fundamental properties as the basis to control their biological behaviour, - Proceedings of the
International Coriference CERAM-200I, Polish Ceramic Bulletin.
8. VA Dubok, N. V. Ulyanchich (1998) Synthesis, properties and using of osteotropic substitute of bone
issue on the base of ceramic hydroxyapatite, OrthopediCS, traumatology and prostheSiS, NB, pp. 26-30.
9. M.Vatanatham, S.Kimura (2001) Hydroxyapatite coating on thermally oxidized titanium substrates, J.
Am. Ceram. Soc, 84 (9), pp. 2135-2137.
10. A.V.Shevchenko, EAShevchenko, VADubok (2001) Bioactive coatings on high strength zirconia
based ceramics, 7th meeting and seminar on Bioceramics cells and tissues, Biomimetic Engineering: a
new role for ceramics, Faenza, Italy, June 13-15.
II. Suk-Woo Ha,R.Rebert, K-L.Eckert et al. (1998) Chemical and Morphological Changes of vacuum-
plasma-sprayed hydroxyapatite coatings during immersion in simulated physiological solutions, J. Am.
Ceram. Soc, 81(1), pp.81-88.
12. JF.Kay (1992) Calcium-phosphate coatings for dental implant - current status and future potential,
Dent. Clin. North Am., 36 (I), pp.l-18.
13. JHuracek, P.Spirig (1994) The effect of hydroxyapatite coatings on the fixation of hip prosthesis. A
comparison of clinical and radiographic results of hip replacement in a matched pair study, Arch.
Orthop. Traum. Surg, III (2), pp.72-77.
14. R.G. Courtney-Harris, M.V.Kayser, S.Downess (1995) Comparison of early production of extracellular
matrix on dense hydroxyapatite and hydroxyapatite- coated titanium in cell and organ culture,
Biomaterials, 16 (6), pp.489-495.
15. S.Hurson Differentiation ofHA coatings, J. Dent.Symp., 1, pp.65-66.
16. CChai, B.Ben-Nissan (1993), Interfacial reactions between titanium and hydroxyapatite, J.Aust. Ceram.
Soc, 29 (112), pp 71-80.
17. L.A.Litvinov (2001) Review Sapphire manufactures for medicine in Functional Matenals for Science
and Technique Inst. of Monocrystals Publisher, Charkiv, pp.188-199.
18. O.N. Gudushauri, B.L. Solomiansky, DA Omiadse (1978) On the materials for the substitution of
defect of bones and for the joining of bones, Orthopedy, traumatology and prosthetiCS 6, pp.l-6.
19. I.S. Kaynarsky, E.Y. Dieghtiaryova, I.G. Orlova (1981) Corundum refractories and ceramics,
Metalurgy, Moscow,.231 p.
20. T.1. Sanduza, L.A. Dierghaputskaya (1990) The research of properties of corundum heat insulating
products. Refractories, 1, pp.14-17.
21. V.V Martynenko, NA Korzh, P.P. Krivoruchko, L.A. Dierghaputskaya, N.L. Pyanyh, S.N.
Shevchenko, G.Kh. Gruntovsky (1999) Aluminium oxide - from porcelain and precious stones up to
endoprostheses for the reconstructive-rehabilitative surgery of human osseous tissues. Ukrainian
ceramics. National Cultural Annual, 4, pp.2S7-261
12

22. Kazutaka Sasaki, Manabu Nakano, Junpei Mimurada, Yiichi lkuhara, Taketa Sakuma (2001) Strain
Hardening in Superplastic Codoped Yttria-Stabilized Tetragonal-Zirconia Polycrystals, J. Am. Ceram.
Soc, 84(12), pp.2981-2986.
23. P.M.Royce, B.Steinman (I993) Connective tissue and its heritable disorders. Molecular. cenetic and
mineral aspects, pp.85-101.
24. G. Runge (1995) Osteoporosis Medicine Publisher, Moscow.
25. VDoubok, NUlyanchich, AZyrin, A.Senkevich, F.Buzaneva(1999) Synthesis of hydroxyapatite
powders with controlled morphology to produce osteotropic granules and ceramics for different
applications in surgery. British ceramic proceedings, 2 (60), pp. 515-517.
26. AB. Brik, N.N. Bagmut, AM.Kalinichenko, O.N.Atamanenko, O.I.Scherbina, VA.Oubok,
N.V.Ulianchich (2000) Characteristics of phosphate paramagnetic centres in natural apatites,
biominerals and their synthetic analogues. ISSN 0204-3548. Mineralogicaljournal, NQ4, pp,8-18.
27. VA Oubok, EAShevchenko, A.Y.Shinkaruk et al. (2002). Regulation and control of structure and
relief of particles' surface of hydroxyapatite nanopowders for usc in different medical applications, E-
MRS 2002 Spring Meeting, Strasburg, June 2002
28. VAOubok, N.V.Ul'yanchich (1998) . Syntesis, properties and uses of osteotropic replacements for
bone tissue based on hydroxyapatite Ortoped, Travmat, Protez., N3, pp.26-30.
29. T.Kanazava (1998) Inorganic Materials, Kiev, Naukova Oumka, 298 p.
30 K.Carison (1998) Biochemical activity of glass and its connection with structure, Fi::.. Khim. Stek!a, 24
(3), ppA05-412.
31. Malushkina S. V., Oedux N.V., Gruntovsky GX et al. (1998), .f of Orthopedy, Travmato!ogy,
ProsthesIS, N3 pp.1 10-114.
32. L.Alvanchenko, E.P.Podruchnyak, O.Yu. Xijun et al. (1999) Energy of electron transition in crystal of
bone hydroxyapatite at people of different age, DAN Ukraine, 3, pp.168-173.
33 LA Ivanchenko, T.I.Falkovska, N.V. Oanilenko et al. (1999) Structure and properties of a high-porous
glass ceramics containing biological hydroxyapatite, Powder Metallurgy and Metal CeramiCS, 38 (911 0).
ppAI8-453.
34 LA Ivanchenko. T.I.Falkovska, ND.Pinchuk et aI. (2002) Production and propertIes of hydroxyapatite
reinforccd by glass phase, Powder Metallurgy and Aleta! Ceramics, 41 (in print).
35 Patent of Ukraine 23250, Priority trom 22.07.97
36 LA Ivanchenko, TLFalkovska, N.Y.Oanilenko et al. (1999) Refractory and ceramics materials
Structure and properties of a high-porosity glass ceramic containing biological hydroxiapatate Powder
metallurgy and Metal Ceramics 38 (911 0) pp. 448-453.
37 EP.Podruchnyak, L.Alvanchenko, AT.Brusko (2000) The perspectives for using of glass-ceramics
containing biological hydroxyapatite at reduction of bone issue II .f of Orthopedy. Travmatology.
Prosthesis, N2, pp.129-130.
38. V.P.Orlovskii, S.G.Kurdyumov, O.I.Slivca (1996) Syntesis, properties and uses of calcium
hydroxyapatite, Stomatologia,.N 5, pp.68-73.
39. LA Ivanchenko, T.I.Falkovska, L.M.Kusmenko et al. (2002) The influence of nanosized structure of
biological hydrohyapatite on mechanical durability composites on its base, Nanocrystalline materials,
Publisher IPM NAS of Ukraine, Kiev, pp.44-49.
40. Z.z.Zuman, VAFilipenko, VARadchenko et al. (2001) Production, properties and orthopedic use of
hydrohyapatite granule, International conference "Advanced ceramics for third millennium", Kiev,
November 5-9, 2001, p.175
41 K.AGross. V Gross, Ch.C.Berndt (1998) Thermal analysis of amorphous phases in hydroxyapatite
coatings, J.Am.Ceram.Soc.,.81. (I). pp.106-112.
42. Kushchevskaja N.F. (1997) The use of ferromagnetic particles for medical application, Powder
metallurgy and Metal CeramiCS, NI1I12, pp.116-120.
43. Kushchevskaja N.F. (1997) Ultrafine ferromagnetic and possible directions of their use in biomedicine
for medical treatment of infection wounds, New technology in surgery, Kyiv,. pp .. 163-165
44. Shvets T.M, Kushchevskaja N.F., Klochko E.V. (1997) Study of ultrafine iron use for directional
transportation oftirocsine to organism, Medical bUSiness, N I, pp.73-75
45. Baraboy V.A., Savtsova Z.O., Shvets T.M., Kushchevskaja N.F (1996) Experimental analysis of tunor
ccus exposure to highly-dispersed ferromagnetic, Experimental Oncology, 18. (4), ppAI3-418.
 13

46. Gvozdyak R.I., Shvets T.M., Kushchevskaja N.F. (1996) Antibacterial activity the compounds with the
ultrafine iron, Microbiology, N6, pp.97-99.
47. Shvets T.M., Kushchevskaja N.F., Grigoryev E.S. (1997) The new materials for neurooncology, The
today's aspects for medical help, Lviv,. pp.163- I 64.
48. Gvozdyak R.I., Shvets T.M, Kushchevskaja N.F. (1997) Magnetic field and its influence on the
biologist objects, Medical business, N 5, pp.15-l7.
CURRENT TRENDS IN BIOMATERIAL COATINGS

S.V. MIKHALOVSKY, M. SANTIN, L.I. MIKHALOVSKA,

A.W. LLOYD, S.P. DENYER
School of Pharmacy and Biomolecular Sciences, University
of Brighton, Cockcroft Building, Lewes Road, Brighton,
BN24GJ, UK

1. Introduction

Our understanding and use ofbiomaterials evolves according to our understanding of the
phenomenon of biocompatibility. In earlier days, biocompatibility was synonymous with
the word 'inertness'. Attempts to create such an inert biomaterial, however, failed. As a
result, the quest for a perfect biomaterial has shifted from seeking an inert one to a
material that interacts with the biosystem it is in contact with in the same way as a
natural material would react in the same place. Respectively, biocompatibility is defined
as "the ability of a material, device or system to perform without a clinically significant
host response in a specific application" [I]. Although bulk propertie~! of biomaterials,
such as mechanical strength and optical properties, are equally important for their
performance, surface properties are most closely related to their biocompatibility.
Complexity of interfacial phenomena at the 'biomaterial - biological medium' interface is
hard to overestimate. They include adsorption (of water, proteins, low molecular solutes,
etc.), ion exchange, cell and bacterial adhesion, corrosion and decomposition of
biomaterial, catalytic effects and so on. Adsorbed proteins can change their conformation
and trigger cascades of events in the body, such as complement activation, blood
clotting, release of inflammatory cytokines and other effects comprising response of the
organism to the foreign material and, in general, to the damage caused by its
implantation.
Thus surface modification of biomaterials can be a powerful tool for improving their
biocompatibility. The most common techniques currently used in surface modification
are as follows:
• disguising foreign material by coating it with (i) a biomimetic substance or a layer of
endothelial cells, or (ii) a layer of more biocompatible (bioinert) material, such as
serum albumin, cellulose, synthetic hydrophilic polymers - PEO, PV A or
polyHEMA, or diamond-like carbon;
• charging surface with a biologically active compound, such as heparin, fibrinolytic
substances or prostaglandins. This compound could be attached to the surface or
loaded on a polymer coating, being slowly released from the latter.

In this paper results of in vitro evaluation of coated biomaterials performed by the

authors, are discussed.
15
Y. G. Gogotsi and l. v.: Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 15-26.
© 2003 Kluwer Academic Publishers.
16

2. Materials and Coatings

2.1 BIOMATERIALS

Three different types of materials have been studied. One, polymethylmethacrylate

(PMMA), is widely used for manufacturing intraocular lenses (IOL); the second,
titanium, has various dental and orthopaedic applications; and the third, stainless steel
316L (8S), is used for making cardiovascular stents.
Although all three materials have a successful track record of their use as
biomaterials, there are still many problems associated with postoperative complications
and long-term presence in the body. In case of IOL, for example, these complications
include post-operative inflammation, bacterial endophthalmitis, posterior capsule
opacification (secondary cataract) and corneal endothelial damage, all of which have
been attributed to poor interfacial biocompatibility of the existing intraocular lens
materials.
Titanium-made hip prostheses normally last for 12-19 years after implantation,
however, due to the ageing popUlation problem, this term is becoming insufficient, as it
requires the second replacement in patients in a very advanced age. By 1995 the
percentage of revision of hip prostheses increased in the UK to 18% [2]. Titanium and
other metal devices fail to integrate completely with the surrounding bone under
mechanical solicitations. Although an improvement of the mechanical grip of the
regenerating bone to the implant has been obtained through the engineering of the
surfaces into porous morphologies, a thin (few hundred nanometers) layer of non-
mineralised, soft tissue still forms at the bone/material interface. Such a soft layer, under
the continuous mechanical stress, is believed to initiate micromotions, which eventually
lead to the mobilisation of the implant and to the fracture of the surrounding bone [3].
Two major complications of cardiovascular stenting are blood clotting and in-stent
restenosis. The first one has been significantly reduced by using anticoagulant drugs and
improving the techniques of stent deployment, whereas restenosis - rapid growth of the
injured arterial tissue through the openings in the stent, poses the major challenge with
up to 30% occurrence of this complication [4]. The exact mechanism of in-stent
restenosis remains unclear, but it is generally assumed that insufficient stent
biocompatibility could be responsible for this complication.
It is possible that in all three cases new biomaterials with superior biocompatibility
will be eventually created or discovered and replace the existing ones, but these events
are difficult to predict, and the alternative option - improving performance of existing
materials by making their surface more biocompatible is currently the approach of choice
offering short-term solution to the problem.

2.2 BIOMIMETIC COATINGS

2.2.1 PC-Coating
An approach pioneered by Chapman et al. was the use of biomimetic coatings to
enhance biocompatibility ofbiomaterials [5]. Chemical composition and structure of the
coatings would mimic the phospholipids in the outer membrane of, e.g., red blood cells.
These cells are perfectly biocompatible, and phosphatidylcholine (PC), or lecithin, is the
predominant component of the outer layer of cell membrane. Chapman's work led to the
development of the phosphorylcholine (PC) technology, which has been successfully
 17

applied by various groups to enhance the biocompatibility of blood contacting and

ophthalmic devices. This technology was originally based on the development of
polymers containing phospholipid-like structures, which could be attached to surfaces by
physical adsorption or covalent bond (Fig. 1). Such PC-coatings have been applied by
our group to the development of novel intraocular lens coatings for the treatment of
cataract [6]. Some results of in vitro biocompatibility assessment of PC-coated PMMA
lenses are presented here. PC-polymer was produced and supplied by Biocompatibles,
Ltd. (UK). PMMA discs of 6 and 13-mm diameters were used in the experiments. Two
techniques of PC-coating were used: (i) dip-coating, or (ii) physical adsorption of PC-
polymer from ethanol solution. A detailed description of the experimental set-up is given
elsewhere [7].

o
-+-----CH2~-+-------------------------------+-

n
Figure 1. Chemical structure of a PC-polymer.

2.2.2 Hydroxyapatite
Another example of biomimetic coating becoming increasingly popular in producing
osseointegrative dental and orthopaedic implants, is hydroxyapatite (HA),
CalO(P04MOH)2. It is the major mineral constituent of hard tissue - teeth and bones in
humans and is considered to be a bioactive material that should be completely
compatible with tissue. The problem, hence, is to create strong adhesion between HA
and the metal implant to prevent defoliation of the HA coating. Recently it has been
suggested to functionalise the metal surface of the implant by gl~nerating surface
functional groups able to bind calcium, such as hydroxy, carboxy, phosphate and
sulphate moieties [8, 9].
Although these methods have offered new routes for implant osseointegration, they
rely on the principle that the stronger the binding of calcium, the higher is the
mineralisation potential of the surface. This approach ignores the natural process of
biomineralisation that takes place in living organism through functional groups with low
calcium-binding aOinity and high calcium-binding capacity [10]. Indeed, the nucleation
molecule, possibly a protein or polysaccharide complex, acts as a catalyst in the crystal
nucleus formation decreasing the crystallisation energy barrier without being included in
the final product. It means that calcium-binding groups should have relatively low
affinity with Ca2+. In an attempt to realise this concept, the early phases of
biomineralisation of the developmental bone were mimicked. In the developmental bone,
as well as in areas of rapidly mineralising bone, matrix vesicles are constantly found.
Their composition is similar to that of the cell membrane and contains phospholipids,
particularly phosphatidylserine (PS). Calcium phosphatelHA crystals form in the vesicle
interior from supersaturated solution in the presence of calcium-binding phospholipids.
18

In our experiments, titanium implant was dip-coated with phosphatidy1choline and

placed into simulated bodily fluid (SBF) for mineralisation.

2.3 BIOINERT COATINGS

Different coatings of approximately 3 Jlm in thickness were deposited industrially by

means of closed field unbalanced magnetron sputtering technique (Teer Coatings Ltd,
UK) on discs of medical grade SS 316L [11]. Metallic (Ti or Zr) and diamond-like (D)
coatings on stainless steel were prepared. Preliminary studies of the biocompatibility of
some of these coatings were encouraging [12]. Before biocompatibility tests, the discs
were thoroughly washed with a detergent, deionised water and ethanol.

2.4 BIOCOMPATIBILITY TESTS

There are no universal biocompatibility tests, as biomaterials are used in different

applications and contact different tissues and media in the body. For materials contacting
bodily fluids (stents and IOLs in this case), adsorption of proteins and cell adhesion are
the most important primary events, whereas rapid development of an HA layer may be
considered as the most important event for osseointegration. For all biomaterials
adhesion of microorganisms is relevant to their performance, because infection caused
by implantation of an artificial organ is a serious post-operative complication. Details of
the biocompatibility tests are discussed in the following sections.

3. Results and Discussion

3.1 PC-COATED PMMA IOLS

Presence of PC-coating on PMMA IOLs was confirmed using XPS analysis, which
revealed P, CI and N in the coated samples (Table 1). Physical adsorption produced more
homogeneous coating than dipping; hence, the former was used subsequently in PC-
coated PMMA preparations. Dynamic contact angle analysis, DCA, using a CAHN
DCA-322 instrument, proved to be a sensitive technique to measure the effect of PC-
coating on the surface properties of PMMA (Fig. 2). PC polymer increased the
advancing contact angle from 82.1±0.6° for uncoated PMMA to 104.7±6.4° for PC-
coated from 10 g L'] solution. PC-coating also increased the hysteresis between
advancing and receding contact angles from 30.0° for uncoated PMMA to 104.7° for
PMMA coated from 10 g L'] solution of PC-polymer. The threshold in the 'contact angle
vs. PC-polymer concentration' relationship occurs in the 0.1 - 0.5 g L-] concentration
range (Fig. 2). The in vitro biological tests performed with the PC-coated PMMA
showed improved biocompatibility of these materials compared to PMMA. These
include: reduced fibrinogen (Fbg) adsorption (Fig. 3), reduced cell adhesion (Fig. 4) and
reduced bacterial adhesion (Fig. 5).
 19

TABLE 1. XPS elemental analysis of PMMA

Sample Element, atomic %

C 0 P Cl N
PMMA (calc.) 71.4 28.6
PMMA(exp.) 60.4 39.6
PC adsorbed· 64.7 35.2 0.D3
PC adsorbed·· 62.4 35.9 0.53 0.26 0.93
PC dip-coated· 59.7 39.9 0.16 om 0.15
PC dip-coated·· 68.1 27.6 1.3 0.60 2.34
• From 0.1 g L·' PC solution in ethanol
•• From 109 L·' PC solution in ethanol

120,1------------------------------------,

60 .~~~---L~~L-~--~~--~~~---L-.~~IJ'
o 0.001 0.01 0.1 0.5 5 10

Initial PC concentration, g C'

L -_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __

Figure 2. Dependence of the average advancing contact angle of PC-coated PMMA on the initial PC polymer
concentration in the solution.
Experimental conditions: PC-polymer is physically adsorbed from ethanol solution at RT, for 5 h.

Fibrinogen is one of the major proteins responsible for the failure of implanted devices,
because it triggers an inflammatory response to foreign materials and initiates the blood
clotting cascade. A low level of Fbg is ubiquitously present in the aqueous humor and
rises to much higher levels following ocular tissue trauma during intraocular implant
surgery. The reduced adsorption ofFbg in vitro should be indicative of generally reduced
protein adsorption to the coated PMMA, which is beneficial for IOLs.
Cell adhesion is a very important index ofthe biocompatibility ofIOLs. The adhesion
of fibroblasts is mediated by a wide variety of proteins, and the significant reduction in
adhesion to PC-coated PMMA may also reflect a general tendency of PC-polymer
coating to suppress protein adsorption (Fig. 4). The reduced adsorption of S. epidermidis
cells to the coated PMMA is of great potential relevance in the prevention of infection,
particularly post-operative endophthalmitis [13].
20

tl 2.5
'2
;:!
2 +-~----------------------------------
§
d 1.5
o
.~
o
.g 0.5
OIl
.D
r....
0.001 0.01 0.1 0.5 5 10

Initial PC concentration, g C l

Figure 3. The effect ofpC-coating on the adsorption of fibrinogen to PC-coated PMMA.

Experimental conditions: Fbg adsorbed from 2 mg L-' phosphate buffer solution, PBS for 2
h at RT. The amount of bound Fbg was determined using ELISA with the chromogenic
substrate ABTS (Aldrich-Sigma) by measuring absorbance at 670 nm.

~ 100%
~
Cl..
~
<>
d 50%
0
'00
<lJ
..<::
"0
-<
0"10
10 5 0.1 om 0.001
Initial PC concentration, g C l

Figure 4. The effect ofPC-coating on the attachment of fibroblasts to PC-coated PMMA.

Adhesion to uncoated PMMA = 100%.
Experimental conditions: 3T3 embryo mouse fibroblasts were seeded on the sample at a
density 1000 cells/sample and placed for 72 h in DMEM + 10% foetal bovine serum. The
attached cells were lysed, and cellular A TP extracted and measured by A TP luminometry
(MicroLite ATP assay, Dynatech)

The evaluation of the biocompatibility of PC-coated and uncoated PMMA revealed a

clear correlation between the physicochemical properties of a PC-coated surface, as
determined by contact angle analysis, and the biological performance of that surface as
measured by bacterial and fibroblast adhesion and fibrinogen adsorption. The threshold
of the contact angle in the 0.1 - 0.5 g L- 1 concentration range seen on Fig. 2, corresponds
to the threshold between PC-coated PMMA samples, which demonstrate markedly
improved biocompatibility and those, which do not differ from uncoated PMMA.
Attempts to find a correlation between the surface energy of biomaterials (that can be
 21

estimated from contact angle measurements) and their biocompatibility have been made
since the work of Zisman [14], but our understanding of this relationship is poor. The
fact that PC-polymer coating of PMMA significantly improved its biocompatibility in
vitro and at the same time actually increased its surface heterogeneity (larger hysteresis)
suggests that this relationship is much more complex than it was previously supposed
[15,16].

<t:: 140"10
~
~
Q.,
?t- 90%
d
0
.:;:;
0)
..<::
~ 40%
03
.£:
E
i;l
co -10% 0.001 om 10
0.1 5
Initial PC concentration, g C1

Figure 5. Adhesion of S. epidermidis cells to the surface ofpe-coated PMMA.

Adhesion to uncoated PMMA = 100%.
Experimental conditions: PMMA samples were incubated in a suspension of 3xlO' cells
mL" S. epidermidis in PBS, for 4 h at 37° C. ATP was extracted from the adherent
bacteria and measured by ATP luminometry.

3.2 HYDROXY APATITE ON TITANIUM IMPLANTS

Dip-coating of titanium implants in solution of phosphatidylserine produced a layer able

to re-arrange itself into a 3-dimensional structure in calcium-rich t:nvironment, thus
mimicking the ossification of bone in vivo. Calcium ions cross-link the phospholipid
molecules thus forming a mesh, in which the concentration of calcium and phosphate
reach supersaturating levels resembling the interior of the matrix vesicles. A positive
feedback in the system is thus created, as phosphatidylserine promotes the binding of
calcium ions at the implant surface and calcium acts as the driving force for the re-
arrangement of the coating into a 3-dimensional matrix prone to mineralisation.
Phosphatidylserine was able to induce the formation of a mineral phase in the simulated
bodily fluid, SBF within 30 min, and the presence of plasma proteins caused only a
slight delay of the mineralisation process at the implant surface, which was in any case
completely coated by a mineral phase after 2 h of incubation (Fig. 6, a and b).
Phosphatidylserine being a molecule recognised by inflammatory cells during apoptosis,
concerns were raised about the pro-inflammatory potential ofthese coatings.
22

Figure 6. Mineralisation of a dental fixture after 2 h Incubation in SBF.

Left-hand photo: implant immersed into SBF; right-hand photo: SEM (20x magnification) showing
details of the mineralised 3-dimensional matrix of PS in the screw trough (white layer). The screw
thread was coated with a thinner coating also containing Ca and P according to EDAX analysis.

In vitro experiments demonstrated that PS-derived coatings inhibited adhesion and

activation of macrophages. Preliminary study of in vivo behaviour of coated titanium
implants in rabbit femour bone at 4 and 8 weeks after implantation seems to confirm the
osseo integrative potential of this type of coatings.

3.3 COATED STAINLESS STEEL STENTS

EDAX analysis (Philips XL20 SEM with an ED AX system, 10 kVaccelerating voltage)

confirmed the chemical composition of metallic coatings. SEM examination also
revealed that Ti coating is more corrosion resistant than uncoated SS or Zr-coated
samples in both human blood plasma and Krebs solution after a week of incubation at
ambient temperature.
Adsorption experiments were carried out at 37° C for 1 h. Preliminary studies on the
plasma protein adsorption from citrated normal human blood plasma onto uncoated and
coated discs using SDS-Page showed relatively low adsorption of plasma proteins. The
total protein amount in the eluates from discs determined by DC Bio-Rad assay was less
than 2.0 Ilg per disc of 10-mm diameter.
A detailed SDS-Page analysis carried out on the eluates suggested that at least three
main proteins, human serum albumin (HSA), fibrinogen (Fbg) and immunoglobulin G
(IgG) were adsorbed by the surfaces. To confirm this conclusion, immunoprobing of
blotted proteins onto nitrocellulose membrane was performed using specific antibodies,
Ab (Western blot). Immunoprobing with polyclonal Ab against Fbg, fibronectin, HSA,
C3c complement protein and IgG proved presence of all these proteins in SDS eluates
(results for Fbg are shown in Fig. 7).
Electrophoresis and Western blot are semi-quantitative rather than quantitative
methods of analysis. Quantification of protein adsorption was performed using human
citrated platelet poor plasma spiked with 125I-labelled HSA or 125I_Fbg (Amersham, UK).
 23

SS Ti Zr D FblZ PI SS Ti Zr PI D

Non-reduced Reduced

Figure 7. Immunoprohing of eluates klr lihrinogen using Western hlot.

Experimental conditions: chemical reduction of the eluates was done by ~-mercaptoethanol;
proteins separated by SDS-Page were electrophoretic ally blotted onto nitrocellulose membranes.
The membranes with blotted proteins were blocked for an hour in 5% skimmed dried milk in
phosphate buffer solution containing 0.1% Tween-80 and further incubated overnight with specific
primary Ab. Ab binding was detected with HRP-conjugated secondary antibodies with 4-chloro-l-
naphthol and H,O, as substrates. In this figure three distinctive bands in the range of molecular
masses between 45,000 and 80,000 (reduced eluates) characteristic of Fbg, are clearly seen. Well
two of the Non-reduced sample was left empty.

1200 + - - - -

800 -j-------~-----------__I

400 -------.-----.--------~-_i

ss Zr Ti D
Materials
~----------------------- ---------~

Figure 8. Adsorption of fibrinogen on uncoated and coated SS sample.

Experimental conditions: plasma was spiked with I"I-Fbg to final activity in the range ofO.734xl06
cpmlmL to 1.406 xl06 cpm/mL of plasma (3.668xl06 cpmlmg to 7.03 lx106 cpmlmg Fbg);
radioactivity of the samples was measured using WALLAC Wizard 1480 Automatic gamma
counter (Wallac, Finland).
24

35
~ 30
'6 25 rt"_
~ 20 f---

::1. 15 -
-i
r/)
10 -
::I: 5 -
o r---, I I J-E\
SS Ti Zr D
Materials

Figure 9. Adsorption of human serum albumin on uncoated and coated SS samples

Experimental conditions: similar to those described in the legend to Fig. 8.

From Fig. 8 it seems that SS with diamond-like coating (D) should be less
haemocompatible than three other materials due to higher adsorption affinity with
fibrinogen; however, D-coating also favours higher serum albumin adsorption (Fig. 9),
which is generally considered to be an indicator of good haemocompatibility.
Apparently, adsorption rate becomes an important parameter in such a case and more
subtle experiments are needed to evaluate biocompatibility.
Cell adhesion is an important contributor to the biomaterial performance in the body.
All four materials adhered comparable number of platelets as estimated by LDH assay,
but from the platelet morphology analysis by SEM it could be seen that platelet
spreading was the highest on SS and the lowest on D-coated material (Fig. 10). Lower
platelet spreading should mean that the surface does not activate these cells reducing the
likelihood of thrombus formation upon biomaterial contact with blood.
The early activation of monocytes in contact with a biomaterial surface within first 5
min was evaluated by a chemiluminescence method. Only polydimethylsiloxane, Sil
used as a reference material showed an oxidative burst by the monocytes significantly
different from the control cells. SS, Zr- and D-coated discs did not induce any significant
activation of the inflammatory cells. However, when the activation of the
monocytes/macrophages was evaluated at a longer incubation time (3 h) in terms of
released amount of TN Fa, it appeared that SS, Ti and Sil surfaces stimulated the cells to
produce levels of this inflammatory cytokine higher than the control. Zr- and D- coated
samples stimulated TNFa production comparable to the control level (Table 2).
The data presented in this paper clearly demonstrate that although diamond-like
coated surface binds higher levels of proteins such as fibrinogen, its thrombogenic and
inflammatory potential is significantly lower than that of SS surface. These findings,
associated with the known lack of corrosion for carbon materials, emphasise the benefits
of using carbon coating in stent manufacturing. However, further, more sophisticated
experiments in vitro are required that take into account conformational changes of the
adsorbed proteins and the relationship between inflammatory cell activation and smooth
muscle cell proliferation to develop an adequate in vitro restenosis model and to
elucidate the molecular basis of this phenomenon thus establishing clinically reflective
screening tests and minimising animal trials.
 25

c d
Figure IV. Platelet adhesion to ullcoated S8 (a), Ti- (b). Zr- (c), and D- (d) coated 8S.
Experimcntal conditions: matcrials were incubated in autologous platelet rich plasma for 1 hour at
37"C

TABLE 2. Material-induced TNFa synthesis by

monocytes/macrophages after 3-h incubation

Sample (n=3) TNFa, ngL- 1 TNFa range, ng L- 1

Control 1.6 0.5-2.9
SS 10.1 4.5-15.3
Sil 7.5 3.0-10.7
D 3.3 0.6-6.9
Ti 5.1 3.9-7.6
Zr 2.8 0.9-5.6

Experimental conditions: The discs were incubated with ]08 cells L" in RPMI-1640 medium for 3 h at
37° C in the humidified air with 5% CO, flow. Tumour necrosis factor alpha (TNFa) level was
measured in each supernatant from the cell adhesion experiments using an Amersham kit (Cat. N
RPN2788).

4. Conclusions

Improvement of the biocompatibility and longevity of implants by developing novel

coatings is considered to be an efficient but short-term trend. It is likely to include the
26
development of bioactive and bioinert ligands for surface engineering of materials for
particular applications, some of which are presented in this paper. Although in the longer
term it is expected that completely novel products for tissue and organ replacement will
be developed using achievements of tissue and genetic engineering, it is unlikely that the
surface modification approach will become obsolete. After all, the surface structure of
cells, blood vessels and other tissues and organs demonstrates that Mother Nature uses
surface engineering to create perfectly biocompatible materials.

References

I. Gurland, HJ., Davison, A.M., Bonomini, V., Falkenhagen, D., Hansen, S., Kishimoto, T, Lysaght, MJ,
Moran, J. and Valek, A. (1994) Definitions and terminology in biocompatibility, Nephrol. Dia/. Tramp/.
9, Supp!. 2,4-10.
2. R&D Priorities for Biomaterials and Implants (1996) UK Department of Health.
3. Steflik, D.E., Sisk, A.L., Parr, G.R, Hanes, PI, Lake, E, Song, MJ., Brewer, P. and McKinney, RV.
(1992) High-voltage electron microscopy and conventional transmission electron microscopy of the
interface zone between bone and endosteal dental implants, J. Biomed. Mater. Res. 26, 529-545.
4. Serruys, P.W., de Bruyne, B., Carlier, S., Sousa, J.E., Piek, J., Muramatsu, T, Vrints, C, Probst, P.,
Seabra-Gomes, R, Simpson, I., Voudris, V., Gurne, 0., Pijls, N., Belardi, J., van Es, GA, Boersma, E.,
Morel, MA and van Hout, B. (2000) Randomized comparison of primary stenting and provisional
balloon angioplasty guided by flow velocity measurement, Circulation 102, 2930-2937.
5. Hayward, JA and Chapman, D. (1984) Biomembrane surfaces as models for polymer design - the
potential for hemocompatibility, Biomaterials 5, 135-142.
6. Lloyd, AW., Faragher, RGA and Denyer, S.P. (2001) Ocular biomaterials and implants, Biomaterials
22,769-785.
7. Lloyd, AW., Oropcova, S., Faragher, RG.A., Gard, P.R., Hanlon, GW., Mikhalovsky, S.v., Olliff,
CJ., Oenyer, S.P., Letko, E. and Filipec, M. (1999) The development of in vitro biocompatibility tests
for the evaluation of intraocular biomaterials,J. Mater. Sci. - Mater. M. 10, 621-627.
8. Kokubo, T (1998) Apatite formation on surfuces of ceramics, metals and polymers in body
environment, Acta Mater. 46, 2519-2527.
9. Nanci, A., Wuest, J.D., Peru, L, Brunet, P., Sharma, V., Zalzal, S. and McKee, M.D. (1998) Chemical
modification of titanium surfaces for covalent attachment of biological molecules, J. Biomed. Mater.
Res. 40,324-335.
10. Mann, S. (1988) Molecular recognition in biomineralization, Nature, 332,119-124.
II. Teer, D.G. (2001) New solid lubricant coatings, Wear, 250, 1068-1074.
12. Lazarenko, O.N., Alexeeva, TA., Mikhalovsky, S.V., Shekera, O.v. and Skiba, I.A. (1999)
Biocompatibility of different intravascular stent coatings with whole blood, BasiC Res. Cardiol., 94, 379.
13. Arciola, CR, Cenni, E., Tarabusi, C, Caramazza, R and Pizzoferrato, A. (1993) Corneal endothelium
cell adhesion on intraocular lenses in vitro, J. Appl. Biomater., 4, 249-252.
14. Zisman, W.A. (1963) Relation of the equilibrium contact angle to liquid and solid constitution, in
F.M.Fowkes (ed.), Contact Angle, Wettability, and AdheSion, Amer. Chem. Soc., Washington, DC, pp.
I-51 (Advan. Chern. Ser., 43).
15. Ruckenstein, E. and Gourisankar, S.V. (1984) A surface energetic criterion of blood compatibility of
foreign surfaces, J. Colloid Interface Sci., 101,436-451.
16. Schrader, M.E. (1982) On adhesion of biological substances to low energy solid surfaces, J. Colloid
Interface Sci., 88, 296-297.

Acknowledgments

This work was supported by EPSRC grant GRIR31584.

INJECTABLE NANOPARTICLE TECHNOLOGY FOR IN VIVO
REMEDIATION OF OVERDOSED TOXINS

R. PARTCH
Engineering Research Center, University of Florida
Gainesville, FL 32611, USA; and
Center for Advanced Materials Processing, Clarkson University
Potsdam, NY 13699, USA
E.POWELL
Department ofChemistry, Clarkson University
Potsdam, NY 13699, USA
Y-H. LEE
Department of Chemistry, Kyungwon University
Sungnam City, Korea
M. VARSHNEY
Department of Chemistry, Hamdara University
New Delhi, India
S.KIM
Material Science and Engineering, University ofFlorida
Gainesville, FL 32611, USA
N. BARNARD AND D. SHAH
Department of Chemical Engineering, University ofFlorida
Gainesville, FL 32611, USA
D. DENNIS and T. MOREY
Department ofAnesthesiology, University of Florida
Gainesville, FL 3261 1, USA

Abstract

Internal and external exposure to excess amounts of natural and synthetic chemicals,
including some therapeutics, can be reversed if general or selective antidotes are
available. The focus of the present study-in-progress is to synthesize, characterize and
evaluate the biological effectiveness of several types of injectable dispersed phases
having potential for binding and deactivating some lipophilic molecules that when
taken internally in excess cause cardiac failure. The types of dispersed phases under
investigation are I )microemulsions, 2)microgels, 3)porous nanoparticles, 4)nanotubes,
5)core-shell nanoparticles and 6)nanoparticles with toxin receptors covalently attached
to their surfaces. In this chapter only types I), 3) and 6) will be discussed.
Biocompatible oil-in-water microemulsions stabilized by co-suriactants have been
27
Y. G. Gogotsi and 1. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 27-40.
© 2003 Kluwer Academic Publishers.
28

prepared which are stable in blood and capable of quickly absorbing quantities of
bupivacaine, cocaine and amitriptyline. Silica particles to be used for reference
purposes have been synthesized with pores templated for bupivacaine and will be
evaluated for adsorption capacity. Silica nanoparticles with attached dinitrobenzoyl or
cyclodextrin binding receptors show rapid and high yield binding by aromatic 7t-7t
complexation or by cavity penetration, respectively. BET, HPLC, NMR, SEM, TEM
and TGA data have been obtained to support the chemical conclusions. Bioassays have
been performed using EKG/QRS and TEG techniques.

Introduction

The risk of exposure to and death from overdoses of many natural and synthetic
chemicals in the environment are problems that must be solved in parallel with
advancing technology. In cases where the chemicals produced are byproducts and not
useful, reduction or elimination by improved "green chemistry" is desireable if not
required in the long term. But many chemicals beneficial to human comfort and
longevity when used in proper amounts become toxic when in excess. Antidote
availability may be the only pathway to remediation in emergency situations (Figure
I). Materials for use in selectively reversing

Most Toxin Molecules are Lipophilic

Drugs of Abuse TherapeutIcal Drugs Organophosphates

• opiates • anti-inflammatories Soman, GD

• cocaine • antibiotics Sarin, GB
• cannabinol • steroids Tabun
• steroids • anti-arrythmics VX
• antidepressants
• anesthetics

Figure I. Example lipophilic molecules having toxicity if overdosed

the lethal effects of some therapeutics responsible for thousands of accidental deaths
annually have previously not been investigated. This chapter serves as an interim
report on efforts by an interdisciplinary and international team to prepare, characterize
and evaluate injectable nanoparticles able to bind and inactivate a selection of
chemicals that can adversely affect heart function (Figure 2). The dispersed phase
nanoparticulates discussed are of three types:
 29

Drug Characteristics
H3C~croH3C~:y~
I "" I • Clinically important drugs
• No known antidotes

CH 3 . • Both highly lipophilic molecules

Bupivacaine
Kill by inhibiting heart function
(Local Anesthetic)

~
~,/CH3 N
JH Amitriptyline
3 (Antidepressant)

Figure 2. Chemical structures and biological activities of example toxic therapeutics

oil-in-water microemulsions, prepared by coauthors Varshney and Shah; templated

porous silica particles prepared by coauthors Partch, Lee and Bamard; and silica
particles with covalently attached electron deficient aromatic rings or cyclodextrln
molecules prepared by coauthors Partch, Powell, Kim and Lee. Coauthors Dennis,
Morey and Varshney performed the bioassays on all the above materials.

Experimental and Discussion

TEMPLATED POROUS SILICA NANOPARTICLES

Procedures for forming molecular voids, otherwise called imprints, in solid inorganic
or polymeric particles or membranes are well known [1-5]. Regardless of the physical
state of the reactants, i.e. vapor, aerosol or liquid, the reactions employ either addition
of templating molecules or ones having similar structure to the reaction medium so
that they are trapped in the matrix being created [3]; or, templating molecules are
incorporated chemically into a reagent before the synthesis of the matrix begins [5]. In
both processes the templates are removed from the matrix to leave the imprinted voids
by dialysis, hydrolysis or thermolysis.

In the present work silica nanoparticles were prepared having high surface area and
pores templated by 2,6-dimethylaniline, including its acetamide and bupivacaine
derivatives, the latter an often used local anesthetic having serious cardiotoxicity if
overdosed. Sol-gel hydrolysis of several alkoxy silanes was carried oul: with either acid
or base catalysis and with one of the dimethylaniline derivatives present. [6]. Figure 3
gives some detail regarding an example reaction and analytical data on the
nanoparticles produced. Based upon literature precedent it is assume:d that the pores
30

remammg after thermal removal of the template molecules from the solids are
approximately the size of the aniline derivative. The affinity for these types of
nanoparticles for several aniline derivatives, including bupivacaine, is under current
investigation using HPLC and proton NMR techniques. Future experiments will focus
on creating pores for selective adsorption in biocompatible core nanoparticles.

TEOS + H20/dilHel + ~ -----+

I mole 3 mole CH,
° CH,
~
24 hr
dilHCl
5 mole
~ templated silica (7 nm)
NH, \047 m2/g, 21 'l

Figure 3. Example procedure for the synthesis of silica using a 2,6-dimethylaniline derivative template

OIL-IN-WATER MICROEMULSIONS

Microemulsions are established media for absorption, transport and release of

biological materials [7,8]. Their application as antidotes for removal of overdoses of
chemicals from blood has not been studied. The interfacial energy of oil-in-water
microemulsions is extremely low and provides a potential method for extraction of
lipophilic molecules into the oil core. The level of solubilization of a drug or toxin into
the oil phase depends on the hydrophilic-lipophilic (HLB) number of the emulsifiers,
the nature of the oil, inter-droplet interactions, and droplet size distribution in the
mixture. In the current study several pluronic and fatty acid salt co-surfactants were
evaluated with ethyl butyrate to derive an optimum microemulsion for absorbing either
of the molecules shown in Figure 2 (Figure 4). Pluronic Fl27 and sodium caprylate
gave the best results. Figure 5 depicts
 31

Surfactant: Pluronic (PEO-PPO-PEO)

L.D:<;:::; 2 to >1 b gikg

I'EO rI'o PFO

r~.n l'f'O 1'':0

• Oil Ethyl butyrate
IV ,njectiop fQr dDgs.

1T! ~222 rng!kg has no effect

Figure 4. Some microemulsion components for absorption of lipophilic molecules

Extraction of Dmg by Microemulsion

Figure 5. Conceptual description ofmicroemulsion structure and drug absorption

the structure of a microemulsion oil droplet with approaching and absorbed drug
molecules. The efficiency with which this dispersed phase removes amitriptyline from
blood is shown in Figure 6, as determined by HPLC.
32

Saline Micelles ME-I ME-II

0% EB) (4% EB)
Figure 6. Efficiency of removal of amitriptyline from blood by a microemulsion. EB refers to ethyl butyrate.

An obvious question when preparing materials for injection into the body is - are the
materials meant to serve as antidotes for other toxins, toxic themselves? Some
encouraging preliminary data regarding this has been collected. First, the
microemulsions are stable and retain their structure in water, blood plasma and whole
blood. Second, even though the pluronic surfactant causes a change in blood clotting
as measured by thromboelastography (TEG), the amount in the micro emulsions is so
low that the small difference in clotting behavior from normal is not a problem,
especially when death due to overdose of a toxin is the alternative. Third, the
EKG/QRS heart function time interval returns to normal after application of the
microemulsion to an isolated guinea pig heart that had previously been caused to beat
slower by contact with a solution containing bupivacaine. The next phase of the
research will focus on making smart oil-in-water microemulsions that may have
selectivity as to what type of lipophilic molecule it will absorb. One approach to
achieving this goal will be to use oil other than ethyl butyrate, which, in addition to
serving as a reservoir, may exhibit ability to form a complex with the toxin.

SILICA NANOPARTICLES WITH ATTACHED TOXIN RECEPTORS

In principle, all three of the well known interactions - hydrophobic, hydrogen bonding
and ionic - can result in binding of a lipophilic molecule to a receptor attached to a
carrier species. For the hydrophobic situation the attached receptors may be similar to
large hydrocarbon C l2 -30 moieties silated to silica stationary phases used for
 33
chromatography. Fundamental principles of this type of toxin absorption are the basis
of quantitative structure activity relationships (QSAR) first developed by Hansch [9].
The hydrogen bonded and ionic situations may occur if the lipophilic molecule and
receptors each have polar functionalities such as amino, carbonyl, carboxyl or
hydroxyl.

Two other types of molecular interactions may also be employed to achieve binding of
a toxin to a receptor. These are 7t-7t charge transfer complexation between an electron
rich (donor) aromatic ring and one that is electron deficient (acceptor)(Figure 7); and,
insertion of a toxin molecule into the cavity of bowl-shaped cyclodextrins.

o
II
~_-- NHCR

Figure 7. Representation of 11-11 interaction between donor-acceptor aromatic rings

The 7t-7t concept has received considerable attention for many decades but has not
been applied as in this work. Early publications focused on 7t acceptors such as
trinitrobenzene and tetracyanoquinodimethane, and the experimental and
computational studies established what types of donors yielded to cr versus 7t
interactions [1O-l2]. More recently 1t-7t complex phenomena have be:en implicated in
such diverse technologies as enantiomer separation [13], supramolecular polymer
stacking [14], assembly of microelectronic circuits [15], and biomolecule analysis [12,
16-18]; and, spectroscopic methods have been developed to determine binding
constants [19-21]. Capture of a toxin molecule by cyclodextrin will be discussed after
describing the 7t-7t approach to removal of overdose toxins.

In order to determine the feasibility of binding a 7t-donor drug to a receptor attached to

a carrier particle for subsequent elimination from the body, it was necessary to
evaluate the interaction using soluble molecules. Proton NMR spectra were recorded
on solutions of donor, of acceptor and of donor-acceptor mixtures (25-50: I w/w)
following published procedures [12, 18-20]. Figure 8 is a reproduction of the aromatic
34

resonances for acceptor N-methyl-3,5-dinitrobenzamide. This molcule was evaluated

because it was to be the receptor aromatic unit for attachment to silica nanoparticles
(see below).

Figure 8. Aromatic proton NMR resonances for uncomplexed It acceptor.

The benzamide aromatic proton doublet and triplet signals appear at 8.9658 and
8.9593, and at 9.1828, 9.1755 and 9.1690, respectively, before addition ofa 1t donor.
When donor is added the peaks shift upfield the magnitude of which is dependent on
the donor concentration (Table I). Benzenesulfonamide 1t acceptors undergo the same
shifts. Note that the aromatic ring portion of the donor molecule used in Figure 8 is
 35

identical to that in bupivacaine (Figure 2). These data strongly support the proposal
that drug molecules like bupivacaine can be bound to a 7t receptor atta,ched to silica.

TABLE I. Upfield proton NMR shifts for three 11-11 donor-acceptor complexes

Donor Proton 11 (ppm) Direction

2,6 0.0874 Upfield
2,6-Dimethylaniline
4 0.0779 Upfield
2,6 0.1584 Upfield
2,6- Dimethylacetanilide
4 0.0055 Up field
2,6 0.0891 Upfield
Bupivacaine
4 0.0275 Upfield

NMR experiments were carried out following methods in the literatun: for determining
binding energies for the 7t-7t complexes formed in this study [18-20]. Plots of the
inverse of the concentration of the donor, versus the donor concentration dependent
upfield shifts of the acceptor aromatic protons, yielded straight lines (R2 = 0.995) with
slopes for the binding constants which were converted into enthalpy values. The
binding energy for 3,5-dinitrobenzamide complexed with 2,6-diIDI~thylaniline was
found to be endothermic by 365 caUmol, that for N-ethyl-2,4-dinitrobenzamide
complexed with 2,6-dimethylaniline was also endothermic by 141 caUmol, but the
binding energy for n-ethyl-2,4-dinitrobenzamide complexed with bupivacaine was
exothermic by 395 caUmol. Thus, while the upfield shifts confirm that 7t-7t complexes
form for all three systems, only the donor bupivacaine appears to do so with energy
decrease as determined by these preliminary results. The larger hydrophobic non-
aromatic portion of the drug may contribute to favorable interaction with the
hydrophobic receptor molecule. Future studies will hopefully clarify this possibility.

The next phase of experimentation involved covalent attachment to the 7t receptor to

silica nanoparticles. Previous experience with microencapsulation and molecular
surface modification of core particles of many compositions, shapes and sizes in
dispersions simplified the task [22-32]. Some of the molecules that have been
covalently tethered to core metal oxides include C60 fullerene, 18-crown six ethers and
electroluminescent organo aluminum complexes. In each example various organo
trialkoxysilanes were prepared and reacted with the cores. Similar procedures were
36

employed in the present work to attach the 1t receptors to the carrier particles. Silica
nanoparticles obtained from Nissan Co. were either reacted directly with 3,5-
dinitrobenzoyl chloride as shown in Figure 9, or with an amino group on a spacer unit
previously attached to the cores [16]. The resulting functionalized nanoparticles were
then immersed in solutions of bupivacaine in saline, blood plasma or blood for a few
minutes and the supernatant analyzed by HPLC for unbound drug. The results are
exemplified by the data shown in Figure 10. Unmodified silica shows no tendency to
bind bupivacaine while increasing concentrations of silica modified as shown in Figure
9 show excellent capacity to form a complex with the toxin. Similar results were
obtained for removing cocaine. From the average size of the particles as determined by
TEM and light scattering methods, and the concentration uptake of bupivacaine by
YL31-1 particles listed in Figure 10, it is estimated that each silica particle has 60-80 1t
receptors on the surface. Work in progress to attach the 1t acceptors to biocompatible
cores, including gold nanoparticles.

°
5:.Cl
"

THF
~ NaHC0 ~ 3,5-dinitrobenzoyl bound
3
12 nm silica + ~
O,N NO, 6
by Nissan Co. silica (12 nm)

Figure 9. Synthetic procedure for preparing silica modified with 11 receptor and TEM of product
 37

Adsorption Binding Isotherm

• Ternplated YL31-1 nanopartides, 0.1% wlv

~ 5000 Templated YL31-1 nanopartides,O.050/0wiv
::2: • Non-templated (blank) nanoparticles, 0.1 %wlv
::L
~ 4000
Q)
c:
.~ 3000
~
.g. 2000
III
-g:::J 1000
o
III 0

o 5000 10000 15000 20000 25000 30000

Free Bupivacaine (~M)

Figure 10. Bupivacaine removal from solutions by silica nanoparticles modified with 1t receptor

Comment is in order regarding the possible modes of binding of a toxin like

bupivacaine or amitriptyline to the 1t receptors attached to a carrier nanoparticle. As
stated above it is believed that each nanoparticle has numerous 1t receptors attached to
the surface. The latter molecules occupy a volume of space, arld have degrees of
conformational freedom dependent on the length of the spacer between the aromatic
ring and the surface of core particle. These steric factors may defme how the incoming
1t donor toxin molecule might bind. Both of the toxin molecules used in this study
have basic amine groups located at some distance from their aromatic rings. The
nitrogen atoms may, therefore, be able to hydrogen bond to any of the relatively acidic
silanol groups at the same time the 1t-1t interaction occurs. This process is depicted in
Figure II as "2 point interaction". The alternative, where steric hindrance inhibits
approach of the basic amine group to the core surface is also shown. Experiments are
underway to determine which binding mode is favored.
38
NO,

HN~~ •

)
NO,
o
2 point interaction
II
OH C-NH
OSi ~ /
H·····:-:N/ -
0It" \

oNO
OHO H t NH-~ '
OSi 0 NO, ~ /
I II N
~HNC I
NO,
1 point interaction
Fig. II. Surface chemical features of a nanoparticle showing covalently attached 1l: acceptor aromatic rings
complexed in two possible ways to toxin 1l: donor aromatic rings.

The last top ic included in this interim report on injectable dispersed phases for in vivo
removal of overdosed lipophilic chemicals is on use of cyclodextrin receptors. These
bowl-shaped oligomeric carbohydrates have dimensions as follows: 0.79nm high and
top openings 0.S7-0.9Snm wide. The outside and top and bottom rims of the structures
have many exposed hydroxyl groups while the conformation of the individual
carbohydrate residues provides hydrophobicity to the inner cavity. Their synthesis,
derivatization and pharmaceutical applications have been extensively reviewed [33-
38]. The most common is 13-cyclodextrin composed of seven carbohydrate residues
arranged in head-to-tail fashion. Cyclodextrins are employed to increase the solubility
of otherwise insoluble drugs for improved delivery. Penetration of even tricyclic
aromatics and polyphenyl ethers into the cavity of 13-cyclodextrin is accommodated.

As with application of the 7t-7t complex concept to fulfill the goal, evaluation of the
ability of 13-cyclodextrin to bind with bupivacaine or amitriptyline in saline, blood
plasma or blood was desireable before attaching the receptor to silica nanoparticles.
However, due to the low solubility of the receptor in aqueous media, 13-cyclodextrin
was reacted by published procedures with tosyl chloride [39] to form the mono-
tosylate at one of the primary hydroxyl groups on the bottom rim of the receptor, for
subsequent attachment to aminosilated silica nanoparticles, and with methylenedi-p-
phenyl diisocyanate [40], followed by hydrolysis, to form an amine derivative for
protonation and improved water solubility. The products were purified by
recrystallization from pyridine and purity determined by DSC melting point.
Preliminary HPLC analyses suggest that bupivacaine is bound to the receptor when the
dilute tosylate is stirred in saline with bupivacaine for two hours at a ratio of
drug:receptor of 1.3:10. It is anticipated that further experiments with a water soluble
 39

form of cyclodextrin will confmn the reality that capture of an overdosed toxin in the
cavity of biocompatible cyclodextrin may be useful.

Conclusions and Prognosis

The original goal of the research was to prepare several types of dispersed phase
nanoparticles for possible deployment as injectable antidotes for overdoses of
chemical toxins. The concepts and data in this interim report reveal that at least the
approaches using microemulsions, or nanoparticles having attached toxin receptors,
are viable candidates that should be useful for in vivo removal of overdosed chemicals
when optimized (e.g. Figure 12). The early experiments employing silica
nanoparticles as receptor carriers have served as models for next generation cores of
more biocompatible material. By combining the concepts of 1t-1t or cavity binding
modes with the properties of oils, and by use of different functional groups such as
trifluoromethyl or cyano attached to 1t receptor benzene rings, it may be possible to

Biodegradable Shell
(stability & ~.. t~,,,.f".'IhI,
Lipid Phase (microemulsion)
(rapid concentration of toxin
from blood)
Molecular Pores
(toxin passage into nanopartide)

Aqueous Pbase with enzyme incorporation

Figure 12. Conceptual structure of an optimal nanoparticle for binding and destroying toxins

obtain smart microemulsions which contain oils having different partition coefficients
for different lipophilic toxins.

Acknowledgments

The coauthors are indebted to the National Science Foundation for funding for the
initial phases of this research through its support of the Engineering Research Center at
the University of Florida, and to the New York State Science and Technology
Foundation for partial funding of the Center for Advanced Technology at Clarkson
University. Facility use in the Department of Anesthesiology and Department of
Material Science and Engineering at the University of Florida, and the Departments of
40

Chemistry at Clarkson University and the State University of New York at Potsdam is
gratefully recognized. The scientific opinions in this chapter are those of the coauthors,
not the National Science Foundation.

References

l. Dickey, F. (1949) Proc. Natl.Acad.Sci. 35,227.

2. Raman, N., Anderson, M. and Brinker, C.J. (1996) Chem. Mater. 8, 1682.
3. Makote, R. and Collinson, M. (1998) Chem. Mater. 10,2440
4. Zimmermann, C, Partch, Rand Matijevic, E. (1991) Colloids and Surfaces 57,177.
5. Katz, A and Davis, M. (2000) Nature 403, 286.
6. Lan, E., Dunn, B. and Zink, J. (2000) Chem. Mater. 12, 1874.
7. Shah, D. (1998) Microemulsions and Monolayers:Science and Technology, Marcel Dekker, New
York.
8. Friberg, S., Mortenson, M. and Neogi, P. (1985) Separation Science and Technology 20, 285.
9. Garg, R., Gupta,S., Gao, H., Babu, M., Debnath, A and Hansch, C (1999) Chem. Rev. 99, 3525.
10. Foster, R. (1969) OrganiC Charge Transfer Complexes, Academic Press, London.
Il. Buncel, E., Crampton, M., Strauss, M. and Terrier, F. (1984) Electron DefiCient Aromatic- and
Heteroaromatic- Interactions. The Chemistry of Anionic o-Complexes, Elsevier, Amsterdam.
12. Dust, J. (1992) Can. J Chem. 70, 151.
13. Stinson, S. (2001) Chem. Eng. News, Dec. 10, pg 35.
14. Brunsveld, L., Folmer, B. annd Meijer, E. (2000) MRS Bulletin, April, 49.
15. Rouhi, A(2001) Chem. Eng. News, July 30, pg 46.
16. Wassef, W., Ghobrial, N. and Agami, S. (1991) Spectrochimica Acta 47A, 623.
17. Goodnow, T., Reddington, M., Stoddart, J. and Kaifer, A (1991) J. Am. Chem. Soc. 113,4335.
18. Fesik, S., Medek, A, Hajduk, P. and Mack, J. (2000) J. Am. Chem. Soc. 122, 1241.
19. Hanna, M. and Ashbaugh, A (1964) J Phys. Chem. 68, 811.
20. Foster, R. and Fyfe, C (1965) J Chem. Soc. 1626.
2l. Neusser, H. and Krause, H. (1994) Chem. Rev. 94,1829.
22. Keklikian, L. and Partch, R. (1989) Colloids and Surfaces 41,327.
23. Partch, R. , Gangoli, S., Matijevic, E. Cai, W. and Arajs, S. (1991) J. Colloid Interface Sci. 114,
27.
24. Partch, R. (1993) Materials Technology 8, 43.
25. Huang, C-L., Partch, R. and Matijevic, E. (1995) J Colloid Interface Sci. 170,275.
26. Matijevic, E. Zhong, Q. and Partch, R (1995) Aerosol Sci. and Tech. 22,162.
27. Avella, M., Martuscelli, E., Raimo, M., Partch, R., Gangolli, S. and Pascucci, B. (1997) J Mater.
Sci. 32, 2411.
28. Dilsiz, N., Partch, R., Matijevic, E. and Sancaktar, E. (1997) J Adhesion Sci. and Tech. 11,1105.
29. Partch, R and Brown, S. (1998) J Adhesion 76, 259.
30. Randive, R., Syvinski, C and Partch, R., Gradient Functionalization of Metal Oxide Particles
with Crown Ether, Fullerene and Electroluminescent Molecules, in preparation.
31. Wang, H., Bachman, S. and Partch, R, Covalently Tethered Vinylic and Peroxido Functionalities
to Silica Particles and Polymerization Coating Thereof, in preparation.
32. Partch, R. (1997), in Materials Synthesis and Characterizations, D. Perry, Ed., Plenum Press,
New York.
33. Croft, A and Bartsch, R. (1983) Tetrahedron 39,1417.
34. Szejtli, J. (1998) Chem. Rev. 98, 1743.
35. Khan, A., Forgo, P., Stine, K. and D'Souza, V. (1998) Chem. Rev. 98,1977.
36. Loftsson, T. and Brewster, M. (1996)J Pharm. Sci. 85,1017.
37. Rajewshi, R. and Stella, V. (1996) J Pharm. Sci. 85, 1142.
38. Gadre, A, Rudiger, V., Schneider, H-J. and Connors, K. (1997) J Pharm. Sci. 86, 236.
39. Matsui, Y. and Okimoto, A. (1978) Bull. Chem. Soc. Japan 51, 3030.
40. Hargitai, T., Kaida, Y. and Okamoto, Y. (1993) J Chromatography 628, Il.
NANOSTRUCTURED MeSi01 (Me=Ag, Cu) COATINGS WITH ANTIBACTERIAL
ACTIVITY

C.C. TRAPALIS 1 *, T. YAIMAKIS2, A. KHARLAMOy3, M. KOKKORIS 4 , G.

KORDAS 1
I Institute of Materials Science. NCSR Demokritos, 153 10, Athens, Greece
2Department of Chemistry, University of Ioannina, 451 11, 10annina, Greece
3/nstitute for Problems of Materials Science, National Acadii.'my of Science, Kyiv
,Ukraine
~lnstitute of Nuclear Physics, NCSR Demokritos, 15310, Athf'ns, Greece

Abstract

In the present work, composite metal containing silicate thin coatings (Me=Ag, Cu) were
prepared on glass substrates by the sol-gel route. The preparation process included
hydrolysis and subsequent polycondensation of corresponding alkoxid~: under refluxing and
addition of soluble salt of antibacterial metal to the resulting sol. The coatings deposited by
dipping process, were thermally treated in oxidative and reductive conditions up to 500°C
for metal nanopartic1es formation. The coating structure and the nanopartic1es formation
were studied by X-ray Diffraction, AFM, UY-YIS and Heavy Ion Rutherford
Backscattering (HIRBS) Spectroscopies. The antibacterial activity against Escherichia Coli
was examined by the so-called antibacterial-drop test. The possible COITelation between the
layer interdiffusion after the thermal treatment and the antibacterial actiivity, was considered
and analyzed. The coatings exhibited a high antibacterial activity, which was enhanced with
the increase of the metal concentration and was decreased with the inc:rease of temperature
of thermal treatment and metal nanopartic1es formation.

Keywords: Nanoparticles, Antibacterial, Sol-Gel, Coatings

1. Introduction

It is widely known that materials containing antibacterial metals like: silver, copper, zink
show antibacterial properties [1-3]. More specifically, metal ions penl!trate in bacteria and
inactivate their enzymes, or can generate hydrogen peroxide, thus killing bacteria [4].
Recently, various inorganic antibacterial materials containing antibacterial metals have
been developed and some of them are already in commercial use [5]. Colorless and more
chemically durable materials especially glasses and ceramics, which slowly release metal

• Corresponding author: C.C. Trapalis

Email Address:trapalis(wims.demokritos.gr
41
Y. G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 41-49.
© 2003 Kluwer Academic Publishers.
42

ions for a long period of time, are the subject of an ongoing investigation evident in
literature [6-8], due to their wide range of applications including ceramic materials, medical
tools and appliances. Such materials can be prepared via the sol-gel route [9-13], although
they sometimes show color changes owing to the existence of metal colloid particles,
especially after heat treatment in reductive conditions.
In the present work, silicate thin coatings, containing metal nanoparticles (Me/Si0 2 ,
Me=Ag, Cu), were prepared on glass substrates by the sol-gel method. The antibacterial
activity against Escherichia Coli was subsequently examined for "as prepared", oxidized
and reduced coatings. The coatings structure and the nanoparticles formation were studied
in an attempt to establish a possible correlation between the layer interdiffusion,
nanoparticles formation and the antibacterial activity.

2. Experimental
2.1. SOL-GEL PROCESSING

Metal containing silica thin coatings (Me/Si0 2 ) were deposited by sol-gel process on
optically transparent microscope glass substrates. This process has been widely used and
almost exclusively applied for the fabrication of initially transparent coatings on glass
surface [14-18 ]. The compositions of the coatings deposited were adjusted to the formula
xMe/(lOO-x) Si0 2 were x=2.5, 5, 10 and 13mol%. For the preparation of the dipping
solutions silicon alkoxide (TEOS - Si(C2H50)4 supplied by Aldrich) was added under
stirring in a mixture of 100ml absolute ethanol were the appropriate amount of metal salt
(AgN0 3 and Cu(C2H302)2H20 - Aldrich) was previously dissolved. To this solution the
stoihiometric amount of water (molar ratio H20ITEOS=4) for hydrolysis and
polycondensation of the SHOEt) groups was added. A few drops of HN0 3 were added to
adjust the pH up to a value of 3. The resulting homogeneous colored mixture was heated at
70°C for 2h. Then, the solution was cooled down to room temperature. After the sol
preparation, metal containing gel thin coatings were deposited onto microscope glass slides
of dimensions (60 mm x 20 mm x 1.5 mm) using the dipping process. The prepared
samples were dried at room temperature and then heated at a heating rate of 5°C/min, in a
horizontal quartz tube furnace (Therma Watt) at 500°C for 30 min under a stream of
oxygen in order to remove the organic components. Finally, the samples treated at first in
an oxidative atmosphere, were additionally thermally treated under reductive atmosphere at
500°C for 2h. A stream of 200 ml/min of a H2 4%-Ar 96 % gas mixture was used for the
reduction of contained metal and formation of metallic nanoparticles. The samples obtained
after heating in reductive conditions were transparent with slight yellow (for silver) and red
(for copper) colors.

2.2 . STRUCTURE ANALYSIS

The samples were characterized by X- ray diffraction (XRD) technique using a SIEMENS
D500 diffractometer with secondary graphite monochromator and CuKa radiation.
The measurements were performed by using the following combination of slits: 1.0°/ 1.0°/
1.0° as aperture diaphragms, 0.15° as detector diaphragm and 0.15° as diffracted
beam monochromator diaphragm. The measured 20 range was scanned in steps of 0.03°
in 10 sec/step. The Ultraviolet-Visible (UV -VIS) spectroscopy was applied to thin coatings
characterization using a UV - 2100 SHIMADZU recording spectrophotometer. Atomic
 43

Force Microscopy (AFM) was applied for surface structure determination using a
Nanoscope II, Digital Instruments microscope in the tapping mode.
The samples were studied using the Heavy Ion Rutherford Backscattering Spectroscopy
(HIRBS) technique. The HIRBS measurements were performed at N.C.S.R "Demokritos",
Athens, Greece, using the 5.5 MV TNII TANDEM Accelerator. The samples were
irradiated using 8 MeV 12C3+ ions for a total accumulated charge of24 !-ICb. One part of the
samples were studied using 1.4 MeV 160 2 + ions, provided by the NEC-5SDH 1.7 MV
Tandem Pelletron Accelerator in Beirut - Lebanon. For the analysis of the spectra the well
known simulation code RUMP [19] was used.

2.3. ANTIBACTERIAL DROP - TEST

The antibacterial activity of the coatings was studied using the known antibacterial drop-
test [19]. E. Coli ATCC 25922 was used as the model microorganism. These were cultured
in the culture medium Brain Heart Infusion Agar (BIOLIFE S.R.L.) at 37°C for 18-24 h.
Cultured bacteria were added in 10 ml saline solution to reach approximately the
concentration of bacteria corresponding to 1 of MacFarland scale (10 8 colony forming
units per milliliter - CFUlml). A portion of the saline solution containing the bacteria was
diluted to 106 CFU/ml in order to run the "Drop-method" antibacterial experiments.

0.15
10CU'90Si02 -prop. 1,.4 O.lU9U75iO, U pr.p.
o,soo'c
1,2
f\ ..--O"UO'C

." 0.10
-H,.soo'c
1\ -H" seo'c
I\
1••
u
c:
co
gl..... - II ... _- Ilan substrate
c
-e0
l 0,'
I \
«'"
.c
0.05
.
0
.a
0.8
I \
" n,4 I
J \
'
••2 ~'/ \ "-
~--:.
....
0.00
300 400 600 700 800 0,0
,.0 41. 500 800 rao
WIMlIa1gIh [m1 Wavelength [om)

Figure 1. UV-VIS spectra of copper containing as Figure 2. UV -VIS spectra of silver containing as
prepared coalings and after thermal treatment in prepared coatings and after thennal treatment in oxidative
oxidative and reductive atmosphere and reductive atmosphere

Three groups of metal containing (Ag and Cu) coatings were studied. Each group
represented coated microscope glasses with differences in their thermal treatment namely,
"as prepared", oxidized and reduced. Each group of samples contained four Me/Si0 2 coated
glasses. Each one of the samples corresponded to different sampling time. For each group
44
of samples one Si0 2 coated microscope glass was used as a blank. The samples were
sterilized with 70% ethanol and placed into sterilized Petri-dishes of 90 mm. Then 100/JI
saline solution with E. Coli concentration of 106 bacteria/ml were added dropwise by an
Eppendorf Reference variabel onto the surface of each Cu/Si02 composite coating. The
samples were left at room temperature for 6, 12, 18 and 24h. After each time period the
bacteria containing drops were washed from the glass surfaces using 5 ml Phosphate Buffer
Solution (PBS) in the sterilized Petri-dish. Then 10 /JI of each solution was dispersed on a
Brain Heart Infusion Agar culture medium. The number of surviving bacteria on the Petri-
dishes were counted after incubation for 24 h at 37°C.

3. Results and Discussion

The coatings obtained were transparent after thermal treatment up to 500°C. Figure 1
presents the UV -VIS spectra of the glass substrate and of copper containing samples treated
at different conditions. It was observed that the glass substrate, the "as prepared" coating as
well as the coating treated in oxidative conditions does not show any significant absorption.
After thermal treatment in reductive atmosphere an absorption band with maximum at
571nm appears, which corresponds to the formation of copper nanoparticles in the silicate
coating. Figure 2 presents the UV-VIS spectra of silver containing samples treated at
different conditions. It is evident that only after thermal treatment in the reductive
atmosphere does an absorption band appear with a maximum at 410nm. This band
corresponds to the formation of silver nanoparticIes in the coating. The above strong
absorptions appear due to the surface plasmon resonance (SPR) of the Cu and Ag metal
nanoparticIes. Similar optical property of Me/Si02 nanocomposites prepared by several
techniques like ion implantation [21], melting [22] and sol-gel processing [23-26] were
reported.
"anoScop. Tapping AFM
SO-illl &lze 1.008 UN
S.tpoi nt 1.803 U
:::!.~a!: .... pl.. O.8~!~ Hz Figure 3. AFM
micrograph,
showing the
formation of large
(10·30 nm) silver
nanoparticles after
thermal treatment
under reductive
0.4
conditions.
UM o.zoo u...... t ...
1.0 100.000 n ......diu

The formation of silver nanoparticles was also observed using the AFM technique. Figure 3
depict a typical three-dimensional representation (Ixl /Jm surface plot) of the film
thermally treated under reductive conditions at 500°C. The film presents a rough surface
texture consisted of isolated particles with large spacing between them. The morphology
and surface characteristic of the film can be described as a discontiguous coating with small
isolated islands of colloidal particles and aggregates. These nanoparticles seems to have an
 45

average diameter of 10-30 nm. Their height threshold was calculated at about 3.30 nm and
their height distribution is fairly narrow. From this fact it is possible to conclude that the
nanoparticles are embedded within the Si0 2 coating.
The nanoparticle formation was also confirmed using XRD technique. Figure 4 shows the
XRD patterns of the samples after heating under reductive conditions. A broad background
contribution from the amorphous silica and glass substrate is observed. Curve 4(a) presents
the XRD pattern of single CU/Si0 2 layer on glass substrate. The peak intensities of the
crystalline structure of Cu nanoparticles is very low due to the presence of the small
volume fracture of the Cu and the limited coating thickness. Howev~:r, clear broad peaks
characteristic of the cubic Cu metal crystals are observed in the graph 4(b) of the same
figure, which presents the XRD pattern of three layer coating.

Figure 4. XRD
patterns of
10Cul90SiO, CuiSiO, (a) single
Cu(lll)
and (b) three layer
coatings after
thennal treatment
m reductive
atmosphere,

35 40 45 50 55

2e

A detailed study of coatings "as prepared" and thermally treated in oxidative and reductive
conditions has been performed using Heavy Ions Rutherford Bacscatering Spectroscopy
(HIRBS). Figure 5 shows the experimental as well as the simulated HIRBS spectra of the
coatings on glass substrates. From the spectra in the graph 5(a), it was found that the thin
coating in the case of the "as prepared" coatings has a thickness of ,,220 nm and a large
surface density of organic constituents. The analysis of the spectra in the second case
(graph 5(b» demonstrated that the organic constituents are removed after heating under
oxidative conditions, thus reducing the coating's thickness by -40%. The glass/coating
interface remains sharp in both cases, indicating that the diffusion of eu ions is practically
negligible. However, in the third case (graph 5 (c» a strong diffusion pattern is evident. The
penetration of Cu inside the glass in this case is of the order of -100 nm and there are
strong indications [18] that the diffusion is enhanced when the sample is heated at higher
temperatures, thus considerably affecting the antibacterial activity of the sample.
Nevertheless a significant amount of copper remains close to the surface of the sample
heated under reductive conditions thus retaining the antibacterial properties of the coating.
46
This copper is in the form of metal nanoparticles as it was shown by UV -VIS and XRD
measurements.
A typical HIRBS spectrum of silver containing samples is presented in fig. 6, using 1.4
MeV 160 2+ ions, provided by the NEC-5SDH 1.7 MY Tandem Pelletron Accelerator. The
two curves presented correspond to two similar thin coatings after heating under oxidative
and reductive conditions at 500°C. The long tail in the low energy part of the spectra
corresponds to the diffusion of the Ag ions inside the glass, up to a depth of 0.6 J.lm, as seen
by the incoming beam. The most interesting experimental result however, is the non-
uniformity of the Ag concentration inside the film. Under oxidative heating, although the
diffusion effect is equally strong, the remaining silver ions in the film seem to have
acquired a trend in their concentration towards the surface, while under reductive
conditions there seems to be an opposite trend, towards greater depths, indicating that at
higher temperatures and longer treatment times the observed diffusion will be enhanced.

!II
;
o
o
b w"
.,'
~'G we
2Il
~
%teredo ·,'4
3Db!
~
I' ~

3D
.
II
=
=
UlDOO

.IID
,.'"
II']h" A~ 0'11>'

rt"biw~
d'd ....
3D '!II 2!IJ 4IIl

Figure 5. HIRBS spectrum, showing the Cu Figure 6. HIRBS spectrum of silver containing
distribution of samples studied: Ca) as prepared, (b) (\ 3Ag87SiO,) thin coatings, after thermal treatment
after heating under oxidative and (c) after heating at at 500°C under oxidative and reductive conditions,
reductive conditions, using 12C3+ ions at 8 MeV. using 160 2+ ions at 1.4 MeV.

The antibacterial properties of the coated glasses with copper and silver containing silicate
coatings were recorded using the antibacterial drop-test [19]. Figure 7 represents the results
of the antibacterial test of 13Ag87Si02 composition against Escherichia Coli bacteria. It is
recorded that after 6h only the "as prepared" coating shows a reduced number of developed
bacterial colonies in comparison to the other three coatings. It is clearly shown that after
12h, the "as prepared" coating eliminates completely the development of the bacteria. The
coating which has undergone thermal treatment at oxidative conditions allows the
development of fewer bacterial colonies. However, the number of bacterial colonies of the
coating which have undergone additional heating under reductive conditions is slightly
 47

greater compared to that of the powders treated under oxidative ones. After 24 h a
complete elimination of bacteria in all coatings except the blank coating was established.
The reduced antibacterial activity of the coatings after heating in reductive atmosphere
could be attributed to the reduced surface area of the total metal contained in the coating,
due to formation of metal nanoparticles, as well as to the partially diffusion of the metal
into the glass substrate. This diffusion of the metal in the glass substrate is negligible in the
case of oxidized samples. In the case of the reduced ones the diffusion is more evident as it
can been seen from the HIRBS results (Figure 5(c) and 6). This is probably related to the
longer time of thermal treatment needed for reduction of the metal in the coatings (2h) in
comparison to the time used for oxidation (0.5h). Nevertheless, the metal remaining on the
glass surface, even in the form of nanoparticles after thermal treatment at reductive
conditions, shows sufficiently antibacterial properties for complete antibacterial elimination
after 24 h treatment.
Figure 8 shows the antibacterial activity of composite coatings with various copper
contents. The compositions shown are xCu/(lOO-x)Si0 2, x=2.5, 5 and 10 mol % . Coatings
with the above metal concentrations were heated in oxidative atmosphere at 500°C for 0.5
h. As it can be seen the antibacterial activity of the coatings increases with the increase of
metal concentration in the coatings. Even the lower concentration of 2.5mol% copper in the
Si02 coatings shows strong antibacterial activity eliminating a significant amount of
bacteria deposited on the coating after 24 h of treatment.

....
>
to

as. • Rd.DllRim
."'-J'
. . . I. I
~

.."
'E
• . 0dcIIaI FIm
III
' ' ' . 1. .
.
-.: 06 • /leAepndFim . •. BII1<
ti
III

....E
'0 04
~

...
Q2

.
z"
>
'".!!. ao . .,-t
II<:
12 18 6 12 18

Time 1.11

Figure 7. Relative number of Escherichia Coli
survived after antibacterial drop-test on the as prepared
l3Ag87SiO, sol-gel thin coatings and after thermal
treatment at 500°C under oxidative and reductive
conditions.
Figure 8. Relative number of Escherichia Coli
survived after the antibacterial drop-test on the
xCuI(IOO-x)SiO, coatings thermally treated under
oxidative conditions contained (a) 2.Smol%, (b)
5mol% and (c) 10 mol% copper.

It is evident that the best antibacterial performance can be attributed to the "as prepared"
sol-gel film, while it gradually decreases for the heated samples under oxidative and
reductive conditions. This is due to the lack of diffusion of metal ions inside the glass
substrate in the former case, as well as, to the ability of metal cations [27] to react with the
48

proteins of the outer wall of microorganisms deposited, thus inhibiting their growth [28].
The issues of the diffusion and mobility mechanisms of metal ions and the antibacterial
activity addressed in the present work will be also the subject of detailed future studies.

4. Conclusions

In the present work it was proved that the Me/Si0 2 Me=Ag, eu sol-gel thin coatings show
high antibacterial activity. The antibacterial capability of fresh "as prepared" coatings has
shown to be more significant than the ones of coatings heated under oxidative and reductive
atmosphere. The coatings thermally treated under oxidative conditions are more active
against bacteria compared to the coatings treated under reductive ones. Nevertheless, these
coatings even at low metal concentrations exhibited sufficient activity for Escherichia coli
bacteria elimination after 24h treatment. It was established that the antibacterial activity
increases with the increase of the metal concentration and decreases with increase of the
temperature of thermal treatment.

Acknowledgments
The financial support by GSRT - Ministry of Development under contract N° PAVET-758
is acknowledged.

References
I. Modak S. M., Fox e. L. J., (1973) Biochem Phannacol 22,2391.
2. Berger T J., Spadaro J. A., Chapin S. E., Becker R 0., (1976) Agents Chemother. 9,357.
3. Williams R. L., Doherty P. J., Vince D. G., Grashoff G. J., Williams D. F., (1989) Crit. Rev.
Biocompat. 5,221.
4. Yoshinari T, Uchida M., (1994)J. Kor. Cerarn. Month. Feb, 119.
5. Oloffs A., Grosse-Siestrup e., Bisson S., Rinck M., Rudolph R., Gross U., (1994) Biornaterials 15,
753.
6. Kawashita M., Tsuneyama S., Miyaji F, Kokubo T, Kozuka H., Yamamoto K., (2000) Biornaterials
21,393.
7. Mitrikas G., Trapalis e. e., Deligiannakis Y., Boukos N. and Kordas G., (1998) J. of Sol-Gel Sci. and
Tech. 13, 1/3, 503.
8. Aslanoglou X., Assirnakopoulos P A., Trapalis e. e., Kordas G., Karakassides M. A., Pilakouta M.,
(1996) Nucl Instr And Meth. B 118,630.
9. Innocenzi P., Kozuka H., (1994) J. Sol-Gel Sci. Tech. 3, 229.
10. Hinsch A.• Zastrow A., J. Non-Cryst. Solids, (\992) 147-148, 579.
II. Breitscheidel B., Zieder J .• Schubert U., (\991) Chern. Mater. 3, 559.
12. Brusilovsky D., Eyal M., Reisfekd R., (1988) Chern. Phys. Lett. 9, 203.
13. Tarnahashi I., YoshidaM., Manabe Y., T. Tohda, (1995)J. Mater. Res. 10,362.
14. Trapalis e.e., Karakassides M.A., Kordas G., Aslanoglou X., (1995) Materials Letter.;, 25, 265.
15. Samuneva B.I., Trapalis e.e., Kozhukharov VS., Kranold R, (1993) Journal of Materials Science 29,
2353.
16. Trapalis e.e., Kozhukharov V.S., Samuneva B.I., Stefanov P., (1993) Journal of Materials Science,
28,1276.
17. Kozhukharov V.S., Trapalis e.e., Samuneva B.I., (1993) Journal of Materials Science, 28,1283.
18. Kokkoris M., Trapalis e.e., Kossionidis S., Vlastou R., Nsouli B., Grotzschel R, Spartalis S., Kordas
G. and Paradelis Th., (2002) Nucl Instr. and Meth. B, 188, no. I, 67.
19. Doolittle L. R., (1985) Nucl Instr. And Meth. B 9, 344.
20. Zhao Z., Hasebe K., Sakagarni Y., Osaka T, (1997) Bul Chern. Soc. Jpn. 1631.
21. Mazoldi P., Arnold G.W., Battaglin G., Bertoncello R and Gonella F., (1994) Nucl Instr. Meth. B,
91,478.
22. Kundu D., Honna I., Osawa T, and Korniyarna H., (1944) J. Cerarn Soc., 77, 1110.
23. Menning M., Schmitt M., Kutsch B., Schmidt H., (1994) Sol-Gel Optics ITI, SPIE 2288, 120.
24. Doremus R, Kao S, andGarciaR., (1992) Appl Opt., 31, 5773.
 49

25. Nogami M., Abe Y., and Nakamura A., (1995) J. Mater. Res. 10,2648.
26. De G, Epifani M., Licciulli A., (1996) J. Non-Crystalline Solids, 201, 250.
27. Maeda H., Iwasaki M., Yasumori A. and Yamane M., (1990)J. Non-Cryst. Solids 121,61.
28. Ghandour W., Hubbard A., Deinstung J., Hughes M, and Poole R., (1988) Appl. Microbiol.
Biotechnol., 28, 559.
LOW FRICTION WEAR RESISTANT SPUTTERED CARBON COATINGS
FOR BIOMEDICAL APPLICATIONS.

S. K. TAYLOR, A.H.S. JONES, D. MERCS, D.G. TEER, M. ELLOY.

Teer Coatings Ltd.,
290 Hartlebury Trading Estate,
Hartlebury, Worcestershire DY10 4JB. u.K.

ABSTRACT

Sputtered carbon coatings have been reported to have excellent tribological properties
(low friction coefficient and wear rate) and therefore have the potential to be used as
hard, solid lubricant, wear resistant coatings. During a research program aimed at
reducing the wear of UHMWPE hip joint prostheses by coating the metal femoral head,
a new type of carbon coating was developed which now has the trade name
Graphit-iCTM. This coating gave exceptional friction and wear n:sults in simple
laboratory tests against coated and uncoated metal counterfaces, but not against
UHMWPE as originally intended.

Research was redirected towards the production of Graphit-iC coated metal-on-metal

THR, and simple configuration laboratory tests predicted a prosthetic lifetime in excess
of 50 years. CoCrMo prostheses were subsequently coated and performed well in
pendulum tests conducted at the University of Leeds. Further tests on a physiological
hip simulator gave mixed results with coating failure detected on some pairs, whilst one
pair showed no detectable wear after considerable testing. Further testillg is required to
fully establish the cause of the failures. Recent work has studied characteristics of the
Graphit-iC coating under a wider range of deposition parameters with a view to
producing increased toughness for softer substrates, or graded coatings which reach
steady state friction values more quickly, or provide different surface eharacteristics to
bulk characteristics. Biocompatibility tests indicate that the Graphit-iC coating may be
a suitable surface for hip joint, and other, implants used within the body.

51
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings jor Biomedical and Sensor Applications, 51-62.
© 2003 Kluwer Academic Publishers.
52
1. Introduction

The most widely used artificial hip joints consist of a metal femoral ball, and an
acetabular cup made from ultra-high molecular weight polyethylene (UHMWPE). Much
attention is currently focussed on the reduction of polymer wear debris, as the worn
particles can collect in the tissue surrounding the joint, causing osteolysis (bone
resorption) and subsequent joint failure by loosening [1,2]. The polymer wear is often
accelerated by third body damage to the metal ball resulting in increased abrasive wear.
Various surface treatments have been investigated with the aim of protecting the smooth
surface of the ball.

An initial programme of work focussed on the potential use of Me:C OLC coatings
deposited by a plasma assisted CVO technique [3] to protect the femoral head.
However, the low friction coefficients obtained between these coatings and metallic
counterfaces were not reproduced when OLC coated CoCrMo was tested against
UHMWPE in water. Friction coefficients obtained were slightly higher than for the
uncoated CoCrMo against UHMWPE in water, and UHMWPE wear was not
significantly reduced [4].

Effort was then concentrated on the production of new carbon-based coatings with
improved friction and wear characteristics, and a coating, now known as Graphit-iC was
developed by sputtering carbon targets in a pure argon atmosphere. Benefits of the new
coating included a greatly increased load bearing capacity and much lower friction
characteristics against metal counterfaces [5,6] compared to previous OLC's.
Performance characteristics when tested against UHMWPE were not greatly improved
however, and further testing of a number of different OLC coatings, obtained from
different laboratories around the world, suggested that a carbon based coating was
unlikely to lower the wear or friction between CoCrMo IUHMWPE couples [7]. The
authors acknowledge however, that Lappalainen et al have since reported a reduction in
friction by using polished amorphous diamond coatings [8].

The excellent performance of the Graphit-iC coatings against metallic counterfaces, in

air and water, suggested that they might provide increased life for metal-on-metal hip
joints, such as the McKee type, rather than the Charnley type (UHMWPE against metal
or ceramic) protheses. Wear rates for these are in the range 2-20 J.lm per year, compared
to 100-200 J.lm per year for UHMWPE against metal or ceramic [9,10]. Reported wear
rates [\1] for CoCrMo tested against CoCrMo in serum of 2x 10- 15 m 3 lNm, were
compared to the specific wear rate of Graphit-iC in ambient conditions of 3x 10-
17 m31Nm, and it appeared that considerably large increases in lifetime might be
achievable. Coatings were applied to metal-on-metal hip prostheses manufactured by
Avatar Medical Ltd. (now Zynergy Orthopaedics Ltd.) for testing at the University of
Leeds. Short-term tests in a pendulum-type simulator were conducted, followed by long
term tests in a physiological simulator. Further study and optimisation of the Graphit-iC
coating characteristics, for this and other biomedical applications, continues.
 53

2. Experimental details

2.1 COATING.

Coatings of 2.51lm were deposited in a Teer UDP650, four magnetron, coating system.
The deposition details have been published previously, [5,6,12]. The initial coating
development results were obtained from coatings deposited onto M42 grade tool steel.
For work focussed on metal-to-metal prostheses, coatings were deposited onto CoCrMo
alloy (ASTM F799-96, Co-27Cr-5.5Mo-0.06C).

Coatings of 21lm thickness were produced later to re-examine the influence of hardness
on performance at the start of testing, i.e. initial friction coefficient and running in time.
These coatings were produced in a Teer UDP650 four magnetron sputter ion plating
system, at different applied bias voltages to obtain the different hardness values.

2.2 TESTING.

Initial tests, during coating development, were performed at Teer Coatings Ltd. (TCL)
on coated M42 tool steel samples, using a uni-directional pin-on-disc tester. Normal
loads of up to 100 N were applied to a 5 mm diameter WC-6%Co ball sliding at a speed
of 200 mms· l , in ambient air. Wear was measured by ball cratering [13,14].

Graphit-iC coatings aimed at metal-on-metal couples were tested in a reciprocating

motion. Uncoated 5 mm diameter, hemispherically ended, CoCrMo pins were rubbed
against Graphit-iC coated, and uncoated, flat CoCrMo samples. Normal loads of 40 N
and 5 N were used respectively, in de-ionised water or laboratory atmosphere (ambient,
20°C, 40 - 60% RH) conditions. Friction was continuously recorded during the tests.

Reciprocating friction tests were performed at the University of Leeds to record friction
between Graphit-iC coated CoCrMo (low carbon content) pins and flat plates. The pins
used were 12 mm diameter, 100 mm radius of curvature of tip with a 3 mm diameter
flat. 20 N normal load was used with a 30 mm track length, 0.5 Hz or 30 mms· l , with a
25% bovine serum lubricant.

Metal-on-metal THRs were manufactured by Avatar Medical Ltd. from wrought

CoCrMo alloy. These had a head diameter of 28 mm and radial clearances varied from
50 Ilm to 550 Ilm. A 53 mm diameter pair with a radial clearance of 50 Ilm was also
manufactured. These were then coated with the Graphit-iC coating at Teer Coatings
Ltd, and supplied for test.

The hip joint couples were used first in a study of friction. Frictional torque was
recorded for the different pairings of head and cup at a constant load of 500 N in a
driven pendulum friction rig. Each pairing was tested in the presence of water, Ringer's
solution, 25% bovine calf serum and with no lubricant (dry). A CoCrMo Metasul
54

pairing was also used as a comparison, plus metallUHMWPE and alumina/alumina

pairs.

The same head and cup pairings were then tested on a physiological simulator which
consisted of two articulating stations and one dynamically loaded creep station. The
prostheses were mounted in the anatomical position and experienced three directional
forces (vertical, anterior-posterior and medial-lateral) and three degrees of motion
(flexion-extension, abduction-adduction and rotational) as defined by Paul [15]. All
tests were performed in 25% bovine serum (0.1 % sodium azide).

Tests performed later, on coatings of different hardness, used a pin-on-disc tester

at Teer Coatings Ltd. Normal loads of 80 N were applied to a 5 mm diameter WC-
6%Co ball sliding at a speed of 200 mms'], in ambient air. Coated substrates were M42
tool steel.

Biocompatibility tests directed at the hip joint application were performed at the
University of Newcastle Medical School and these results have been published
elsewhere [16]. Biocompatibility tests to study the influence of Graphit-iC on cell
adsorption and growth, for the potential use of the Graphit-ic coating in other
applications, are ongoing at the University of Brighton, UK [17].

3. Results and discussion

The carbon coating, Graphit-iC, was originally developed to give improved

performance over conventional hydrogenated DLC in wear applications. This was
achieved using conventional wear tests in ambient air, oil and water against hard
counterfaces such as WC or Cr-steel balls. The coatings were found to have good
friction characteristics and excellent wear resistance in these situations, as shown in
figure 1, (a) and (b). The coatings had a plastic hardness (analogous to Vickers) of
around 1600 kg mm- 2 at a load of 50 mN, and excellent adhesion provided by a Cr
interlayer, shown by the Rockwell indent in 1 (c). The deposition technique utilising
multiple magnetron sputtering sources allowed the films to be deposited as pure carbon,
amorphous carbon/chromium or as carbon/chromium multilayers. Cross-sectional TEM
images of coatings containing chromium, show multilayers or structures with a
periodicity ofless than lOnm [7,20].
 55

.,----------_._----

(a)

Figure t. Graphit.jC wear characteristics.

(a) Friction trace during an 80 N ul1i·diroctional pill-<)U-disc lest pertormed in air.
(b) Ball crater on the w~.ar track of the pin-on-disc test shO\\'n in (a)
(c) Rockwell C indent used to indicate coating adheskm

A series oftests performed at Teer Coatings Ltd, using Graphit-iC coated CoCrMo pins
and flats, sliding in water at 3 mm 5. 1, gave very promising results_ A load of 40 N on a
5mm diameter hemisphericaily ended pin was chosen to exert the highest possible
contact pressure without causing severe plastic deformation of the CoCrMo substrate
material. This provided a mean HertLian contact pressure of 1.4 GPa and did cause a
little plastic deformation as can be seen in figure 2a. The wear scar on the coated pin
(figure 2b) was measured every 10,000 cycles which allowed the progressive wear of
the cooted pin to be measured.

Figure 2: Wear on the Graphit·iC coated CoCrMo pin and plate during testing in de-ionised water.

(a) Ball crater on the flat wear track after 50,000 cycles.

(b) Wear scar on the coated pin after 50,000 cycles.

56

After 40,000 cycles the coating on the pin had worn through to the Cr interlayer, and the
CoCrMo was exposed after 50,000 cycles as can be seen above. The friction coefficient
during the tests was stable at - 0.03 and the steady state wear rate of the coated pin was
5 x 10. 18 m3 N·lm· l . The wear of the coated flat was not measurable but there is nothing
to suggest that it would be any worse than that of the pin, so a combined wear rate of
1 x 10. 17 m3 N·lm- I could be assumed.

Tests using uncoated CoCrMo pins and flats at a reduced load of 5 N (mean Hertzian
contact pressure of 0.7 GPa) in water produced a friction coefficient of 0.4, but the load
was high enough to deform the CoCrMo and make wear quantification by geometric
measurement very difficult. Streicher et al [11] reported that the best wear and friction
results were found with high carbon content (0.2 - 0.3 wt.% C) fine grain structured
(wrought as opposed to cast) CoCrMo alloys running in serum and providing average
friction coefficients of - 0.2 (0.1 < /l < 0.45) and specific wear rates of -2 x 10- 15 m 3 N-
1m- I • They used Hertzian contact pressures of -1 GPa and 10 mm S-I sliding speed -
perhaps explaining the lower coefficients of friction, but they also reported that the
lower carbon content (0.05-0.08 wt.% C) material did not show significant differences
in friction, and found very little difference in friction and wear when using serum or
water.

This suggests an improvement in wear of 400 times and a reduction in friction to one
quarter when comparing the Graphit-iC to the CoCrMo directly. If the CoCrMo heads
and cups wear from 2 - 20 /lm per year [10] the Graphit-iC coated heads and cups could
wear as little as 0.05 /lm per year and give potential lifetimes of up to 50 years. This
would provide a considerable reduction in wear debris which would be mainly carbon
and therefore inherently safe within the human body. A reduction in friction would also
reduce the possibility of component loosening by reducing the frictional torque on the
component fixturing down to one quarter.

Friction measurement tests performed at the University of Leeds, to determine the

performance of Graphit-iC in serum, recorded a friction coefficient of 0.25, at a contact
pressure of 2.83 MPa. (A load of 20 N was applied to Graphit-iC coated pins with a 3
mm diameter flat that provided conformal contact). Friction was higher than that found
for coated hemispherical pins in water (/l - 0.03, 1.41 GPa mean Hertzian contact) and
was thought to be due to the serum and the conformal contact. Tests in serum,
performed at TCL with coated hemispherical 5 mm diameter pins (non-conformal
contact) at loads of 40 N and 4 N produced friction coefficients of 0.07 and 0.12 at
mean Hertzian contact pressures of 1.41 GPa and 0.65 GPa respectively. This
demonstrates a variation of friction with contact pressure, as has been shown for solid
lubricants [18], and also a difference between water and serum. It was therefore
concluded that it would be necessary to measure the friction of the heads and cups in
serum to determine the exact level of friction.

Tests were performed on a series of heads and cups coated with Graphit-iC. The pairs
had been manufactured to have varied radial clearances (50, 100, 150,300 and 550 /lm),
to provide different contact geometries. A larger 53mm diameter pairing with 50 /lm
radial clearance was also included. Figure 3 below shows the results when the two
 57

extreme clearances of the 28 mm diameter pairs, and a 53 mm diameter pair, were used
to record friction in the Leeds driven pendulum test, and also compared to other
material pairings.

I
I

(a) (b)

Figure 3. Friction values in the driven pendulum test rig of (a) Graphit-iC coated and a commercial uncoated
CoCrMo hip in different lubricants, and (b) Graphit-iC with different radial clearancl:5 and other material
pairings in serum.

It can be seen in 3a that the Graphit-iC coated pairs produced lower fril;tion values in all
fluids. It also shows that the 550 J.1m clearance (highest contact pressure) gave the
lowest friction in all fluids except the serum. The serum appears to contradict the trend
showing lowest friction in the 28 mm diameter 50 Ilm clearance pair. It also benefits
the CoCrMo pair, providing more lubrication than the other fluids. The difference in
friction between these tests and the pin-on-flat tests suggest that the contact pressure
and geometry should be matched to that found in heads and cups, to enable different
material pairings to be compared in simple bench tests.

When the same hip joint couples were tested in the physiological simulator the coating
performance was not as expected. The tests were stopped prematurely as there was
evidence of failure in all but one pair, indicated by the serum turning dark in colour due
to small particles of coating. These failures occurred after differing numbers of cycles.
There was no way of determining the exact moment of failure during the test, as the
tests were checked after approximately every 100,000 cycles. The duration up to the
last check was the only estimate that could be made of the coating durability, and these
values are shown in table 1.
58
Table l. Lifetimes ofGraphit-iC coated heads and cups in simulator tests

Radial Clearance Period before failure Final condition of

(IJ.m) (xlOO,OOO cycles) coating.
50 0 Coating removed
100 4 Coating removed
150 <1.92(no failure) Intact
300 1 Coating removed
550 5 Coating removed

The heads and cups with 150 J.lm and 300 J.lm clearance were run together and both
stopped after 192,000 cycles as the 300J.lm clearance pair appeared to have failed. The
failed head was ball cratered at the edge of the worn area and the fractured coating can
be seen in Figure 4a. The 150 J.lm clearance head showed no signs of wear, without
even the slightest scratch. This head was ball cratered, figure 4b, which allowed the
coating thickness to be measured: no reduction in thickness was found.

Figure 4. Ball craters on (a) the failed coating of the 300 I'm clearance head, and (b) the lOtact coating ofthe
150 Ilm clearance head, both after 192,000 cycles.

On the failed head, delamination of the coating near to the substrate interface (shown in
figure 4a) suggested an adhesion failure rather than progressive wear of the coating, as
was seen on simple configuration tribological testing (figure 2a). However, the
adhesion of the coatings was not found to be inferior as determined by a standard
scratch test performed on an area of coating next to the failed area on the 300 J.lm
clearance head.

This suggested a type of failure not previously observed for the Graphit-iC coating. The
mean Hertzian contact pressures for the heads and cups did not appear to be excessive
for the coating. The standard pin-on-disc tests, at TeL, using coated M42 steel and WC
balls produce Hertzian contact pressures in excess of 3 GPa, whereas the coated
CoCrMo heads and cups produce a maximum contact pressure of 230 MPa on the 550
 59

)lm clearance pair, and 47 MPa for the 50 )lm clearance pair (calculated using the baIl-
on-plane model with an equivalent radius [19]. If there was misalignment of the head
and cup it may have been possible for the contact stress to exceed these values, and
those permitted by the coating, resulting in the type of failure shown. The coating on
CoCrMo would not be as compliant as UHMWPE or uncoated CoCrMo, where some
misalignment could be tolerated via bulk deformation.

Another possible cause of failure could have been due to the repetitive cyclic loading
produced in the simulator, resulting in a type of fatigue failure near to the
coating/substrate interface. However, the cyclic nature ofthe load was not thought to be
dissimilar to that found during a high load pin-on-disc test, where the coating and
substrate experience elastic deformation as the pin moves around the wear track. The
short duration of some of the tests was not thought to be long enough for fatigue failures
to occur.

The nature of the failures made it impossible to draw any conclusions. on the effect of
radial clearance in the heads and cups tested, as failure was not c:onsidered to be
dependant on this parameter.

These results highlight the differences between accelerated bench tes.ts and simulator
testing. Further development of the Graphit-iC coating will require longer duration
wear testing with geometries closer to those found in hip joints, to try to induce the type
of failure shown here. The simulator testing procedure used, designed for testing
Charnley type hip joints, did not allow the performance of the coatl!d bearing to be
monitored, which made detailed analysis after the failure more difficult. It is difficult to
measure wear following catastrophic coating failure. Once the mode of failure is
understood steps can be taken to improve the performance of the coatings in the metal-
on-metal hip prostheses application. Residual stresses within these coatings are not
typically high. The presence of chromium within the coatings reduces internal stresses
and enhances the toughness of the coating, demonstated by the high load bearing
capability of these coatings on flat substrates. Stresses within the coating are
compressive and hence convex surfaces do not cause problems.

Further testing on simulators designed for metal-on-metal hip joints is the focus for the
next stages of the work. Currently, further study of the Graphit-iC coating is aimed at
improving tribological performance further at the very start of testing. The Graphit-iC
coating exhibits very low steady state friction values ()l = 0.04 to 0.08) and low wear
rates at loads up to lOON, but friction is slightly higher at the start of testing ()l=0.1 to
0.2 typically). This value falls quickly over a short running in period of a few minutes.
Further optimisation of the coating is aimed at reducing the initial friction coefficient
further and eliminating the short running in period. Figure 5 below shows the variation
in running in time when coatings of different hardness were tested against the same
counterface material.
60

80 N (22 ·C, 32 %RH)

2000

1800
Hplast (kg/mm')
1600
+ 40 N (22 ·C, 32 %RH)

1400

1200

1000

800
o ~ ~ ~ D 1~1~1~1~

Running in duration (s)

Figure 5: Effect of the graphit-iC hardness on the running in duration

Pin on Disc testing at 80 and 40N

Figure 5 shows that the softer coatings reach a steady state friction value more quickly.
However, past work [19] shows that softer coatings have higher wear rates. Current
work is aimed at the production of a coating with a graded structure to tailor the surface
properties to the application further, whilst retaining the low wear rate of the Graphit-
iC.

Biocompatibility tests at the at the University of Newcastle Medical School showed the
Graphit-iC coating to compare very well with the currently used and accepted prosthetic
materials such as CoCrMo, titanium and 316L stainless steel in all of the tests used (15).

4. CONCLUSIONS
Simple laboratory pin-on-flat wear tests indicated a higher coefficient of friction for the
carbon coating Graphit-iC sliding against UHMWPE in water than for uncoated
CoCrMo.
 61

Pin-on-flat tests have shown that Graphit-iC/Graphit-iC bearing contacts can produce
friction coefficients of 0.03 and a wear rate of 5 x 10.18 m3 N- I m- I when tested in water
compared to a coefficient of 0.2-0.4 and a wear rate of 2 x 10- 15 m3 N- I m- I for
CoCrMo/CoCrMo (10). When tested in serum the friction coefficient was increased to
0.07.

Graphit-iC coated CoCrMo heads and cups produced lower coefficients of friction than
metal-on-metal, metallUHMWPE and ceramic/ceramic material pairings when tested in
serum.

Physiological simulator tests produced catastrophic coating failure in some hip joint
pairs, but one pair survived and showed no signs of wear whatsoever.

The wear resistance and low friction still provide excellent potential for the carbon
coatings in the application of hip prostheses, but new simple configuratilon test methods
are required to further improve the coating performance before full term simulator tests
can be utilised to the full.

5. Acknowledgements
The authors would like to thank the DTI, DoH and EPSRC for support as part of LINK
Surface Engineering Projects, Prof. Santu Datta and Redha Touaitia at the University of
Northumbria for structural analysis of coatings, Prof. John Fisher, Paul Firkins, Devon
Derby and Hanna Marrs of the University of Leeds for testing, John Egan and 3M
Healthcare for materials supply and wear testing, and Avatar Medical Ltd. for providing
the metal-on-metal hips.

6. References

I. Wang, A, Stark, C & Dumbleton, IH. Mechanistic and morphological origins of ultra high molecular
weight polyetbylene wear debris in total joint replacement prostbeses. In Advances in Medical
Tribology, MEP Ltd. 1998, pp.l-15.
2. Willert, H.G. and Buchhom, G.H., Particle disease due to wear ofUHMWPE. Biological, material
and mechanical considerations of joint replacements (Ed. B.F. Morrey), 1993,pp.87-102 (Raven Press,
New York)
3. Monaghan, D.P., Teer, D.G., Logan, PA, Efeoglu, L & Amel1, RD. Deposition of wear resistant
coatings based on DLC by unbalanced magnetron sputtering. Surf Coat Tech., 60 (1993) 525-530
4. J. Pickard, J. Fisher, D Gibbons, H Koybayashi, A Ito, T Tateishi, A Jones, D. Te,:r & J. Egan,
"Multi-centred Tribological Testing ofUHMWPE against DLC coatings on cobalt chromium al1oys."
Presented at the World Biomaterials Congress, Hawaii 2000, to be published in J. Biomed. Mater. Res.
5. Jones, AH.S., Camino, D., Jiang, J. & Teer, D.G. Novel high wear resistant coatings deposited by
magnetron sputtering of carbon targets. J. Eng.Tribol. 1998,212,301-306.
6. Camino, D., Jones, A.H.S" Meres, D. & Teer, .D.G. High performance sputtered carbon coatings for
wear resistant applications. Vacuum, 52, (1999) 125-131.
7. Jones, A.H.S., Taylor, SK, Teer, D.G., El1oy, M., Friction and wear Testing ofDLC Type Coatings
on Total Hip Replacement Prostbeses. Presented at 'Friction, Lubrication and Wear of Artificial
Joints-Tribology meets Medical Engineering' (Leeds, UK Nov. 2000) To be published (I Mech E).
62

8. Lappalainen, R., Anttila, A & Heinonen, H. Diamond coated total hip replacements. Clinical
Orthopaedics and Related Research. No. 352. Lippincott-Raven publishers 1998, pp 118-127.
9. Semlitsch, M & Willert, H G. Clinical wear behaviour of UHMWPE cups paired with metal and
ceramic ball heads in comparison to metal-on-metal pairings of hip joint replacements, In Advances in
Medical Tribology, MEP Ltd. 1998,165-180
10. Riecker, c.B., KOllig, P., Schon, R, Windler, M., and Wyss, V.P. Clinical wear performance of
metal-on-metal hip arthroplasties. In Alternative bearing surfaces in total joint replacement, ASTM,
STP, 1346, J.J. Jacobs and T L.Craig, ASTM, 1998.
II Streicher, RM, Semlitsch, M, Schon, R., Weber, H., Rieker, c., Metal-on-metal articulation for
artificial hip joints: laboratory study and clinical results. In Advances in Medical Tribology, MEP
Ltd., 1998, 83-92.
12. Jiang, J., Arnell, R.D. & Tong,J. Some special tribological features ofDLC coatings on soft substrates
Wear, 211 (1997) 254-
13 V. Bellido-Gonzalez, Hampshire, J., Jones, AH.S., Witts, J., Allen, T.J., Teer, D.G. & Pierret, B.
Advances in the analysis and characterisation of DLC coatings. Surface and Coatings Technology 98
(1998) 1272-1279.264
14. 1 von Stebut, H. Michel, V. Bellido-Gonzalez, AH.S. Jones, "Ball Cratering: a Versatile Tool for First
Order Surface Coating Characterisation". Paper prerented at the PSE Conference, Garmisch
Partenkirchen, Germany, 1996.
15. Dowson, D. and Jobbins, B. Design and development ofa versatile hip joint simulator. J. Engng. In
Medicine, 1988, 17, 1-6.
16. Goodacre, J., and Mellor, G,. In preparation, available as a private publication from Prof J.Goodacre,
Lancashire Postgraduate School of Medicine and Health, UK
17. Mikhalovska, LI., Santin, M., Denyer, S.P., Lloyd, AW., Faragher, RG.A., Teer, D.G., Taylor, S.,
Mikhalovsky, S.V. Fibrinogen adsorption to metal and ionorganic coatings. Pesented at the
International Conference on Advances of Biomaterials for Reconstructive Medicine. (Capri, Italy Jun
2002) To be published.
18. Bowden, FP., and Tabor, D. The Friction and Lubrication of Solids, Part 2, (Clarendon Press,
Oxford, 1964), P214.
19 Z. M. Jin, D. Dowson & 1. Fisher. Analysis of fluid film lubrication in artificial hip joint replacements
with surfaces of high elastic modulus. Proc.I.MechE 1997, vol.2ll, Part H, 247-256.
20. Yang, S., Camino, D, Jones, AH.S., Teer, D.G. The Deposition and Tribological Behaviour of
Sputtered Carbon Hard Coatings. Surf. Coat. Techno!. 124 (2000) 110.
MECHANISM OF INTERACTION OF SILICON NITRIDI~ POWDERS
WITH BIOCHEMICAL MEDIA AND THEIR TOXIC EFFECT

V.A.LA VRENKO, N. V.BOSHITSKA YA, G.N.MAKARENKO

Institute for Problems ofMaterials Science, National Academy of
Sciences of Ukraine, 3, Kr=lzy=hanovsky str., Kiev-142, 03680 Ukraine

For a simulation of toxic effect of silicon nitride powders (in different dispersion
state) on human organism, the rates of conversion and mechanisms of interaction with
physiological Ringer-Lock solution, gastric juice, blood serum and inorganic media
with the same pH were studied. For the alkali and neutral biochemical media, the
mechanism of bioinorganic catalysis was established, the Ca2+- ions play the role of
catalyst. The results of study were confirmed by toxic tests using a priming of white
rats with Si3N4 powders and the further analysis of their blood indexes as well as
morphological state oflungs and liver after 1,3 and 6 months.

1. INTRODUCTION

In up-to-date technologies a great enough place belongs to thl~ nitride-based

ceramics. They proved to be the best high-temperature materials capable for a long-
time working in the aggressive gas and metallic melts media, preserving their
exceptionally high corrosion resistance. Besides, these materials have an
exceptionally high level of mechanical properties (hardness, strength, wear- and
erosion-resistance). The powders of silicon nitride are widely used in the materials
science and industry - for the manufacturing of different parts of high-performance,
tool and refractory application, using the powder metallurgy methods, mainly, the
pressing and sintering. Hereby the plant and laboratory personal is, undoubtedly,
subjected to an unhealthy action of fine-dispersion substance.
Therefore it is very important to study the toxic action of Si3N4 powders on the
human organism. Hereby not only the direct tests on the rats must be carried out, but
also a simulation of an interaction of these particles with both biochemical and
inorganic control media.
The study of toxic properties of Si3N4 powders, obtained at nitridation of silicon in
the resistance furnaces as well as using the low-frequency plasma discharge [1],
showed that, first of all, a failure of animal's liver and kidney functions took place.
The structure of cell membranes was injured, and a disorder of circulation of blood
and seat swelling of liver cells occurred [2]. Herein the toxicity ofSi 3]'1;4 nanopowders
proved to be approximately by 7 times more than that of large-dispersion ones.
Because the powders have been introduced into the stomach and lung
63
Y. C. Cogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 63-70.
© 2003 Kluwer Academic Publishers.
64

directly, the straight mechanical failure of liver and kidneys was excluded.
Obviously, the changes in an organism were associated with an action of products of
Si3N4 dissolution in the living body media [2]. In this work an attempt was made to
ascertain the mechanism of Si3N4 powders interaction with different biochemical
media.

2. MATERIALS AND METHODS

The following kinds of Si3N4 powders were studied: 1) technical with ex- Si3N4
content of 78% and average size of particles a till 5 Ilm, the specific surface area S =
4.3 m2/g; 2) Si3N4 fibers (ex-Si3N4 > 90%, d = 1-3, I till 20 Ilm); 3) nanopowder
obtained by plasma synthesis (ex-Si3N4 - 50%, S = 42.0 m2/g, a till 0.02 Ilm).
As biochemical media the following ones were studied: 1) the physiological salt
solution of Ringer-Lock (pH 7.0) that as to its osmotic pressure was near to a blood
plasma; 2) gastric juice (pH 2.0); 3) a blood serum with pH 7.35-7.47. As the control
media the distillated water (pH 7), hydrochloric acid (pH 2) and a diluted sodium
hydroxide (pH 7.4) were used. All the tests were carried out in a Goeppler
thermostat at 37°C during 100 h under constant mixture. The initial powders and
non-soluble residua were investigated with the aid of chemical and XRD analyses,
for nanopowder the ~-potential was measured under electrophoresis in the constant
electric field. For the determination of nitrogen in the samples the modified Kieldal
method was applied. The rate of conversion is defined as ex = (Co - C)/C o x 100%,
where Co - nitrogen concentration in the initial sample (mass.%), C - nitrogen
concentration in the solid remainder (mass.%). Besides, the corresponding
experiments on the animals priming with Si3N4 powders were carried out with SEM
control of their organs damaged tissues and determination of appropriate indexes of
blood biochemical analysis.

3. RESULTS AND DISCUSSION

In all the cases the Si3N4 nanopowders interacted with biochemical media with
the greater rate than the large-dispersion ones. In the physiological solution they
were strongly decomposed while in the gastric juice the powders were more stable
(Table 1). The interaction of Si3N4 powders with a blood serum also proceeded
more intensively than with a gastric juice (Tables 1, 2).

Table 1 Rate of conversion, a (%)

Type of Neutral medium Acid medium Alkali medium
powder
Phys. H 2O Gastric Hel Blood NaOH
sol. juice serum
ShN4 techno 28.0 7.8 12.2 0.5 13.8 1.3

Si3N4 fibers 6.0 4.0 10.0 1.45 26.3 0.8

nanopowder 86.5 16.9 13.7 7.8 67.6 11.5
 65

Table 2. Results of filtrate analysis (mass.%) after interaction ofShN4 powders with
biochemical media

Kind of Physiological Gastric juice Blood Sl!rum

powder solution
N Si N Si N Si

Techn. 1.4 0.055 2.2 0.21 Not 0.73

determ.
Fibers 2.8 0.069 2.2 0.14 0.67
Not
Nanopow- 0.8 0.189 2.1 0.22 determ. 0.69
der

At dissolution of ShN 4 nitrogen, mainly, educes in molecular form, and silicon

remains in the non-dissolved residue as a gel of silicic acid. Hereby the Si content in
the filtrates is determined by H2 Si03 solubility. The acceleration of interaction of
Si3N4 particles with blood serum and physiological solution may be explained by a
presence in these media of catalytically active Ca2+-ions [3]. So, the interaction with
OH- - ions in alkali and neutral biochemical media finally occurs according to the
reaction

but for the intermediate stage it is determined by interaction with a bioinorganic

catalyst K according such scheme: A + K = AK (formation of an intermediate
substance); AK + B = AB + K where A is a medium component, in our case OH--
ion, B is Si3 N4 , K - catalyst (Cal,). Hereby in the acid medium (gastric juice) the
interaction is implemented by

(2) reaction.

The results of study of initial interaction of Si3N4 nanopowders with different

electrolytic media proved to be very interesting. Although the dissociation of formed
silicic acid according to the first stage

(3)

is thermodynamically more probable, the partial dissociation by the second stage is

possible as well:
(4)

The Si3N4 nanoparticles (a size of 0.02-0.04 11m) can be considered as the colloidal
mycelia formed as a result of initial adsorption interaction of fine-dispersion phase
with an electrolyte. In particular, for a neutral medium the scheme of a double
electric layer formation around of the ShN4 particles may be presented as Fig. 1.
66

The adsorption layer consists of outer and inner parts. In the outer part there are
the hydrated H30+ - ions which are kept near an interface with the aid of electrostatic
forces. Obviously, in its inner part there are the HSi0 3- - and SiO/ - ions
chemisorped on the solid surface, the latter having the same charge that a solid phase
and partially destructed hydrated shells. It was established that the Si3N4 colloidal
mycelia in our case had a total negative charge.
One can see that for the initial interaction - according to the mechanism of
bioinorganic catalysis with the participation of OH- -ion of the medium - the
participation of outer Si3N4 mycelia layer containing H+ - particles is demanded.
Only under these conditions the negative part of nanopowder mycelia may be
connected with the positive Ca2+ - ions.
The toxic, cyto-toxic, fibrogeneous and allergologic effect of Si3N4 powders
having the different dispersion state and particle morphology was determined using
the tests on the animals. First of all, the total toxic effect of powders was
investigated. The average half-death total-toxic doses of the powders above were
determined using the tests on the white rat males. The corresponding powder
suspensions in the physiological solution (a concentration - Ig per 1 kg of animal
weight) were brought in a rat peritoneum with the aid of a syringe. After this the
observation was held during 14 days. The half-death dose was calculated using a
Prozorovsky's method_ The powder leading to more than 50% cases of an animal
lethal outcome was considered as the toxic one.
One can see in Fig. 2 the microstructure of rat's lung tissue after nanopowder
priming. At first (during 1 month) only some changes of lung tissue are observed,
then the nucleation of new-formation takes place. At last, after 6 months, the
collagen fibers are observed, such pattern is characteristic for a pneumonia
development.

HSiO;
HSiO~SiO;
HSiO; ~ HSiO; H+
HSiO; HSiO;
HSiO;

H+
Figure 1. Scheme of double electric layer formation on the surface of ShN 4
nanopowder (~= - 40.1 mY).
 67

Figure 2. Microstructure of rat's lung tissue after priming witb Si3N4

nanopowder: (a) - after 1 month (x 10000); (b) - 3 months (x 7000); (c) - 6
months (x 7000).
68
Table 3. Activity of ferments in blood serum and biochemical indexes of rats' lung
tissue resulting from animals' priming with Si3N4 powders

Test term, Kind of Alkali Alanyn- Aspartate- Lipids Oxypro-

months powder phos- aminotran- aminotran- (mg/l00 line (Ilg/
phatase, spherase, spherase, mgof 100mg
units/l Ilmol/ml..h Ilmollml. h dry of dry
weight) weight)

1 SJ3N4 312.6 2.21 0.75 23.8 1145.6

3 techno 298.2 3.14 0.98 21.6 925.3
6 260.4 3.02 2.05 22.3 875.5
control 216.5 1.98 1.93 17.3 812.3

1 Si3N4 613.8 4.99 1.8 26.9 1112.2

3 fibers 1593.2 3.87 2.5 27.3 1284.4

6 501.3 2.68 2.0 22.4 1125.2

control 268.3 2.25 2.3 21.7 982.4

1 Si3N4 593.9 6.25 0.57 23.9 935.2

3 nanopow 410.1 4.72 1.47 20.3 1284.1

6 365.4 3,23 1,58 16,7 966.3

control 254.5 2.11 1.82 16.9 717.9

The aspartate-aminotransferase, alanyn-aminotransferase and alkali

phosphatase amounts were determined in the rat's blood serum in I, 3 and 6 months
after priming with Si3N4 powders, in comparison with priming by significantly less
toxic AlN powders (Tables 3 and 4). One can see that almost all the biochemical
indexes of rats' blood after priming proved to be significantly changed only in the
case of ShN 4 powders. This especially concerns the alkali phosphatase index that is
responsible for the animal's cancer susceptibility. This index was increased almost
by 6 times after 3-month priming of rats with Si3N4 fibers whereas it remained
practically the same as for control rats in the case of AIN priming. It is known that at
the acute and chronic diseases accompanied with a destruction of any cells of an
organ or a tissue, the going out of transaminase from an affection seat to a blood
takes place. The output of enzymes from the liver and lung tissues to the blood
 69

serum occurs on the account of an injury of a permeability of cell membranes. This,

in one's tum, testifies to a damage of a functional state of an appropriat1e organ.

Table 4. Change of biochemical indexes in blood serum and rats' lung tissue
resulting from animals' priming with AIN powders

Kind of Test tenn, Alkali Alanyn- Aspartate- Lipids Oxyproline

powder months phos- aminotranspherase aminotranspherase (mg/lOO (~g/I00
phatase (~mol/ml. h) (~mol/ml. h) mgof mgofdry
(units/I) dry weight)
weight)
AIN 16 212.5 0.58 \.68 17.4 823.3
techno control 201.3 0.45 \.63 15.8 914.8
254.1 2.11 1.82 16.9 717.9

A1N 16 234.1 \.00 \.94 16.6 955.9

fibers control 225.8 0.67 \.87 15.9 984.7
268.4 2.68 2.02 17.4 812.3

AIN 16 200.1 0.39 \.23 19.4 924.2

nano control 188.6 0.43 1.25 14.6 1250
223.4 \.98 \.94 16.0 768 ..3

CONCLUSION

1. The stability of ShN4 powders in the corresponding biochemical media

essentially decreased with a diminution of particle sizes.
2. The silicon nitride powders proved to be less stable in the biochemical media
than in the inorganic ones, obviously, due to a protein-calcium salt catalytic
effect (the mechanism of intermediate catalysis).
3. It has been shown that the toxic effect of Si3N4 powders on the living organism
was generalized, it was realized through the circulation of the blood.
4. It has been established that the technical ShN4 powders were characterized by a
comparatively moderate toxic and fibrogeneous action on the living organism.
In the same time the plasma-chemical Si3N4 nanopowders led to a change of a
rat lung tissue, there were the regions of collagen new-formations. The silicon
nitride fibers had a large enough cyto-toxic action: they, mainly, provoked the
pneumonia of animals.
70

Thus, the simulating physical-chemical study as to interaction of silicon nitride

powders and fibers with different biochemical media has been entirely confirmed by
the real toxic tests with the animals.

REFERENCES

I. T.Ya.Kosolapova, T.V.Andreeva, T.S.Bartnitskaya et al. Non-metallic

refractory compounds, Moscow, Metallurgy, 1985 (in Russian).
2. A.M.Shevchenko, G.A.Shkurko, O.V.Zagorina, N.V.Boshitskaya, Powder
Metallurgy, 1/2 (1994), 117-120(inRussian).
3. K.B.Yatsimirsky, Interaction into the bioinorganical chemistry, Kiev, Naukova
dumka, 1976 (in Russian).
MICROSTRUCTURE AND STRENGTH PROPERTIES OF GERMANIUM
MICROWIRES FOR BIOMEDICAL DEVICES

D. GRABCO, M. DYNTU, D. MEGLEI, O. SHIKIMAKA

Institute ofApplied Physics of Moldavian Academy of

Sciences, 5 Academy str., Chisinau, MD-2028, Moldova,
e-mail: mech.prop@phys.asm.md

Abstract

The qualitative gennanium fibres in glass isolation adequate for practical application
have been prepared. Their microstructure (micrograininess, dislocation density),
electrical parameters (conductivity and specific resistance) and mechanical properties
(microhardness, tensile strength, bending critical radius) have been studied. A special
techniques for the fabrication of microelectrodes ensuring qualitative microcontact with
biological objects to be applied in devices at biomedical investigations has been
elaborated.

Keywords: gennanium microwires / microstructure / microhardness / tensile strength /

Ge microelectrodes for biomedical devices

1. Introduction

The progressively deep penetration of scientific and engineering achkvements into the
various branches of modem physiology is marked last decades. Herewith,
microelectrodes are a special instrument which ensure the reliable electrical contact
with biological cells without infringement their normal function. Microelectrodes on the
base of Ge microwires are used for the producing of electrical signals when measuring
of the biopotentials of human and animal tissues by their incorporation into tissue.
That is why the investigation of these materials is of a great interest for the both
physical science and medicine. In particular, practice of the utilization of gennanium
microwires inside a glass isolation for the electro-measuring engineering put forward
the necessity of their strength characteristic investigation.
71
Y. G. Gogotsi and I. V Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 71-76.
© 2003 Kluwer Academic Publishers.
72

2. Experimental details

Qualitative germanium fibres in glass isolation (having diameter d=(20-300) 11m and
sufficiently low dislocation density N= I 0 2_10 3 cm- 2) suitable for practical application as
microelements in biomedical investigation are fabricated using the Ulitovscky
techniques. The microstructure perfection of Ge wires have been studied and some
mechanical parameters such as microwire bending radius (rer), tensile strength (cr) and
Vickers microhardness (H) have been determined. Electrical characteristics
(conductivity, specific resistance) as a function of the wire diameter and length were
also measured. A techniques of preparation of the specimens adequate for device
application was worked out.
The measurements of deformation by bending technique were made using a
modified setting, described in [1]. The tensile strength was determined by loading
techniques on apparatus for thread-like crystal rupture. Specimens undergone to rupture
represented the oriented microwires about the 2P mm length having the glass cover as
isolation. The tests were made on the 6-10 specimens with the equal geometric
dimensions: wire diameters and isolation thickness. The average fracture stress proper
to the middle wire diameters and cover (isolation) thickness was determined as a
arithmetical mean. A formula

a = PIS, (I)

was used for the tensile strength calculation. Here P is the load of the specimen failure,
S is its cross-section area:

(2)

where R is the total radius of wire with insulating cover.

The microwire rupture surfaces were investigated by optical and electron microscopy
methods. Microhardness of the Ge-wires was measured by the microhardness tester
PMT -3 supplied with the low temperature research device. The Vickers pyramid was
used as indentor. Microhardness values were determined using the usual formula [2].

3. Results and discussion

Investigations of the relative change of the fabricated wires specific resistance in

dependence on the same of the parent bulk Ge crystals were performed; the distribution
of the specific resistance along the sample length and the conductivity dependence on
wire diameter were also estimated. It was shown that maximal specific resistance of
wires did not exceed value of p = 4 Om·cm decreasing with its length increase and
gradually increasing with diameter diminution.
The obtained germanium wires having diameters d = (20-300) 11m possess cylindrical
form (Fig. 1), crystalline structure and relatively low dislocation density: N = (10 2- 10 3 )
cm- 2 .
 73

Figure 1. Germanium microwire in glass isolation (cross-section).

The test of microwire bending was proceeded up to rupture of wires and no the
diameter reduction was observed in fracture regions, i.e. the break had a brittle nature.
The glass isolation also suffered the brittle fracture and its failure took place
immediately as soon as the critical bending radius was achieved. Results of bending
parameters measurements are given in table 1.

TABLE 1. Critical radius of various microwire bending

Germanium Nichromium Stainless steel

d, 11m D,l1m fcr,mm d,llm D,flm fcr,mm d,l1m D,1lffi fcr,mm
1.4 27.9 0.25 7.6 46.7 0.5 3.1 46.1 0.25
8.3 36.5 0.75 11.0 46.1 0.6 5.2 48.0 0.45
15.7 40.7 1.00 14.9 59.4 0.9 7.3 47.1 0.45
18.1 41.9 1.75 19.5 74.7 1.2 12.3 55.7 1.00
22.3 40.8 2.00 28.3 99.6 2.3 15.7 62.3 1.70
94.5 119.5 4.20 35.6 100.5 4.0 20.9 ~·6.0 1.50

In accordance with table 1 the critical bending radius decreases with the wire
diameter diminution that is indicator of the improvement of microwire strength
properties if wire diameter becomes more thin. Similar dependence was noted for
microwires of other composition (nichromium, stainless steel, etc.).
The results of tensile strength determination also confirm that strength properties of
germanium microwires improve with wire diameter diminution. It can be seen in table
2. For comparison the data for other studied materials are shown in the table as well.
Optical and electron microscopy investigations of the microwires break surface show
that the destruction in this case has a brittle nature as well (Fig.2). The separate blocks
and grains disoriented each in regard to other on several degrees are clearly seen on the
cross fault surface. Hence, the material fracture takes place on the grain boundaries
existing in thread-like crystals in the regions where the most strong concentrators of
internal stresses are created under deformation. The number of grains and blocks as
74
well as their disorientation degree decrease by crossing from thick wires to thin ones.
Simultaneously with this the dislocation density is also reduced. All these contribute to
the microwire improvement.

TABLE 2. Tensile strength of various microwires

Germanium Nichromium Stainless steel

d, gm D,f!/lI a,MPa d,gm D,gm a,MPa d,gm D,gm a,MPa
2.1 18.8 427 2.1 43.9 673 3.l' 46.1 861
4.2 24.1 344 5.2 41.9 532 4.2 45.0 622
5.2 28.3 330 73 51.3 265 5.2 48.0 662
6.3 33.0 313 8.4 58.6 238 15.7 69.1 507
73 26.2 223 13.6 60.7 207 16.7 60.7 450
8.4 78.5 227 19.9 74.3 124 20.0 97.4 283
19.2 57.6 122 21.9 75.4 144 27.0 101.4 238
25.0 98.4 100 38.7 128.8 68 34.0 136.0 193

a b

Figure 2. Cross-section image of Ge microwires after tensile testing

Measurements of microhardness of the germanium wires were carried out on the

longitudinal cut and the cross-section. The sample preparation was performed using the
specific techniques (Fig. 3). It was found that microhardness of Ge-wires is higher than
microhardness of bulk crystals (Hb c). At room temperature H wires are equal to to 900-
1000 MPa in comparison with Hbc = 725 MPa. The microhardness magnitude is varied
in dependence on sample diameter: H increases with diameter decrease. Microhardness
also increases at temperature lowering: H77K>H 293K and ll.H is equal about 30%.
Presented results evidence that the obtained Ge microwires are the qualitative
material suitable for the practical application. In this connection a special techniques for
the preparation of spiky tips of microwires was worked out. For achievement of the
good microcontact with the biological object the electro-chemical etching was applied
in the devices fabricated on germanium wires base. The aqueous solution of sulphuric
and hydrochloric acids was used for chemical treatment of wires. This technique allows
to get the electrodes with various degree of sharpness (Fig. 4).
 75

a b

Figure I Micwnardnl%s indenUltkms made on the (a) cross-section and (b)longituduna! cut of Gc wires.

a b

c d

Figure4. a, b - The spicate tips of the gennanium microwires obtained by the electro-chemical technique.
C,d - The image ofGe microwire before (c) and after (d) electrolytic sharpening prepared for
the biomedical application.

A special techniques for fabrication of the medicine microdroppers with built-in

microelectrodes was worked out for the application in medico-biological investigation
(Fig. 4, c,d). A schematic picture of such dropper is given in Fig. 5.
76

I • Mi¢_1=tt_
,.lruli_ .....
2-Ej><>xW" _ "

4""O¢'m~'jnt
$.M""~
6 - O:"".."',,*km

FigureS. A schematic piCUlfe of biomedical dropper.

The technological improvement and worked out of the electrolytic sharpening

method made it possible to get the cone-shaped tips of microelectrodes with diameter of
0.5-1.0 flm and outgoing from dropper no more than 50 fJffi. These microelectrodes
passed the test and shown good quality.

4. Conclusion

1. The qualitative germanium fibres in glass isolation adequate for application as a

microelements in medico-biological investigations are prepared in this work.
2. Their electrical parameters ( conductivity and specific resistance) are determined as
a function of diameter and length of microwires.
3. Microstructure and mechanical properties: microhardness, tensile strength, aptitude
to bending, dislocation density have been studied. It has been shown that
mechanical properties of germanium fibres improve with microwire diminution.
4. The techniques to yield the qualitative microcontacts with biological objects to be
applied in devices for biomedical investigations has been elaborated.

Reference

[1]. BANARI, V.F., DYNTU, M.P. Tensometric properties of thread-like PbTe crystals. In Semimetals and.
narrowgap semiconductors and devices on its base , editor Shtiintsa, Kishinev, 1986.
[2]. BOYARSKA VA. Yu. S., GRABCO, D.Z., KATS. M.S. (1986) Physics o!microindentation processes,
Shtiintsa. Kishinev.
THE RESEARCHES OF PROPERTIES OF BIOMA TERIALS BASED ON
BIOLOGICAL HYDROXYAPATITE IN SYNTHETIC AND NATURAL
PHYSIOLOGICAL MEDIUMS

*IVANCHENKO L.A., **SUL YMA V.S., *PINCHUK N.D.,

*Institute for problems of material science of NAS of Ukraine
3 Kr=hy=hanovsky Str., Kyiv 03142 Ukraine
**Institute of traumatology and orthopaedics of NAS of Ukraine
27 Vorovsky Str., Kyiv 01601 Ukraine

Abstract

Biomaterial "Osteoapatit keramichesky" (OK) based on biological

hydroxyapatite (BHAp) under the trademark "Osteoapatit" were studied. A silicate-
borate-sodium glass amount in the OK samples composition reached 85 % (mass.). The
degree of biochemical activity was appreciated on a loss of sample mass in blood
plasma and normal saline solution. An influence of biomaterial on activity of humane
bone marrow osteogenics cells-precursors was investigated. The experimental
researches of interaction "in vivo" between biomaterials and bone tissue were conducted
in 35 adult mongrel dogs. A new bone formation and rebuilding of OK granules after
implantation in bone defect have been studied.

1. Introduction

The biomaterials based on hydroxyapatite (HAp) involved serious attention of many

scientists. This material has high affinity and ability to be substituted by new bone
tissue. They are widely applied in surgery, orthopaedics, traumatology and stomatology
for elimination of bone defects. The materials differed by preparation technology,
composition and features of biological interaction. The general term "calcium-phosphate
ceramics" combines there [1,2].
We have developed the bioactive glass-ceramics - "Osteoapatit keramichesky"
(OK) (State registration of Ukraine N~628/2002 from 16.01.2002) as a searched result of
new manufacturing technology and biomaterials with given properties [3-5].
The main purpose of this work was an experimental substantiation of clinical
application these biomaterials in treatment of the patients with the various parentage
bone defects and, particularly, with the chronic osteomyelitis.
77
Y.G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 77-82.
© 2003 Kluwer Academic Publishers.
78

2. Materials and methods

We have utilized implants based on BHAp according to demands ASTM F1185-88

(conclusion by Ministry of health of Ukraine N2 5.04.03/769 from 16.03.2000). The
sodium-boro-silicate bioglass was added in fusion mixture OK in quantity up to 85 %
(mass.) according to technology. Also two-stage sintering was exposed at preliminary TJ
temperatures (below and above 900°C) and final T2 (below 900°C) [4].
The composition of OK samples was supervised by infrared (IR) spectrometry
analysis, chemical and X-ray-phase methods. The crystalline frame of HAp was revealed
in OK samples by all methods. The degree of biochemical activity of biomaterial was
appreciated on loss of samples in a blood plasma and normal saline solution at T = 36-
37°C during two day.
Biocompatibility properties ofbiomaterials were studied using the cultivation the
humane bone brain's osteogenics cells-precursors. There were investigated 94 cultures
of fibroblasts from bone autospongious in 10 patients with orthopaedic pathology or
chronic osteomyelitis. It was determined the dependence of cloning efficiency of colony-
creating of bone marrow fibroblasts units on the biomaterials granules with size from
50 micron up to 3000 micron, and quality of biomaterial composition. In all series of
experiments the colony-creating of bone marrow fibroblasts units was cultivated as
control in standard conditions without BHAp addition. The results were estimated on
fibroblasts cloning efficiency among 10 5 of nucleus cells [7]. The opportunity of
collimating of antibacterial properties to biomaterials by saturation of antibacterial drugs
(ABO) have studied.
Character of interaction between OK samples and ABO was studied using there
infrared spectrometry after saturation during 15 and 30 minutes and desiccations during
1 hour at T=11O-130 OC. The thirty bacteriological examinations on determination of
ABO (Ceftriaxoni) acting duration are made by agar diffusion method (method of disks)
[7].
Interaction between biomaterials and bone tissue has studied experimentaly in 35
adult mongrel dogs and 10 rabbits, according to "General Ethical Principles of Animals
Experiments", ratified on the First National Congress of bioethics (Kiev, 2001).
Biomaterials OK 100 (without bioglass) and OK 2.15 (15 % bioglass) were mainly
explored (tab. I). The experiment consists of mechanical building of epimethaphiseal
bone defects, which substituted by biomaterials granules or cylinders. The infected bone
defect models an experimental animal with the subsequent realisation ofnecroctomy and
filling by granules with ABO were also constructed. All operations carried out under the
intravenous narcosis. Terms of observations were equal to 3, 5, 7, 13, 15, 26 and 30
weeks. Colouring of microsections was executed by haematoxylin and eosin, pikrofuxin
by van Gizon. The formation and rebuilding of bone tissue was studied using the X-ray,
histomorphological and morphometrical methods.
 79

3. Results

3.1. EXPERIMENTS "IN VITRO"

The results of OK samples solubility (X) are submitted in the table 1. The dependence of
solubility from the contents of BHAp was not fixed. It is probably caused by particle
solubility of bioglass. We have revealed the tendency to intensifY the solubility
depending on porous size of OK samples.
Sometimes we detected the OK mass samples amount increasing in a blood
plasma, and we had to study sorption OK samples ability in different mediums.

TABLE 1. The solubility of OK samples in normal saline solution

Material brand % (mas.) BHAp Tl,oC Pores volume, % X, % mass/cm2.da~

OK 100 ;?:95 <900 0.0085
OK 2.15 ;?:85 <900 28-30 0.095
OK 1.6 50 ;?:900 23-27 0.0224
OK 1.6 50 ;?:900* 39-44 0.0254
OK 1.6** 50 ;?:900 30-35 0.0156
OK 1.5 45 >900 28-35 0.0083***
OK 1.2,5** 29 ;?:900 37-44 0.0048***
The note: * Enlarged duration of the first sintering;
** A material has a residual Carbon;
*** Tests are conducted in a blood Elasma

The results of studying the biomaterial ability to adsorb the synthetic normal
saline solutions and antibacterial drugs solutions are submitted in the table 2.

TABLE 2. The solubility of OK in antibacterial drugs solutions

Material brand Sorption, % masslcm 2

10 % an aqueous solution of 7.7 % solution "Ceftriaxoni" in
Acidi ascorbinici, normal saline solution,
15 minutes 15 minutes
OK 2.6 0.0409 0.0503 0.0396 0.0404
Pores volume, % 43-45 45-50 43-45 45-50
OK 2.6* 0.5785 0.8318 0.2677 0.4866
Pores volume, % 63-64 67-69 63-64 67-69
OK 2.12 0.0935 0.1928 0.0636 0.01406
Pores volume, % 46-47 49-50 46-47 49-50
* Material has a residual Carbon
An opportunity of chemical interaction between OK particles and antibiotic
solution was studied using IR spectrometry. The OK samples were there in the ABO
solution. Than we extracted the hydroxyapatite particles being on the surface and in
80

depth of OK samples, and studied them (tab. 3).

The minimum IR spectrums of a transparency confonned to maximal absorption
of a drug. It was caused by optical fluctuations of complexes (P-O) and hydroxyl
bunches (OR) from composition of OK samples or antibiotic. The main bands of IR
absorption for ABD were closed to 1740 cm'\ 1649 cm'I, 1538 cm'\ 1398 cm'I, 507
cm'I (table 3).

TABLE 3 Spectral positions of bands of IR-absorption maximum of OK saturated by

ABD ceftriaxon

Sample 1 Sample 2 Sample 3 ABD Absorbtion

m=1.57 g m=1.64 g, m=0.77 g ceftriaxo maximum
d=1.5 cm d=1.5 cm d=0.6 cm, n
,I change tw,
~m=2.92 %(mas.) ~m=3.45 %(mas.) ~m=4.02%(mas.) v,cm cm,I
vmax, cm'I vmax,cm,I vmax, cm,I
Surface Volume Surface Volume Surface + volume
358.04 364.36 358.00 358.04 358.21
469.71 463.62 471.11 465.48 472.37 507.51 5-7
570.63 570.69 569.14 570.20 570.07
602.88 602.88 602.86 602.83 603.03 644.64
959.65 959.56 958.57 959.38 959.00 889.04
1043.81 1041.94 1041.54 1041.00 1044.84 1032.49
1413.53 1416.11 1411.37 1415.79 1413.54 1367.55 2-4
1458.55 1460,04 1459.14 1460.08 1460.10 1398.00
1542.58 1542.82 1546.74 1537.94 4-8
1606.36 1605.63 1606.23
1751.24 1763.92 1740.00 1740.02 11-23
2015.73 2016.34 2015.81 2016.16 2016.48
2202.19 2202.83 2202.22 2202.71 2202.49
3417.00 3400.00 3400.66 3423.00 3423.00 3254.93
3569.92 3571.12 3569.56 3570.97 3572.06

These little bit changed spectral IR strips occurred in IR transparency spectrum

(T (v» the hydroxyapatite particles on the OK samples surface after saturation. These
strips in spectrums of sample 1 are poorly appreciable.
The dislocation of the antibiotic IR spectral strips of sample 2 in some cases
reaches 17 or 24 cm'I (for example, strip close 1763,93 in spectrums of a sample 2
instead of a strip 1740 cm,I for an antibiotic). For other spectral strips of an antibacterial
drug this dislocation compounds 1-7 cm,I, The most appreciable transformation is
characteristic for hydroxyl bunches (OR) fluctuations - about 3570 cm ,I.
These changes of IR spectrums transparency of the hydroxyapatite particles taken
from the surface of OK samples may be linked with the mechanical sorption of ABD
solution, or with chemical interaction between the components. This process passes
about 15 - 30 minutes and takes part mainly on the surface of a sample. There were two
stages in chemical interaction: the first stage - partial dissolution of the sample surface,
 81

the second - sorption of the ABD solution. It is probably caused by the large saturation
of the sample 2 surface, than sample lone. The partial sample 1 increased the saturation
rate of sample 2 by the same solution. We linked with the presence of calcium ions or
other ions in the solution after sample 1 had dissolved.
The duration of OK antimicrobic action, packed by an antibiotic" does not depend
on dimension of beads and porosity of a biostuff, that is proved to bacteriological
methods of researches. The active influence of an antibiotic proceeded within day with
gradual decrease of its concentration.
We have studied influence of a biomaterial on cloning efficacy of the humane
bone brains cells-precursors. OK does not show inhibiting influence on a proliferation
"in vitro" the stromal fibroblasts extracted from the autospongious of the patients with
an orthopaedic pathology or chronic osteomyelitis. The cloning efficiency depends on
granule's size. The OK granules with the dimensions smaller then 160 microns inhibit
the body growth of colonies. It can be caused by infringement of three-dimensional cell
orientation, which is necessary for their body growth and differentiation. Probably,
shallow OK particles blocking cellular receptors, interfere with building of intercellular
connections and block cell ability to form colony. The OK granules with the dimensions
from 500 up to 3000 microns do not retard body growth of the cells-precursors and
multilayer colonies formation.
Multilayer colonies of the cells-precursors formed more actively using OK
samples with carbon content (samples OK 1.2,5 c). This fact testifies about a stimulation
of this specified biomaterial the structural-forming potential of the cells-precursors.

3.2. EXPERIMENTS "IN VIVO"

The high biocompatibility of OK and bone tissue is proved to X-ray and
histomorphological researches of the implantation zone. The osteogenic tissue
germinates into the granules with gradual formation of a dense bone beams network.
The granules fragmentates and new bone tissues substitutes them. The high adhesion
compatibility between new bone tissue and granule's surfaces without forming fibrous
capsules and absence of an inflammatory reaction, testifies an adequate interaction
among biomaterial and surrounded tissues.
We detected irregular inspissations of bone tissue with indistinct contours and
new-formed network of the bone tissue in 3 weeks after implantation. The vessels
germinated into the biomaterial and formed bone tissue and were in tight
interconnections with the granule's surfaces. The germination depth was about 1,5-2,0
mm. Some granule'S exceeded density of the bone in 26 weeks. The thickness of bone
layer on the granule's surfaces depended on their size. A new-formed bone layer was
thicker around OK granules size from 160 up to 500 microns, than around granules size
froml000 and 3000 microns.
Dystrophic and the necrobiotic changes were observed in a new-formed bone
tissue. The most part of granules degraded, fragmented and was gradually substituted by
a new tissue. The zones of a yellow bone marrow were formed in intratrabecular
interspaces, which were filled with fibro-reticular bone marrow. The border of an
implantation range was absent.
The X-ray and histomorphological changes of an implantation zone of granules
82
without bioglass (OK 100) was similar. The new-fonned bone tissue quantity around
granules without bioglass in early tenns of observation (3 months) was authentic less in
comparison with its quantity around granules with bioglass. The described researches
have allowed us to use biomaterials based on natural hydroxyapatite and saturation by
the antibiotics for substitution in 24 patients with chronic osteomyelitis. The
biomaterials have passed clinical approbation successfully as a plastic material for
substitution of the bone defects of a various parentage in department of bone-purulent
surgery.

4. Conclusion

1. The solubility rate of the OK samples biomaterials in simulated physiological media

by a complicated way depends on a mass relation BHAp/glass. The solubility is
accelerated with increasing total amount of pores in samples of the same composition.
2. The availability of sorption effect decelerates of solubility rate of the OK samples
biomaterials in the indicated mediums and simultaneously allows to samples to save the
absorbed antibacterial substances within 24 hours and more.
3. It is fixed from experience "in vitro" and "in vivo" that the growth rate of a new
created bone tissue is enlarged at implantation of the OK samples biomaterials in
granules form with the dimensions exceeding 500 microns, and with the presence of
residual Carbon in the composition of utilised biomaterials.

5. References

1. Pinchuk N.D. and Sulyma V.S. (2000) Biomaterials for osteoplastics, Problemy
osteologyi 3,37-41.
2. Podrushnyak E.P., Ivanchenko L.A., Brusko A.T. (2000) Advanced use of composite
biological hydroxyapatite glassceramics for a bone tissue regeneration Ortopedija,
travmatologija i prote=irovanie. - 2. - C. 129-130.
3. !.A.lvanchenko, Podrushnyak Ye.P., T.I.Falkovska et al. (1997) Bioactive materials
for use in biology and medicine Mater. coni: "Osteoporo=: epidemiologiya, klinika i
likuvanniya" Lviv, may 1997, Kyiv, Institute of Gerontology AMS Ukraine, 45-46.
4. Ivanchenko L.A., Falkovska T.!., Danilenko N.V. et al (1999) Structure and
properties of a high-porosity glass ceramic containing biological hydroxyapatite,
Powder Metallurgy and Metal Ceramics 38,418-453.
5. Podrushnyak Ye.P., Ivanchenko L.A., Falkovska T.!., et al. (1998) New
biocomposites based on bone hydroxyapatite and possibility of its use in biology and
medicine II Problemy osteologyi 1, 102-104.
6. Podrushnyak Ye.P., Ivanchenko L.A., Brusko A.T. et al. (2000) Biocompability with
bone tissue and osteotropy of compo used materials based biological hydroxyapatite -
"Osteoapatit" Problemy osteologyi 3, 89.
7. Sulyma V.S. and Panchenko L.M. (2002) Immunological substantiation possibility of
using of materials based biological hydroxyapatite - "Osteoapatit" for bone defects
substitution Galickiy likarskiy visnik 9, 87-89.
PREPARATION OF HYDROXYAPATITE NANOPARTICLES USING A
MODIFIED PRECIPITATION METHOD

G. e. KOUMOULIDIS\ C. e. TRAPALIS z, A. T. SDOUKOSI, and

T. C. VAIMAKIS 1
}Department of Chemistry, University of loannina, P.o'Box 1186,
JOANNINA 451 10, GREECE.
2«Demokritos» National Centre for Scientific Research, Institute of
Materials Science, 153 10 ATHENS, GREECE

Abstract

Hydroxyapatite (HAp) is the primary constituent of the human hard tissues. The
preparation of fine and well sinterable HAp powder is the most important step in
fabrication ofbioceramics. HAp nanoparticles were prepared using rapid increase of the
pH value, by adding an amount of concentrated NH 40H solution, of a well-mixed
solution of Ca(H zP04h'HzO and CaCh with stoichiometric atomic ratio of Ca/P. This
modified precipitation method is called "pH shock wave method". Tht: XRD and FTIR
analyses indicated that the precipitated powder was crystallized HAp. The HAp was
stable during calcination at 550°C, while a significant amount of I3-CaJ(P0 4h appear at
800°e. The TEM micrograph showed arbitrary morphology of nanoparticles, with
average grain size of -80-200 nm in length, and of -10-30 nm in width and thickness.
The adsorption isotherm of HAf was typical for nonporous particles. The specific
surface area of powder was 74 m 'g'l. The chemical analysis indicates the formation of
nonstoichiometric hydrated HAp with empirical formula
Ca886(HP04)128(P04k72(OH)·7.50H20 and atomic ratio of Ca/P equal to 1.477. The
lattice parameters were different from standard ones, due to the presence of HPO/'
groups in the powder. The sizes of the crystallites responsible for the Bragg reflection of
(002) and (300) planes were 31.1 and 14.9 nm, respectively, indicating anisotropic
growth with maximum along the <001> direction. The HAp particles form aggregates,
which average diameter is 1-5 /lm. In the interspaces between the aggregated particles,
the water remained entrapped after drying. The thermogravimetric analysis confirmed
the above mentioned conclusions. The formation of I3-Ca3(P04h took place at about
780 D e. Sintering of a unixially compressed cylindrical specimen, took place mainly in
the temperature range from 800 to 1150 DC.

Keywords: 1. Hydroxyapatite, 2. Nanoparticles, 3. Bioceramics ..

83
Y. G. Gogotsi and l. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 83--90.
© 2003 Kluwer Academic Publishers.
84

1. Introduction

Bioceramic hydroxyapatite [HAp, CalO(P04MOHh] is the primary constituent of the

human hard tissues. The artificially prepared HAp, due to the same structure of the
natural bone, has good compatibility with the human organism, and for this reason, is
widely used in medical applications as implant or as coating on prostheses [1-4]. Many
of these medical applications require highly densified implants with high enough
mechanical strength for load-bearing, which are prepared via conventional pressureless
sintering of fine ceramic powder. Their mechanical properties are largely dependent on
their microstructural features, such as sintered density, grain size, grain size distribution,
and microstructural defects. Therefore, the preparation of fine and well sinterable HAp
powder, is the most important step in the fabrication of ceramic HAp with high sintered
density and desirable microstructure. In order to prepare fine HAp powder many
chemical processing routes have been employed, including co-precipitation,
hydrothermal reactions, sol-gel synthesis, pyrolysis of aerosols and microemulsion.
Recently [5] HAp nanoparticies were prepared using rapid increase of the pH
value, by adding of concentrated NH40H solution, of a well-mixed solution of
Ca(H2P04)2·H20 and CaCh with stoichiometric atomic ratio of CaJP. This modified
precipitation method is called "pH shock wave method".

2. Experimental
2.1. EXPERIMENTAL PROCEDURE

An 800 mL solution of 0.0538 mol Ca(H 2P0 4h.H20 and 0.1254 mol CaCh with molar
ratio CaJP=1.67 (the stoichiometry of HAp) was transferred into a reactor vessel and
was heated at temperature 97±1 °c with air flow rate of 15 L·h- 1 for 30 min. The
rotation speed of stirrer was adjusted at 1000 rpm and then 10 mL of condensed
ammonium solution (25 % w/v) was added in drops during two (2) min by a syringe.
The pH of the solution was increased from 3.22 to 8.79 and white slurry was produced.
The above described modified precipitation method, we called as «pH shock wave
method». The slurry produced was aged in the above mentioned conditions for 30 min.
Subsequently, the mixture was cooled to room temperature, filtrated, washed using
distilled water and dried at 90°C for 3 h. Three parts of as prepared powder were
calcined at 550, 800 and 1400°C, respectively, for 3h.

2.2. CHARACTERIZATION

Calcium and phosphorus contents of resultant product were assayed by ICP-AES using
a Perkin Elmer, model OPTIMA 3000 spectrometer. The standard deviation of the
measurement values was -0.05ppm. XRD technique was carried out using a SIEMENS
D500 diffractometer with secondary graphite monochromator and CuKIl radiation
("-=0.15418 nm). The measured 29 range was scanned in steps of 0.03° in 10 sec/step.
The lattice parameters from the XRD data were calculated using the FlRESTAR
software [6]. The size of the crystallites responsible for the Bragg reflection of the (002)
and (300) planes were determined using the well-known Scherrer relationship [7]:
 85

d = /W(bcos8) (l)
where d is the crystallite diameter in nrn, k the shape constant (-0.9),). thf: wavelength
in nm, 8 the Bragg angle in degrees, and b the observed peak width at ha.lf-maximum
peak height in radians. FT-IR was performed using a Perkin-Elmer, SPECTRUM RX I
FT-IR spectrophotometer. The KBr disk technique was employed using 2 mg of HAp
powder in 200 mg of spectroscopic -grade KBr (MERCK), which had been dried at 100
°c. Infrared spectra were recorded in the 4000 - 400 cm· l region, with resolution of 4.0
cm ·1 .
The textural studies were performed by nitrogen adsorption porosimeter (FISONS
Instruments, model SORPTOMATIC 1900), TEM was examined under a CM20 Philips
TEM, operated at a 200 kV accelerating voltage. Differential thermal analysis (DT A)
and thermogravimetry (TG) of the powder were performed using a. NETZSCH
simultaneous TG-DTA apparatus (Model STA 449C, Jupiter) from ambient temperature
to 1400 °C in nitrogen atmosphere (flow rate 20 mL·min· l ) with 10 "C·min· l . For
sintering analysis, the as prepared powder was uniaxially pressed in a cylindrical die,
with diameter of 6 mm, under a 350 MPa compressing stress. The linear shrinkage was
determined by a NETZSCH horizontal dilatometer (Model TMAIDIL 402C), using the
same thermal cycle as the one used for TG-DTA techniques up to 1400 0c.

3. Results and discussion

The CaO and P20 5 contents were 46.07 and 39.48 % respectively, lower than the
theoretical values of pure hydroxyapatite (55.82 % CaO and 42.39 % P 20 5). The atomic
ratio of Ca/P was about 1.477, which is smaller to the theoretical one (1.67). These
results indicate the formation of nonstoichiometric hydrated HAp, whic:h formula is
Ca8.86(HP04k28(P04h.n(OH)·7.50H20 according to the literature [8]. The acidic group
content and the bounded water of the HAp powder are significant.
Figure I shows the XRD pattern of as prepared HAp powder. As it can be seen,
the XRD pattern containing strong peaks correspond characteristically to crystalline
HAp with no extraneous peaks (File No. 73-0293, International Center for Diffraction
Data, ICDD). The lattice parameters were in the range of 0.9427 ± 0.0001 nm for (a)
and 0.6865 ± 0.0001 nm for (c) axis. These results were different from the standard File
Card values of a = 0.9418 nrn and c = 0.6884 nm. This is due to the presence ofHPO/
groups in the powder, that is demonstrated in the IR characterization bellow. From the
chemical analysis and XRD results, we suggest that the rapid change of the pH value of
a well-mixed solution, results in supersaturated HAp solution and finally well
crystallized HAp powder.
The crystallite size of the HAp powder varied from 14.9 to 31.1 nm for (300) and
(002) plates respectively. The crystallite growth was anisotropic in space and the (c)
axis direction was about 2 times longer than (a) axis direction (DOO21D300 =2.09).
Figure I shows, also, the XRD patterns of calcined HAp powder at 550, 800 and
1400°C. As it can be seen, the XRD pattern of the calcined powder at 550"C, is the same
with as prepared HAp. With the increasing of calcination temperature up to 800°C, a
substantial amount of ~-tri-calcium phosphate (~-TCP, ~-Ca3(P04h File No. 70-2065,
ICDD) is recorded. The XRD pattern ofsintered specimen at 1400 °C shows that the a-
86

.§Zl

1§ "-.J\i..-,,-_.J

o
j

20 30 40 50
29

Figure 1. The X-ray diffraction patterns of as prepared (uncalcined, u.c.) and calcined
powders. The calcination temperatures are indicated. Solid square- HAp, open circle- 11-
TCP, and solid triangle- u-TCP

tri-calcium phosphate (a-TCP, a-Ca3(P04h, File No. 70-0364, ICDD) as the main
phase, proving that the ~ -form converts to a-form at 1120-1180 °C.
Figure 2 shows the FTIR absorption spectra of HAp powder, as prepared and
calcined powders at 550, 800 and 1400 °C. In general, the spectra are in accordance
with those in the literature [9, 10). The broad band at 3100-3500 cm·! corresponds to
adsorbed hydrate [II], decreases while temperature increases. The total intensity of this
band for calcined samples at 550 and 800°C is about the same, and can be attributed to
same amount of water absorbed from environment. A weak sharp peak at ~3570 cm-!
corresponds to the stretching vibration of the lattice OH- ions [II). This peak almost
disappeared at 1400°C, which indicates the minimization of the amount of HAp phase.
The weak shoulder at ~1208 cm·! and a weak peak at 868 cm·! can be assigned to p-o-
H in plane and out of plane deformation modes respectively. This fact indicates the
presence of the ionic HPO/ group, in other words, the formation of nonstoichiometric
HAp. These peaks, of course, appeared only in the uncalcined sample. The
characteristic bands for pol of HAp appear at 473,564,602,962,1033 and 1093 cm-!.
The medium sharp peak at 633cm·! can be assigned to the O-H bending deformation
mode, and indicats the presence of HAp phase. This peak decreases substantially after
calcination at 800°C, while disappeared at 1400°C. The bands for pol of the calcined
powder at 800°C appear at 430,474, 567, 604, 946, 979, 1043, 1093 and 1126 em·!,
which can be attributed to HAp, and to ~-TCP phases. The corresponding bands for
 87

14OO"e

F I I I I
4000 3500 1400 1200 1000 800 600 400

Figure 2. The FT-IR spectra of as prepared (uncalcined u.c.) and calcined powders. The
calcination temperatures are indicated.

calcined powder at 14000 e appear at 448, 474, 563, 599, 957, 986, 1027, 1045, 1063
and 1090 cm-', which can be attributed to HAp and u-Tep phases.
Typical TEM bright field micrograph of as prepared HAp sample can be seen in
the Fig.3. It is obvious that the HAp morphology varies. Particles of arbitrary shapes
and mean size of -80-200 nm in length, and of -\ 0-30 nm in width and thickness are
observed. The HAp particles form aggregates, which average diameter is 1··5 11m. In the
interspaces, between the aggregated particles, probably water remained enclosed after
drying, and could explain the significant amount of the bounded water in 1he empirical
formula of the HAp powder.
Figure 4 shows the adsorption isotherms of the as prepared HAp and the calcined
at 550 and 8000 e powders. All the adsorption isotherms have the same shape attributed
to type III being given by nonporous solids [12]. The specific surface area of as
prepared HAp powder was 74m2 •g". The adsorption isotherms of as prepared and
calcined at 550 0 e have a small hysterisis loop, which indicated the appearance of
mesopores due to the interspaces between the particles in the aggregates.
Figure 5 shows the TG and DT A curves, as well as the linear shrinkage curves of
the as prepared HAp. The thermal decomposition is very complicated procedure,
without distinguishable endothermic stages. In the temperature range from ambient to
about 300 °e the removal of absorbed and/or crystalline water are observed. The
transformation of acid-phosphate to pyro-phosphate and consequently to OItho-
88

Figure 3. TEM bright field micrograph of as prepared HAp particles.

0.0 0.2 0.4 0.6 0.8 1.0

PIPo

Figure 4. Adsorption isotherms of as prepared (uncalcined u.c.) and calcined powders.

The calcination temperatures are indicated.

phosphate occur at about 520 and 780°C, respectively. A continuously but slow weight
loss from 800 to 1400 °C is attributed firstly to the dehydroxylation and consequently
decomposition of HAp to tri-calcium phosphate. The linear expansion up to 650°C
 89

I
./\DTA
I
,.
•

i '

r-~--.-~---r--~~--~-'--~--r-~--'-~---I
o 200 400 600 800 1000 1200 1400
Temperature ('C)

Figure 5. The TO, DTA and linear shrinkage cnrves ofas prepared powder.

indicates no contraction (Fig. 5). But next, shrinkage (~LlLo) takes place mainly from
800 up to I 150°C. A sudden decrease in shrinkage, as well as a sharp endothermic peak
on DTA curve at 1200°C, can be attributed to the formation of new phases of tri-
calcium with lower density (a-Ca3(P04)2 No. 70-0364 JCDD, Dx=2.863g·cm· 3) than
HAp (No. 73-0293 JCDD, Dx=3.l47g·cm·3), that is, product with larger volume than
reactant.

Conclusion

The XRD and FTIR analyses indicated that the precipitated powder was crystallized
HAp. The HAp was stable during calcination at 550°C, while a significant amount of~­
Ca3(P04)2 appears at 800°e. The TEM micrograph showed arbitrary morphology of
nanoparticles, with average size of -80-200 nm in length, and of -10-30 nm in width
and thickness. The adsorption isotherm of HAp was a typical for nonporous particles.
The specific surface area of as prepared powder was 74 m2.g. 1• The chemical analysis
indicates the formation of nonstoichiometric hydrated HAp with atomic ratio of CaIP
equal to 1.477 and empirical formula Cas.s6(HP04)1.2s(P0 4)4.n(OH)·7.50H 20. The
lattice parameters were different from standard ones, due to the presence of HPO/·
groups in the powder. The size of the crystallites responsible for the Bragg reflection of
(002) and (300) planes were 31.1 and 14.9 nm correspondingly, which indicates
anisotropic growth with maximum along the <001> direction. The HAp particles
formed aggregates whose average diameter is 1-5 11m. In the interspaces, between the
aggregated particles, the water remained enclosed after drying. The thermogravimetric
90
analysis con finned the above mentioned conclusions. The fonnation of ~-Ca3(P04)2
took place at about 780°C. The sintering, of unixially compressed cylindrical specimen,
took place mainly at temperature range from 800 to 1150 °C.

Acknowledgements
The authors are grateful to the Ring of Laboratory Units and Centers of University of
Ioannina for Thennal Analysis.

References

l. Hench, L.L., (1991) Bioceramics: From Concept to Clinic, J. Am. Ceram. Soc., 74, 1487-1510.
2. Suchanek, W., and Yoshimura, M., (1998) Processing and Properties of Hydroxyapatite ~ Based
Biomaterials for Use as Hard Tissue Replacement Implants,J. Mater. Res., 13 [1]94-117.
3. Suchanek, W., and Yoshimura, M., (1998) Preparation of Fibrous, Porous Hydroxyapatite Ceramics
from Hydroxyapatite Whiskers, J. Am. Ceram. Soc., 81, 765-767.
4. Hong, L., Hengchang, X., and de Groot, K., (1992) Tensile Strength of the Interface Between
Hydroxyapatite and Bone, J. Biomed. Mater. Res., 26, 7-18.
5. Koumoulidis, G.e., Vaimakis, T.e., Sdoukos, A.T., Boukos, N.K., and Trap.lis, e.C., (2001)
Preparation of hydroxyapatite lathlike particles using high-speed dispersing equipment, J. Am. Ceram.
Soc., 84,1203-1208.
6. Ferguson, I. F., and Rogerson, A. H., (1984) A Program for the Derivation of Crystal Unit Cell
Parameters from X-Ray Powder Diffraction Measurements, Comput. Phys. Commun. 32, 95-107.
7. Smith, D. K. (1989) Modem Powder Diffraction, in D. L. Bish, and J. E. Post (Eds.), Reviews in
Mineralogy, Vol. 20, Mineralogical Society of America, Washington, DC, , pp.183.
8. Elliot, J.e., (1994) Structure and Chemistry of the Apatites and other Calcium Orthophosphates.
Studies in Inorganic Chemistry 18, Elsevier, Amsterdam, pp 149.
9. Yamashita, K., and Kanazawa. T., (1989) Hydroxyapatite;, in T. Kanazawa (ed), Inorganic Phosphate
Materials. Materials Science Monography, Vol. 52, Elsevier Tokyo, Japan, pp 15-54,.
10. Cheng, Z.H., Yasukawa, A., Kandori K., and Ishikawa, T., (1998) FTIR Study on Incorporation of
CO, into Calcium Hydroxyapatite, J. Chern. Soc. Faraday Trans., 94, 1501-1505.
II. Blakeslee K.e., and Condrate, R.A., (1971) Vibration Spectra of Hydrothermal Prepared
Hydroxyapatites, J. Am. Ceram. Soc., 54, 559-563.
12. Gregg, S.J., and Sing, K.S.W., (1982) Adsorption, Surface Area and Porosity, 2nd Edition, Academic
Press, London, pp 248.
BIOMINERALIZATION OF HYDROXY APA TITE I TIT ANIA
COMPOSITES PREPARED BY EXPLOSIVE SHOCK WAVES

HANAN BEHERI*, WAF A I. ABDEL - FA TT AH* AND ROLF

PRUEMMER**
* National Research Centre, Ceramics Department Dokki, Cairo, Egypt
** Ulliversitiit Karlsrhuhe, 1llstitut/iir Werksto/fkullde, Karlsruhe,
GermallY

Abstract

The detailed investigations of hydroxyapatite/titania (HA/Ti0 2) composites

including densification, structural characterization and degradation in human serum
were performed. It was concluded that the higher the firing temperature the higher
activity of the biocomposites at similar microstructural parameters. Pioneering
research of Pruemmer and coworkers proved the formation of various levels of
induced stresses and activation of alumina powders. The presenl work is dealing
with the effect of explosive shock waves on the biomineralization of HA/TiO z
composites after immersion in human serum. Two ceramic composites were
prepared and were detonated along with TiO z for comparison. The obtained samples
were immersed in human serum for 5 sequential periods between 3 and 72 hours at
room temperature without agitation. The released calcium, inorganic phosphorus
ions and protein in the serum left after withdrawal of solids were measured using
relevant biological kits and spectrophotometry at each period. The solids obtained
after 72 hours were analyzed by IR spectroscopy. The biomineralization was proved
for the higher titania content. The formation of dellouvo bone was assessed through
scanning electron microscopy (SEM) supplemented by EDAX. The formation of the
bio-layer was discussed in view of the activity product and Ca/P rario besides the Ti-
OH- species formed on immersion.

1- Introduction:

Bone is a composite material consisting of hydroxyapatite [CalO (PO~)6 (OHh1

crystals as a main phase embedded in a biologically produced organic matrix. The
immune system of the human body rejects any material foreign to the body.
Synthetic biomaterials need HA in their structure to avoid rejection by living bone.
HA ceramics are reported with osteconductivity that is capable of supporting bone
apposition and forming a chemical bond with bone [11- Certain hydroxyl groups such
as Si-OH- and Ti-OH" remaining in SiOl and TiO z prepared by sol-gel methodology
and AbO] materials were reported to promote HA generation by providing sites for
the calcium phosphate nucleation [2].
91
Y.G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 91--101.
© 2003 Kluwer Academic Publishers.
92
The HA I Ti0 2 composites are characterized by their ability to form biologically
equivalent hydroxyapatite in vitro and in vivo [3]. The formation of carbonate-
hydroxyapatite C-HA is associated with a series of chemical interactions between
these composites and human serum. The need to optimize the interaction between
ceramic and interfacial tissue has led to the development of so-called surface
reactive biomaterial, a class of bone implants designed to be anchored in bone by
establishing a physicochemical bond between HA/TiO, composites and tissue in
experimental animals (dogs) [4). When hydroxyapatite was subjected to human
serum to mimic naturally occurring reactions, calcium and phosphate ions were
leached and formed a calcium phosphate rich layer on the ceramic surface [5). This
layer is believed to impart the bone bonding ability to HA ceramic. The interfacial
interactions between human serum and the bioceramic are basically between protein
and calcium and phosphate ions. The biocompatibility of an implant is often
determined, to a large extent, by protein adsorption behaviour. Only through
understanding these interactions, and the factors that regulate them, the biochemical
properties of a material can be optimized. Protein adsorption is a dynamic process
and in the context of a biological environment is a very complex phenomenon. The
effect of surface adsorption on the protein structure attracts increasing interest in the
biological, medical and industrial fields. The structure of plasma protein adsorbed
on TiO, and HA was examined by differential scanning calorimetry (DSC). The
results allowed an estimation of the overall structure of the protein in the adsorbed
states. The electrostatic interactions were proved to be the main mechanism
controlling the adsorption of protein to TiO, and HA composites [6].
Explosive consolidation is a powder compaction process that involves high
dynamic plastic deformation of particles and densification [7]. The direct
compaction method was modified to a double tube set up where higher
compaction pressure is generated due to collision of the two tubes. The process
produces extremely high pressure within few microseconds and finely grained
material can be obtained.
The objective of the present study is to prove the bone bonding ability of the pure
TiO:, or HA/TiO:, composites and to assess applicability of the activation explosive
compaction. The in vitro evaluation of their consequent bioactivity is discussed.

2- Experimental:

Pure anatase form titania (Ts) and two HA/TiO:, composites (TI and T~) were
prepared and sintered by explosive compaction [7). The method of direct contact
between pressed samples with a detonator and explosive booster was applied at
room temperature. The explosive parameters such as mass ratio of explosive to
powder and detonation velocity were kept constant. The pressurized discs were
retrieved from experiments and examined. The composites have molar ratios of TiO,
of 0.67 and 4 forTI and T 4 , respectively, and the rest is calcined bovine HA to
compare with the previous results. The phase structure of the shocked composites
and TiO, were assessed using X-ray powder diffraction (Philips P.W. 1730) with ell
KlX target and Ni filter at scanning speed of28 =2 o/min The biocompatibility of the
samples was assessed by the extent of their degradation in human serum up to 72
 93

hours. The serum left after withdrawal of the solids was analyzed using UV -VIS
spectrophotometry (Jenway 6100, England), and biochemical kits for Ca, inorganic
phosphorus (lP) and total protein after each period of immersion (3, 6, 24, 48 and 72
hours) at room temperature [5]. The solids withdrawn were washed, dried and kept
for analysis. The microstructure of Ti0 2 and composites pre- and post-immersion
were examined using SEM (Philips S.L. with EDAX voltage o30keV). The infrared
(IR) spectral analysis (lR PU 97/2 Philips) was used to supplement biological data
and to characterize the functional groups of the samples pre-and post- ill vitro test,
using the K Br disc technique (2mg of sample to 198mg of K Br) in the range 4000-
200cm- l . The optical density (O.D.) was calculated for all functional groups
applying Beer Lambert's law.

3. Results and Discussion:

3.1. Phase analysis:

The main phase of the shock wave activated S WA Ts sample is anatase, recorded
at d=1.89A, 2.38A and 3.52A (Fig. I) along with rutile, the other phase of Ti0 2,
recorded at d= 3.2SA, 2.48A and 1.69A. This is a proof of the phase transformation
of the metastable (anatase) to the stable rutile form as a result of explosive
preparation. It was proved that pulsed compression revealed polymorphous
transformation which tended to increase the yield of the more stable modification
[8]. The composites TI and T.j have the main phase hydroxyapatite at d=2.81A.
2.73A. and 2.79A along with anatase as the second phase. As the rutile phase is not
detected in any of them. it seems that HA prevents the phase transition of Ti0 2 in the
composites and stabilized its anatase form.

noz(s) o T;02 (Anat"'"')

aA = 3.52,1.89, 2.3B
A 0
A T;Oz (RutH.. )
dOA = 3.Z5, 1.69, Z .9

I
,.
.;;;
c
~

1
70 60 50 .0 30

Fig. I. XRD Tracings of Shock Wave Activated Ts

94
3.2 Biochemical analysis:

3.2.1. Control Serum:

The values of Ca"-, IP and total protein in the control serum are within the
nonnal ranges throughout the whole measuring periods (Fig. 2).

E
0
B
0
~

".

• •
C.

• •
6
5 ~. ..,.. ~ .. ~
..
c
9 .- .f
7. . .",
! 1--';;';';:_·'
..."
E
41 2 -If:)
<J
c: :t.u.
0
y
0
-11"<.

:l.
3 6 24 48 72
Tim~ (hvlifsj

E
0
C>

~ 6
8

... ~.--
·f" f
.", • f
c: -"LI•
.9
~ .j.
C 4 !'-'-"_ent"d
'"
u
C
I".>;
0 ~ ·4'.!O,
V ;','5:
Q. 2
3 6 24 48 72
T;ffie (hOi...i'S)

14
12

t"· .. <to • •
c:
.g
:!
-40
t -

C §a
'" ~6
v
C
0 ~·""1'1I1

V
~
54 r ~~r;>

'"
U 2 flUI

0
3 6 24 48 72
Time jhoursl

Fig. 2. Biochemical Behavior of BA ComposItes Compared to Control Serum

 95

3.2.2 Tile serum after immersion:

The Ca2~ values are significantly reduced in the serum left after withdrawal of
activated samples being Ts< T~ samples, being lower in the serum in contact with Ts
than in the serum in contact with T~ (8.56 and 9.19 mgl IOOml respectively).
Therefore, calcium deposition is more pronounced for the activated Ts. The IP levels
have a similar trend to Ca2~. On the other hand, the Ca'~ level corresponding to the
activated T I (with reduced Ti0 2%) revealed higher Ca2~ values in serum compared to
the others, particularly at 3 and 6 hours immersion, denoting its initial limited
degradation. The values of IP are higher at all periods of immersion except for 3
hours immersion denoting the release of IP. The order of higher Ca2~, and IP
deposition decreased in the row Ts> T ~> T I especially after 72 hours immersion
proving the higher bioactivity of Ts (anatase and rutile composite). Therefore the
activity of the samples is controlled by the presence of Ti0 2 and its content.
The protein (Pr) values are reduced in the serum for the Ts sample revealing
continuous deposition as early as at 3 hours after immersion which continued up to
72 hours. The composite T ~ also deposited proteins but to a lower extent than T, at
times up to 24 hours. The complement factors were denoted to be adsorbed by the
surface via two routes. [i] Directly to the crystal surface and bound to Ig G absorbed
to the surface, [9] and [ii] further binding of the complement filctors to calcium
phosphate implants by releasing ions, which react with various proteins and denature
them [10].

3.3 Surface Morphology:

SEM of shock wave activated Ts preimmersion reveals a very dense structure of

rounded particles. In some other areas, there are several fused grains along with
agglomerates of smaller ones. The biologically activated (BA) Ts withdrawn from
serum after 72 hours reveals network of branches along with spongy nodules
denoting the new mineral deposition (Fig.3a). Similar structure was reported by
Ravaglioli et al [11]. The nodules reveal needles in a flower shape and spongy
deposition inside the flower (Fig. 3b). Its EDAX reveals Ti element as the major
component along with Ca and P (Fig. 3c).
The SWA composite TI possesses a collection of several morphologies of
rounded grains, needles and some grown plates embedded in the surface. The BA
composite T I reveals the collection of hexagons, plates and globules of rounded
particles (Figs. 3. d, e). EDAX results support typical Ca/P ratio for the apatite
associated. Traces of silicon and chlorine originate from the serum. The presence of
Ti is also indicated (Fig. 3f).
The SWA composite T~ possesses highly dense structure with fused rounded
grains and some agglomerated globules, with high optical density and with light
aggregated globules of Ti0 2. The BA composite T~ possesses a spongy structure of
initial minerlization (Fig. 3g). Some granules are coated with a thin film of protein
along with filaments of well developed crystalline phase (Fig. 3h). EDAX proves
that the structure is enriched with the characteristic Ca/P ratio of apatite and
comparable Ti (Fig. 3k).
96

(e) (d)

fb)

(e)

(a)

Figs. 3. SEM and EDAX ofTs (a, b, c) and T] (d, e. f) samples

 97

(h)

(g) (k)

Fig. 3 SEM (g. k) and EDAX (h) ofT, BA Samples

3.4 IR Analysis:

3.4.1 Ts composite
The newly formed carbonated apatite covers the S WA sample, especially the Ts.
This was proven by the specific bands of OR at A = 3692,3571 and 633 cm-' CO,""
bands at A = 1550 and 1460 cm-' and the characteristic bands of apatite structure at A
01050,602 and 570 cm-' for pot stretching and bending mode of vibrations (Fig. 4)
[12]. Therefore, the extensively precipitated new C-HA on the surface denotes the
biomineralization from serum.

3.4.2 TJ Composite:
The biomineralization of SWA T" records a peak associated with an organic
phase, i.e. the peak at A 0 1650 cm-' which is likely to be due to the amide I along
98

with the bands attributed to collagen (combination of C=O stretching, C-N

stretching and N-H bending modes [13]. However, an overlap with H20 species A. 0

1600 cm- I has to be considered [14]. An increase of 0.0. of OH- bending mode
vibration occurs in the low energy region at A. 632cm- l . All these changes indicate
0

the modification of the original structure and the formation of a new carbonyl
apatite precipitate on the TI composite. The OH- stretching mode vibration appears
at A. = 3694 cm- I for the original structure, and shifts to 3671 cm- I denoting
reorientation of OH- due to the interactions of original structure with the serum (Fig
5).

co/" groups: The 0.0 of the CO/ stretching mode in the high energy region
between A. = 1414 -1460 cm- I increases compared with the original composite pre-
immersion. The weak peak of CO/" out of plane of bending mode at A. = 877 cm- I
shifted to A. =880cm-l_ A new band of CO/", symmetric stretching mode of vibration
is recorded at A. = 1550cm- 1 denoting the new deposition on the activated surface of
type A carbonated apatite, therefore replacing OH- by new CO/ [15].
Phosphate groups: The asymmetric stretching mode of PO/- U3 between A. = 1057-
1092 cm- I is shifted from I, =1120 to 1058 cm- I denoting the reorientation of PO/-
due to the interaction with the original surface. Also this region becomes broad for
the immersed composite denoting partial dissolution. The band at A. = 692 cm- I is
assigned to the asymmetric stretching mode of PO/" of U I. The bands of U3 bending
mode of PO/ vibrations recorded at A. = 603 and 575cm- 1 are slightly shifted with
higher 0.0. after immersion for 72 hours.

SWA

" (5)

QH- 3694, 3511,632 em-I

CO u.60. 1!o1L.,677
PO 1092. 1057. 962, 603, 569 .

After

OH' ]510.31.03, 633 em-I

BA cot
Amid 2927. 2368, 2002, 1656 ..
1550. 1460, 1411.. ..
PO,('OS8. 602. 569 ..

3000

Fig. 4. IR Spectra of SW /I and BA Ts Composite

 99

SWA

T "I.

TiOZ A = 687.606. 586.567. 538. 445.341 eni l

ow A = 3692.3570. 633. em-I

NH- A= 2367. 1650 ..
BA CO\A= 15,,0
PO 4 ,\ = 10"9. 60Z. 570

4000 3000 zooo 1000

Fig 5. IR spectra ofSWA and BA T, Composite

3.4.3 T~ Composite:
Hydroxyl groups: The SWA composite T~ has a band assigned to the asymmetric
stretching vibration mode of the OH- groups recorded at A = 3570-3364 cm- I • and
after immersion it shifts to A = 3569 cm- l with increasing 0.0. The 0.0. of the band
at A 632 em- l is reduced after immersion denoting that the environment of OH- ions
e

at the composite surface leads to minerlization due to its interaction with the serum.
The asymmetric carbonate stretching mode of vibration of cot at A = 1460 and
1412 cm- l is recorded before and after in-vitro test respectively indicating the
modification of original structure. (Fig 6).

Organic Species : The peak associated with organic phase is recorded at A=

1654cm- ' and is likely to be due to the amide I band of collagen, (a combination of
c=o and C-N stretching, and N-H bending mode) denoting the biomineralization of
dellouvo bone on the surface [16J.

Phosphate groups: The bands between A = 1091 to 962 cm- I with the center at A =
1049 cm- I are assigned to the asymmetric stretching mode of PO/, \)3 . A weak
shoulder at A - 962 cm- I is assigned to the asymmetric mode of stretching of PO.t.
This band becomes broader especially at its base with high 0.0. values due to
biomineralization. The PO/, bending mode of vibration in the low energy region of
A between 603 and 571 cm- I also becomes broad with reduced 0.0. after ill vitro
test due to leached ions.
100

Table (1) Compaction data, activity product and CalP ratio of the composites
after 72 hour immersion
;
Samples Ti02~/O *p g/cm' -%Pth p g/cm [Ca][IPl-\ CallP ratio
T, 100 209 55.38 073 102 128
T, 5.07 178 55.47 076 262 103
T, 24.17 1.99 60.12 076 1.44 144
Serum - - 1.63 1.63
*p = denSity of samples, -% P th is based on mixture rule, p = density of explosive,
~ = IO-6 M2

Therefore, the sites of P0 4 3 - react differently when subjected to the serum

environment. where the bending P0 4 3 - has reduced 0.0 .. while stretching P0 4 ,- has
increased 0.0,

OW 3570. 3425,632 em-I

SWA CO 1460.1415 ..
T"'" PO 1091,1048,961,602,571,413 ..

After

BA OW 3569,3312,632 em-I
A~de 2921,1655 ..
C0:,i 1550, 1460 ..
P0I;- 1091, 1048,961, 602, 571,413 ..
A,cm- I
)000 2000 tOOO cnl l

Fig.6. IR-Spectra ofT, Composite

An essential requirement for an artificial material to show bioactivity is the

formation of a biologically active bone-like-apatite layer on its surface in the body.
It was reported that the dissolution process of HA after implantation is very sensitive
to the defect structure and the control of the density as well as morphology of those
lattice defects is important [17]. As the consolidation parameters in the present
experiments are constant, individual differences in lattice defects and stored energy
are functions of material characteristics. Therefore the internal energy of the
compacted material should be increased because the shock wave induces a large
 101

density of lattice defects leading to enhanced rigidity of the substances. The stored
crystal strain after consolidation and stored energy are expected to affect the
degradation of the studied biocomposites.
The crystallization of HA occurs when ionic the activity product is 3 x 10-6 !vf
and the Ca 2 - lIP ratio = 2. Hence no crystallization is recorded in serum [18]. The
ionic activity products (Ca 2- & IP) for bioactivated composites (Table I) reduce with
the time of immersion, but they are constantly of higher value for the T I composite
compared to others. The crystallization of the surface with HA is expected and
confirmed by SEM and IR for bioactivated Ti0 2 . Intermediate lower values for Ts
and T ~ composites at longer immersion times are a consequence of microstructure
and environment of the serum as a result of ion exchange equilibria.
The oscillating phenomenon of precipitation and dissolution of the in vitro Ca 2 -j-
and IP are due to metastable phases formed before precipitation of another
subsequent Ca/P phase. The phase transformation of Ts partially from anatase to
rutile phase is expected to assist to chemical bonding. It was reported that in
mixtures of HA and Ti0 2 conventionally prepared, Ti0 2 reacts with hydroxyapatite.
Titanium enters the lattice in two forms Ti 3 - or Ti 4 - in cationic site as well as Ti l \
oxy-anion groups in anionic position for PO/-. The introduction of titanium is
connected to the large number of Ti-OH groups remaining in HA. The activated
Ti0 2 attracts calcium and phosphate from the body fluids and forms biologically
equivalent HA on its surface which is detected by IR and SEM. The surface
transformation to HA ensures that the activated Ti0 2 binds with the bone. with the
possibility to realize the charge balance.

4. Conclusion:

The present work was designed to induce high stress concentration and stored
energy via explosive compaction to an osteoconductive material. Additionally
biofunctionalization of the surfaces with serum protein was followed to assess the
deposition of a bone-like apatite by biomimetic approach. Upon detonation, anatase
transformed into a composite of anatase ! rutile proving induced high stress and
stored energy. Serum immersion proved epitaxial layer of C-HA on anatase/ rutile
composite more than anataselHA composites_ Such layer is a function of higher
Ti0 2 content which was justified by the rich protein layer and HA deposition. These
are proved by SEM/EDAX and IR detection of CO/" - apatite as well as the reduced
activity product of calcium and phosphorus in serum at the higher titania content.
The organization of collagen fibrils and their mineralization represent a main sign of
bone fracture repair.

Acknowledgement:

The authors acknowledge with appreciation the support for this research by the
National Research Centre in Cairo. Egypt and Forschungszentrum JUlich Gmbh,
Germany through the bilateral scientific agreement since 1998.

References:

I. deGroot, K. (1981) Bioceramics consisting of calcium phosphate salts. J

102

Biomateria!s 1,47-50.
2. Li, P., Kokubo, T., Nakanish, K. Soga, N. and deGroot, K. (1994) "A role
of hydrated silica, titania and alumina in forming biologically active bone
like apatite on imp!ant" J Biomed Mater, Res. 28, 7-15.
3. Wafa 1. Abdel-Fattah, Hanan H. Beheri, Lobna Agameya and Imanzaki
(2001) "Triphasic phosphate / titanate composites for preprosthetic socket
augmentation (Experiment study) Key Engineering Materials vol 192-
195.
4. Wesam Zahran K. (2002) Evaluation of hydroxyapatite blocks as bone
graft substite in reconstructive sur gery experimental study dogs, Master,
Oral surgery. Fac. Of Desntistry, Alex. Unrv.
5. Wafa Abdel Fattah and Hanan. H. Ben (J 998) "New Biomaterials, Basic
and Applied studies, Gos press, Amsterdam, P. 91.
6. Chen-YL, Zhan~ - XF, Gong, YD, Zhao, NM, Zcng TX, Song, XQ
(1999) J C%id and lnterfa Science 214, 38-45.
7. Kochiek D., Prummer, R., Brunold A. 1994. Explosive reaction pressing
of intermetallic compounds from stoichiometric powder mixtures.
ceramic processing science and technology, Edit H. Hausner, G.L.
Messing and S. Hirano. Ceramic Trans. Vol, 51,615-619.
8. Ivanov, V., Paranin, S., Khrustov, V., Medvedev, A., Shtots, A.
"Processing of nanostructured oxide ceramics with magnetic pulsed
compaction technique" (2002) Key Eng. Mal. 206-213,377-380.
9. Hulbert. S.F.. Morvison. S1. and Klewittcr. JJ. "Tissue reaction to three
ceramics of porous and non porous structures" (1972) 1.Biomed. Mat.
Rcs; 6, 347-374.
10. Kochwa, S., Litwak. R5., Rosenfeld. R.E. and Leonard, E.F. "Blood
elements at foreign surfaces: a biochemical approach to the study of the
adsorption of plasma proteins" (1997) Ann NY Acad Sci.,283,37-49.
11. Ravaglioli, A. and Krajcewski, A. (1992) Bioceramic Materials properties
and Applications", Chapman and Hall 18.
12. Doyle B.B., Bcndit, E.G. and Blout. E.K. (1975) Biopo(ymer 14,933-957.
13. Nadir,S., Lacout, JL, Ferhat, M. (1991) J de. Ch. et Ph. Ch. Bio!. 88,1999-
1924.
14. Li, P. Ohtsuh C, Kokubo. T.. Nakanishi, K., Soga, N .. Kanamura, T.
and Yamamuro, T. (1992) "Apatite formation induced by silica gel in a
simulated body fluid. I. Am. Ceramic. Soc.75, 2094-97.
15. Dillavcnkatesa, A., Condrate. R.A. (1998) "Sol-gel processing of
hydroxyapatite" 1. Materials science 33, 4111-4119.
16. Condrate, R.A. (1999) private communications.
17. Daculsi, G., Lcqeros S.P. and Mitre, D. (1989) Calc. Tissue Int. 4595.
18. Tak Cheng PEL And Pritzkcr. K.P.H "Solution CaiP ratio affects calcium
phosphate - crystal phases" (1983) Calcif. Tissue. Int. 35, 596.
PROSPECTS OF USAGE OF MATERIALS WITH FULLERENE COATINGS
FOR ENDOPROSTHESIS OF JOINTS

LASHNEVA V.V., TKACHENKO YU.G., DUBOK V.A., SCHUR D.V.,

SYCHEVV.v.
Frantcevich Institute for Problems of Materials Science of NASU, Kiev,
Ukraine,
Matveeva L.A.
Institute of Semiconductors Physics of NASU, Kiev, Ukraine

1. Introduction

Research of physical-chemical properties of fullerene based materials which

was actively conducted last year has opened a lot of new possibilities for practical
applications of these materials. Among the most promising applications of these
materials is its usage as coatings in fricton pairs to increase wear resistance [1] that is
the topical task for some medical devices.
Modem design of hip prosthesis is usually based on usage of titanium alloy Ti 6AI4V
and polyethylene with superhigh molecular weight «Chirulen». Friction pair of such
endoprosthesis consists of titanium head and polyethylene cup. One of the shortcomings
of the endoprosthesis is insufficient tribological properties of the device. For example,
polyethylene wearing of such friction pair makes from 75 up to 250 Iffil per year [2].
This process is accompanied by derivation of wearing products, which are collected in
ambient tissues and cause inflammatory, allerginic and other negative processes in an
organism, and also displacement wearing surfaces relative to one another. It results in
derivation of undesirable gap between the surfaces, and further - to loosening of the
endoprosthesis and loss of its functional properties. Therefore, due to loosening, 3-5
years after operation almost all existing types of hip prosthesis should be substituted
completely or need replacements of some components. Wearing rate of components in
ceramics - polyethylene friction pair is about an order of magnitude lower. But due to
low ceramics toughness usage of ceramic heads in hip prosthesis is limited [3-5].
To improve tribological properties the ceramic - polyethylene friction pairs are used
where the wearing of components is about tenfold lower. However due to low
toughness usage of ceramic heads is limited. Therefore search of new designs and
materials for manufacturing of friction pairs gains the increasing urgency.
This study is dealing with fullerene based materials as coatings to wearing surfaces of
hip endoprosthesis. The physico-chemical properties of fullerene C6Q on coating on Ti
6AI 4V titanium alloy were studied and wear resistance of polyethylene «Chirulen»
coupled with this alloy coated or uncoated by fullerene was measured. For comparison
the wear was measured for pair Al 203 medical ceramics - «Chirulem>.
103
Y. G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 103-109.
© 2003 Kluwer Academic Publishers.
104

2. Materials and methods

The high purity C60 fullerene powder was produced by the salting-out method described
at [6]. Fullerene-containing soot was made by electric arc method by the evaporating of
graphite electrodes in helium. Fullerene separation of soot was carried out in continuous
Socklet reactor. As extractive solvents were used high-purity benzol and toluene. The
fullerene salting-out was made by ethyl alcohol from solutions, which contain 1.2 - 3.0
g of fullerene in I I of solution. Produced sediment was washed out by ethyl alcohol
°
several times, then dried in vacuum furnace at 250 C, then again dissoluted in high
purity toluene and salted-out. The toluene for final dissolution was three times distilled.
The described method permits to manufacture high purity fullerite C60 powder without
solvent impurities.
Fullerene coating was made by thermal evaporation of the produced C60 powder in
vacuum on nonheated polished Ti 6AI 4V substrate from effusion tantalum cell at 470
°c.
Samples of high density AI 203 ceramic in the form of disks with diameter 25 mm and
thickness 4 mm were produced by semi-dry pressing with uniaxial two-side load at
pressure 100 MPa followed by sintering in a vacuum at 1650 °c during I hour [7].
Crystal structure of the coating was studied by XRD method with Cu Ka radiation by
two-crystal spectrometer. To analyze the coating nanostructure the Raman spectroscopy
was used. The measurements were made by dual monochromator ,n:<l>C -24 at room
temperature. To excite Raman spectra the radiation of Ar+- laser was used at wavelength
488 nm. The signal was registered by cooled photomultiplier in photon counting mode.
The composition of fullerene coating was studied also by FTIR spectroscopy. The
reflection spectra were measured at wavelength 400 -5000 cm,l with the help of
standard reflection attachment for Fourier - spectrometer Bruker IFS-48. The light
angle was 16,5 0
•

The nanomorphology of fullerene coating surface and its microrelief were studied by
atomic-force microscopy with Nanoscop Sha instrument in periodical mode regime. The
silicium needles were used with rounded radius less than 10 nm.
To measure adhesion strength of the coatings the method of normal tearing was used.
To apply tearing force the steel cylindrical rods were glued to the surface of coating by
epoxy resin.
The thickness of the coating was measured by interferometer MI1I1-4 and also from
profilograms.
Wear testing was made with pin-on-disk device [8].
According to this method the load which creates contact force is applied to a fixed
cylindrical rod and the disk (the rider) is rotating with constant speed. Thus at
unidirectional rotary motion the alternating deformations can be realized for surface
layers due to the difference in the dimensions of contacting bodies [8]. This method
does not simulate the process which takes place during motion of an endoprosthesis
joint but can be used for comparative assessment of wear of different materials rubbing
against polyethylene rod.
The measurements were conducted at such conditions: constant load - 5 N/mm2, slip
 105

velocity - 0.1 mis, environment - body simulating fluid with NaCI content of 9 gil
(Ringer solution), temperature - 37 °c, The wear was measured by volume of
polyethylene, which was lost due to friction after every 5-hour's cycle of friction.
Simultaneously the weight wear of AI 20 3 and linear wear ofTi 6AI 4V disks (the riders)
were controlled with C 60 coating and without coating. The total term of wear tests was
20 hours.

3, Results and discussion

Initial roughness of studied samples were R. ~ 0,05 !ill1 due to manual grinding and
polishing of the samples with abrasive and diamond pastes.
The thickness of fullerene coating was 1 -1.5 !ill1.
XRD pattern of fullerene coating on Ti 6AI 4V alloy indicated the presence of face-
centered cubic (FCC) lattice in the film. The narrow intensive line, which was registered
at the angle 28 = 32,78° of the diffractogrilIll, can be confidently attributed to the FCC
lattice of fullerite.
Raman spectrum of the fullerene coating is presented at Fig. 1.

1.4

1,2

g 1,0

~
0,8

~ 0,6

'~
0,4
.E
0,2

0,0

1100 1200 1300 1400 1500 1600 1700 1800

Raman shitf, cm,1

Figure 1. Raman spectrum of the fullerene coating

In the high-frequency area of the spectrum there are lines at 1424, 1472 and 1580 cm'\
which certainly can be related to vibratory modes of fullerene molecules with a
symmetry Hg, Ag and Hg accordingly [10,11], It is necessary to mark considerable (3
and 4 cm'l) high-frequency shift of vibratory modes Ag and Hg. Besides the Raman
spectrum, additional lines were found at 1186, 1232, 1448, 1540 1566 cm'l.
106

FTIR-reflection-spectrum offullerene coating on Ti 6Al4V alloy is submitted at Fig. 2.

1,0

>.
+-' 0,9
"en
C
Q)
+-' 0,8
C

0,7

0,6

0,5
0 1000 2000 3000 4000 5000 6000
-1
V, em

Figure 2. FTIR-reflection spectrum offullerene coating on Ti 6AI4V alloy

The spectrum contains lines at 527, 576, 1183 and 1429 cm,l that are reference for
fullerene C60 [5J. Besides the intensive additional lines were observed at 610, 740, 888
and 1107 cm'l. Appearance of additional vibrational bands in Raman and FTIR-
reflection spectra can be attributed to forming of carbides of the metal.
Results of investigation of the coating with atomic force microscope are submitted at
Fig. 3a and 3b.
 107

%, Y - 05ll1l!kill
Z "10K rno :!l!n

Figure 3a. Atomic force three-dimensional image ofnanomorphology ofa surface of the fullerene coating,

Section analysis
alft

45.D

-45.0
o 1.DO 2.DO

Figure 3b. Atomic force surface microrelief of the fullerene coating,

108

Fig.3 (a) presented the three-dimensional image of nanomorphology of a surface of the

fullerene coating, and fig. 3 (b) a microrelief of the surface. From a fig. 3 it is visible,
that the surface of the fullerene coating is homogeneous and smooth enough. The
altitude of a grain does not exceed 20 nm. The adhesion strength of the coating on Ti
6AI 4V alloy is rather high also and makes 0,8 GPa, that is comparable to adhesion
strength of 0,84 GPa, found for fullerene coatings on titanium alloy 90 % Ti + 6 % Al +
4 % V [12].
At FigA the results are presented of measurements of polyethylene "Chirulen" wear in
contact with titanium alloy Ti 6AI 4V without coating (curve I) and the alloy coated
with C60 fullerene (curve 2), as well as wear in friction pair "Chirulen" - AI 20 3 ceramics
(curve 3).

0.040
v[
0,D35
...__0-
v
E 0.030 v
"ro ~~ I
0.025
Q)
/
V - 0 - Ti(BT6)
~ 0.020 f--- f--- /'
I---
c: - 0 - BT6+C60)
Q) 0/
>. 0.015 -- - A - AI203 ceramics f---
J::
Q;
~ 0.010
o
a..
0.005
0
A
Q-
0,000

10 12 14 16 18 20 22

Term of friction, hours

Figure 4_ «Chirulem> polyethylene wear (cml) due to friction with different materials

Presented results reveal that titanium alloy with fullerene C60 coatings has
approximately the same wear resistance as polished alumina ceramics and about tenfold
more than wear resistance of the the same alloy without coating,
It is necessary to note that process of wearing for titanium alloy Ti 6AI 4V
without coating was accompanied by considerable increasing of roughness of contacting
surfaces, blackening of the friction track on titanium disk and its wearing on about 5 - 8
~m during every cycle of testing. Simultaneously the black embedding appears on the
contacting surface of polyethylene rod, which regrettably was not investigated.
The least wear was found after friction with ceramic disk. Due to high
hardness of AI 20 3 based ceramics its wear was not found. Roughness of the both
coupled surfaces was closed to the one found for the initial surfaces.
 109

The wear at the friction pair - Ti 6Al 4V alloy coated by C60 was approximately
tenfold less than the one for the same pair, but without the fullerene coating and was
comparable to the wear in the pair polyethylene - AhO) based ceramics. After friction
during 20 hours the surface roughness of polyethylene and the C60 coated Ti 6AI 4V
disk were preserved almost the same as before friction. There were also no surface
scratches, alteration of color or diminishing of the thickness ofTi 6A14V.
Thus after clinical tests the titanium alloys with fullerene C60 coatings can be
successfully used in friction pairs of endoprosthesis of different joints: hip, knee,
shoulder joints, etc.

References
I. Jazrawi, L.M., Kummer, F.J.and DiCesare, P.E. (1998) Alternative bearing surfaces
for total joint arthroplasty, J. Am.Acad.Orthop.Surg. 6, 198-203.
2. Sherepo, A.M. (1988) Wear of total endprosthesis' heads, Orthpaedy, traumatology
and prosthesis No.6, 25-27.
3. Wenda K., Ritter C., Rudig L. (1993) Zur Therapie bei einem geplatzten
Huftprothesen-Keramikopf, Akt. Traumatol. 23, No.6, 294-296.
4. Simon J.A., Dayan AJ., Ergas E., Stuchin S.A., DiCesare P.E. (1998) Catastrophic
faillure of the acetabular component in a ceramic-polyethylene bearing total hip
arthroplasty J. Arthroplasty No.13, 103-113.
5. Schur D.V., Dubovoi A.G., Anikina N.S. et al. (2001) The production of utrafine
powders of fullerites by the salting out method, in ICHMS'200I VII International
Conference «Hydrogen materials Science and Chemistry of metal Hydrides», 16-22
September 2001, Alushta-Crimea-Ukraina, pp. 478-484.
6. Lashneva V.V., Kryuchkov Yu.N., Sohan S.V. (1998) Alumina based bioceramics,
Steklo and keramika No. 11, 26-28.
7. Coli B.T., Jaquit P. (1988) Surface modification of medical implants and surgical
devices using TiN layers, Suiface and coatings technology, 36,867-878.
8. Evdokimov V. D. (1977) Reversibility offriction and machines' quality, "Technika",
Kiev.
9. Vasin L.A., Matveeva L.A., Yukhimchuk V.A. and Shpilevskii E.M. (2001) The
structure of fullerene C60 films modified by helium glow discharge plasma, Technical
Physics Letters 72, 65-75.
10. Makarova T.L. (2001) Electrical and optical properties of monomer and
polymerized fullerenes, Physics and technics ofsemiconductors, 35, No.5, 257-293.
II. Shpilevski M.E., Shpilevski E.M., Stelmah V.F.and Matveeva L.A. (2001)
Titanium-fullerene coatings for endoprosthesis, in Fullerene and fullerene-containing
materials, BGU, Minsk, pp.106-116.
12. Vul A. (2001) Fullerene applications. Last news, in ICHMS'200i, Vii International
Conference «Hydrogen materials Science and Chemistry of metal Hydrides», 16-22
September 2001, Alushta-Crimea-Ukraina, pp. 496-497.
NANODISPERSED DIAMOND ADSORBENTS
FOR BIOLOGICAL SOLUTION CLEANING

G.P. BOGATYREVA, M.A. MARINICH, N.A.OLEYNIK,

G .A.BAZALIY
V.N. Bakul Institute for Superhard Materials of the National
Academy of Sciences of Ukraine,
2 Avtozavodskaya St., Kiev, 04074, Ukraine

1. ABSTRACT

Modified nanodispersed diamonds are efficient adsorbent to clean the

blood plasma of toxins of organic and inorganic origins. Nanodispersed
diamond adsorbents are characterized by a selective adsorption
mechanism: a hydrophobic surface mainly adsorbs cations of heavy
metals and a hydrophilic surface adsorbs organic molecules.

2. Keywords

Diamond, adsorbents, modification, thermal and electrochemical

treatment, biological solution.

3. INTRODUCTION

To remove toxins of organic and inorganic origins from blood plasma,

coal adsorbents are used [I, 2]. The main disadvantage of these adsorb-
ents is the impossibility to prevent their surface layer from failure, which
can result in impermissible pollution of blood. It is known that nanodis-
III
Y. G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 111-118.
© 2003 Kluwer Academic Publishers.
112

persed diamond powders unlike coal adsorbents are chemically inert,

stable and can be easily recovered.
A large body of research carried out at the Institute for Superhard Mate-
rials have shown that diamond possesses unique adsorption properties [3
- 5] and that a purposeful modification of nanodispersed diamond pow-
der surface affects essentially physicochemical properties of diamond [6,
7].
Our present work has been aimed at developing modified nanodispersed
diamond adsorbents to cleanse biological solutions of both organic and
inorganic toxins. To modify the diamond adsorbent surface, thermal and
electrochemical methods have been used.

4. EXPERIMENTAL

Experiments were carried out on nanodispersed diamonds films. The

films were prepared by applying a mixture of diamonds with fluoroplas-
tic lacquer in the ratio 10: I (the lacquer being dissolved in acetone) to
an isotropic pyrographite substrate, thus making an electrode. In an elec-
trochemical cell, this electrode is a cathode. A platinum electrode was
used as an auxiliary one, and a chlorine-silver as a reference electrode.
The experements were carried out on a P-5848 potentiostat in an argon
atmosphere at room temperature, the rate of the potential increment
being 0,5 mV/s. A 0,9% NaCI solution has been used as a working elec-
trolyte. Diamond adsorbents were subjected to special chemical and
electrochemical treatments.
The adsorption activities of diamonds were evaluated by measuring the
solution conductance and cathode potentiodynamic pulses. The latter
method consists in assessing variations in charge spent for adsorption of
hydrogen when additives of organic and inorganic origins are introduced
into the base solution. The basic indicator of adsorption activity of a
diamond adsorbent is the occupancy of the adsorbent surface (8) by or-
ganic and inorganic substances.
 113

where QI is the charge spent for hydrogen adsorption from the base solu-
tion, Q2 is the charge spent for hydrogen adsorption from the base solu-
tion with an organic or inorganic substance added.
Thermodesorption of water was assessed from the mass- and thermode-
sorption spectra in the 20 - 600 °c temperature range taken on a MU
1201 mass-spectrometer.
To obtain a hydrophobic surface, diamond adsorbent was heat-treated in
a hydrogen flow. Hydrophilic surface was produced by electrochemical
treatment using platinum electrodes. The treatment involved the action
of active substances formed in the electrolysis of low-concentration so-
lutions of hydrochloric acid with an addition of hydrogen peroxide.

5. RESULTS AND DISCUSSION

We have studied the adsorption of ions of the 0.9% NaCl solution as

well as components of blood plasma: potassium chloride, organic (albu-
min, phenol, ethyl alcohol, acetone, glucose) and inorganic (chromium,
iron, nickel, lead) toxins in the following concentrations: KCI 0.9%, al-
bumin 10%, phenol 0.01%, ethyl alcohol 0.01%, glucose O. 1%, ions of
hexavalent chromium I 1 x 10-3 mol/I, ions of iron 1 x 10-3 mol/I, ions of
nickel 1 x 10-3 mol/I. ions of lead 1 x 10-6 molll.

The adsorption activity of nanodispersed diamond adsorbents for organic

substances under study decreases in the following series: phenol, albu-
min, ethyl alcohol, acetone, and glucose and for ions of heavy metals in
the series: iron, nickel, hexavalent chromium and lead.

I The solution for studying the adsorption of hexavalent chromium ions was prepared

by dissolving Cr03 in the 0.9% NaCI solution. In our opinion, the adsorption process
has several studies: (a) adsorption ofCr20/ ions, (b) formation ofH30+ from Cr20l-
and hydrogen, (c) adsorption ofCr/.
114

100

80

60
-;!2..
CD~

40

20

2 3 4
Figure 1. Occupancies of nanodispersed diamond adsorbent surface by
toxins of biological media: 1 - albumin; 2 - ions of hexavalent chro-
mium; 3 - ions of nickel; 4 - ions of iron.
c=:J Initial surface; _ hydrophobic surface;
~ hydrophilic surface

We have also studied the effect of the hydrophilicity-hydrophobicity ra-

tio on the adsorption properties of diamond.
Figure 1 shows the occupancies e of the diamond adsorbent surfaces by
the studied substances, including toxins of organic and inorganic origins
in the concentration of albumin 10% and of ions of heavy metals 1 x 10-3
moUl. It is seen from the figure that ions of heavy metals are adsorbed
by hydrophobic surfaces (positions 2, 3, 4), while albumin - by the
hydrophilic surface (position 1). Thus, the type of the adsorbent surface
modification affects essentially the selective nature of adsorption. For
ions of lead, the occupancy e ranges from 30 to 50%.
 115

3,0

2,4

1,8 3

..
1.2

0.6
~,."
o lOO 200 300 400 500 600
Temperature, °c

Figure 2. Thermodesorption spectra of water from the surface: 1 - the

initial nanodispersed diamond, 2 - the diamond after thermal treatment,
3 - the diamond after electrochemical modification.

The heat treatment removes impurity atoms and groups from active sur-
face centers and the diamond surface acquires hydrophobic properties.
By the interaction between active oxygen and CIO- 1 anions, the electro-
chemical treatment purifies the diamond adsorbent surface and induces
its active saturation with hydroxyl groups. The surface acquires hydro-
philic properties.

Figure 2 shows the thermodesorption spectra of water from initial and

modified surfaces of nanodispersed diamond taken in the 20 - 600°C.
Below 150 °C the physical desorption of water takes place. It is seen
from fig.2 that the treatment of diamond decreases (curve 2), while the
116

electrochemical modification increases the amount of physically ad-

sorbed water (curve 3).
The table shows the adsorption characteristics of various adsorbents (ad-
sorption potential, occupancies of the adsorbent surface by Cr6+ and
molecules of albumin).

TABLE. Adsorption characteristics of various adsorbents

Occupancy of the Occupancy of the

Adsorbents surface by chro- surface by mole-
mium ions cules of albumin

e,% e,%
Acetylene black 65 50
Silard 44 -
Kaolin 12 -
N anodispersed 52 72
Diamond
N anodispersed Dia-
mond Adsorbent:
ACY,n:O-<P 52 46
ACY,n:O-B 30 72

It is seen from the tabulated data that nanodispersed diamonds are effi-
cient adsorbents.
The selective nature of adsorption implies that cations of heavy metals
are mainly adsorbed by a hydrophobic surface, while organic substances
are adsorbed by a hydrophilic surface.
 117

6. CONCLUSIONS

Thus, our findings show that nanodispersed diamond adsorbents are effi-
cient adsorbents of blood plasma toxins of organic and inorganic origins.
Thermal or electrochemical modification of the diamond adsorbent sur-
face increases substantially the diamond adsorptivity. Diamond adsorb-
ents are characterized by a selective adsorption mechanism: a hydropho-
bic surface adsorbs mainly cations of heavy metals and a hydrophilic
surface absorbs organic molecules.

7. REFERENCES

1. X. Kinle and E.Bader(1984) Activ Coals and their Indastrial Appli-

cation [Russian translation], Khimiya, Leningrad, pp, 215.
2. J. E. Neimark (1982) Synthetic Minerals Adsorbents and Catalyst
Carriers [Russian translation], Naukova Dumka, Kiev, pp. 214.
3. Bogatyreva G. P., Marinich M. A., Gvyazdovskaya V. L. (2000)
Diamond - an Adsorbent of New Type, Diamond and Related. Ma-
terials, 9,2002 - 2005.
4. Bogatyreva G. P., Voloshin M. N., Marinich M. A., Malogolovets V.
I., Gvyazdovskaya V. L., Gavrilova V. S. (1999) Surface and Elec-
trochemical Properties of Nanodispersed Diamond, Superhard Mate-
rials, 6,42 - 46.
5. Bogatyreva G. P., Marinich M. A., Gvyazdovskaya V. L. (1997) The
Use of Cathode Potentiodynamic Pulses for a Quantitative Assess-
ment of Efficient Conducting Centers at the Surface of Diamond. In:
Proc.7-th European Conference on Applications of Surface and In-
terface Analysis, Goteborg.
6. Novikov N. V., Bogatyreva G. P., Marinich M. A., Bazaliy G. A.
(2000) Method of Treatment of Superhard Materials, patent of
Ukraine 99727263 of24.05.2002.
7. Bogatyreva G. P., Marinich M. A., Baza1iy G. A., Gvyazdovskaya V.
L. (1999) Studies of the Effects of Electrochemical Treatment on
Adsorption and Structural Characteristics of Diamond Micron Pow-
118

ders. in: Proc. of the Ukrainian Electrochemical Congress [in Rus-

sian], Dnepropetrovsk, pp. 48 - 50.
TRIBOLOGICAL CHARACTERIZATION OF CARBIDE-DERIVED
CARBON (CDC) FILMS IN DRY AND HUMID ENVIRONMENTS

B. CARROLL and Y. GOGOTSI*

Department ofMaterials Engineering
Drexel University
Philadelphia. PA 19104

A. KOV ALCHENKO and A. ERDEMIR

Energy Technology Division
Argonne National Laboratory
Argonne. IL 60439

MJ. McNALLAN
Department of Civil and Materials Engineering
University ofIllinois at Chicago
Chicago. IL 60607

*Corresponding author: gogotsi@drexel.edu

The problem of good wear resistance coupled with low friction coefficient has
been studied extensively. Carbide-derived carbon (CDC) films have been
demonstrated to show excellent friction and wear properties in air. In the
present work, we examine the effect of humidity on the tribological behavior of
CDC films prepared under various experimental conditions. We produced the
films by high temperature chlorination of sintered silicon carbide, characterized
them by Raman microspectroscopy and nanoindentation, and carried out pin-on-
disk tribological tests in air and dry nitrogen (0% humidity) using silicon nitride
counterbodies. Our results indicate that CDC, unlike graphite or glas:"y carbon,
does not fail in dry environments. Moreover, it performs better in dry nitrogen
than in humid laboratory air. The CDC coating on SiC can work for hours in
dry nitrogen showing the friction coefficient of less than 0.05. Chlorination
conditions and the surface condition of the test piece are other important
parameters in tribological performance. These coatings may be used in dynamic
seals and other tribological applications.

1. Introduction

Metal carbides such as SiC, TiC, WC, etc., are often used as wear resistant
materials in machining and sliding contact applications in both the bulk form
119
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 119-130.
© 2003 Kluwer Academic Publishers.
120

and thin films due to their high hardness and excellent wear and corrosion
resistance [I]. SiC is generally used in mechanical seal applications while TiC
and WC are preferred for thin coatings and bulk materials used in machine tools
[2]. However, most of these materials are not lubricious and often suffer from
high frictional losses when used in sliding contact applications. The friction
coefficients of metal carbides may vary between 0.25 and 0.80 [3]. The
frictional behavior of these materials is strongly dependent on the presence or
absence of an adsorbed film on the sliding surfaces of these materials. Further,
most carbides may undergo tribochemical reactions during dynamic sliding in
open air at high temperature or high sliding velocity in spite of their excellent
corrosion resistance under static conditions [4]. The thin oxide films formed on
the sliding surfaces of these carbides may in turn dominate their tribological
behavior. Reduced friction on SiC has been observed at elevated temperatures
mainly because of the formation of a thin layer of carbon on the sliding contact
surface [5]. In short, despite their excellent hardness, high chemical inertness,
and wear resistance, most metal carbides suffer from high frictional losses when
used in sliding contact applications.
Thin diamond and diamondlike amorphous carbon films have been
examined by several researchers as a means of lubricating metal carbides, with
limited success. Specifically, crystalline diamond and highly disordered or
amorphous carbon/graphite layers were formed on carbides such as SiC and WC
to achieve low friction [6, 7]. Diamond films produced by chemical vapor
deposition (CVD) processes were extremely rough and highly abrasive, and
hence useless for sliding contact applications [8]. However, films of amorphous
carbon or nanocrystalline diamond with very smooth surface finish were found
to have very low friction coefficients and long wear lives on sliding SiC and WC
surfaces [9]. WC-based tool inserts coated with diamond films are now widely
used in the machining of aluminum and its alloys, while diamond-like
amorphous carbon (DLC) films are well-suited for bearings, gears, and seals.
An alternative approach to achieving low friction and high wear
resistance on sliding metal carbide surfaces is the carbide-derived carbon (CDC)
process [10-12]. In this process, a low-friction carbon layer is obtained on the
surface of silicon carbide by reacting the carbide with chlorine or chlorine-
hydrogen gas mixtures at about 1000°C in a sealed reactor. The Si atoms react
preferentially with Cl2 and leave the system as SiC4 mainly because SiCI4 is
much more thermodynamically stable than CCI4 at 1000°C. The carbon atoms
rearrange into a very adherent carbon film on the surface of the SiC substrate
[13]. Acting as a solid lubricant, these carbon layers can be useful in
applications such as bearings, SiC-based microelectromechanical systems
(MEMS), electrical contacts, biomedical implants, and mechanical seals.
The tribological performance of carbon is strongly affected by the
environmental conditions of the test or application. In graphite and other Sp2
modifications of carbon, the presence of water vapor is required to obtain
lubricity, rendering these materials useless in dry or vacuum environments [14].
However, diamondlike carbon films have been observed to show an opposite
effect, i.e., ultralow friction coefficients in vacuum and moderately good friction
 121

behavior in humid air [15]. Carbide-derived carbon has been demonstrated to

show good frictional properties in room air [II] and to contain nanocrystalline
diamond [16]. The goal of this study was therefore to evaluate the tJibological
behavior of CDC in humid and dry environments.

2. Experimental

Polished and unpolished sintered silicon carbide (Hexolo/M ) disks 30 mm in

diameter and 3 mm thick were obtained from Saint-Gobain Advanced Ceramics.
These specimens were cleaned ultrasonically in acetone and placed on a quartz
sample holder that was in turn placed in the hot zone of a horizontal quartz tube
furnace. Gases were supplied from cylinders of pure Ch, pure Ar, and
5%H 2ibalance Ar. These gases were passed through anhydrous CaS04 and
concentrated sulfuric acid, H2 S04 , to remove water vapor and were mixed in a
packed glass bead mixing column. Gases were ported through PTFE tubing and
glass, and flow rates were monitored with calibrated flowmeters. The gases were
metered into the furnace to supply the indicated C1 2 :H2 ratio at reaction
temperature and atmospheric pressure.
For each experimental run, the furnace was raised to operating
temperature at a rate of 5°C/min under an argon purge. Once stable at operating
temperature, the reaction gas mixture was introduced for the specified time.
Experiments were performed at Drexel University in gas mixtures containing
various amounts ofCh, H2, and Ar at temperatures from 950°C to I 100°C.
Typical gas mixtures included pure Ch, 9%:4.5% Clz:H 2, and 11.8%:4.4%
Clz:H2 in balance argon. Experiments were run for periods ofthree to six hours.

20000 • Graphite
18000 o a-SiC o
16000
~14000

=
g .12000
o
U lOOOO
Amorphous
8000 o o
carbon
6000
4000
2000
O+-----~-----r----~-----,------~--~T_~
10 20 30 40 50 60 70
29
Figure 1. Characteristic X -ray diffraction pattern of carbide-derived carbon showing amorphous and
graphitic phases.
122

8000
,-,
.......'" 7000
=
= 6000
.ci 5000
=
I-

£4000
'-'

'" 3000
....= 2000
~

....=
1000

0
800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800
Raman Shift (em-I)
Figure 2. Deconvolution of Raman spectrum of CDC showing D and G bands. Peak heights are
compared for the lolIG ratio.

At the end of each experimental run, the furnace and reaction gas mixture were
secured and an argon purge was run through the furnace during the cool down
period.
The treated specimens were analyzed by micro-Raman spectroscopy
using an Ar ion laser (514.5nm line) at 500x magnification with a spot size of
-211m. X-ray diffraction was performed using a Siemens diffractometer with
CuKa radiation. An MTS Nano Indenter XP equipped with a Berkovich
indenter and utilizing a load of 25mN was used to characterize the Young's
modulus and hardness of the polished surface of the coatings.
Friction and wear testing was carried out at Argonne National
Laboratory in a ball-on-disk apparatus under loads of 2-20N. Treated and
control (untreated SiC and commercial graphite) flat specimens were tested
against the 9.55-mm diameter ShN 4 balls with a surface finish better than
O.05Ilm centerline average (CLA). All tests were carried out at 60rpm
(corresponding to approximately 6.3 cmls linear velocity) for 20 minutes unless
otherwise noted. Tests were carried out in room air at -11% humidity and dry
nitrogen, and samples were tested in both the as-treated state as well as with
post-treatment polishing of the carbon layers.

3. Results

Figure 1 shows a typical X-ray diffraction pattern of CDC. This shows two
broad peaks corresponding to amorphous carbon, two lines of graphite, and the
lines corresponding to the a-SiC under the carbon layer. Little difference
between treatment conditions can be seen in X-ray diffraction, and formation of
amorphous carbon and nanocrystalline graphite can be assumed based on XRD
data. Figure 2 shows a typical Raman spectrum and peak deconvolution for the
 123

14
l12 •
I.'l
..='"
';'10
•
-=..
8

=
;;e 6
'" 4
:;=
-= 2 Hardness •
• • •
Q
~
0
1.5 1.6 1.7 1.8 1.9 2
(a) I./IG ratio

0.25 Air

..=
.... • :::
:•
0.2
•
.;;:: i
IS.. 0.15
u
Q
Dry nitrogen

....= 0.1
.S!
u
'1:
0.05
""
0
1.5 1.6 1.7 1.8 1.9 2
(b) Iof~ ratio
Figure 3. Mechanical properties including hardness and Young's modulus as measured by
nanoindentation (a) and friction coefficient (b) increase with 10110 ratio

CDC material. CDC exhibits the G band at -1600cm,1 and the D band at
-1340cm'l, corresponding to the relative amounts of graphitic and
disordered/amorphous carbon, respectively. The comparison of these two bands
as the 10110 ratio allows a semi-quantitative determination of the degree of
graphitization (amount of graphite in the sample) and the approximate size of
the graphite crystallites according to the Tuinstra-Koenig equation:

/(0) C(A,)
/(G)
= (1)

where La is the size of the crystallite and C(A) is a function of the excitation
wavelength [17]. Based on these spectra, the treated layers were found to be
composed of mostly disordered or amorphous carbon with some nanocrystalline
graphite.
124

2.4
2.2

S 2
=
'"' 1.8 l>.
..
..." 1.6
.~------------ ----------~
~
1.4
1.2

2:0 2:0.75 2:1

(a) C1 2:H2 Ratio

2.4
i -+- As treated • .a. • N2 wear track
2.2 L -i!r- Air wear track, 2N lC Air wear track, 20N

........= 2
=
'"' 1.8
..."
t:::!Q 1.6
1.4

1.2

2:0 2:0.75 2:1

(b) C1 2:H2 Ratio
Figure 4. lo/lG ratio measured before and after wear testing as a function of treatment
conditions. No significant change occurs in the strucrure of the material during wear at treatment
temperatures of either IOOO°c (a) or I 100°C (b)

The mechanical properties as determined by nanoindentation of the

polished layers were compared to the 1014; ratio in Figure 3a. Hardness and
modulus both increase with the IDIIG ratio and have average values of about 2
GPa and 18 GPa, respectively. In Figure 3b, the friction coefficient is
demonstrated to follow a similar trend, with a slight increase with ID/4; ratio.
The IdIG ratio reveals no trend at 1000°c (Figure 4a), and a very weak
dependence on hydrogen content at 1100°C (Figure 4b). Post-polished surfaces
and wear tracks show no significant difference in the ID/IG ratio, indicating no
significant change in the degree of graphitization during wear. However, direct
comparison of the Raman spectra shows that there is a fine change during wear
which is independent of the IdIG ratio and instead manifested in the broad peaks
at 1125 and 1500 em-I.
 125

=08
~
'Cj
151L6
Q,j

8 Graphite
=0.4 ,(
~OlJk::::~~~============================-~
'1: CDC (llOOOC, pure C12)
~ o+-----~--~~--~~--~~----~--~
o <at lIro
Time (sec)
Figure 5. Comparison of CDC to silicon carbide and commercial graphite in pin-on-disk
testing (2N load) in room air. CDC (II OO°C, pure CI,) is substantially improved over silicon carbide.
The frictional behavior is similar to that of graphite with a slightly reduced friction coefficient.
O~~----------------------------------,

0.25
...
=
II>
·C 0.2
e
...
Q,j
~ 0.15
~ =
...
'-C
0.1
...
·C
2N 5N
0.05 ION 20N

0
o 500 toOO 1500 2000
Time (sec)
Figure 6. Friction coefficient as a function ofload. Stepped increases in the loading during tests in
room air show progressive decreases in the friction coefficients, with end values of about 0.13 .

....
=
.~ 0.8
15
~ 0.6
.§'" 0.4
~ 0.2 -IF==:::::'!!!:========:::\===
~ O~~~~~~~~~~~~
o 200 400 600 800 1000 1200
Time (sec)
Figure 7. CDC shows decreasing friction coefficients when tested in dry nitrogen, dropping to -0.13
by the end of the test. High-quality pyrolytic graphite shows a very low friction coefficient initially
but fractures easily after a short time under these conditions.
126

Figure 8. Friction coefficients are consistently lower in dry nitrogen than in air. Polishing reduces
friction coefficient in air and can increase it in dry nitrogen.

In room air, CDC coatings (f-0.25) were shown to have considerable

improvement over untreated silicon carbide (f-0.75). Marginal improvement
over commercial pyrolitic graphite (f-0.3) was also obtained. This coating was
produced at I 100°C in pure chlorine and is a representative tribological result
(Figure 5).
Two tests were run to determine the effect of a stepwise increase in the
load on the CDC layer in a room air environment (Figure 6). These results show
that the friction coefficient tends to reduce with increasing load, in the case of a
pure Cb treatment, or remain steady, as with a 2: I CIz:H 2 ratio. The end friction
coefficients in both cases were approximately 0.13.
In dry nitrogen (Figure 7), CDC shows a decreasing friction coefficient
as the coating wears in, with an end-of-test friction coefficient of about 0.15.
Under these
conditions, the pyrolitic graphite has initially a very low friction coefficient but
fractures and gives very unstable performance after a short time.
Figure 8 is a chart showing the effect of treatment conditions and post-
treatment polishing on the tribological performance of the CDC material. The
friction coefficient is consistently lower in dry nitrogen than in room air. In the
polished condition, the smoothness of the carbon film decreases the wear of the
counterbody, especially in the case of harder CDC films produced in CI21H2
environments. Moreover, polishing also corresponds to reduced friction
coefficient. The unpolished specimens show reversing trends in air and nitrogen.
Moreover, Figure 9 shows the results for a single sample wear tested in room air
before and after polishing. The friction coefficient becomes lower with sample
polishing and shows a decreasing trend rather than increasing. This corresponds
to minimal break-in and buildup of wear material as shown by the optical
micrographs of the Si3N4 balls (Figure 10).
 127

4. Discussion

-=
.st
0.4
0.35
0.3
Unpolished

=
(,J

....~ 0.25
= 0.2

-
(,J

=
.S 0.15 Polished

.~ 0.1
~ 0.05

0 ~----~----~----,------.-----.----,

0 200 400 600 800 1000 1200

Time (sec)
Figure 9. Effect of polishing on the friction behavior a CDC sample prepared at II OO°C for three
hours in 2:0.75 CI,:H,. Friction coefficient is improved from about 0.36 to about 0.2 and is
decreasing in the polished state rather than increasing when tested at 2N in room air.

Figure 10. Optical micrographs of the worn surfaces ofSi,N4 counterbodies: (a) As-treated CDC
with high loIla ratio; (b) polished CDC showing no visible wear and a slight transfer film.

The results for Raman spectroscopy and X-ray diffraction are in

agreement that the structure of the CDC layer is amorphous and graphitic
carbon. Degree of graphitization is controlled by the process param~:ters. Since
graphite is the thermodynamically stable phase of carbon at room pressure,
higher temperatures tend to induce the formation of more graphite by driving the
process more quickly toward equilibrium. Hydrogen, on the other hand, tends to
stabilize the Sp3 hybridization of carbon and preferentially etch off g11lphite. This
effect is balanced by the enhanced surface diffusion which is also fadlitated by
hydrogen and drives graphitization in the CDC material. As seen in Figure 4, the
end result is a small but steady trend in increased graphitization at higher
temperatures and with lower hydrogen levels. The mechanical behavior of the
CDC coating is dependent upon the degree of graphitization. The amorphous
128

modification of CDC is harder and stiffer than graphite due to the presence of
tetrahedral carbon. The obtained values are similar to those of glassy carbon.
The observed trend, a small increase in hardness and modulus corresponding to
increasing IdIa ratio, bears out this observation. Since the friction coefficient
follows the same trend, graphitic carbon is the major contributor to the
tribological performance of CDC, as expected for this soft and lubricious
modification of carbon.
However, many carbons are known to fail in dry environments such as
nitrogen or vacuum due to the wearing off of adsorbed oxygen or water on the
surface of the material. Without this stabilizing layer, the dangling sigma bonds
on the surface can covalently bond to the counterbody and cause adhesion and
high friction [18]. This characteristic of carbon films requires that an alternate
explanation for the behavior of the CDC films be proposed.
It is known that onionlike nanostructures of materials such as MoS 2
show very promising tribological behavior in both humid and dry environments
[19]. Carbon onions are known to form in CDC coatings. These could stabilize
the coating and, by closing off the edge planes of graphite, allow the CDC to
perform well in dry nitrogen.
Structural changes, characterized by a marked increase in the Raman D
band intensity, are known to occur in graphite when it is used in sliding couples
[20]. Prior work has shown that the CDC materials self-adjust in a ductile
fashion without measurable wear in the track [II). Given that the Raman spectra
show little change between worn and unworn surfaces, it is logical to conclude
that the CDC materials currently under study show similar characteristics.
Rearrangement of carbon onions and densification of porous CDC occurs in the
wear track.
Hydrogen content in the reaction mixture tends to increase both the
ID/la ratio and the friction coefficient in air, and reduce the friction coefficient in
dry nitrogen. This effect is probably due to the hydrogen termination of carbon
atoms, generating a structure somewhat like the highly hydrogenated
diamondlike carbon films. Such films were shown to provide superlow friction
and wear properties especially in inert test environments [18]. It is reasonable to
suppose that CDC coatings operate by a similar mechanism.
Initial surface roughness can have a considerable effect on the
tribological properties. Buildup of wear debris in the track and roughening ofthe
counterbody, as seen in Figure 10, can significantly impair tribological
performance. Post-polishing of the CDC layer eliminates the need for a wear-
intensive break-in period in the friction test.
The graphite used for the test was highly oriented pyrolitic graphite and
showed a friction coefficient of 0.3 (Fig. 5) in both air and nitrogen. In most
tests in dry environments, it failed after 5 minutes or less. The graphite shown in
Figure 7 was the sample with the lowest friction coefficient we observed, due
probably to adsorbed moisture on the sliding surface. It performed well (f-O.l)
in dry nitrogen for less than 20 minutes before the adsorbed water molecules
wore off completely. The irregular performance of the graphite even before
 129

fracture supports this view. CDC material can outperform high-grade graphite in
both air and nitrogen, making it a good candidate for tribological applications.
The ability of CDC to tolerate increasing loads during and after break-
in attests to the potential for CDC to be used in applications that include variable
loads or spikes. Moerover, the coatings are thick enough (tens to hundreds of
microns) that rapid wear-through does not become an issue as can be the case in
thin DLC coatings. No spallation of CDC coatings have been observed [21]. The
layer thickness has been demonstrated to follow linear reaction kinetics. This is
due to the nanoporosity of the layer, which also accounts for the damage
tolerance of this layer and minimizes residual stresses. These effects may
account for the excellent versatility of the tribologica1 properties of the material.

5. Conclusions

Tribological CDC films have been synthesized on silicon carbide. These films
show mechanical and room-air tribological properties consistent with degree of
graphitization. In dry nitrogen, the films show remarkably low friction
coefficients with minimal wear. These coatings hold potential for tribological
applications in both dry and humid environments.

6. Acknowledgements

The authors wish to thank Dr. Alexei Nikitin for experimental assistance.
Financial support was provided by the Department of Energy Industrial
Materials for the Future program.
130

7. References

I. Gogotsi, Y. and Andrievski, R.A., eds (1999) Materials Science of Carbides, Nitrides
and Borides. Kluwer: Dordrecht, NL.
2. Hutchings, LA. (1992) Tribology: Friction and Wear of Engineering Materials. Boca
Raton, FL: CRC Press.
3. Bhushan, B. and Gupta, B.K. (1991) Handbook of Tribology: Materials, Coatings and
Surface Treatments. New York: McGraw-Hill Inc.
4. Gogotsi, Y.G., Kossko, I.M., and Kovalchenko, A.M. (1992) Tribocherrucallnteractions
of Boron Carbides Against Steel, Wear, 154215-227.
5. Miyoshi, K. (1991) Uses of Auger and X-ray Photoelectron Spectroscopy in the Study of
Adhesion and Friction, Advances in Engineering Tribology, 31 3-12.
6. Erdemir, A. and Fenske, G.R. (1996) Tribological Performance of Diamond and
Diamondlike Carbon Films at Elevated Termperatures, Tribology Transactions, 39 787-
794.
7. Miyoshi, K., Wu, R.L.e., and Garscadden, A. (1992) Friction and Wear of Diamond and
Diamondlike Carbon Coatings, Surface & Coatings Technology, 54/55 428-434.
8. Hayward, I.P., Singer, LL., and Seitzman, L.E. (1992) Effect of Roughness on the
Friction of Diamond on CVD Diamond Coatings, Wear, 157 215-227.
9. Erdemir, A., Halter, M., Fenske, G.R., Krauss, A., Gruen, D.M., Pimenov, S.M., and
Konov, V.L (1997) Durability and Tribological Performance of Smooth Diamond Films
Produced by Ar-C60 Microwave Plasmas and by Laser Polishing, Surface & Coatings
Technology, 94-96 537-541.
10. Gogotsi, Y.G., Jeon, I.-D., and McNallan, M.J. (1997) Carbon Coatings on Silicon
Carbide by Reaction with Chlorine-Containing Gases,). Mater. Chem., 7(9) 1841-1848.
II. Ersoy, D.A., McNallan, M.J., Gogotsi, Y., and Erdemir, A. (2000) Tribological
Properties of Carbon Coatings Produced by High Temperature Chlorination of Silicon
Carbide, STLE Tribol. Trans., 43(4) 809-815.
12. Gogotsi, Y.G. and Yoshimura, M. (1994) Formation of Carbon Films on Carbides under
Hydrothermal Conditions, Nature, 367 628-630.
13. Ersoy, D.A., McNallan, M.J., and Gogotsi, Y. 1998 High Temperature Chlorination of
SiCfor Preparation ofTribological Carbon Films, in Proceedings of the Symposium on
High Temperature Corrosion and Materials Chemistry, P.Y. Hou, MJ. McNallan, R.
Oltra, E.J. Opila, and D.A. Shores, (eds). The Electrochemical Society, Inc., Pennington,
NJ, USA. pp. 324-333.
14. Yen, B.K. (1996) Influence of water vapor and oxygen on the tribology of carbon
materials with sp2 valence configuration, Wear, 192 208-215.
IS. Donnet, e., Fontaine, J., Le Mogne, T., Belin, M., Heau, e., Terrat, J.P., Vaux, F., and
Pont, G. (I 999) Diamond-like carbon-based functionally gradient coatings for space
tribology, Surface & Coatings Technology, 120-121 548-554.
16. Gogotsi, Y., Welz, S., Ersoy, D.A., and McNallan, M.J. (2001) Conversion of Silicon
Carbide to Crystalline Diamond-Structured Carbon at Ambient Pressure, Nature, 411
283-287.
17. Ferrari, A.e. and Robertson, J. (2000) Interpretation of Raman spectra of disordered and
amorphous carbon, Physical Review B, 61(20) 95-107.
18. Erdemir, A. (2001) The role of hydrogen in tribological properties of diamond-like
carbon films, Surface & Coatings Technology, 146-157292-297.
19. Chhowalla, M. and Amaratunga, G.A.J. (2000) Thin films offullerene-like MoS,
nanoparticles with ultra-low friction and wear, Nature, 407164-167.
20. Langlade, C, Fayeulle, S., and Olier, R. (1993) Characterization of graphite superficial
thin films achieved during friction, Applied Surface Science, 65/66 83-89.
21. Ersoy, D.A., McNallan, M.J., and Gogotsi, Y. (2001) Carbon Coatings Produced by High
Temperature Chlorination of Silicon Carbide Ceramics, Mat. Res. lnnovat., 55-62.
NANOSTRUCTURED CERAMIC COATINGS PRODUCED BY MAGNETRON
SPUTTERING

VASCO TEIXEIRA
University of Minho, Physics Department, GRF-Functional Coatings Group
Campus de A=urem, 4800-058 Guimaraes, Portugal
fa.x:+ 35 1-2535 10401, tel: +35 1-25351 0465, email: vasco@fisica.uminho.pt

Abstract

Zirconia based coatings, because of optical properties (high refractive index, low
absorption over a broad spectral region from near-UV to mid-IR, high pulse laser damage
threshold), thermal properties (low thermal conductivity and high thermal expansion
coefficient), high dielectric constant, mechanical properties (high fracture toughness,
thermal shock resistance) and corrosion-resistant properties at high temperatures, have
many important scientific and technological applications. Physical Vapor Deposition
(PVD) techniques, in particular magnetron sputtering are suitable technologies to deposit
ceramic coatings with tailored structure (nanocomposite and multilayered coatings).
Zirconia-Alumina (ZrO z-AI2 0 3 ) nanolayered thin coatings are a new method to stabilize
the zirconia high temperature tetragonal phase at room temperature. Zr0 2-A1 2 0 3
transformation-toughening nanolaminates were prepared by reactive magnetron
sputtering. In this contribution it is studied the structural properties of Zr0 2-A120 3
nanostructured coatings. The paper begins with a brief overview of nanolayer concepts
and nanoscaled coating architecture, properties of zirconia ceramics and structural
stability study of zirconia coatings. Coatings of pure (undoped) zirconia presents a
monoclinic phase with traces of tetragonal. The nanostructured coatings present a Zr02
polycrystaUine phase (monoclinic and tetragonal phases depending on the ratio of
thickness in the nanolaminated structure) and an Al 2 0 3 amorphous phase. The zr0 2 high
temperature tetragonal phase content increases, as the nanolayers get thinner. After
annealing in air at lOOO°C the alumina is preserved in amorphous state and the quasi-

131
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 131-147.
© 2003 Kluwer Academic Publishers.
132

amorphous tetragonal zirconia nanosized grains crystallizes to tetragonal phase without

any monoclinic transformation.

I-Introduction

Conventional materials have grain sizes ranging from microns to several millimeters and
contain several billion atoms each. Nanometer sized grains contain only about 900 atoms
each. As the grain size decreases to the nanometer range, there is a significant increase in
the volume fraction of grain boundaries or interfaces. A nanostructured crystalline
material is one in which the spacing between lattice defects approaches inter-atomic
distances. These characteristics strongly influence the chemical and physical properties of
the material. For example, it has been found that nanostructured ceramics are sometimes
tougher and stronger than the coarser grained ceramics and nanophase metals exhibit
significant increases in yield strength and elastic modulus. It has also been shown that
other properties (electrical, optical, magnetic, etc) are influenced by the fine grained
structure of these materials. Physical Vapor Deposition (PVD) processes, such as
magnetron sputtering have shown to be powerful methods for the synthesis of
nanostructured ceramic thin coatings [1-6]. Magnetron sputtering is a non-equilibrium
process induced by collision processes of high-energy particles [7]. This is a complex
method to deposit thin films and coatings, because small changes in deposition
parameters induce radical changes in physical properties of coatings [7-9]. For the
deposition of composite multi phase coatings this technique allows to control at an atomic
level the addition of the elements to the coating matrix, and thus providing an improved
uniformization of the phases, e.g. nanocrystalline phases embedded in an amorphous
matrix such as Zr02-A1203 [10,1] or TiN-Si3N4 [3] coatings.
Interfaces and grain boundaries are key parameters in designing nanolayered coatings.
Internal interfaces in materials are extended defects including grain boundaries and
interphase boundaries, found in almost every engineering materials. Interfaces play a
crucial role for the performance of layered composite coatings. Extreme service
conditions sometimes prevail such as high temperatures, aggressive surrounding media,
interdiffusion, wear, etc. It follows that the structure of interfaces must be regarded from
the point of view of the whole complexity of the thermodynamic conditions during
fabrication and exploitation of coated components, frequently in thermodynamic non-
equilibrium. Therefore, interface engineering is an important field of coating materials
science and engineering. For many technological applications, the control of internal
interfaces, including the orientation relationship, the interfacial misfit, dislocations,
segregation, and of interface kinetics playa crucial role.
 133

Besides the class of super-hard coatings based on nitride nanocomposite layers, oxide
coatings like Zr02 and Ah03 are very interesting for hard and protective applications at
high temperatures [12-15] and optical applications, e.g. where high laser power is used
[16]. Zr02 bulk ceramics and thin coatings are very interesting materials because of their
outstanding mechanical, thermal, optical and electrical properties. Zirconia has a high
melting point, high resistance to oxidation, low thermal conductivity, high hardness and
high thermal expansion coefficient. These ceramic coatings are widely use in many
technological applications such as heat resistant layers and TBC's [12,14], optical
coatings [16], for memory cells [17], buffer layers for growing superconductors [18],
chemical sensors for automotive applications , e.g. oxygen sensors and ion conductors
[19-21], etc.
Advanced ceramic coatings based in stabilized zirconia, such as multilayered, composite
and graded coatings, are commonly used as protective coatings for advanced power
engineering applications to improve performance, e.g. thermal barrier coatings (TBC's)
deposited by plasma spraying or electron beam evaporation (EB-PVD) techniques are
currently applied on aero and land based gas turbine blades and diesel engine components
[14,22,23]. A conventional gas turbine based approach to TBC's involves the use of a
two-layer coating consisting of an oxidation resistant metallic layer (bond coats, typically
NiCr-alloys) and a ceramic top coat (partially stabilized zirconia), which serves the role of
a thermal barrier. The metallic bond coat: i) provides a surface texture to promote
adhesion of the ceramic top coating; ii) reduces the coefficient of thermal expansion
(CTE) mismatch between the TBC and the metallic substrate; and iii) provides oxidation
resistance to the superalloy substrate. A NiCoCrAIY alloy bond coat with a zr02Y20 3 top
coat is commonly used for a conventional duplex TBC system [12-14,22,23].
Within the frame of this contribution an overview of nanocomposite coatings for
protective applications it is outlined with a particular example of the development of
stabilized zirconia coatings. With the objective of developing transformation-toughening
high temperature ceramic coatings we deposited multilayers of Zr02-Ah03 in which the
layer spacing was scaled to ensure nanosized zirconia crystallites. The paper presents the
study of microstructure and x-ray structural properties of Zr02"AI 20 3 nanostructured thin
coatings deposited by DC reactive magnetron sputtering using Ar+0 2 atmosphere.

1.1. NANOCRYSTALLINE AND NANOSTRUCTURED MATERIALS

Nanostructured materials and coatings offer the potential for significant improvements in
engineering properties based on improvements in physical and mechanical properties
resulting from reducing microstructural features by factors of 100 to 1000 times compared
to current engineering materials. Decreasing the grain size of a material to the nanometer
134

range leads to a drastic increase in the number of grain boundaries reaching typical
densities of 10 19 interfaces per cm3 • The fraction of atoms located at grain boundaries
scales with the reciprocal grain size. Nanocrystalline materials have attracted considerable
scientific interest because of their unusual physical properties. Such materials are
characterized by their small crystallite size, which is in the range of several nanometers.
The grains are separated by high-angle grain or interphase boundaries [1,2].

Alif/yed Multilayered isotropic

TiA!N TiN-Mo TiN·Si,N4
Si-DLC TiN-Al,OrTiC TiC-a-C
ZrOZY20J Zrf)z-AlzO J TiC-DLC
NiCoCrAIY Z1'02 YzOrAlz(h ZrOrAI203
Figure! - Exmnp/i!s ofcomposite ami l!a1WS!nJctured coatings jor engineering ,nuiace applications

A nanoscaled coating architecture can be described as multilayers with individual layer

thicknesses of few nanometers, dispersion particles in the range of few nanometers and
also structure modulations on the nanometer scale. Nanoscaling of protective and
functional tribological coatings represents a promising mean for tailoring the coatings
with respect to specific applications. Often such coatings can only be deposited by the
combination of different PVD processes. In tribological applications, sequences of hard
coatings for cutting tools with the same or different bonding type or structure are used for
multilayers to obtain specific improvements like TiN, TiB2 and TiAIN [2-5]. On the other
hand, for high temperature applications, sequences of multilayers of A1 20 3, zr0 2 and
Zr02Y203 can be designed as nanostructured coating systems with the objective to
produce coatings with different physical and mechanical properties as well more reliable
and high performance coatings. However a rating of the properties of different coating
systems is extremely difficult, because they depend not only on the materials and their
chemical distribution but most often even more on the deposition processes and
parameters.
 135

A composite coating usually consists of two or more phases combined either as different
layers (multilayer or nanolaminate) or as homogeneous isotropic mixture of different
phases (multiphase or dispersion coating) as can be represented in Fig. 1. Adding
different elements to an existing coating is a way to control some of its properties to a
desired level according to the specific technological application. Some of the affected
basic physical properties that alter the final coating system performance can be: hardness,
lattice constant, thermal expansion coefficient, elastic modulus, residual stress, grain size,
texture, crystallographic phases, oxidation behaviour, chemical reactivity, optical and
electrical properties, etc ... Among these parameters, residual stresses are a key concept to
develop feasible PVD coatings. Due to the problems of stresses in the mechanical
integrity of such composite coatings a methodology for thermo-mechanical design in
layered and graded coatings should be used that complements experimental procedures
for evaluation of coatings [15,23,24].
Isotropic (dispersion) coatings are characterized by the presence of a second phase in
form of small particles embedded in the coating matrix (see Fig. 1 and 2). Physical vapor
phase deposition results in such coatings when two materials are deposited without
forming a complete mutual solid solution. This especially holds for materials which either
are of very different types or have different lattice structures. This is, for example, the
case for the Zr0 2-A1 20 3 or TiN-Si3N4 coating system, where depending on the deposition
parameters, chemical composition and grain sizes different physical properties, residual
stress and microhardness can be observed [3,10,11,25,26].
a-ShN4
(or a-AI 20 J )

TiN
(or zr02 )
crystalli tes

Figure 2-Example of an isotropic coating composed by a nanocrystalline phase embedded in an amorphous

matrix, ego TiN/a-Si,N, or ZrO/a-AI,03
136

Another class of multilayered coatings is a functionally graded coating, in which the

composition profile is designed to develop the aimed performance with respect to thermal
and mechanical properties [15,24,27,28]. The functionality and reliability of coated
engineering components (frequently involving metal/ceramic interfaces) are strongly
related to residual stresses of thin coatings. Because of different thermal expansion and
temperature gradients, residual stress develops in the ceramic-metal system which can
cause failure by a delamination process or fracture of the material. To avoid coating
failure it is necessary to control the level ofthermal stresses. These stress reduction can be
done by replacing the sharp interface with an intermediate composite layer within which
the composition (and hence the thermo-mechanical properties) are smoothly varied from
the ceramic to the metal material. This graded multilayer system is known as a
Functionally Gradient Material (FGM) and has many of technological applications
[15,25] such as the functionally graded thermal barrier coatings for use in gas turbine
components [22,23]. For the fabrication of a FGM the gradients can be continuous from a
microscopic point of view, or they can be laminates comprised of graded metal/ceramic or
gradients in porosity, etc. [23,24,27]. By varying the composition, the coefficient of
thermal expansion of each layer is changed to some intermediate value between that of
the metal substrate and ceramic, thereby relieving the thermal mismatch and the resulting
thermal stress.
The potential benefits of nanoscaled architecture include higher hardness and strength in
metals and cermets resulting from reduced grain size and slip distance, respectively. In
ceramics, higher hardness and toughness may be accomplished with reduced defect size
and enhanced grain boundary stress relaxation, even at room temperature. These coating
materials with nanometer-scale grain or particle sizes can be also a new and effective
strategy for either increasing or reducing heat transfer rates. Diffusivity is greatly
increased, associated with a larger volume of grain boundaries. Thermal conductivity may
also be reduced because of enhanced phonon scattering from grain boundaries and other
nanoscale features. For example, nanocrystalline yttria stabilized zirconia (YSZ) coatings
were recently observed to exhibit a strong reduction in thermal conductivity at room
temperature when the grain size was reduced below 40 nm [29]. The small spacing
between oxygen vacancies and associated small phonon mean-free-path is responsible for
the low thermal conductivity of stabilized zirconia and is one reason that this material is
used in many thermal barrier applications. While grain boundaries may have only a small
effect on the phonon mean-free-path in zirconia, phonon scattering at interfaces provides
an additional mechanism for significantly reducing thermal conductivity with decreasing
grain size, thus the interest in nanocrystalline zirconia coatings as a way to reduce thermal
conductivity. It should be noticed that these mechanisms are less important at high
temperature, where radiative heat transfer contributions are significant for thermal barrier
 137

applications. Benefits would still be obtained by minimizing conductive heat transfer

[30,31 ].
The ionic conduction in the zirconia-based ceramics is also an interesting physical
property for electrochemical devices which present a grain size-dependent behavior [32-
34]. In order to achieve the optimized use for these materials in electrochemical devices,
such as solid oxide fuel cells, oxygen sensors and as a constituent layer in electrochromic
devices, a strong demand for clarifying the source of the grain size dependence is
emerging. An enhanced ionic conductivity attributed to the interfacial effect related to
grain boundaries was recently found in the nanocrystalline rare earth stabilized zirconia
coatings [32-35]. As the grain size is reduced from micrometer to nanometer, the defect
properties (defect formation, lattice defect and band structures) are considered to playa
crucial role in defining the ionic and electronic conduction in the nanosize regime [34,35].

1.2-ZIRCONIA BASED COATINGS FOR HIGH TEMPERATURE APPLICATIONS

A coating material to be used as protective layer at very high temperatures should have
the following requirements:
1)high melting temperature;
2)low thermal conductivity;
3)chemical inertness;
4 )good thermal shock resistance;
5)high hardness and fracture toughness;
6)low permeability and diffusion for oxygen and other reactive, species
(to prevent internal substrate oxidation/corrosion);
7)hot corrosion resistance;
8)high density (to avoid gas flux through open pores to the substrate),
9)stress free or in a state of compressive stress at working temperature,
IO)good adherence to the substrate;
The items from 1 to 7 are easily obtained by selecting an appropriate material, which is
typically stabilized zirconia. The coating requirements from 8 to 10 are achieved through
the selection of a deposition technique, where magnetron sputtering is a very powerful
technique to deposit several types of coating materials with a good control of the
microstructure, stress and adhesion.

Table 1- Crystallographic data/or the three main phases a/zirconia

Crystal Structure Monoclinic phase Tetraxonal phase Cubicyhase
Space group P2 l/e P4 2inrnc Fm3m
a (nm) 0.5156 0.5094 0.5124
b(nm) 0.5191 -- --
c(nm) 0.5304 0.5177 --
Il (0) 98.9 -- --
Density (g/cm ) 5.83 6.\0 609
138

Zirconia is also the essential component of all functional bulk transformation-toughening

ceramics. Transformation-toughening depends on the retention of the tetragonal
crystallographic phase at room temperature, and the ability of this phase to transform
locally to monoclinic in response to applied stress. The defect structure established within
the monoclinic phase and at the monoclinic/tetragonal phase boundaries toughens the
material [36,37].
It is well known that pure Zr02 crystallizes in different polymorphs under different
conditions of temperature and pressure (see Fig. 3 and table 1). Three main equilibrium
solid phases have been reported: monoclinic (baddeleyite) phase, tetragonal phase, and
cubic fluorite phase [14,38,39]. The monoclinic phase of zr0 2 exists in a temperature
range from room temperature to about 900-1250°C, tetragonal phase between 1250°C-
2370°C and cubic phase for T>2370°C. The tetragonal to monoclinic transformation is
accompanied by a volume expansion of approximately 3-4%. This transformation may
lead to failure ofthe coating by cracking and is intolerable in technological applications at
high temperatures. The stability of high temperature polymorphs of zirconia (tetragonal
and cubic) is known to be critically dependent on the number of oxygen vacancies in the
system. When cations such as Y, Mg and Ce substitute for Zr, oxygen vacancies are
generated to maintain charge neutrality and, thus the high temperature phases are
stabilized. The cubic phase is unstable at room temperature, but can also be stabilized by
doping zr0 2 with cations such as y 3 +, Ca 2+, Ce2+, etc [12,14, 39-41].The tetragonal phase
can also be stabilized at room temperature by incorporation of cationic impurities [42] .
• Zratom
CD Oatom

i) ii)

Figure 3- Unit cell structure for ZrO" i) monoclinic (baddeleyite) and ii) tetragonal phase
 139

The transformation oft-Zr02 to m-Zr02 can be suppressed by alloying with Y20 3, Ce02.
Y203 and Ce02 form a solid solution with zirconia and change the phase boundaries.
Anions such as N also generate oxygen vacancies which will act also as stabilizer [43,44].
A certain degree of control of the crystallographic cubic/tetragonal phase of Zr02 is also
possible by energetic ion bombardment (Ar+) using ion beam assisted deposition
processes [45].
Besides these mechanisms to retain the high temperature tetragonal phase two others can
be employed: mixture of zr0 2 with an oxide such as AI 20 3 [14,46] and by decreasing the
particle size of the crystalline domains, i.e., using small zirconia crystals with radii lower
than 6 nm (the surface energy of the tetragonal phase is lower than the one of the
monoclinic phase which results in stable tetragonal crystal at room temperature)
[10,37,47,48]. The ZrOz-AI203 bulk composite is a classic model of the transformation-
toughening ceramic system. AI 20 3 is not soluble with zr0 2 and alters the stability through
another mechanism called "constraint". Alumina has a Young's modulus higher than
zirconia and it is suggested that it forms a rigid matrix around the zirconia crystals which
causes a local compressive stress and hinders the mechanism of the martensitic phase
transformation. Essentially, the critical nucleus size of zirconia crystallites is increased by
a constraining matrix. [10,14,46,49,50].
Several authors also reported that there is a critical crystallite size that is bigger than the
observed for sputtered coatings and could also stabilize the tetragonal phase at room
temperature. These dimensions are 30 nm observed in powder [51], 50 nm observed in
evaporated films with heat treatment [52] and 7.5nm-165 nm observed in CVD films [53].

2. Experimental details

Zr02, Zr02 Y203, AI 2 0 3 and zrOz/AI 20 3 coatings (as single pure layer, nanocomposite
and nanolaminate structure) were prepared by DC reactive magnetron sputtering in an Ar
and O2 gas mixture and were deposited on glass, lnconel 617 and Hastelloy X (Ni-alloy
substrates). All coatings were deposited at constant temperature, bias voltage, sputtering
power, and target-substrate distance. The alloy substrates were polished, and
ultrasonically cleaned before the deposition process. Glass substrates were also
ultrasonically cleaned. Before deposition the vacuum chamber was evacuated to 2*10. 4
Pa. Both targets and substrates were pre-cleaned at the same time in argon atmosphere for
1200 s. The addition of Y203 to the Zr02 was done by putting small discs of Y on the top
Zr target (within the main area of erosion). During deposition of the laminated coatings,
the substrates change position, rotating between two metal targets (Zr and AI) with a
purity of99.5%, stopping in front of targets during the estimated deposition time for each
140

ceramic nanolayer. The total time (including the rotation from one target to another) to
deposit the nanolayered coatings was 425 minutes. (see table 2 with the main sputter
deposition conditions).
Table 2- Sputtering deposition conditions for the nanolaminated ZrOrAl10 3 coatings

Deposition porameters ZrOl layer Al~3layer

Base pressure (Pa) 2x10-4 2x10-4
Sputtering power (W) 1000 1000
Bias voltage (V) -50 -50
Target-substrate distance (m) 0.06 0.06
Deposition temperature (K) 573 573
Oxygen pressure (PaJ. 5.8xlO-" 7.8xlO-
Argon pressure (Pa) 4.6xlO- 4.2xlO-
Total working pressure (Pa) 5xlO-' 5xlO-
Deposition rate (nmls) 0.214 0.146

The surface microtopography, microstructure, coating thickness and chemical

composition was analyzed by Atomic Force Microscopy (AFM), using a Digital
Instruments Nanoscope in non-contact mode, by a Leica Cambridge S360 Scanning
Electron Microscopy (SEM) and Energy Dispersion X-ray Spectroscopy (EDS) from
Oxford Instruments. X-ray diffraction (XRD) measurements were used to characterize the
film structure and to study the influence of nanolayer thicknesses on the high temperature
phase stabilization. X-ray diffraction method was carried out using a Philips PW-171O
Diffractometer to determine the structure of the films. Peak position, peak intensity and
full width of peak at half of maximum intensity (FWHM) were obtained by fitting the
measured peak with two Gaussian curves in order to find the true peak width (FWHM)
and the intensity corresponding to monochromatic Cu Ku, radiation. The average coating
crystallite dimension, D, was calculated by the formula D=O.9IJBcose, where "- is the X-
ray (Ku,) wavelength and e is the Bragg diffraction angle and B is the FWHM after
correction for instrument broadening.

3-Results and Discussion

3 .l.-COA TING MICROSTRUCTURE

Nanocomposite coatings (mixed oxides such as Y203 orland Al 20 3 with the Zr02) and
nanolaminate coatings of ZrOrAl203 with nanosized layer thicknesses were deposited by
DC reactive magnetron sputtering (see table 2). In table 3 and 4 are presented the
 141

structures and thicknesses of the coatings studied. All deposited coatings were well
adherent to the substrate and present compressive residual stresses typical for sputtered
ceramic coatings on metallic substrates [10,11,54]. The coating microstructure was
studied by SEM. The usual columnar and dense structure found in sputtered coatings
prepared at low temperature [7-9] was clearly identified in the SEM analysis. The AFM
analysis presented in Fig. 4 shows the surface roughness for the Zr02-A120 3
nanolaminated sample with a thickness relation of 6nml7nm. From the surface
topography that can be seen in Fig. 4 it is also evident the top of the (:olumns developed
during the sputtering deposition. The zr02y 203 composite coatings also show a columnar
structure after a detailed analysis by SEM. The fracture cross-section of the composite
coatings with alumina (zr02Y203AIz03) analyzed by SEM revealed a glassy, amorphous
and dense microstructure.

Figure 4- AFM images shoWing, at different magnifications. the surface topography of or nanolaminated coating
of ZrO,-AI20-, (sample Z6A 7).

Table 3- Sample structure/or zirconia based sputtered composite coatings

Sample structure zrO, zrO,Y,O, zrO,Y,O, zrO,Y,O,AI,O,

Sample code ZIAO ZlAOY5 ZIAOYII ZlA25Y5
%wt Y,O, 0 52 11 11
%wt. AbO, 0 0 0 25
Nurn her oflayers 1 I 1 1
Layer thickness (nm) 950 2000 6750 800
Deposition rate (nmls) 0.320 0.224 0.272 0.089
142

Table 4- Structure ofnanolayers and coating thickness for the nanolaminates

Sample structure ZrO,IAI,O, ZrO,IAI,O, ZrO,/AI,O,

Sample code Z3A3 Z6A7 Z12A14
Number of layers 375/375 188/188 94/94
Layer thickness (nm) 3/3.5 6n 12/14
Total thickness (nm) 2440 2440 2440

3.2-STRUCTURAL ANAL YSIS

After sputter deposition all the coatings were analyzed by XRD to study the phase
composition and crystallite sizes. The average grain size was determined from the X-ray
diffraction peak width. For the pure zrOz coatings without dopant the main phase
presented is monoclinic with traces of tetragonal. Unstabilized zirconia exists in the
monoclinic (m) phase with traces of tetragonal (t) 'on all substrates. Annealing in air at
1273 K does not modify the phases presented. The coatings are textured with the (111)
direction normal to the substrate surface. A peak at approximately 300 is attributed to
diffraction from (111) planes of the tetragonal phase of zirconia and the peaks at -28 0 and
-3 10 correspond to the monoclinic phase (see Fig. 5).
A non-transformable tetragonal phase of zrOz can be produced by adding 5wt.% to
12wt.% of YZ0 3 to ZrOz. The coatings, which are, as-deposited, in the tetragonal phase,
show also a preferred orientation with the (111) crystallographic plane parallel to the
surface. The average crystallite size, determined from the XRD peak width, is about 30
nm (see table 5), without any significant changes after high temperature exposure to air
[14]. The zrO zY203 coatings are stable even at high temperatures, presenting the
tetragonal phase with a predominant (111) peak
Zr02 Y203AIz03 nanocomposite coatings were amorphous after deposition with evidence
of zirconia tetragonal phase and amorphous alumina. After annealing in air at 1273 K the
tetragonal phase of Zr02 crystallizes with the alumina particles showing an amorphous
structure. The crystallite size was much lower than in the Zr02 YZ0 3 coatings and was
found to be about 15 nm. The coatings with alumina may found some interest in high
temperature protection of Ni-alloys used in gas turbine applications because they proved
to reduce more effectively the oxidation of metallic substrates at high temperatures
[12,14,55].
The nanostructured coatings with the lowest layer thickness of ZrOz and AIz03
(3nm/3.5nm) has a quasi-amorphous X-ray structure with evidence of tetragonal phase.
The broadening of the diffraction peaks clearly indicates the presence ofnanocrystals. For
 143

higher nanolayer thickness the tetragonal phase is found clearly with the (111)
crystallographic plane parallel to the interface (see fig. 5). We found that most tetragonal
crystallites grow with their (111) planes parallel to the growth interface as was the case
for the nanocomposite coatings of Zr02 Y203. The t( Ill) planes are the most densely
packed in tetragonal zirconia, and as such are thermodynamically favored to grow parallel
to the substrate [14,46,48,56]. The estimation of the nanocrystallite size of the zirconia-
alumina multilayer films by the XRD peak broadening indicates that thl:y have an average
crystallite dimension approximately equal to the nanolayer thickness (see table 5).
Table 5- Grain sizes o/the zirconia composite coatings
Roughness
Sample Structure Coating type Phase r.m.s. Grain size
Thickne (nm) (om)
ss(nm)
ZIAO 950nmlOnm Single layer 950 m(II-I)* 25 52
ZIAOYlI 6750nmlOnm Nanocomposite 6750 t(J II) 34.8 34
ZIAOY5 2000nmlOnm Naoocomposite 2000 t(J II) -- 29
ZIA25Y5 800omlOnm Nanocomposite 800 amorphous -- 15
Z12A14 12nmll4om Nanolaminate 2440 t(1II)** 9.3 12.8
Z6A7 6nml7nm Nanolaminate 2440 t(l1l) 14.3 7.3
Z3A3 3nml3.5nm Nanolaminate 2440 amorphous 10.5 4.9
*the main phase is monoclinic m(ll-I) with traces o/tetragonal t(II 1)
•• the main phase is t(I I 1) with some m(II-I)

As it was discussed before, alumina is not soluble with zirconia and according to Lange
[57] alters the stability through constraint. The structural stability of tetragonal zirconia
has been studied by Lange and his theory shows that essentially the critical nucleus size is
increased by a constraining matrix having a Young's modulus higher than that of zr02
(this is the case for Al 20 3 which has a Young's modulus of 390 GPa, approximately twice
that of Zr02 [12,14,56,57]). The excellent mechanical properties of partially stabilized
zirconia, such as the good toughness are associated with the martensitic t => m
transformation which increases the toughness by two distinct mechanisms. Firstly, if a
restricted number of particles undergo the transformation during cooling from the
fabrication temperature, a fine distribution of microcracks is produced, which increase the
toughness. The stress field at a crack tip can induce a metastable t-particle to transform
into monoclinic. This is the basis of the second toughening mechanism, transformation
toughening, where the propagation of a crack is hindered by both the transformed
particles in the crack wake [36,37,56]. This transformation toughening technique was
applied recently to thin coatings [10,11,48,58] and is discussed in the frame of this work.
144

We applied the model presented in ref. [48] to predict the layer thickness at which
tetragonal phase is produced in any zirconia-based nanolaminate, independent of the
materials of the starting nanolayer provided its interface with the growing zirconia
crystallites is incoherent. The stable tetragonal zirconia phase is produced when the
thickness of each zirconia layer is less than the radius at which an unconstrained,
unstressed hemispherical tetragonal zirconia crystallite spontaneously transforms to
monoclinic at the growth temperature [48,58]. Aita et al. [48] derived, using an end-point
thermodynamics analysis, the expression for the critical radius, Re, at which a t7m
zirconia growth transformation occurs:
Re=3.79 [1 - (T/1448)r l nm [1]
Where T is the absolute temperature of deposition in K. For the growth temperature used
in our experiments equation (I) gives Re=6.3 nm.
We have deposited multilayers of zirconia and alumina in which the layer spacing were
scaled to ensure nanosized zirconia crystallites, thus depositing nanolaminates with
amorphous or tetragonal structure at room temperature without the use of dopants. At the
growth temperature used in our experiments this gives a critical radius of about 6 nm.

sample Z12A14

sampleZ6A7

m(-111) 1(111)

26 28 30 32
Diffraction aOI!I. (29)
Figure 5- X-r<ry diffraction patterns of the ZrOrAI,O, nanostructured coatings.

The nanolaminate coatings fabricated with zr0 2 nanolayer thickness of 3 nm, 7 nm and
12 nm are in good agreement with the model described in ref. [48]. In tact, for the
samples where the layer thickness of Zr02 in the laminate structure is 12 nm, the XRD
 145

pattern presents the tetragonal peak and the monoclinic peak, which means that there are
some monoclinic grains in the tetragonal matrix (see Fig. 5). With layer thicknesses lower
than this a quasi-amorphous or tetragonal phase is retained at room temperature. Our
previous studies [10], show that the layer thickness of AI20 3 can also influence the critical
thickness of the zr0 2 layer at which the monoclinic phase starts to appear. In the present
study we use the same thickness for both Ab03 and Zr02 layers, but with a constant
Al 2 0 3 layer thickness (8 nm) it was found that zr02 layers with thickness up to 20 nm
have retained the tetragonal phase and did not show any transformation to monoclinic
when annealed in air at 1273 K for 24 h [10,11].

4-Conclusions

It was demonstrated that magnetron sputtering is a suitable technique to produce dense,

adherent and stabilized zr0 2 coatings employing three different mechanisms:
i) by substitution of some Zr atoms by Y;
ii) by mixing AI 2 0 3 with zr02Y20 3 ;
iii) by decreasing the zr02 crystallite dimensions through the growing of nanosized
laminates where the zr0 2 nanolayers are constrained between AI2 0 l nanolayers.
The nanostructured coatings present a zr0 2 polycrystalIine phase (monoclinic and
tetragonal phases depending in the ratio of thicknesses of the nanolaminated structure)
and an Ah03 amorphous phase. The Al 2 0 3 nanolayers are used to stabilize the zr02
tetragonal phase at room temperature. The content of the zirconia high temperature
tetragonal phase increases as the nanolayers in the structure get thinner. After high
temperature annealing at 1273 K alumina is preserved in the amorphous state and the
quasi-amorphous tetragonal zirconia with nanosized grains crystallizes to tetragonal phase
without any monoclinic transformation. The nanocrystallite size of the zirconia-alumina
nanostructured coatings was found to be close to the nanolayer thickness.

Acknowledgments
This work was partially financially supported by FCT- Fund~a:o para a Ci!ncia e Tecnologia-FCT under the
project POCTIlEMEJ39316/2001: "PVDCOAT-Composite and multilayered protectivI: coatings for efficient
energy systems" and European Commission-DG-XII under contract: COST 522, WP2/SP2-1999/02: "Residual
stresses and failure in multilayered and functionally graded coatings for advanced energy systems".
The author gratefully acknowledges to NATO science program and INVOTAN (Portugal) for the financial
support to present this invited lecture at this NATO ARW, Kiev 2002. The author would like to thanks to the
master student A. Monteiro and PhD student A. Portinha who did many of the experimental work presented.

References
[I] W. Lojkowski, H. J. Fecht, Progress in Materials SCience, 45, p. 339-568, 2000
[2] H Hhan, R. S. Averback, Nanostr. Mater., I, p.95, 1992
146

[3] S Veprek, Surf Coat. Technol., 97, p.I5-22, 1997

[4] Y.G. Gogotsi, RA. Andrievski (Eds.), Materials Science of Carbides, Nitrides and Borides Kluwer,
Dordrecht, NL, 1999
[5] P. e. Yashar, W. D. Sproul, Vacuum, 55, p. 179-190, 1999
[6] M. Cassir, F. Goubin, C. Bernay, P. Vernoux, D. Lincol, Appl Surf Sci., 193, p. 120-128,2002
[7] M. Ohring, The Materials Science of Thin Films, Academic Press, I 99 I
[8] V. Teixeira, M. Andritschky, NATO Series E-Appl. Sci, Kluwer Acad. Publ., Vo1.234, p.I21-125, 1992
[9] J. A. Thornton, Ann. Rev. Mater. Sci., 7, p.239, 1977
[10] P. Gao, L. J. Meng, M. P. Santos, V. Teixeira, M. Andritschky, Vacuum, 64, p.267-273, 2002
[II] V. Teixeira, A. Monteiro, J. Duarte, A. Portinha, R Vassen, D. Stoever, Vacuum, 2002,in print
[12] V. Teixeira, M. Andritschky, W Fischer, H. P. Buchkremer, D. Stoever, Surf Coat. Technol, 120-121, p.
103-111,1999
[13] e. Funke, J.e. Mailand, B. Siebert, R Vassen, D. Stover, Surf Coat. Technol, 94-95, p. 106-111, 1997
[14] V Teixeira, M. Andritschky, High Temp.-High-Press., 25, p.213, 1993
[15] K. Kokini, Y.R Takeuchi, B. Choules, Surf Coat. Technol, 82, p. 77, 1996
[16] T. Sikola, J. Spousta, L. Dittrichova, 1. Benes, Nuc/. Instr. Methods Phys. B., 1(4),673,1999
[17] M.S.R Rao, e.p D'souza, P.R. Apte, R. Pinto, L.e. Gupta, J. Appl. Phys, 79, p.940, 1996
[18] Y. Komatsu, T. Sato, S. Ito, K. Akadi, Thin Solid Films, 341, p.132, 1999
[19] G. Z. Cao, HW. Brinkman, J. Meijerink, KJ. Vries, A. J. Burggraaf, J. Am. Ceram. Soc. 76, p.2201, 1993
[20] A. Bastianini, G. A. Battiston, R. Gerbasi, M. Porchia, S. Daolio, J. Phys. IV 5, p.525, 1995
[21] J. Zosel, F. Blauwe, U. Guth, Adv. Eng. Mater., 3(10), p. 797-801,2001
[22] M. Alaya, G. Grathwohl, J. Musil, Proc. 3rd Int. Symp. on Sructural and Functional Gradient Materials,
ed. I1schner & Cherradi, Switzerland, p.405, 1994
[23] V. Teixeira, M. Andritschky, D. Stover, "Modelling of Thermal Residual Stresses in Ceramic Coatings
with a Graded Composite Interlayer",(invited paper) NATO-ARW, Multilayered and Fibre-Reinforced
Composites: Problems and Prospect, NATO series E-Appl Sci., Kluwer Acad. Publ., vol. 3143, p. 393-408, 1998
[24] T. Hirano, K. Wakashima, MRS Bulletin-Functionally Gradient Materials, vol. xx, I, p. 40, Jan. I 995
[25] F. Vaz, L. Rebouta, P. Goudeau, J. Pacaud, H. Garem, J. P. Riviere, A. Cavaleiro, E. Alves, Surf Coat.
Technol., 133-134, p.307-313, 2000
[26] C J. Tavares, L. Rebouta, J. P. Riviere, J. Pacaud, H. Garem, K. Pischow Z Wang, Thin Solid Films,398-
399,p397-404,2001
[27] K. A. Khor, Y. W Gu, Z. L. Dong, J. Therm. Spray Technol., 9(2), p.245-249, 2000
[28] R. Watanabe, MRS Bulletin, vol. XX, p.32-34, 1995
[29] G. Soyez, J. A. Eastman, L. J. Thompson, R J. DiMelfi, G. R. Bai, P M. Baldo, Appl. Phys. Lett, 77(8), p.
1155,2000
[30] H. S Yang, G.R Bai, L. J. Thompson, J. A. Eastman, Acta Materialia, 50, p.2309-23 17, 2002
[31] D. Stoever, e. Funke, J. Mater. Process. Technol., 92-93, p.195-202, 1999
[32] Y. W. Zhang, S. Jin, Y. Yang, G. B. Li, S. J. Tian, J T. Jia, C. S. Liao, e. H. Yan, Appl. Phys. Lett., 77,
p.3409,2000
[33]1. Kosacki, T. Suzuki, V. Petrovsky, H. Anderson, Solid State lonics, 136-137, (2000), p.1225
[34] H. L. Tuller, Solid State lonics, 131, p. 3409,2000
[35] Y. Zhang, S. Jin, e. Liao, e. H. Yan, Mater. Lett., 2002 (in print)
[36] E C. Subbarao, in Advances in Ceramics-Science and Technology of Zirconia, American Ceramic Society,
Columbus, Ohio, 1981
[37] 0. Grigorev, N Orlovskaya, Y. Gogotsi, V. Yaroshenko, G. Khomenko, G. Krivoshey, Dispersion-
Strengthened and Transformation-Toughened AI,O,-Based Ceramics, in Ceramics Today - Tomorrow's
Ceramics, ed. by P. Vincenzini, Elsevier, Amsterdam, p. 2821-2830,1991
 147

[38] H. G. Scott,1. Mat. Sci., 10, 15227, 1975

[39] E. M. Levin, N. F. McMurdie, in Phase Diagrams for Ceramists, ed. M. K. Reser, American Ceramic
Society, Columbus, OH, p. 76, 1975
[40] P. Gao, L. 1. Meng, M. P. Santos, V. Teixeira, M. Andritschky, Thin Solid Films, 377-378, p.32-36, 2000
[41] E. V.Stefanovich, A.L.Shluger, CR.A.Catlow, Phys.Rev.B 49(17),11560, 1994
[42] P.D.L. Mercera, J.G.van Ommen, E.B.M. Doesburg, AJ. Burggraaf, J.R.H. Ross, Appl.Catal.78, 79, 1991
[43] T. J Chung, J. S. Lee, D. Y. Kim, JAm. Ceram. Soc., 84(1), p. 172-178,2001
[44] S. CoUard, H. Kupfer, W. Hoyer, G. Hecht, Vacuum, 55, p. 153-157, 1999
[45] J. P. Holgado, J. P. Espinos, F. Yubero, A. Justo, M. Ocai'la, J. Benitez, A. R. Gonzalez-Elipe, Thin Solid
Films, 389, p.34-42, 2001
[46] S.B. Qadri, CM. Gilmore, C Quinn, E. Skelton, CR. Gosset, J. Vac. Sci. Techno!. A7(3), 1220-1224, 1989
[47] B. E. Yoldas, 1.. Mat. Sci., 21, (1986),1080
[48] CR Aita, M.D. Wiggins, R. Whig, C.M. Scanian, M. Josifovska,J. Appl. Phys., 79(2), p.1176-1178, 1996
[49] A. Nazeri, S. B. Quadri, Surf Coat. Technol, 86-87,p. 166-169, 1996
[50] M. Andritschky, I. Cunha, P. Alpuim, Surf Coat. Technol,94/95, 144, 1997
[51] RC Garvie, J. Phys. Chem., 82, p. 218,1978
[52] IA EI-Shanshoury, VA Rudenko, IA Ibrahim,1.. Am. Ceram. Soc., 53,p. 264,1970
[53] JS. Kim, HA Marzouk, P. J. Rreucroft, Thin Solid Films, 254, p. 33,1995
[54] V. Teixeira, Thin Solid Films, 392, p. 276-281,2001
[55] M. Andritschky, V. Teixeira, L. Rebouta, H. P Buchkremer, D. Stover, Surf Coat. Technol., 76n7, p.IOI-
105,1995
[56] R C Garvie, M. V. Swain,1.. Mater. Sci., 20, p.1 193, 1985
[57] F. F. Lange, J. Mat. Sci., 17, p. 225,1982
[58] M. Gajdardziska-Josifovska, C R. Aita, 1. Appl. Phys., 79, p. 1325,1996
ANALYSIS OF THE MICROHARDNESS AND THE POROSITY IN GRADED
THERMAL BARRIER COATINGS

A. PORTINHA a, V.TEIXEIRA"·, J. MARTINSb, M.F. COSTAe, R.

VASSENd, D. STOEVERd
• University of Minho - Physics Department, GRF-Functional Coatings
Group, Campus de A=urem, 4800 Guimariies, Portugal
b University of Minho - Mechanical engineering Department, Campus de
A=urem, 4800 Guimariies, Portugal
C University of Minho - Physics Department, Campus de Gualtar, 4700
Braga, Portugal
d Institute for Materials and Processes in Energy Systems 1,
Forschungs=entrum Julich GmbH, D-52425 Julich, Germany
·Corresponding author: fax: +351-253510401, tel.: +351-
2535104651400, email: vasco@fisica.uminho.pt

Abstract

Thermal barrier coatings (TBC's) are used in a number of energy-related applications,

such as protective layers in aero and land based gas turbine components at very high
temperature.
In this paper we present a study on the relationship between the porosity and the
microhardness in zr02 - 8 wt% Y203 thermal barrier coatings (TBC's) prepared by
atmospheric plasma spraying. In order to produce different porosities, with a variation
along the cross section the plasma spraying parameters were varied during each
deposition cycle.
The microhardness was measured with a Vickers indenter and 0.981 N load. The
micro hardness has been evaluated for coatings in as-sprayed condition and after
annealing at 1100 °C during 100h. The results show a fast increase of the hardness after
annealing.
Scanning Electron Microscopy (SEM) was used to acquire the micrographs from
polished cross-sections for image analysis with appropriate software. Differences in
grey levels allowed us to estimate porosity as well as the porosity profile through the
thickness of the coating.
Absolute opened porosity was measured with a mercury porosimetry and by image
analysis. The second technique was also used to estimate the porosity variation along
the cross section. The variation of the microhardness with porosity for various coatings,
and along the cross section for each coating is presented.
149
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 149-158.
© 2003 Kluwer Academic Publishers.
150

1. Introduction

Zirconia (Zr02) crystallizes in three distinct polymorphs according to the temperature

and pressure conditions. The principal phases of zr02 are the monoclinic phase stable at
room temperature, the tetragonal phase stable between 1170 °C and 2370 °C and cubic
phase stable for 2370 to 2680 °C [1-5]. When the applications for these coatings involve
operation in range of temperatures that involve phases transformations it is necessary to
stabilize the high temperature phase at room temperature because this phase
transformation is accompanied by a 3 to 5% volume expansion, and this volume
expansion can cause high residual stresses and micro cracks leading to delamination
and spallation of coatings.
To prevent this transformation, it is necessary to stabilize the high temperature phases at
room temperature. For this purpose these phases can be stabilized by doping zrO z with
other oxides such as YZ03, ceO z, MgO, CaO or Alz03 [1-6].
Ceramic coatings of engineering materials such as zirconia partially or totally stabilized
are used for a variety of technological applications requiring thermal insulation, wear
and erosion resistance or protection from oxidation, sulfidation and hot corrosion. These
kinds of coatings have been applied as Thermal Barrier Coatings (TBCs) for protection
ofmetaIlic components in gas turbines (vanes, blades, shrouds, etc.) and diesel engines,
and improve performance at high temperatures [7, 8, 18]. The TBC concept allows
increase in operating temperature and/or reducing the cooling systems due to the
temperature gradient across the thick ceramic coating, that permit better thermodynamic
performance, lower emissions without requiring major alloy development. Due to the
good combination of properties, zirconia coatings have many others technological
applications such in components at high temperature and adverse corrosive
environments, oxygen sensors [9], optical coatings [10], etc.
TBC's, traditionally, consist in a thick stabilized ZrOz top coating commonly deposited
by atmospheric plasma spraying (APS) on superalloys precoated with a metallic bond
coat (NiCoCrAlY) produced by vacuum plasma spraying (VPS) [11]. The zr0 2 top coat
has a porous and laminar structure and consists of splats with cracks perpendicular to
the surface, this porous structure allows the increase in the thermal isolation and the
cracks permit better stresses accommodation [7, 12, 13, 19]. The metallic bond coat
reduces the diffusion of contaminants and the mismatches of the thermal expansion
between top coat and the substrate and this leads to an increase of the lifetime of
operation for the components.
Nowadays higher operation temperatures are required and in order to obtain systems of
coatings that allow its range of temperatures we need to develop new materials for
coatings or new architectures for the existing materials. These new concepts of TBCs
should have lower thermal conductivity, and be stable at higher temperatures than the
conventional ones [11, 14].
In order to obtain better thermal insulation, in this contribution we present a new
concept of TBC. It consists in a conventional bond coat and a graded Zr02- 8
wt'%Y20] top coat that is graded in the porosity (see figure 1).
In this paper we report on the study of structural properties of Zr02 Y203 multilayered
coatings focusing on the porosity of the micro layers. In order to increase the efficiency
 151

of the thermal barrier different layers with different porosities increasing towards the
surface were studied.
The SEM images are processed using dedicated routines, in order to measure the
porosity of the coatings [15, 16].
Not only the porosity values for each layer
were obtained but also it was evaluated the
way the porosity changes along the coatings
cross-section. To obtain the total porosity,
we also determine the porosity by mercury
intrusion [16].
The micro hardness was measured with a
Vickers indenter in as-sprayed condition and
Figure I. Model of TBe graded in porosity
after annealing. The values of microhardness along cross section. The graded coating was
were thus related with the deposition divided in 4 layers. The porosity was
conditions, heat treatment and porosity increasing from layer I to 4.
variation [17, 20].

2. Experimental Details

The thermal barrier coatings (TBCs) investigated were produced by plasma spraying. A
Sulzer Metco AG vacuum plasma spraying (VPS) system was used to deposit a
NiCoCrAIY bond coat (Ni 192-8 powder by Praxair Surface Technologies Inc,
Indianapolis, IN) on a square plates of a nickel super-alloy Inconel 738 LC previously
sand-blasted with alumina particles which are 0.71 mm average size and ultrasonically
cleaned in acetone. The substrate dimensions were 40 x 40 x 3 mm.
The ceramic powder used for top coats was Zr02-8wt% Y203 (Metco 204 NS Sulzer
Metco GmbH, Germany) and was sprayed by atmospheric plasma spraying (APS) using
a Sulzer Metco Triplex gun. Deposition conditions are presented in Table 1. For
characterization of the as-sprayed condition of the top coat steel substrates were coated
at the same time.
The samples were annealed at 1100 °C for 100 h in air.
The microhardness was measured with a Vickers indenter and 0.981 N load and was
measured along the whole cross section.
Total porosity levels were evaluated in as-sprayed condition by mercury intrusion after
removal of the steel substrate from the coating with hydrochloric acid.
The morphology of the coatings was analyzed by optical and scanning electron
microscope (SEM). Micrographs with two magnifications (400x and 500x) from
polished cross- sections were used for image analysis for the determination of the total
porosity and the porosity profile through the cross section.
152

3. Results and Discussion

3.1. MICROSTRUCTURE

In figure 2 is presented SEM cross-sectional micrograph showing the microstructure

after annealing for our thermal barrier coatings. The thickness was determined by SEM
analysis (see table I) of all coatings, and combining this analysis with deposition
parameters was estimated the thickness of each micro layer for the different porosities
along the cross section. The coatings present a porous and lamellar structure which is
characteristic for this kind of coatings, and it is possible to see the variation of the
porosity for different coatings [21]. Additionally, small microcracks with diameters of
about 200 nm are observed. These cracks are originated from the thermal stresses which
arise from the rapid cooling during the spray process (quenching stresses).

Figure 2. SEM micrographs showing microstructure and porosity on

Zr02 - 8wt %Y203 coating produced by APS in annea1edc ondition,
in which the TOO (Thermal Grown Oxide) is clearly visible.

After annealing in air all coatings present a sintered structure and consequently a
reduction in the porosity levels [22]. A thermally grown oxide (TGO) observed between
bond coat and top coat which is shown in (figure 2). Energy dispersion x-ray
spectrometry (EDX) analysis reveal that the TGO is predominantly aluminum oxide.

3.2. POROSITY ANALYSIS

Porosity ofTBC's can be characterized qualitatively by microstructure observation and

quantitatively by mercury intrusion porosimetry (MIP) technique besides coating
density measurement. The direct examination of coatings microstructure from cross
section of coatings using a (SEM) gives comparative information about porosity for the
different coatings. In conjunction with a backscattered electron (BSE) detector, the
chemical composition of the microstructure is represented in the images by gray level
Table I. Deposition parameters

GPI GPII Grm All Samples

15 1.5 15 1.5

10 10 10 10

8.5 85 8.5 8.5 I <U)

90 90 120 90 I 120 215

aIm atm atm atm 60

423 473 - 423 453 423

,...
Ul
l#.l
154

variation. Pores appear very dark, which permit them to be distinguished and quantified
by image analysis. By this method we can't obtain information about the 3-D pore
network or connectivity between them [15,16].
For our analysis two series of images were acquired, one with 400 x magnification and
the other with 500 x magnification.
Using the MIP it is possible to obtain measurements of total porosimetry for open pores
and the evaluation of pore size distribution. MIP does provide information about the
connectivity of the pores and microscopy reveals information about pore geometry, so
there is interest in combining these two techniques for a more complete analysis.
Mercury intrusion porosimetry is based on the premise that a non-wetting liquid (one
having a contact angle greater than 90°) will only intrude capillaries under pressure.
Mercury must be forced using pressure into the pores of a material. The pore size
distribution is determined from the volume intruded at each pressure increment. Total
porosity is determined from the total volume intruded.
The digital micrographs were evaluated on a Matrox II program for image analysis.
The pores were identified by thresholding the brightness of the pores to produce a
binary image. The dark area fraction in the binary image was evaluated and the
percentage determined.
The corresponding porosity values for the different coatings are presented in Table 2.

Table 2. Coating porosity measured by image analysis and Hg-porosimetry

Samples HP GPI GPII GPIII

Hg -Porosity (%j 14.75 1531 15.29 1338

Image Analysis (%j' 11.79 13.08 15.48 1334
Image Analysis (%jb 1Ll5 10.76 12.73 934
Image Analysis after annealing (%j' 8.44 833 9.90 10.57
Image Analysis after annealing ('lot 6.94 7.99 7.40 8.27
'Porosity with small cracks and ribbons.
'Porosity without ribbons.

We can see in Table 2 a considerable difference between the measured Hg-porosities

and the porosities evaluated by image analysis. Also a reduction was found in porosity
values after annealing for all samples, its reduction is mainly due to the sintering
effects. While the Hg-porosimeter gave reliable results for small pores and
microcracks, it failed for pores with radii larger than 80 I!m. Contrarily the image
analysis is a technique to analyse porosity in which the contribution of small pores and
small microcracks between and through the lamellas within the plasma-sprayed
coatings is difficult to measure in contrast to the big pores. This microcracked
microstructure leads to relatively low thermal conductivity values in APS TBCs. [n
agreement with SEM analysis and deposition parameters we estimate the thickness of
each micro layer in the graded coatings and by image analysis the porosity values of
each layer was determined. In figure 3 the porosity variation along the cross section for
these three kinds of coatings is presented.
 155

The porosity increases from the bond coat interface towards the surface of the top coat.
In as-sprayed condition the variation is from 3% to 5% and in annealed condition we
observe a significant reduction in its
variation (from 1% to 2%). This reduction
is due the sintering effects at high
temperature [22, 23].
The pore size distribution determined by
Hg-porosimeter can be seen in the figure 4
o 50 100 150 200 250 300
for all coatings. The typical bimodal
Thlcknes. [mm)
distribution for the pore radii is found.
One fraction represents microcracks and

L_------
~ 20 Sample GPII _. ___ d
the other, larger one open pores. The e:,17 ..~ .....
pores with radii lower than 0.2 f.lm have ~ 14
lower contribution than other porosity to ~ 11
the total porosity. o 8
D.

50 100 150 200 250 300 350

18
Thicknes!1 [mm)

E
16

14
I=: ~:I
GPII 20
'""l! ~ I

"'
GPIII
~ 17 Sample GPIII
~f
14
'II ~

~ 11~~=-A-~-W__-~-­

-
U :.. 8
\\.
""'lIII o 50 100 150 200

~100 Thickness [mm)

L-____________________________ ~

0001 001 01 10
Pore nldiu.. [pm]

Figure 4. Pore size distribution of as-sprayed Figure 3. Porosity distribution along

coatings obtained by Hg intrusion. the cross section determined by image analysis:
b, d, f - porosity with small cracks and ribbons;
a, c, e - porosity without ribbons.

3.3. MICROHARDNESS

Figure 5 shows the microhardness distribution of the ceramic top c:oatings for as-
sprayed condition. It can be observed that the microhardness decrease towards the
surface in graded samples (GPI, GPII, GPIII) and slightly increased in case of the
constant porosity. For the sample with constant deposition parameters we observe a
little decrease in the porosity values towards the surface that might be explained by the
increase of the surface temperature during deposition, and justifies the small
microhardness increase on sample HP. The decrease in microhardness for the graded
samples is due the increase in porosity along the cross section and the sample GPII has
less decrease because the variation in porosity is minor too. Also it is observed that the
156

microhardness values have a considerable standard deviation, corresponding to the

inhomogeneous, porous microstructure present in thermal barrier coatings obtained by
plasma spraying.
After annealing all coatings show higher values for the microhardness (see table 3) and
maintain gradients from the bond coat interface to the surface. Also the microhardness
of the bond coat was measured and gave constant values along the cross section.
The higher values for the annealed coatings are due the sintering effects at high
temperature [22, 23].
900
880
860
840
Table 3. Microhardness values for
~.2O the same position of each coating in
i 600 as sprayed condition and after
1! 760 annealing.
:!! 760
740
720
Microhardness values [Hv1
Sample HP GPI GPII GPIII
o 50 100 150 200 250 300 350 .we 450 500 550 As sprayed 791 781 810 776
Positionfl.m]
Annealed 1002 907 954 949

Figure 5. Microhardness measurements in

as-sprayed coatings, along the cross section.

4. Conclusions

• All as-deposited coatings have a lamellar structure which is a characteristic for this
type of deposition technique.
• Changing the power of the plasma gun during deposition allows the production of
coatings graded in porosity along the cross section.
• Small cracks and interlamelar defects gave a rather small contribution to the total
porosity in image analysis measurements, in contrast to Hg-intrusion investigations.
• The typical pore size distribution is bimodal.
• The microhardness decreases from the interface to the surface and increase after
annealing. As expected the microhardness decreases when the porosity increase.
Non-graded, as-sprayed HP sample showed an increase in the hardness from the
interface to the surface which was explained by the increase of surface temperature
during deposition.
• The porosity increases towards the surface in graded coatings, decreases with
increasing plasma gun power and increases with increasing working distance.
• The annealed coatings showed reduced porosity levels due the sintering effects.
 157

Acknowledgments

This work was partially financially supported by FCT- Funda9ii.o para a CHlncia e
Tecnologia-FCT under the project POCTJIEME/39316/2001: "PVDCOAT-Composite
and multilayered protective coatings for efficient energy systems". The cooperative
work is also supported by German-Portuguese Co-operative Programme ICCTI-DAAD
and European Commission-DG-XII under contracts: ICCTJ-DAADI 423/2000,
"Composite Coatings for high temperature applications" and COST 522, WP2/SP2-
1999/01: "Residual stresses and failure in multilayered and functionally graded coatings
for advanced energy systems" (projects leader: V. Teixeira). A. Portinha is gratefully
for the Research Grants supported by F.C.T.- Portuguese Foundation for Science and
Technology.
The authors acknowledge to Alcino Monteiro for some hardness measurements.

References

[I] Teixeira, V., Andritschky, M., (1993) Residual stress and corrosion in high-temperature zrO, coatings,
High Temperature-High-Pressures, 25, 213-219.
[2] Kih, J. S., Marzouk, H. A, Reucroft, P J., (1995) Deposition and structural characterization of zrO, and
yttria-stabilized zrO, films by chemical vapor deposition, Thin Solid Films, 254, 33·38.
[3] Guinebretiere, R, Soulestin, B., Douger, A, (1998) XRD and TEM study of heteroepitaxial growth of
zirconia on magnesia single crystal, Thin Solid Films, 319, 197-201.
[4] Aita, e.R., Wiggins, M. D., Whig, R, Scanlan, C. M., (19%) Thermodynamics of tetragonal zirconia
formation in nanolaminate film, J. Appl. Phys., 79 (2), 1176-1 178.
[5] Ji, Z., Haynes, J. A., Ferber, M. K, and Rigsbee, J. M., (2001) Metastable tetragonal zirconia formation
and transformation in reactively sputter deposited zirconia coatings, Surf. Coat. Tech., 135 (2-3), 109-
117
[6] Gao, P., Meng, L. J., Santos, M. P., Teixeira, V., Andritschky, M., (2000) Study of zrO,'Y,O, films
prepared by rfmagnetron reactive sputtering, Thin Solid Films, 377-378,32-36.
[7] Teixeira, v., Andritschky, M., Fischer, W., Buchkremer, H. P., Stover, D., (1999) Effects of deposition
temperature and thermal cycling on residual stress state in zirconia-based thermal barrier coatings, Surf
Coat. Tech., 120-121, 103-111.
[8] Jeanine T. DeMasi-Marcin, Dinesh K Gupta, (1994) Protective coatings in the gas turbine engine, Surf.
Coat. Tech., 68/69, 1-9.
[9] Schmidt-Zhang, P., Sandow, K -P., Adolf, F., Gopel, W. and Guth, u., (2000) A novel thick film sensor
for simultaneous 0, and NO monitoring in exhaust gases, Sensors and Actuators B: Chemical, 70 (1-3),
25-29.
[10] Ghanashyam Krishna, M., Narasimha Rao, K and Mohan, S., (1990) Structural and optical properties of
zirconia thin films, Thin Solid Films, 193/194,690-695.
[11] Va~en, R, StOver, D, (2001) Conventional and new materials for thermal barrier coatings, Functional
Gradient Materals and Surface Layers Prepared by fine Particles Technology, Kluwer Academic
Publishers, Netherlands, 199-216.
[12] Teixeira, V., Andritschky, M., Fischer, W., Buchkremer, H. P., StOver, D., (1999) Analysies of residual
stresses in thermal barrier coatings, J, of Materials Processing Technology, 92-93, 209-216.
[13] Funke, e., Mailand, J.e., Siebert, B., Vassen, R, Stover, D., (1997) Characterization of ZrO,-7 wt.%
Y,O, thermal barrier coatings with different porosities and FEM analysis of stress redistribution during
thermal cycling ofTBCs, Surf Coat. Tech., 94-95,106-111.
[14] Stover, D., Funke, e., (1999) Directions of the development of thermal barrier coatings in energy
applications, J. of Materials Processing Technology, 92-93, 195-202.
158

[15] Mailhot, K., Gitzhofer, F., Boulos, MI., (1998) Absolute coating porosity measurement using image
analysis, Proceedings of the 15 th International Thermal Spray Conference, 25-29 May, Nice, France,
917-922.
[16] Abell, A. 8., Willis, K. L, and Lange, D. A., (1999) Mercury Intrusion Porosimetry and Image Analysis
of Cement-Based Materials, Journal of Colloid and Interface Science, 211 (1),39-44.
[17] Valent, T., (1997) Statistical evaluation of Vicker's indentation test resultsfor thermally sprayed
materials, Surf Coat. Tech., 90, 14-20.
[18] Johner, G., Schweitzer, K. K., (1985) Flame ring testing of thermal barrier coatings and correlation with
engine results, 1. Vac. Sci. Techno!., A3 (6), 2516-2524.
[19] Lech Pawlowski, Didier Lombard and Pierre Fauchais, (1985) Structure-thermal properties-relationship
in plasma sprayed zirconia coatings, J. Vac. Sci. Techno!., A3 (6), 2494-2500.
[20] Lin, e. K., and Berndt, e.e., (I 993) Microhardness variations in thermally sprayed coatings, Proceedings
of the 1993 National Thermal Spray Conference, Anaheim, CA, 7-11 June, 561-568.
[21] Li, C.J., He, Y, Ohmuri, A., (1998) Characterization of structure of thermally sprayed coating,
Proceedings of the 15 th International Thermal Spray Conference, 25-29 May, N ice, France, 717-722.
[22] Siebert, 8., Funke, e., Vaj3en, R., Stover, D., (1999) Changes in porosity and young's modulus due to
sintering of plasma sprayed thermal barrier coatings, J. of Materials Processing Technology, 92-93, 217-
223.
[23] Vallen, R., Czech, N., Mallener, W., Stamm, Wand Stover, D., (2001) Influence of impurity content and
porosity of plasma-sprayed yttria-stabilized zirconia layers on the sintering behaviour, Surf. Coat. Tech,
141 (2-3), 135-140.
NANOSTRUCTURED MESOPOROUS SILICA FILMS

M. CLARA GON<;ALVES*, GEORGE S. ATTARD

Department o/Chemistry, Southampton University, Southampton
S017 lBJ, UK.
* Departamento de Engenharia de Materiais, 1ST, Universidade
Tecnica de Lisboa, Av. Rovisco Pais, 1000 Lisboa, Portugal

ABSTRACT

The lyotropic liquid crystalline phases of surfactants have been used as templates for the
synthesis of mesoporous nanostructured materials. To achieve direct templating by
liquid crystalline phases, surfactant concentrations in excess of 30 wt% in water are
used. In the case of silicas and metallosilicates the materials are obtained as monolithic
objects with typical dimensions greater than 5 mm. Here we report on the processing of
thin films of nanostructured mesoporous silicas by dip coating from mixtures that
contain high surfactant concentrations. We find that the addition of methanol to the
reaction mixture facilitates the formation of uniform, crack-free films. By altering the
surfactant to water ratio we were able to obtain films that had micellar cubic (I I), normal
topology hexagonal (HI)' or lamellar (L,,) organization. Calcination of these films
afforded adherent films that in most cases retained the long-ranged architecture of the
liquid crystalline phase. The methanol concentration of the dip-coating mixture was
observed to have no effect on the structural parameters of the resulting c:alcined films.

1. Introduction

The seminal work conducted by researchers at the Mobil Oil Corporation in the early
1990's on the synthesis of mesoporous silicates has led to a number of syntheses in
which surfactants are used as templates [1-4]. Typically syntheses of mesoporous
oxides employ surfactant concentrations less than 7 wt% with respl:ct to the water
content of the reaction mixture. The materials produced by using surfactant templates
are characterized by a re~ular system of uniformly sized pores and extremely high
surface areas (900-1500 m g-I in the case of silica). In most cases, mesoporous materials
are obtained as fine powders as a consequence of the biphasic nature of the reaction
medium. However, it has been shown recently that reaction mixtures with low
surfactant concentrations can be used to process fibers and supported thin films [5-7].
In 1995 we reported the use of homogeneous (i.e. monophasic) lyotropic liquid
crystalline phases as templates for the synthesis of silicas and metallosilicates from sol-
gel precursors such as tetramethyl orthosilicate (TMOS) [8]. To achieve homogeneous
phases we used surfactant concentrations that were higher than 30 wt%. One of the
advantages of using homogenous phases as templates is that the nano-architecture of the
159
Y. G. Gogot..i and I. V. [Jvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 159-168.
© 2003 Kluwer Academic Publishers.
160

reaction mixture is retained throughout the condensation and gelation process and hence
the nanostructure ofthe material can be determined a priori. For example, the synthesis
of silica in the normal topology hexagonal phase (HI) of a surfactant leads to silica with
cylindrical pores of uniform diameter and indefinite length disposed on a long-ranged
hexagonal lattice. This silica is denoted by HI-Si0 2 to show that its nano-architecture is
derived from that of the HI phase. The diameters of the cylindrical pores are controlled
by altering the length of the hydrocarbon chain of the surfactant or by adding
hydrophobic compounds such as n-dodecane or mesitylene. II-Si02 , Ia3d-Si02 and L,,-
Si0 2 have been obtained from the corresponding phases [9]. A consequence of using
homogeneous lyotropic phases as templates is that the resulting calcined materials are
monolithic in nature but have the high surface areas associated with mesoporous silicas.
Although the production of thin films of mesoporous silicas by dip-coating from
reaction mixtures containing low surfactant concentrations has been reported, we were
interested in the ability to process analogous films from the reaction mixtures that are
used to produce materials from homogeneous liquid crystalline phases. Thin films of
meso porous nanostructured silica are of interest for a range of applications, in particular
integrated sensors. Here we report on studies of the relationships between the
composition of the reactive mixture and the nature and quality of the films that can be
obtained by dip-coating.
The reaction mixtures we employed consisted of Brij56™ as the non-ionic surfactant,
water acidified to pH 2 with HCI and TMOS as the silica precursor. At pH 2 the
hydrolysis and subsequent condensation of the TMOS are decoupled temporally, with
hydrolysis occurring in a few minutes with condensation taking between 6 and 12
hours. Upon hydrolysis of the TMOS methanol is released which prevents the formation
of a liquid crystalline phase in the preparation of monolithic mesporous silica. However,
the rheology of the liquid crystalline phases is such that it is impossible to process them
into thin films by dip coating. Hence in our experiments the methanol released by the
hydrolysis of the TMOS was not removed. Indeed we chose to add differing amounts of
methanol to investigate its effect on the quality of the films and the nature and regUlarity
of the nanostructure.

2. Experimental

2.1. MATERIALS

The silica precursor, tetramethyl orthosilicate (TMOS), Si(OCH 3)4, and the non-ionic
surfactant Brij56™ were purchased from Aldrich Chemical Co. and were used as
received. The pH of the water used in the dip-coating mixtures was adjusted to pH 2 by
using hydrochloric acid.

2.2. TEMPLATING MIXTURES

The procedure used to prepare the mesoporous silica films was based on the method
described by Attard et. al. [8]. To determine the effect of methanol on the dip-coating
process and on the properties of the resulting silica films, 2.5 g of methanol was added
to each of the compositions following the first dip coating. After each subsequent dip-
coating further aliquots of 2.5 g of methanol were added to the mixture. Thus for each
 161

composition, films were produced from mixtures containing 7.S g, 10.0 g, 12.5 g and
IS.0 g of methanol in batch mixtures of 6 g of BrijS6™ I TMOS I acidified water. All
the films were processed within IS minutes from adding the acidified water to the initial
BrijS6TM/TMOS/methanoi mixtures.
Two series of mixtures were investigated. In Series 1 the mass ratio of TMOS to
BrijS6™ was kept equal to 9:S. Varying amounts of acidified water wl~re added to this
mixture to produce the component mixtures of this Series. In Series 2 the mass ratio of
acidified water to BrijS6™ was kept constant and equal 1:1. Varying amounts ofTMOS
were added to this mixture to make the components of this Series. Details of the
compositions of the mixtures we investigated are shown in Table 1.

BrijS6 . ! water
(wt"/n (mole)
Id) 1 7.(}8x104
Ie) 5 3.27x1O"' 5.232 0.2930
It) 10 0.510 0.918 6.02xlO·3 4.578 0.2560 42.52 6
19) 20 0.882 1.590 1.04_10.1 3.528 0.1980 19.04 6
Ih) 30 1.170 2.106 I.3S_1O·2 2.730 0.1530 1109 6
Ii) 40 1.398 2.514 1.65",10" 2.094 0.1170 7.09 6

2c) 2.20 1.428 3.144 2.10xHY' 1.428 0.080 3.809 6

2d) 2.60 1.302 3.390 2.20xlO'" 1.302 0.073 3.318 6
2e) 3.00 1.200 3.600 2.36"10" 1.200 0.067 2.839 6
21) 4.00 1002 4.(102 2.63-10" 1.002 0.056 2.129 6

2.3. DIP·COATING

Microscope slides (7.S cm x 2.5 cm) were used as substrates for dip-coating. These
were degreased with a laboratory detergent, rinsed repeatedly first with distilled water
then with acetone, and subsequently dried in an oven at SO°C. To dip-coat a film the
slide was immersed to a depth of ca 3.5 cm into the reaction mixture, :allowed to stand
for 1 minute and then withdrawn at a speed of 0.1 cm S·I. The films were cured in an
oven at 40°C for S days. Surfactant was removed from the films by calcination in a tube
furnace. The films were heated to 450°C under nitrogen for 8 h follow{:d by oxygen for
12 h.

2.4. FILM CHARACTERIZATION

The structures of the as deposited cured films were investigated by polarized light
microscopy and by low angle X-ray diffraction (XRO). The structures of the calcined
162

films were investigated by polarized light microscopy, low angle X-ray diffraction and
transmission electron microscopy (TEM). The macroscopic morphology and quality of
the calcined films were also assessed by scanning electron micrscopy (SEM). The
optical anisotropy of the as deposited cured films and the calcined films was
investigated by polarized light microscopy using an Olympus BH-2 polarized light
microscope.
Scanning electron microscopy studies were carried out using a JEOL JSM-6400
Analytical Scanning Electron Microscope operating at a voltage of20 kY. Transmission
electron microscopy investigations were conducted with a JEOL 2000FX Transmission
Electron Microscope operating at a voltage of 200 kY. Samples for TEM were ground
and suspended in water. The suspensions were sonicated for 25 min and then applied to
a 400 mesh carbon-coated copper grid.
Low-angle X-ray diffractograms were recorded over the range - 0.7 to - 3.0 28 degrees
(Cu Ka radiation) using a Siemens 05000 diffractometer.

3. Results and Discussion

The results from XRD and TEM studies on the dip-coated films produced from Series I
are summarized in Table 2, for the most representative samples. Data obtained from the
calcined films are shown in the shaded rows to distinguish them from the data
pertaining to the as-synthesized cured films.
In the templating mixtures in Series 1 the ratio of Brij56™ to TMOS was kept constant
and the only variables were the amounts of water and methanol present. Mixture Id
contained the highest amount of water while mixture 1m the lowest. Table 2 shows that
by decreasing the water content of the templating mixture, films having micellar cubic
(II), hexagonal (HI) and lamellar (La) nano-architectures could be obtained irrespective
of the amount of methanol present.
 163

Brij56 (wt%)

(] ld - II (very ard.)
LJ Ie-II
00 If· IJ+HJ (very ord.)
Ii] Ig II +}-fl
00 Ih -Il+HI
o Ii -J;;I (very ord.)
• Ij -I-U+La
• lk - H1+La (very ord.)
o 11 -1:n+La
* 1m-La
• 2a (without order)
• 2b (without oIder)
• 2c-HI+La
* 2d-La
* le-La
* 2f-La

100 90 80 70 60 50 40 30 20 10
0.5 M He, ("",/.)

Figure 1. Map of the nanostructures of the calcined films on a ternary phase diagram.

The final nanostructures of the calcined films prepared from Series 1 and Series 2 are
mapped onto a ternary phase diagram in Figure 1.
Representative X-ray diffraction data obtained from Series 1 films prepared from the
precursor mixtures with 15,0 g of methanol addition, are illustrated in Figures 2. In the
case of the monophasic systems, the X-ray data suggest that the lattice parameters of the
as synthesized cured films fall in the range 5.5 nm to 4.8 nm. The data from the as
synthesized films that are in an homogeneous HI phase, and those in an homogeneous
La phase, suggest that the wall-to-wall distance decreases from - 6 nrn to - 5.2 nm on
traversing the phase diagram. Since it was not possible to assign the structure of the
micellar cubic phase because of insufficient X-ray reflections, Wt: are unable to
determine whether the wall-to-wall distance in the I, phase is larger than that of the HI
phase. However, both the decrease in this parameter on going from the H, to the La
phase, as well as the magnitUde of the change, are comparable' with previous
observations on non-ionic surfactants.
164

<a)

. "
, m
~ BR!J 56
11·70 II R I J 56

, • 60 B R J J 56

-~

f-J~
-A a
" " R I J 56

A ,. ·40 B R I J S 6

"

0
. B R I J ~ 6
"

~ ". lOB R I J 56

/\. ,. ·5 HRJJ 56

- 1 BRIJ 56
"
~

'.5 '.5

(b)

" .
______________________ ~~.-~O~BRIJ56

--____ ~~~==~ ____________________ ~~.~'~08RIJ56

, d ., B R I J 5 6

_.5 '.5

Figure 2. Small angle X-ray diffractograms obtained from dip-coated films processed from the mixtures in
Series 1, for \5,0 g of methanol addition. (a) as synthesized films, (b) calcined films.
 Table 2. Summary oflow angle X-ray diffraction data and observations from TEM for films prepared from mixtures of Series 1.

dspllCitig Fruther p(l!'e to pore

MeOH Peak width at Poro-pore distance
SerillS of oHlrs! fnrfmit peaks lI!Id Structure dl$lallCe from
addition half height of from TEMfnlU
label Brij56 reflect. ref!~'t indexing 12.& llSsignrnent XRJ)illll1
Ig l"lll.......fi l~
first reflect r 14. n Is::?) i.,i"J)<i:: ..........,\ (±05mn)
iwl'% I"""
7.5
1<1
15.0

7,5
If 10
1;,(

75
30
15.(

75
11 41)

Ij 50

!l 70
15,0

0-
VI
166

Calcination of the films in Series 1 leads to pure silica films that retain the nano-
architectures of the cured films. Representative transmission electron micrographs are
shown in Figure 3. A significant contraction in the lattice parameter determined by
XRD is observed following calcination. The contraction is typically a factor of 0.7 for
the II materials, 0.8 for the HI materials and 0.8 for the La materials. If this shrinkage is
of equal magnitude along each of the three cartesian co-ordinates, then it would
correspond to the volume of the calcined film shrinking by nearly a factor of 3
compared with the as synthesized film. This magnitude of shrinkage is comparable with
our previous observations on HI-Si02 and is significantly lower than the volume
shrinkage observed typically with sol-gel synthesis of silica monoliths [8, 10].
The presence of methanol in the precursor mixtures of Series 1 does not appear to affect
the lattice parameters of the as synthesized films, at least within the limits of
experimental reproducibility. By contrast the calcined films do show a significant
decrease in the lattice parameter as a function of increasing methanol content. For
example, in the case ofthe calcined II films the lattice parameter decreases from 3.5 nm
for films prepared with the lowest methanol addition of 7.5 g, to 2.5 nm for films
prepared with 15g methanol. A similar decrease is observed for the calcined HI films,
but the lattice parameter for the calcined La films is essentially unaffected by methanol
concentration

Figure 3. Transmission electron micrographs obtained from calcined dip-coated films processed from the
mixtures in Series 1.

In the case of the dip-coated films the birefringent optical textures were not sufficiently
resolved to allow an unambiguous assignment of the nano-architecture. The only
significant observation that we were able to make was that films that were biphasic
either in their as synthesized state or following calcination were more likely to develop
cracks after calcination.
In the templating mixtures of Series 2 the surfactant to water ratio is constant, but the
amount of TMOS, as well as methanol, is varied. Mixture 2a has the highest
water:TMOS molar ratio (8.6) while mixture 2f has the lowest (2.1). Of all the
compositions investigated, mixture 2c is closest to the stoichiometric composition for
the hydrolysis ofTMOS (1 mole equivalent TMOS to 3.8 mole equivalent water). In the
as synthesized and cured films of Series 2 we found that increasing the amount of
TMOS in the precursor mixture lead to the progressive formation of a lamellar
nanostructure; this is preceded by a broad biphasic domain.
X-ray diffractograms for both as synthesized and calcined films were obtained from
reaction mixture Series 2. For the as synthesized films the lattice parameter does not
 167

appear to change as the amount ofTMOS is increased for the single phase systems. This
observation applies irrespective of the amount of methanol in the templating mixture.
By contrast, calcined films with an L" nanostructure do show some dependence of their
lattice parameters on the amount of methanol. For example, the lattice parameter for
mixture 2m and with 15 g of methanol is 3.1 nm while for the lowest methanol addition
7.5 g is 3.6 nm. Overall the lattice parameter decreases by a factor of ca 0.7 following
calcination.
As in the case of Series 1, we found that biphasic films from Series 2 were more likely
to produce cracked films following calcination than were homogeneous films. This
observation suggests that the polycrystalIine domain morphology of the biphasic films
could lead to unequal shrinkage and the build-up of stress that leads to fracturing and
delamination during calcination.
It is observed from SEM images that the thickness of our films varied in the range 300
nm to 600 nm. Previous work on dip-coated sol-gel mixtures suggests that crack free
films should be achievable with ease provided the cured film thickness is less than ca
l!lm [11]. In these cases shrinkage tends to occur primarily in the direction orthogonal
to the film plane. In both series the thickness of the calcined films was found to
decrease in proportion to the amount of methanol present in the reaction mixture. The
thickness shrinkage is typically a factor of 0.5, decreasing from a typical value of l.l
!lm for the lowest methanol addition to 0.65 !lm, for 15.0g of methanol contend. The
uniformity of the films was observed to improve significantly as a function of the
methanol concentration of the dipping solution.
The widths of the XRD peaks, illustrated in Table 2 for the most significant samples of
Series 1, provide qualitative insights into how the regularity of the mesoporous
architecture is affected by the composition of the precursor mixture. In the case of
Series 1, there was very little change in the diffraction peak width of the as synthesized
films on going from the homogeneous II phase, through the HI phase and into the La.
phase. Similarly, there were no significant changes in peak width between samples with
different methanol concentrations. The calcined films from Series 1 showed an increase
in the peak width compared with the as synthesized films. This is consistent with
previous observations and indicates that calcination introduces a degree of disorder in
the structure, presumably as a consequence of the significant contraction in volume [8].
We did not observe any systematic changes in peak width as a function of methanol
concentration or phase architecture. The films in Series 2 showed similar behavior to
those in Series 1. We did however note that in the case of the samples templated in a
lamellar phase, there seemed to be a bigger increase in the peak width following
calcination than was the case for films templated in either the HI or II phases.
Previous studies on meso porous silica films prepared by dip-coating from precursor
solutions that contained low surfactant concentrations «7 wt%) showed that the
mesoporous nano-architecture of a dip-coated film will often vary across the film [12].
Although at a molecular and meso scopic level dip-coating involves a number of
complex and poorly understood processes, it is clear from earlier work that the rapid
evaporation of water and methanol immediately following film deposition can lead to
severe concentration gradients across the film [12]. We expect this to be the case for
most of our films, but without any direct evidence. However, the observation that the
overall film quality, as judged by optical microscopy, improves with increasing
methanol concentration provides some support for this expectation. Furthermore, the
apparent insensitivity of the lattice parameters in the as synthesized cured films to the
168

methanol concentration of the precursor mixture is a significant observation as it

implies that the final structure adopted by the films is essentially determined by the
relative amounts of surfactant, TMOS and water. In other words our data suggest that
the methanol plays no role in the templating process; its presence ensures that the
solution has the appropriate rheology for dip-coating. The main consequences of having
methanol in the precursor mixture are a more dramatic decrease in the lattice parameters
upon calcination. Presumably this reflects the lower density of the incipient silica
formed in the presence of large amounts of methanol. When the film forms the
methanol flash-evaporates from the film, however, the films retain the structure of the
templating lyotropic phase. In most cases there is further continuity with the structure of
the calcined material.

4. Conclusions

We have demonstrated that it is feasible to use precursor mixtures that contain high
surfactant concentrations to dip-coat thin silica films onto glass slides. The macroscopic
uniformity of the films appears to improve when methanol is added to the mixture,
although methanol plays no role in the templating process. Our data indicate that
amount of methanol in the dip coating mixture has little effect on the structural
parameters of the as synthesized films, but does lead to more significant volume
shrinkage on calcination.

5. Acknowledgements

This work was supported by the Engineering & Physical Sciences Research Council of
the UK. The authors would like to thank Dr. Barbara Cressey for her help with the
TEM. M. Clara Gonr;alves would like to thank Fundar;ao para a Ciencia e a Tecnologia,
who supported the work by a grant FMRHIBSAB/203/0.

6. References

I. Beck, J.S., Vartuli, J.C, Roth, W.J., Leonowicz, M.E., Kresge, C.T., Schmitt, K.D., Chu, CT.-W,
Olson, DR, Sheppard, EW., McCullen, S.B., Higins, J.B., and Schlenker, J.L. (1992), J. Am. Chem.
Soc. 114, \0834.
2. Kresge, CT., Leonowicz, ME, Roth, WI, Vartuli, J.C, and Beck, J.S. (1992), Nature 359,710.
3. Bagshaw, SA, and Pinnavaia, T.J. (1996), Angew. Chem. IntI. Ed. Eng/. 35, 1 \02.
4. Zhao, D., Huo, Q., Feng, J., Chmelka, B.F., and Stucky, G.D. (1998), J. Am. Chem. Soc. 120,6024.
5. Bruinsma, P.J., Kim, AY., Liu, J., and Baskaran, S. (1997), Chem. Mater. 9,2507.
6. Lu, YE, Ganguli, R., Drewien, CA, Anderson, M.T., Brinker, CJ., Gong, W.L., Guo, YX, Soyez, H.,
Dunn, 8., Huang, M.H., and Zink, J.I. (1997), Nature 389, 364.
7. Lu, YF., Yang, Y, Sellinger, A, Lu, M.e., Huang, J.M., Fan, HY., Haddad, R., Lopez, G., Bums,
AR., Sasaki, D.Y., Shelnut, I, and Brinker, CJ. (2001), Nature 410, 913.
8. Attard, G.S., Glyde, IC, and Goltner, CG. (1995), Nature 378,366.
9. Attard, G.S., Edgar, M., and G<lltner, CO. (1998), Acta Materialia 46, 751.
10. Coleman, N.R.B., and Attard, G.S. (2001), Micropor. 7 Mesopor. Matis. 44-45, 73.
11. Brinker, cr, Hurd, AJ., Frye, G.C, Ward, K.J., and Ashley, CS. (1990), J Non-Crystalline Solids
121,294.
12. Brinker, CJ., Hurd, AJ., Schunk, P.R., Frye, G.C, and Ashley, CS. (1992), J. Non-Crystalline Solids
147&148,424.
THIN FILM CRYSTALLISA nON BEHAVIOURS FOR OXIDES OF RARE-
EARTH METALS ON AMORPHOUS SILICA

A.F. ANDREEVA, A.M. KASUMOV

Institute for Problems of Materials Science, National Academy of
Sciences of Ukraine
3, Kr=hi=hanovsky St., 03142 Kyiv, Ukraine

1. Abstract

Crystallisation behaviours of rare-earth metal oxides (REMO) were investigated. By

electron beam evaporation of the rare-earth metals in 0 1 at a pressure of (2 - 6)· 10.3 Pa
their oxide films were obtained on SiOl films that served as substTaites. Thin REMO
films deposited on the amorphous silica surface without any special heating the sub-
strate were amorphous. When the condensation temperature increased, crystallisation of
REMO films began. For different REMO films the start of crystallisation was observed
at a substrate temperature of 80 - 200°C. A characteristic of the initial crystallisation
stages was the appearance of high-temperature REMO phases (monoclinic phase of
SC203, hexagonal phase of Gd 20 3, Er203 etc.) alongside with the C-cubic form. For the
C-cubic form a high degree of the coherent dispersion blocks texturing was typical.
With a small increase in the condensation temperature, the high-temperature phases
disappeared in the film, and crystallites reoriented. In increasing the condensation
temperature, nanocrystallites rebuilt repeatedly in the film. This was accompanied with
reorientation of film texture in the film growth direction. It seems that amorphous silica
surface is to influence on the phase transitions at a growth boundary during the initial
stages of the REMO film crystallisation.

2. Introduction

The REMO films have high thermal and chemical stability, mechanical strength and
adhesion to various substrates as well as good dielectric and optical characteristics, and
stability of parameters under radiation. This makes usage of the REMO films in micro-
electronics, optics, computer engineering and other fields very promising.
The literary data on REMO film structure and physical characte:ristics are discor-
dant, which can be explained by the fact that condensate characteristics depend upon the
processing technology used. Today the REMO films are mostly obtained by means of
vacuum evaporation of the corresponding oxides with electron beam, or by the high-
frequency sputtering of oxides, and also by oxidation of a metal milTor in the oxygen
atmosphere [1-3].
169
Y. C. Cogotsi and I. V Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 169-174.
© 2003 Kluwer Academic Publishers.
170

These methods have shortcomings connected with the following technological diffi-
culties: dissociation of refractory REMO resulting in change of the condensate stoichi-
ometric composition in evaporating; contamination of films with impurity gases during
the ion-plasmous sputtering due to the extremely low rate of film growth; incompatibi-
lity of the process of metal mirror oxidation with other technological processes in the
production of integrated circuits or optical instruments.
An effective method to obtain the REMO films is so-called reactive synthesis, which
is vacuum evaporation of a metal in the oxygen atmosphere. Due to the high reactivity
of rare-earth metals and their ability for monoatomic evaporation (no clusters) it is pos-
sible to obtain REMO films having a composition close to the stoichiometric one at a
low temperature and high deposition rate [4, 5].
REMO films obtained by the reactive synthesis method proved to be promising for
the applications of masking layers in photomasks, sensors for high-power heat flows,
dielectric layers in capacitors and multilevel integrated circuits, and reflecting and anti-
reflecting coatings [6, 7].
The possibility of using films of some REMO as high-refractive layers in dielectric
laser mirrors shows considerable promise. The interference laser mirrors, which are
essentially a number of alternating coatings of REMO and amorphous silicon oxide, are
leaders as for the width of operating optical interval, stability under high-power
radiation flows, and a small value of light scattering [8].
A distinctive feature of multilayer coatings for laser mirrors is an essential depen-
dence of their light scattering value upon the condensation temperature of layers. Even a
small rise in the condensation temperature results in deterioration of mirror light
scattering by several times.
On considering all the information stated above, it seems interesting to explore the
effect of condensation temperature on the structure of REMO films that are deposited
on the films of amorphous silica.

3. Materials and methods

The REMO films were obtained using the electron-beam evaporation of the rare-earth
metal in the Oz atmosphere under a pressure of (2-6)'10- 3 Pa. The evaporation chamber
was pumped to less than 1'10.4 Pa, before the oxygen was admitted there. The oxygen of
99,5 % purity was used as a reaction gas.
The REMO films were deposited on the amorphous Si0 2 films that had been preli-
minarily obtained on the surface ofNaCI monocrystal by the electron-beam evaporation
of Si0 2 • The Si0 2 films had been separated from NaCI after the REMO films were
deposited on Si0 2 ones.
The thickness of both SiO z and REMO films was 20-40 nm. The studies were made
with the oxides of Sc, Y, Gd, Ho, Er, and Lu. The condensation temperature was varied
in the range 50 - 400D C. The film structure was examined in the transmitted rays with
the EG electron diffraction camera.

4. Experimental results

According to the electron diffraction data, when the temperature specified for REMO
film condensation, Tc, was in the range 50 - 80D C, all the condensates were amorphous.
 171

At Tc ~ 80-100°C the crystallisation process was observed in the REMO films of

Gd20 3, Er203, and LU203; at Tc ~ 200-250°C - in the films ofY203 and HQz03; and at Tc
~ 330°C - in SC203 films.
A specific feature for all the oxides considered was the formation of REM
sesquioxides like Me203'
For a number of the examined REMO, at the initial stages of crystallisation, the
formation of high-temperature unstable phases was observed alongside with the cubic
C-form (the lattice type Mn203). In the electron diffraction patterns of condensates the
peak that corresponds to the strongest <101> for the hexagonal structure for Gd20 3 and
H020 3, and to <201> for the monoclinic structure for SC203 was detected. The presence
of the high-temperature monoclinic phase was also observed for the neodymium oxide
that was obtained in the work [9] by the reactive synthesis method.
With an increase in the condensation temperature by 1O-20"C, the high-temperature
phases faded out, and only the cubic structure (C-form) formation was observed.
The electron diffraction patterns from condensates at the first stages of C-form cry-
stallisation have the following features. Unlike the tabular data, the first diffraction pe-
aks appeared were in most cases those which are to be of small intensity in the corres-
ponding REMO crystal structure. The diffraction peak <211> of medium intensity was
displayed as well.
For example, at the initial stages of Gd20 3 crystallisation besides the peak corres-
ponding to the strongest peak from the crystallographic plane <101> in the high-tem-
perature hexagonal phase, two intensive peaks from <321> and <530> planes of the
cubic phase were detected though by tabular data these intensities are to be small.
With an increase in the condensation temperature up to 250°C, the peaks corres-
ponding to the initial stages of crystallisation disappeared, and the most intensive peaks
for cubic C-form appeared.
For the initial stages of H020 3 crystallisation the two intensive peaks were observed:
<101> from the hexagonal phase, and <521> from the cubic phase; the intensity of the
<521> peak, according to the tabular data for the structure of H0 20 3, is to be small.
The crystallisation behaviours for Er203 and for LU203 are similar to the described
above.
As an example, Fig. 1 shows the diffraction pattern from the SC203 films obtained at
different condensation temperatures. The condensates deposited at Tc=50-330"C had the
amorphous structure. When Tc was increased up to 340°C, in eh:ctron diffraction
patterns from the films two rather sharp and intensive peaks were observed. These peaks
correspond to the low-temperature cubic (C-form, <211», and to high-temperature
«201» monoclinic structures of SC203. With the increase in Tc up to 350°C, the
diffraction peak from the crystallographic plane <201> disappeared; and the peak from
the <211> plane became weaker and more blurred. The peaks corresponding to the
intensive in the C-form peaks <222>, <332>, and <411> appeared in the electron
diffraction patterns. As for the tabular data, the intensity of the <411> peak is to be
small.
172

(222) [440)
a
..
[211]
~

=
=
.. b
Q,j
~

CU

~
---c
......
I..

___""--r--_ _ _ _ _d

..==
~

~ --AA-.. e
....
f
o 10 20 30
2 e, grad
Figure. Diffraction pattern from Sc,OJ films obtained at various T,oc:
a - 380; b - 370; c - 360; d - 350; e - 340; f - tabular data

At the initial stages of crystallisation, the location of the peak corresponding to the
<211> plane is noticeably different (by 4-5 %) from the tabular values given in
reference books, approaching them with rising the condensation temperature.

5. Discussion

It seems very likely that the amorphous silica surface at the growth boundary influences
the features of initial stages of the REMO films crystallisation.
The Si0 4 tetrahedrons oriented in a particular way are the basis for all the structural
modifications of silicon oxide (cristobalite, tridymite, and others). In amorphous silica
the close order is almost coincident with that in crystalline Si0 2 •
Each atom of silicon is enclosed into tetrahedron formed by four atoms of oxygen,
whilst each oxygen atom is linked with two silicon atoms, forming a flexible bond,
which exhibits some spread in the Si-O-Si angle a, ranging from 120° to 180° and
having the maximum at a = 144°. The possibility of varying the bond angle a in
 173

amorphous quartz results in some elasticity of the Si-Si distance, which for u=144°
makes about 0.31 nm.
The interatomic distances for both Si0 2 crystalline and amorphous modifications
[10] are presented in Table 1.

TABLE I. The interatomic distances for crystalline and amorphous

fused quartz and amorphous SiO, films, nm

Q- /3- Amor- Amorphous

Cristobalite Tridymit phous SiO,films
e fused obtained by
quartz sputtering
Si l -0 1 0.154 0.154 0.162 0.160
01 - 0, 0.252 0.252 0.265 0.265
Si l - Si2 0.308 0.308 0.340 0310
Si l -0 2 0.388 0.388 0.400 0388
01 - 0, 0.436 0.436 0.450 0.436

The misfit between the interatomic distances in REMO lattices with nearest for them
interatomic distances in Si0 2 was evaluated by the formula:

fi=2(a;-bJI(a; + bJ.

where a; is the interatomic distance in the i-th direction in the REMO lattice, and b, is
the nearest interatomic distance in Si02 [11). The results are listed in T.able 2.

TABLE 2. The misfit values for the interatomic distances in dift:erent

directions in REMO lattices to the nearest ones in Si02, %

Oxide <222>, <400>, <622>,

Si l - Si" 0 1 - 0" Sh ··01,
Sc2O, 8.7 7.4 7.5
Y20, 1.3 0.15 0.18
Gd2O, 0.4 1.7 1.6
H<nO, 1.6 0.2 0.3
Er20, 1.9 0.7 0.7
Lu 2O, 3.3 1.9 2.1

As we can see, the interatomic distance between <222>, <400>, and <622> planes in
the REMO are close enough to the corresponding distances Si} - Si 2 , O} - O2 , Si} - 01 on
the silica surface.
It is known that for systems having a strong interphase interaction and a small misfit
for interatomic distances, the formation of the substrate-to-the film interface structure
occurs by Frank's and Van der Merwe's mechanism and is accompanied with forming
an interstitial pseudomorphous transitional layer [12]. The condensate in the latter is
coherently adjoined to the substrate, and misfit becomes complete:ly eliminated on
account of the film homogeneous deformation.
It seems that just availability of the pseudomorphous, amorphous layer in the REMO
is responsible for a small value of the light dispersion loss in laser mirrors based on
174

these oxides.
With increasing condensation temperature, the coherent conjugation of condensate
to substrate is getting broken which is accompanied with active crystallisation and
refining the growing layer crystalline structure; the light dispersion of the pack of
Si0 2-REMO layers becoming significantly deteriorated.
Hence, degradation in the optical properties of dielectric laser mirrors with a small
increase in the condensation temperature (by 50-100°C) is connected with the
breakdown of the structure coherence between the Si02 surface and the REMO layer,
which is caused by active crystallisation processes in the rare-earth metal oxide films.

6. References

I. Rainer, F., Lowermilk, W.H., Milan, D. et al. (1985) Materials for optical coatings in the ultraviolet, Appl.
Optics, 24,496-500.
2. Serebrenikov, V.v, Kozik, VV., Sergeev, A.N., Jakunina, G.M. (1979) Obtaining, properties and
application of the films based on the REE oxides, in Refractory compounds of the REE, Nauka,
Novosibirsk, pp. 117 - 120. (in Russian)
3. Vdovin, OS., Kirijashkina, Z.I. et al. (1988) Thefilms of the rare-earth metals oxides in MDM and MDS
structures, Publ. of Saratov State University, Saratov. (in Russian)
4. Andreeva, A.F., Gilman, I.Ja. (1983) Crystalline structure and some properties of the Yttrium oxide films,
Izvestija of USSR Academy ofSciences, Inorganic materials, 19,125-127. (in Russian)
5. Andreeva, A.F., Sisonyuk, A.G., Himich, E.G. (1994) Growth conditions, optical and dielectric properties
of Yttrium oxide thin films, Phys. St. So/., AI4S, 441-446.
6. Andreeva, A.F. (1994) The rare-earth metals oxides films (the growth features, structure, phySical
properties, areas ofapplication), Publ. of IPMS ofNAS of Ukraine., Kyiv. ( in Russian)
7. Andreeva, A.F. (1998) Obtaining and properties of the rare-earth metals oxides films, Powder metallurgy,
1-2,107 - 111. ( in Russian)
8. Andreeva, A.F., Kasumov, A.M. (1999) The structure and optical properties of Scandium oxide, The
questions of the atomic science and technics, 2 (9), 79-82. (in Russian)
9. Takagi, K., Fukazana, T., Susa, K. (1987) Epitaxial growth of neodymium sesquioxide films on calcium
fluoride by reactive evaporation method, J. App/. Phys. 61, 1030 - 1034.
10. Revesz, A.G. (1980) The defect structure of vitreous SiO, films on silicon. I Structure of vitreous SiO,
and the nature of the Si - 0 bond, Phys. St. Sol, AS7, 441-446.
II. Van der Merwe, lH. (1964) Interfacial misfit and bonding between oriented films and their substrates, in
M.H. Francombe and H. Sato (eds.), Single-crystal films, Symposium Publications Division, Pergamon
Press, pp. 139-163.
12.levlev, V.M., Trusov, L.I., Holminskij, VA (1982) The structural transformations in thin films,
Metallurgia, Moscow. (in Russian)
STRUCTURE AND MECHANICAL PROPERTIES OF NANOSTRUCTURED
METAL-CARBON FILMS

Y. PAULEAU!, F. THIERY!, J. PELLETIER!, V.V. UGLOV 2 , A.K.

KULESHOV 2 , S.N. DUB 3 , and M.P. SAMTSOV 2
I Centre National de la Recherche Scientifique, LEMD. Grenoble, France.

2Belarussian State University, Minsk, Belarus.

3/nstitute for Superhard Materials, Kiev, Ukraine.

Abstract

Nanocrystalline copper/hydrogenated amorphous carbon (nc-Cula-C:H) films have been

deposited on Si substrates by a microwave plasma-enhanced chemical vapor deposition
process combined with a sputter deposition process. The deposition rate of films was
found to vary in the range of 15 to 25 nm/min depending on the composition of argon-
methane mixtures. These films were analyzed by Rutherford backscattering
spectroscopy and Raman spectroscopy. The crystallographic structure, nanohardness,
and friction coefficient of films were investigated as functions of the carbon content.

1. Introduction

Amorphous carbon films such as diamond-like carbon (DLC), amorphous hydrogenated

carbon (a-C:H) and i-C films produced by carbon ion beam depositiion may provide low
friction coefficients and reduced wear rates for sliding tests conducted under various
conditions Pl. However, the magnitude of compressive residual stresses developed in
these films may reach several GPa and the adhesion of films is nolt sufficient to ensure
the reliability desired for mechanical components in industrial applications. In addition,
the friction coefficient of carbon films increases as the relative humidity in the
environment increases [2]. Recently, novel materials have been proposed to overcome
these difficulties, in particular DLC nanocomposite or DLN films composed of
amorphous carbon (DLC or a-C:H) and a-Si:O vitreous phase [3,4]. The hardness of
these DLN films between 12 and 17 GPa is less than the hardness of DLC films (15 to
25 GPa). However, the residual stresses are significantly lower (0.5 to 1 GPa) and these
DLN films were found to be tightly adherent to various types of substrates.
The objective of the present work is to prepare nanostructured copper-amorphous
carbon (nc-Cula-C:H) films exhibiting tribological properties similar to those of DLN
films. The deposition of films was carried out using a new plasma reactor based on
multipolar magnetic confinement, named as distributed electron cyclotron resonance
(DECR) microwave plasma [5]. The composition of films was determined by
175
Y.G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 175-\80.
© 2003 Kluwer Academic Publishers.
176

Rutherford backscattering spectroscopy (RBS) and Raman spectroscopy. The

crystallographic structure, nanohardness, and friction coefficient of films were
investigated as functions of the carbon content.

2. Experimental procedure

Copper-carbon composite films have been deposited on (lOO)-oriented single crystal

silicon substrates mounted on a water-cooled substrate holder. The deposition technique
consisted of a combination of the plasma-enhanced chemical vapor deposition
(PECVD) of carbon from argon-methane mixtures associated with sputter deposition of
copper from a copper target [5J. The microwave power was adjusted at 400 W. The d.c.
voltage applied to the copper target of 100 mm in diameter was maintained at - 300 V.
The sputtering power was in the range 45-75 W. The substrate holder, at 13 cm from the
target, was maintained at the floating potential. Prior to the gas phase ionization, the
total pressure in the reactor was fixed at 0.13 Pa while the CH4 concentration in the gas
phase was varied from zero to 100 %. The maximum flow rates of pure argon and CH4
required for a total pressure of 0.13 Pa were 8 and 6 cm3 /min, respectively.
The composition of films was obtained from RBS measurements using a particles of 2
MeV or protons of 1 MeV depending on the film thickness; the incidence and detection
angles were 90° and 160°, respectively. The films were also analyzed by Raman
spectroscopy using a Spex 1403 spectrometer with a photomultiplier tube (R828) cooled
to 243 K. An argon ion laser beam at a wavelength of 488 nm performed the excitation
with a power on the sample of 0.3-0.35 W. The structure of films was investigated by x-
ray diffraction techniques using an Inel diffractometer in 8-28 position, with a
collecting angle varying from 5 to 120° and the usual copper Kal radiation (A. = 1.5406
A). The position and the integral width of x-ray diffraction peaks were determined to
identify the crystalline phase and to calculate the grain size from the Scherrer equation.
The mechanical properties of films were determined by nanoindentation using a Nano
Indenter II equipped with a Berkovich indenter. The load was varied in the range of
0.25-50 mN. The hardness and modulus of elasticity were found at the minimal load
using the Oliver and Pharr model [6]. The average indentation contact pressure (Meyer
hardness) was found by the method proposed by Dub et al. [71. The pin-on-surface
tribological tests were carried out by means of a TAy-tM tribometer with a pin made of
BK-8 hard alloy (hardness of 87.5 HRC). The sliding speed was maintained at 4 mmls.
The load was either 0.5 or I N.

3. Experimental results and discussion

Pure copper and carbon containing copper films have been deposited on Si substrates
covered with a native oxide layer under various experimental conditions. These films of
0.1 to 0.5 /lm in thickness, tightly adherent to Si substrates, exhibited a shiny surface.

3.1. DEPOSITION RATE AND COMPOSITION OF FILMS

The deposition rate of Cu-C films decreased very slowly from 260 to 210 Almin as the
methane concentration in the gas phase increased up to 50 %. The minimum value of
 177

the deposition rate of 140 A/min was reached for Cu-C films produced from argon-CH4
mixtures containing 60 % of methane. The deposition rate of films was approximately
160 Almin for films produced from pure methane microwave dischrurges.
The concentration of Cu and C atoms in the films determined by RBS measurements
was found to be dependent on the CH4 concentration in the gas phase. The argon
content in the pure copper and Cu-C films was less than the detection limit (1 at %) of
RBS analysis. The carbon content reached 20 at. % (and the coppc!r content was 80 at.
%) in Cu-C films deposited from argon-CH4 mixtures containing 10 % of CH4 • The
carbon content in the films increased slightly from 20 to 25 at. % (and the copper
content decreased from 80 to 75 at. %) as the CH4 concentration in the gas phase
increased from 10 to 60 %. The increase in CH4 concentration from 60 to 70 % resulted
in a large increase in C content from 25 to 60 at. %. Then, the C content in the films
increased progressively from 60 to 75 at. % (while the Cu content decreased from 40 to
25 at. %) as the CH4 concentration in the gas phase increased from 70 to 100 %.
Typical Raman scattering spectra (RS) of Cu-C films are given in Fig.1. These spectra
are characterized by the presence of two peaks at 1340 cm· 1 (broad D-peak) and 1560
cm- I (G-peak). According to Theye et aI. [8], the sharp G-peak in Ithe spectra of carbon
films defines the scattering by optic zone center phonons of graphite (Sp2 bonded carbon
atoms). The D-peak in RS spectra of carbon films is usually diffused and is
characteristic of the scattering from «disordered» carbon clusters (Sp3 bonded carbon
atoms). This kind of RS spectra with two smeared D and G peaks is characteristic of
DLC films [8,9].

1200 1400 1600 1800

RAMAN SHIff (em-I)
Fig. I - Typical Raman spectra of Cu-C films produced from Ar-CR, mixtures of various CR, concentrations.
The C content in the films was 25, 65 and 75 at. % for 60,80 and 100 % of CH4 in the gas phase, respectively.

3.2. CRYSTALLOGRAPHIC STRUCTURE OF ALMS

Typical XRD patterns of films produced from argon-methane mixtures of various

compositions are given in Fig.2. The XRD patterns of carbon containing copper films
178

displayed only diffraction peaks at diffraction angles corresponding to those of pure

f.c.c. copper phase [10]. Therefore, these films consisted of polycrystalline copper and
amorphous carbon phase with various metal and carbon concentrations. The position of
diffraction peaks in the XRD patterns was independent of the methane concentration in
the gas phase; as a result, the crystal lattice parameter of the copper phase was constant
and independent of the composition of Cu-C films. However, the intensities and widths
of diffraction peaks were found to be dependent on the methane concentration in the gas
phase, Le., composition of films. Depending on the carbon content in the Cu-C films,
two types of films can be distinguished on the basis of XRD data. For films containing
less than 25 at. % of carbon produced from argon-methane mixtures with a methane
concentration below 60 %, the XRD patterns were similar to that of pure copper
reported in the JCPDS card. A number of diffraction peaks corresponding to the f.c.c.
copper phase were not detectable in the XRD patterns of Cu-C films containing more
than 55 at. % of carbon produced from argon-methane mixtures with a methane
concentration higher than 70 %. In addition, the diffraction peaks broadened
progressively as the carbon content in the Cu-C films increased. For instance, the XRD
pattern of Cu-C films produced from pure methane microwave discharges exhibited
only the (111) diffraction peak ascribable to the f.c.c. copper phase.

Fig.2 - Typical X-ray diffraction patterns of Cu-C films of various C contents produced from various C~
concentrations in the gas phase. The (hkl) Miller indices are those of the f.c.c. Cu phase.

For Cu-C films produced from argon-methane mixtures with a methane concentration
lower than 60 %, the copper grain size was in the range 150-300 A.. A large variation of
grain size was observed as the methane concentration in the gas phase was increased
from 60 to 70 %, Le., as the carbon content in the films increased from 25 to 60 at. %.
After this transition, for methane concentrations higher than 70 % or carbon content in
 179

the films in the range 60-75 at. %, the copper grain size was lower than 50 A. These Cu-
C films produced from argon-methane mixtures containing more than 70 % of methane
consisted of nanosized copper crystallites (40 to 30 at. % in the films) dispersed through
an amorphous carbon matrix (60 to 70 at. % in the films).

3.3. MECHANICAL AND TRIBOLOGICAL PROPERTIES OF FILMS

The nanohardness and elastic properties of Cu-C composite films were found to be
dependent on the methane concentration in the plasma or carbon content in the films
(Fig.3). The hardness varied from 2.5 to 1.9 GPa with increasing methane concentration
up to 40 %; then, the hardness increased from 1.9 to 2.6 GPa as the methane
concentration increased from 40 to J00% in the gas phase. The magnitude of the Young
modulus of composite films decreased from 160 to 30 GPa with increasing methane
concentration in the plasma or carbon content in the films.
U 20 40 60 80 100
3.0 180

~----~
160 -<
~2.5 140 g
Z
8 I ,
' L I
120 Cl
'"'"~ 2.0 - ...¢'
100
~
§
Ci 80 c:::
0:: r
c:::
~ 1.S 60 '"
'0
40 ~
1.0 20
0 20 40 60 80 100
METHANE CONCENTRATION (%)
Fig.3 - Hardness (curve in dashed line) and Young modulus (curve in solid line) of films produced from
various CH4 concentrations in the gas phase.

The friction coefficient for pure Cu and Cu-C films was observed to be dependent on
the sliding distance (Fig.4). Three different regions can be distinguished by examination
of wear tracks in the films by optical methods. A pure Cu film deposited on !! Si
substrate can be considered as a reference sample (Fig.4, curve B). For a sliding
distance varying from 0 to 4 m where the friction coefficient increases nearly linearly
with increasing sliding distance, the friction arises only from thl~ pin sliding on the
copper film. Assuming that the film has a uniform hardness depth distribution and that
the pin does not change its form while rubbing, the increase in the friction coefficient
may be attributed to a gradual increase in the real contact area [11]. A transition region
from 4 to 6 m corresponds to the pin friction both in the Cu film and on a very smooth
Si surface. In the region from 6 m to higher sliding distances, a great variation in the
magnitude of the friction coefficient can be observed and corresponds to the penetration
of a greater part of the pin into the substrate.
The best wear resistance was obtained for Cu-C films containing approximately 20 at.
% of carbon with a maximum copper grain sizes of 20 nm (Fig.4, curve C). The
180

comparison between the tribological behavior of pure copper and Cu-C films reveals
that the major improvement of friction properties in terms of friction coefficient and
wear rate is obtained from Cu-C films containing 20 to 25 at. % of carbon. However, the
friction coefficient and wear rate of these films are still relatively high with respect to
data given in the literature for DLC films.
0.8.-~~--r--~--'r--~-"'-~---r--~--'

~
0 0 .6
~
8 0.4
Z
o
6 0.2
~
2 4 6 8 10
SLIDING DISTANCE (m)
Fig.4 - Friction coefficient vs sliding distance for : bare Si substrate (curve A), pure Cu film (curve B), Cu-C
films containing 20 at.% of C (curve C) and 25 at.% ofC (curve D).

Acknowledgements
The partial financial support of this research project by the NATO Scientific and
Environmental Affairs Division, Brussels, Belgium, from a Collaborative Linkage
Grant, is gratefully acknowledged.

References
I. Grill, A., Patel, V., and Meyerson, B. (1991) Tribological behavior of diamond-like carbon: effects of
preparation conditions and annealing, Surf Coat. Technol. 49, 530-536.
2. Franks, J., Enke, K, and Richardts, A. (1990) Friction coefficient of DLC films, Met. Mater., Nov., 695.
3. Dorfman, V.F. (1992) Diamond-like nanocomposites (DLN), Thin Solid Films 212, 267-273.
4. Neerinck, D., Persoone, P., Goel, A., Venkatraman, c., Kester, D., Halter, c., and Bray, D. (1996)
Tribological properties and structural investigation of diamond-like nanocomposites, "Proceedings of the
5th International Symposium on Trends and New Applications in Thin Films", Le Vide: Science,
Technique et Applications, WSup.(279), 67-69.
5. Pichot, M., and Pelletier, J. (1992) Distributed electron cyclotron resonance (DECR) plasmas, in M.
Moisan and J. Pelletier (eds.), Microwave Excited Plasma, Plasma Technology, 4, Chap. 14, Elsevier,
Amsterdam, The Netherlands, pp. 419-434.
6. Oliver, W.C., and Pharr, G.M. (1992) An improved technique for determining hardness and clastic modulus
using load and displacement sensing indentation experiments, 1. Mater. Res. 7(6), 1564-1583.
7. Dub, S., Novikov, N., and Milman, Yu. (2002) The transition from elastic to plastic behaviour in AI-Cu-Fe
quasicrystal studied by cyclic nanoindentation, Phil. Mag. A 82(10), 2161-2172 ..
8. Theye, M.-L., Paret, V., and Sadki, A. (2001) Relations between the deposition conditions, the
microstructure and the defects in PECVD hydrogenated amorphous carbon films; influence on the
electronic density of states, Diamond Rei. Mater. 10(2), 182-190.
9. Robertson, J. (2001) Ultrathin carbon coatings for magnetic storage technology, Thin Solid Films 383,81-
88.
10. JCPDS Data Cards, 04-0836 (International Center of Diffraction Data, Swarthmore, PA, 1999).
II. Wang, F.D., and K Kato, K (2002) Effect of coating thickness on friction for carbon nitride films in
repeated sliding against a spherical diamond with nano-scale asperities, Wear 252(3-4),210-219.
THIN FILM METAL MATRIXES FOR BIOCORROSION STUDIES

O.K. DYOYNENKO, I.A.KOZLOYA, Y.M. STATSENKO

Institute for Problems of Materials Science, National Academy of
Sciences of Ukraine
3, Krzhizhanovsky St., 03142 Kyiv, Ukraine

1. Abstract

The new approach for analysis of steel construction biocorrosion with using of thin film
metal matrixes (TFMM) is discussed. With the help ofTFMM, it is possible to study the
activities of the sulphate reducing bacteria and their metabolism products, and to
determin quantitative indicators of their aggressiveness and effectiveness of various
inhibitors as well.

2. Introduction

Microbial induced corrosion of underground metal and ferroconcrete: constructions is

the actual ecological and economical problem of the current time. The corrosion refusal
of gas main, which in 80-90% of cases is stimulated by active lif.: of the sulphate
reducing bacteria, leads to the considerable financial expenditure. One: outbreak of gas
main is approximately priced as $6-8 millions of the economical loss and is
accompanied with unrecoverable ecological disturbances of the environment. Therefore,
the problem of analysing the mechanisms of the bacteria and metal interplay, purposed
to prevent the biocorrosion, is being actual and timely.
The investigation of mechanism of the microbial induced corrosion in biofilms is
being actively carried on from the end of the 1980s, and the dominant attention is being
paid to the sulphate reducing bacteria [1, 2].
In spite of the some positive results, a considerable part of questions, including the
composition of bacteria, which formed the biofilm, quantitative indicators of their
aggressiveness, and also the role of their metabolism products in the biofilm formation
and its functioning, remains unexplored till this time.
In this work the new methodical approach for analysis of the microbial induced
corrosion is proposed.
As the model objects are offered to use thin film metal matrixes,. with the help of
which it is possible, using the modem research techniques of a surface, to study
successfully an activity of biofilms, also an aggressiveness of their components to the
metal and an influence of various inhibitors.
181
Y. G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 181-185.
© 2003 Kluwer Academic Publishers.
182

The effectiveness of biocorrosion studying on the model iron and steel matrixes is
greatly determined by a possibility to produce the metal layers, which on structure and
chemical composition are close to the gas main.
There are known many methods for preparing the iron and steel films. The most
widely used are: the vacuum evaporation by electron beam or laser, the magnetron or
high frequency sputtering of the targets with the corresponding composition, the
electrolytic deposition etc. Among these methods the magnetron sputtering represents
the most perspective to deposit the thin iron and steel films.
The chemical composition of films obtained through the magnetron sputtering does
not differ practically from the composition of target, the condensates properties are close
to the properties of a bulk material on a determined thickness of layer. The films,
obtained by a magnetron sputtering, have a small porosity and a high adhesion to
substrate.
It is difficult to prepare the magnetron sputtered iron films because of shunting the
magnetic field by the material of target. Authors of work [3] partially diminished the
shunting by means of the special target construction' and heating it by 10-15 °C over the
magnetic transition temperature.

3. Materials and methods

In the present work was applied a target of the original construction that consists of
several rings. One of these rings was here as a principal working area, its surface was
sputtered, the other ones served to reduce the shunting and for cooling of the working
ring [4].
The deposition of metal matrixes was performed in the vacuum device of
URM 327.90.36 mark with a magnetron built-in unit U-4, under the following
parameters: the pressure of argon 1.3 Pa, the voltage 300-400 V; the cathode current
2.2 A. The rate of deposition was 1.5-2 nm·s- I .
The films were deposited on substrates of optically clear quartz glass of the K-8
mark. The thickness of films was varied in the range 0.03-1.6 1-lJ11. As a material of the
target were used the iron and Fe360-B steel, which mainly used for construction of
welded and jointless gas lines.
A crystalline structure of films and of the target was studied with the electron
diffraction apparatus of EG mark and with the X-ray DRON-3 diffractometer, a surface
morphology - with the microscopes MIM-lO and MII-4.
A thickness of the iron films, steel and of the biological layer that is formed as a
result of interaction of the bacteria metabolism products with iron were measured on the
interference microscope MIl-4.
The decrement of iron mass, caused by the microbial induced corrosion, was
determined through measuring change in optical density on the "Specord UV-Vis"
spectrophotometer and using the post processing of data with the aid of specially
designed computer program.
The following strains of sulphate reducing bacteria were explored: the Indonesia
(products of corrosion of the oil ship in the Indian Ocean), the Kyiv-lO (corrosion
 183

products of ferroconcrete reinforcement, Ukraine), the Portsmouth and the Alaska

(corrosion products of the metal constructions in Portsmouth (England) and in Alaska
respectively).
The derivation of extracellular polymers of the listed above bacilliculture was
prepared with the procedure described in the work [5].

4. Experimental results

The structure of thin layer iron films was amorphous. With increase of thickness to up
-30 run, was observed a crystallisation of condensates with the formation of two phases
ofiron: bcc (a) and fcc (y).
On the further increase of the film thickness the maximums, which correspond to the
fcc structure, have disappeared on the electron diffraction patterns, and here remained
the maximums only that correspond to the a-phase.
On X-ray photograph from the films with thickness > 0.6 ~ were present the a-
iron phase lines only. According to the data of X-ray phase analysis, the structure of
films with thickness> 0.6 ~ coincides with the structure of the sputtered target.
In the Fig. 1 are shown the photos of surface of different thickness films, which were
obtained by the magnetron sputtering, and also a surface ofthe plate of Fe360-B steel.

Fig. I. Morphology of the steel films surface depending on their thickm'ss:

I - 0,6 I'm; 2 - 0,8 I'm; 3 - target.

The surface of the films and of the plate was preliminary processed with the etching
reagent for detection of a surface of grains, on the procedure described in [6].
As it is seen, the grain size has depended on the thickness of condensates. With
increase in film thickness, a texture of the surface was becoming more relief,
approaching in appearance to the bulk material.
The size of coherent dispersion blocks, which was determined on the Seljakov's [7]
procedure, depends on the thickness of condensates h and varies from 1,8 10-5 cm (for
h = 0.5 Ilm) up to 3.2'10.5 cm (for h = 1.6 ~). For a surface of the steel plate, the size
of coherent dispersion blocks was 4.41'10-5 cm, the average grain size, determined on
184

MIM-lO with the aid of digital camera and "Olympus DD-soft" program, was
24.10. 5 cm. The size of coherent dispersion blocks for the plate of Ft360-B steel has
turned up as 5.5 times smaller than its average grain size. Figure 2 presences the
dependence of coherent dispersion blocks size from film thickness.

36

.
"{~ 32
~

~ 2.

~
~ 2.4
~

I
~
2.0

16
04

Figure 2. Dependence of coherent dispersion blocks size from film thickness.

These data are in agreement with work [7], on whose data, in martensite steel and in
some other metals also, the areas of coherent dispersion there are importantly smaller of
the grains.
From the data shown, one can see that marked off with the most corrosively are the
extracellular polymers of the Indonesia strain (3.8.10- 5 g), the activity of the Alaska
strain is lower almost by the order (5-4·10-6 g).
Extracellular polymers are known as characterised by their feature to trap the ions of
iron and to accelerate thus an anodic reaction [2]. It is seen from the figure that
extracellular polymers of Kyiv-l 0 and Alaska strains trap the ions of iron on the large
area, forming the lace-like pattern. An extracellular polymer of the Portsmouth strain
produces a clear-cut contour, absorbing all the ions of iron within boundaries of a drop.

5_ Discussion

The obtained results confirm on the whole the known from literature the characteristics
of aggressively of the bacteria explored strains. The most active agents of corrosion
turned out the Indonesia and the Kiev-lO, less active was the Alaska. A high activity has
been surprisingly exhibited by the extracellular polymers of the Portsmouth strain,
contrary to the notions, prevailing about them, as the poorly aggressive. However, the
conception of aggressiveness of some bacteria or other should be in the further
specialised, while will take place the progress of investigations of the hydrogenous
activity role, the structure of sulphides and various extracellular polymers as the factors
of aggressivity.
 185

Then, the approach proposed is the perspective to carry on the comparative studies
of the bacteria metabolites contribution to the overall process of the microbial induced
corrosion.
For the identification of the results obtained on thin film metal matrixes with those
for a massive steelworks it is necessary to develop the further examinations. It seems
interesting to explore the influence of a crystallite size, porosity, and of content of
foreign substances on surface of metal on the corrosion activity of the particular strains
of bacteria and their metabolites. It will be possible in this case to extrapolate on the
massive steel objects those quantitative results, which are obtained on the thin film
model matrixes, which will allow achieve an essentially greater accuracy in the
determining of quantitative indicators of corrosion activity.

6. References

I. Ford, T, Mitchel, R. (1990) Advances in Microbial Ecology, Plenum Press, New York, 11, pp. 231-262.
2. Hamilton, W.A, (1991) System. Proc. Int. Workshop. Stuttgart. Sept. 13-14. Springer-Verlag, pp.187-193.
3. Danilin, B.S., Syrchin, V.K (1982) Magnetron sputtering systems, Radio i svjaz', Moscow, 72 p. (in
Rassian)
4. Dvoynenko, O.K et al. (2002) Structure of films obtained by magnetron sputtering, in Skorokhod, V. V.,
Ragulja, AY, Uvarova, LV. (eds.) Nanocrystalline materials. The publ. of IPMS of NAS of Ukraine.
Kyiv, pp. 138-141. (in Rassian)
5. Kozlova, I. et al. (2000) New approach to the study of the microbial induced corrosion in biofilms. Phys.-
Chem. Mech. Mat. Special Issue, 2, N. I, pp. 626-629. (in Rassian)
6. Kovalenko, V.S. (1973) Metal/ographic reagents. Handbook, Metallurgija, Moscow, pp. 8-9. (in Rassian)
7. Lysak, L. I. (1954) Investigation of the intragranular mosaic structure of metal by the width of X-ray
interference lines. Questions o/metal physics and metallurgy,S, pp 45 -60. (in Rassian)
SYNTHESIS AND PROPERTIES OF NANOSTRUCTURED CARBON
MATERIALS: NANODIAMOND, ONION-LIKE CARBON AND CARBON
NANOTUBES

V.L. KUZNETSOV and Yu.V. BUTENKO

Boreskov Institute ofCatalysis, Lavrentieva 5, Novosibirsk, 630090 Russia

1. Introduction

Recent interest in nanomaterials (typically on the scale of 10.7 _1O.9m) is stimulated

by the fact that, owing to the small size of the building blocks (particles, grains, or
phase) and the high surface-to-volume ratio, these materials demonstrate unique
mechanical, optical, electronic and magnetic properties [1-3]. The properties of
nanomaterials are mainly determined by interplay of: (1) fine building block size «100
nm); (2) the chemical composition of constituent phases; (3) the presence of interfaces,
more specifically grain boundaries, heterophase interfaces, or free surface; (4)
interactions between the constituent blocks.
One of the most intensively developin~ fields of nanomaterial technology is related
to carbon nanostructures (nanocarbons - C). Numerical theoretical and experimental
works provide information on the unique properties of nanostructures, mainly nanotubes
[4-7] and nanodiamond [8, 9] Synthesis of new composite materials based on NC can
provide the development of a broad ra',!le of materials with new and improved
properties [6]. However the prediction of C composites properties requires intensive
experimental work.
While NC are intensively investigated, their biological aspects are not as well
developed as nonbiological ones. The most important features of biological aspects of
nanotechnology concern the surface interactions and interfaces between biomolecules
and substrates (here NC). Thus they determine possibilities to develop new:
nanotechnology building blocks, drug delivery systems, membranes for biological
separations, supports for enzymes and bacterial cells, opltoelectronic devices,
nanomachines, ceramics, assembling and templatin~blocks.
Here we have considered the properties of C studied in the aspects of their
incorporation into the biological active systems and medical devices. The most
important requirements for the biomedical utilization of materials are their non-toxicity,
essential surface chemical groups allowing bonding with biomolecules, chemical
stability and mechanical strength etc. However for nanotechnology there are additional
requirements concerning the optimal particle size distribution, porosity, handling
facilities.
187
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 187-202.
© 2003 Kluwer Academic Publishers.
188

2. Nanodiamonds

2.1. METHODS OF NANODIAMOND PRODUCTION USING EXPLOSION

ENERGY

Several methods of ND (2-160 nm) synthesis using explosion energy are known.
These methods are based on an impact of high temperatures and pressures on carbon
materials with rapid cooling following to stabilize the diamond phase. The recent
reviews summarize results on ND synthesis and properties [8, 9]. The methods of ND
production can be divided into three main groups [9].
i) Mixture of carbon materials (graphite, carbon black) with metal powder (Ni, Cu, AI,
Co) placed in a capsule undergoes by shock wave compression induced by solid
explosive detonation [IO, II]. The diamond-bearing material produced contains up to
4.8 % of diamond phase by weight of the initial mixture.
ii) ND synthesis is performed by the detonation of a mixture of solid explosives with a
carbon material (carbon black or graphite) in a blasting chamber in inert atmosphere.
Depending on detonation conditions the produced diamond-bearing material contains up
to 3.4% by weight of the initial mixture [12, 13].
iii) The diamond containing solid carbonaceous products (detonation soot) are produced
by the detonation of pure and composite explosives with the negative oxygen balance in
a hermetic tank filled with different gases (N2, Ar, H20, CO2). The diamonds are
generated in a detonation wave zone at pressures of 20-30 GPa and temperatures of
3000-4000 K during the time closed to the microsecond range. The gas in the tank
decreases the effective temperature of detonation products and, thus, prevents the
conversion of diamond obtained into other carbon forms. Diamond yield is up to 10% of
the explosive weight [II, 14-21]. This approach is the most productive because the
continuous automatic process can be realized (the prototypes were developed in Russia
[9]).
ND is isolated from detonation products by combination of mechanical and
chemical methods. The impurities in the detonation soot present in the forms of metal
containing compounds and nondiamond carbons (graphite, black, amorphous carbon).
After mechanical removal of process admixtures a complex of chemical oxidative
operations directed to remove nondiamond forms of carbon and inorganic impurities are
used. Following sedimentation, filtration and centrifugation are used for ND separation
into several fraction containing particles with different size [9]. At the present moment
industry can produce tens of tons ofND per year.

2.2. PROPERTIES AND APPLICATIONS OF NANODIAMOND

Properties of ND are widely discussed in recent reviews [8, 9]. A HRTEM

micrograph of the purified initial ND is given in Fig. I (l). The dark contrast lines
belong to the (Ill) crystallographic diamond planes. XRD analysis revealed that the
distance between the (Ill) diamond planes is 0.206 nm (for bulk diamond this distance
is 0.205 nm) [19].
 189

Figure 1. HRTEM micrographs of the ND sample annealed under vacuum at (1) 1170,
(2) 1600, (3) 1900, (4) 2140 K. The dark straight contrast lines in micrographs (1) and
(2) correspond to the (Ill) crystallographic diamond layers. The distance between these
lines is 2.06 A. The dark lines in Figs. (3) and (4) correspond to the (0002)
crystallographic graphite layers. The distance between these lines is 3.5 A. The diamond
weight fractions (x) of the samples are also presented within each image.

Average diameter of ND obtained by the detonation method is 4 - 5 nanometers.

Most particles are between 2 - 10 nanometers. Fig. 2 presents small angle x-ray
scattering data (SAXS) on ND particle size distribution. One can see that the
distributions have clear cutoffs at both the lower and upper ends. [22].
According to XRD data the size of coherent scattering ofND particles varies from 2
to 5 nm. However, these values actually do not reflect the true size of particles
aggregates. The sedimentation analysis indicates that the basi<: ND structure is an
aggregate with a size from tens of nanometers to several microns [21]. The aggregates
consist of several particles. In this context the problem to disintegrate the ND
aggregates to individual particles is the most interesting for nanotechnological
applications.
The chemical analysis revealed that ND contains the heteroatoms such as 0, N, S,
H, which form different surface and bulk groups. The carbon content depends on the
purification procedure and varies within 85-95 %. The unit surfaGe area determined by
190

Brunauer-Emmet-Teller isotherms from the

thermal desorption of argon is in the range
from 200 to 450 m2/gr. The true density of
the initial ND (3.0-3.25 g/cm3 , helium
pycnometry), is significantly lower than that
of the natural crystal diamond is 3.515
g/cm 3• This difference may be attributed to
the relatively large contribution from the
surface layer of the smaller particles.
ND is widely applied [8, 9], namely for
precise polishing (ND allows to reduce
o 1 2 3 4 5 6 7 8 9 10 surface roughness to a few nanometers, or
radius, nm less); for production of wear-resistant and
Figure 2. The radial distribution corrosion resistance coatings; for production
function (1), and the granulometric of ceramic and composite materials; for the
function (2) of the original ND particle development of metal-matrix composites and
(average size is 4.7 nm) sample also for polymer and rubber based
determined by SAXS. composites, anti-friction lubricants and
additives to oils for industrial equipment etc.

2.3. SURFACE CHEMISTRY OF ND

The surface chemistry of ND is the most important for biological applications. It

was extensively studied by the TPD, XPS and FTIR spectroscopy in numerous papers
[8, 9, 23-25]. It was shown that the surface of oxidized ND isolated from detonation
soot are covered to a large extent with different oxygen-containing groups: -COOH
(carboxylic), -O-C=O (lactone), >C=O (carbonyl), -C-O-C- (ether), -OH (hydroxyl).
For such samples ratio of oxygen and surface carbon atoms is close to 0.5. The oxygen-
and hydrogen-containing groups at diamond surfaces saturate the dangling bonds of
carbon atoms and stabilize the diamond surfaces, prevent surface graphitization.
Temperature-programmed desorption (TPD) study demonstrates that concentration and
type of diamond surface groups can be influenced by thermal treatment. Thus Fig. 3
shows the temperature regions of the decomposition of the diamond surface groups.
Note that the hydrogen evolution at 1070-1420 K in vacuum from the diamond surface
coincides with the graphitization of diamonds surfaces. At these temperatures the
formation of closed curved graphite-like shells on the diamond surfaces was observed
[22]. Partial graphitization of ND opens an additional way for modification of the
diamond surface properties and synthesis of new type of carbon/diamond
nanocomposites. The Fig. 1(2) presents HRTEM micrographs of initial states of the
transformation ofND surfaces to curved graphite-like shells.
The rich surface chemistry of ND, the absence of toxic impurities, optical
transparency of the ND, small size (can penetrate narrow gaps) make them useful for a
variety of biomedical applications. The basic applications of ND surface
functionalization range from altering the wetting or adhesion characteristics and
 191

ether, carbonyl groups

improving the nanoparticles dispersion in
carboxyl, hydroxyl matrices to enhancing the catalytic properties
and lactone groups and ordering the interfacial region, and such.
C-Hxgroups
The diamond surface includes a wide
spectrum of functional chemical groups with
Nanodillmonds directly linked carbon structures. Thus Fig.
4. presents the possible functional groups
and molecular fragments that can be
produced on a diamond surface and used for
the creation of different building blocks and
nanostructures (it should be mentioned that
the same functional groups can be created on
the surface of any carbon materials). It opens
the great possibilities for chemical reactions
on the surface and functionalization of
diamond surfaces. The creation of specific
surface sites on ND fi:>r selective molecular
attachment is considered a promising
approach for their applications in
nanofabrication, self-assembly, nanosensors,
bioprobes, drug delivery, pigments, etc.

300 500 700 900 1100 1300

Temperature, K 3. Onion like carbon (OLC)
Figure 3. Temperature-programmed
OLC is one of the nanostructured forms of
decomposition of surface functional
carbon consisting of fullerene like shells
groups from a ND and micron diamond
enclosed each into another. The first time OLC
powder treated by a mixture HCI04 and
were observed by Iijima in arc discharge
H2 S04 • The temperature regions of the
products in early 1980's [26). However, the
decomposition of the diamond surface
importance of this observation was realized
groups are presented.
only after discovery of fullerene [27).

3.1. METHODS OF OLC PRODUCTION

At present several methods of OLC production are known. They are based on two
principally different approaches. First approach deals with a condensation of initially
evaporated atoms or small clusters of carbon. The processes can be carried out in gas
phase, liquid or even within metal matrix. For carbon evaporation arc-discharge, laser
ablation or carbon ions beams are used. A general accretion mechanism has been
proposed in which the graphitic sheets bend to eliminate the highly energetic dangling
bonds present at the edge of growing graphitic structures. OLC growth has been
proposed to be generated by spiral multiple shell particle growth in gas phase [28, 29].
192

Functional Chemical bridges between NC It should be mentioned that in the case of

groups and another molecules gas phase condensation of carbon atoms,
OLC is a byproduct and only small
quantities can be produced. Higher yields
are obtained when carbon atoms or clusters
are condensed in metal matrix or water.
Significant amounts of OLC, allowing to
investigate their optical properties, were
produced using 120 keY carbon ion
implantation at 500°C into silver matrix [30,
31]. It has been proposed that due to
weakness of C-metal (Cu, Ag) bulk
segregation of carbon and metal occur
Figure. 4. Examples of organic groups leading to the formation of carbon onions.
that can be used for functionalization of Higher production rate of 3 mg/min was
nanocarbons. achieved using arc discharge between two
graphite electrode submerged in water [32].
This approach allows producing nano-onions with the average diameter of 25-30 nm
(range 5-40 nm). Mechanism of this process is not understood, however one can
propose that condensation of carbon atoms or small clusters proceeds at high
temperature in steam bubbles. The high yield of carbon onions in comparison with the
yield of amorphous carbon can be explained in terms of the relatively high reactivity of
amorphous carbon to the water gas shift reaction.
Another approach to the OLe synthesis is based on the transformation of
condensed carbon under high-energy treatment. Ugarte has found the curling and
closure of graphitic networks of carbon soot into OLC particles under electron beam
irradiation [33]. This approach allows synthesis of perfect onions containing up to 70
shells. Unfortunately the method may be applied only within an electron microscope or
electron accelerator so the yield of onions is very low.
Later Banhart and coworkers have found mutual transformation of OLC and
nanodiamond under intensive electron beam irradiation. They have developed non-
equilibrium phase diagram of carbon, in which graphite-diamond phase boundaries are
given in terms of temperature and irradiation intensity, where the latter is characterized
by the rate of atomic displacements in graphite. All experimental observations of
different groups on electron induced graphite-diamond mutual transformations are in
good agreement with the predictions of proposed theoretical model [34-36].
We have developed a new process of generating hundreds of grams of OLC in the
laboratory. It consists of thermal annealing of nanodiamonds with a size of 3-6 nm [37,
38]. Because this method made it possible to produce onion-containing material in
macroscopic quantities, the samples could be studied by various spectroscopic methods:
Raman [39], X-ray photoelectron [40, 41], and electron energy-loss [42]. Results of the
studies ofOLC produced via ND annealing are presented below.
Macroscopic samples of hollow OLC can be also produced by heating of carbon
soot in inert atmosphere at 2270-2670 K. Thus heating of carbon soot generated in an
 193

arc discharge yielded closed-shell graphitic particles containing a few shells 3-10 nm in
cross-section. A number of shells can be controlled to some extent. Thus 2-3 and 5-8
shell particles can be selected [43]. However, thermal annealing of soot allows
producing only hollow polygonized onions.

3.2. OLe FORMATION VIA ND ANNEALING

Using the various independent methods (HR TEM, Raman spectroscopy, X-ray
emission spectroscopy, electrical resistivity measurements, ESR, kinetics measurements
and molecular modeling) intermediate products of ND annealing have been
systematically investigated allowing information on OLe structure and the mechanism
of OLe formation to be obtained.
HR TEM studies detected the intermediate products of the diamond surface annealing
[22, 37,44]. HR TEM images ofND annealing products are presented in Fig. I.
The most important observations from these investigations are: the following:
(a) ND graphitization under annealing begins from the surface towards crystal bulk
(b) The transformation rate of the (111) diamond planes to graphite-like sheets is
higher than this of other planes.
(c) The edges of exfoliated graphite-like sheets merge with the upper untransformed
diamond planes.
(d) The distance between the inner graphite-like sheet and the upper untransformed
diamond layer does not exceed 0.35 nm, indicating an interaction between the
diamond and graphite layers.
(e) At the interface between the diamond and the evolving graphite, the formation of
two curved graphite sheets from three (111) diamond planes was observed.
From our study of the kinetics of ND graphitization at temperatures ranging from
l370 to 1860 K, we have obtained kinetic parameters [22] that differ significantly from
those obtained by Davies and Evans for the graphitization of diamond in the 2100-2300
K range [45, 46]. We conclude that there are at least two temperature regions where the
graphitization occurs by different mechanisms. The Debye temperature for diamond,
1910 K, appears to serve as the boundary between these regions. At and above this
temperature all oscillating freedom degrees of the diamond are excited. In the
temperature range 1370-1860 K, graphitization proceeds from the surface towards the
diamond particle bulk. The onset temperature - 1370 K for the process is associated
with the temperature at which various the chemical groups stabilizing the surface are
completely removed (see Fig. 3). The kinetic parameters describing the graphitization of
UDD in the temperature range l370-1860 K, activation energy E=45±4 kcaVmoI, and
pre-exponential factor A=74±5 nm/s, yield rates which agree well with the HRTEM
data related to the graphitization of submicron diamond. This suggests that the kinetic
parameters of the NO graphitization can be used to estimate rates for the graphitization
of larger diamonds. The knowledge gained about the graphitization of diamond and the
determination of the kinetic parameters can help one control the modification of the
diamond surface by the growth of fullerene-like carbon states on it. These carbon states
are of interest for carbon chemistry (the chemistry of non-planar carbon materials).
194

The formed during annealing THE INITIAL ND The hollow Onion-Like

cu rved sheets Carbon particles

§
~
~
The closing of the ~

Q"P","~

~-
The diamond cores ONION-LIKE CARBON (OLC)

Fif{Ure 5. Scheme of Onion-Like Carbon formation.

The general scheme of production of OLC and nanocomposites with different

portion of diamond and nanosized graphitic sheets is presented in Fig. 5. Thus the
annealing of ND at moderate temperature makes it possible to produce carbon
nanocomposites with controlled ratio of Sp2/Sp3 state.

3.3. STRUCTURE AND PROPERTIES OF CARBON ONIONS AND

Sp2;sp3 NANOCOMPOSITES

As it was mentioned that carbon onions, produced with different methods, are
characterized with a number of parameters, namely particle size, number of closed
shells, and presence of different defects. Two different types of OLC carbon can be
produced by mentioned methods: one contains spherical shells another one consists of
polygonized shells. It was proposed that the latter onions contain fullerene like spheres
with twelve pentagons, which are preferentially responsible for the formation of
polygonized structure (Fig. 6 (1». The method of production is the most crucial for the
structure of onions. The perfect spherical onions are produced with electron irradiation
of ND [36] and carbon soot [33-35, 47]. Spherical onions preferentially form via ND
heat treatment at temperatures lower than 2000 K. At the same time polygonized OLC
is usually formed after high temperature treatment (>2000 K) of carbon soot [43] or
spherical OLC [40].
To explain the sphericity of onions formed under electron beam irradiation of
carbon [47] the structural models containing combinations of pentagons and octagons
have been proposed by Terrones (Fig.6 (2» [48]. However real OLC are usually
imperfect and contain high concentration of defects. Thus particles shows the extensive
regions in which graphene network is either missing or highly distorted. In general,
defects of all kind occur and detailed study of real carbon particles result in more
rounded forms.
 195

(1) (2) (3)

Figure 6. (1) The polygonized C240 molecule (Oh) with twelve pentagons; (2) The
spherical C240 molecule (Ih) with 6 octagons and 24 pentagons [48]; (3) Spherical holed
cage C 1SO, obtained by removing twelve pentagons from C240 [40].

Thus X-ray emission spectroscopy study ofND annealing products (see Fig. I and
the scheme in the Fig. 5) detects the difference in the electronic state between carbon
shells generated below 2100 K and both graphite and diamond particles [40, 41]. An
intensive maximum that found in the high-energy region of the C Ka spectrum indicates
significant localization of weakly bonding electrons in OLC. It was proposed that this
localization might result from defects in the curved graphitic networks, namely, the
holes, which appear through a deficit of surface diamond atoms during ND annealing
that are required to form an ideal spherical shell. Quantum-chemical calculation on the
holed cage (see Fig. 6 (3» explains the crucial increase of high-energy intensity by the
additional contribution of the electrons from dangling bonds on the zigzag edges. The
relative chemical inertness of such bonds might be explained by the steric problems
occurring when reagents penetrate deep into the OLC through rather small size of holes.
The proposition on the formation of OLC with a holed structure is in a good agreement
with ESR study of similar samples. Thus 10 unpaired electrons per one carbon onion
were registered in [49], where onions were prepared at temperature lower then 2100 K.
The holed structure of onions can be annealed at temperatures higher then 2100 K,
however this procedure leads to the formation of bigger jolygonized hollow onions.
Resistivity measurements of OLC and Sp2/Sp nanocomposites (see Fig.5)
demonstrate electrical properties that differ from properties of well-known graphite
materials [50] For typical graphite materials (carbon black, graphitized soot, graphite
powder) the 3-dimensional electrical conductivity is usually observed. The common
graphite materials usually have the metallic type of conductivity, which decreases with
increasing temperature. At the same time OLC and intermediate samples of UDD
annealing demonstrate the variation of the electrical conductivity dimensionality from
0.5 to 1.5. Their temperature dependence of electrical resistivity is typical for systems
with variable hopping-length hopping conductivity. It can be explained by specific type
of the composite structure formed with highly curved graphene sheets, carbon onions
and diamond core.
196

Fig. 7 summarizes the data on the

conductivity of OLC [51]. The carbon
powders obtained by annealing of ND
consist of the OLC aggregates. Within
the aggregates onions are linked to
each other with defect graphite-like
sheets and C-C bonds. Some part of the
onions is presented by elongated
particles with linked external graphitic
layers and closed quasi-spherical
Free path I of electrons shells. The free path length of electrons
between defects within
single graphitic nanoshell
within OLC particles is comparable
(12-18 A) with the size of graphitic fragments
within the holed structure of single
Figure 7. The scheme of three scales of OLC onion (I = 12±2 A). When their holed
particle organization. The hard solid lines structure is annealed at temperatures
correspond to current conduction tracks. higher then 2100 K with the formation
of hollow polygonized structures the
free path length of the electron
increases up to 18±2 A. The size localization length (60-100 nm) of current carriers is
comparable with the size aggregates of OLC particles (Fig. 7).
These results allow proposing the mechanism ofOLC formation via ND annealing.
At first steps of nanodiamond annealing formation of curved graphite like sheets are
exfoliated from diamond particles. Closure of these sheets leads to the formation of
spiral-like multi shell particles and finally OLC forms The formation of relatively high
concentration of unsaturated carbon is detected before closure of graphitic shells. The
formation of such species stabilized within nanoparticle volume can provide the
appearance of new unusual properties of carbon materials (electronic, magnetic and
optical properties).
Properties of OLC are poorly studied, mainly due to their low availability for the
experiment. In spite of that, these materials stimulate a great interest. They are related to
the interstellar dust problem, at least they could contribute to large UV adsobtion band
at the 217.5 nm wavelength [31]. Due to a very efficient optical limiting action of OLC
they are good candidates towards photonic applications [52]. OLC are potentially
perfect solid lubricant as hollow WS 2 nanoparticles. OLC demonstrate high selectivity
and catalytic activity in the oxidative dehydrogenation of ethyl benzene to styrene [53].
The last example demonstrates that OLC like ND can be incorporated into chemical
reactions. Thus onion's surface can be functionalized allowing their further
incorporation in biosystems. However extensive experiments for the preparation of
narrow fraction of onions and their functionilization are required.
 197

4. Bio-related applications orNe

Carbon supports are traditionally used for immobilization of biomolecules,

enzymes, adsorption of bacteria's and cells. Carbon nanomaterials also can be
perspective for the development of new types of adsorbents and biocatalysts. This
region is rather new and only few results are known.
Kovalenko at el. have developed number of supports based on catalytic filamentous
carbon (related to nanotubes), which are perspective for the immobilization of different
substances of biological origin, namely amino acids (L-tyrosine), proteins (bovine
serum albumin) enzyme (glucoamylase) and non-growing cells of microorganisms
(Escheria coli, Bacillus subtilis, Rhodococcus sp.) [54]. This approach was also used for
the development of carbon containing macrostructured cemmics [55]. Biocatalysts on
the base of these carbon filament containing supports and immobilized glucoamylase
and yeast invertase were stable and effective for continuous stan:h1dextrines hydrolysis
and sugar inversion respectively.
The first few papers [56-59] describe the immobilization of proteins, iodinated and
platinated oligonucleotides on the surface of carbon nanotubes. Thus multi wall carbon
nanotubes, have been used for the immobilization of proteins and enzymes. Davis at el.
demonstrated, for the first time, direct imaging by HR TEM of Zn2Cds-metallothionein,
cytochromes C, c3 and ~-lactamase 1. This was achieved, without modification, because
the biomolecules encapsulated within nanotubes appear to be shielded from the
consequences of exposure to the intense electron beam. The re:sults indicate that the
internal surface of the nanotubes interacts strongly with the enzymes resulting in their
immobilization . In some cases, the proteins are seen to be distorted giving a concave
meniscus inside the tubes. Single protein molecules, their dimers, tetramers and higher
oligomers were observed inside the centml cavity. Comparison ofthe catalytic activities
of immobilised ~-lactamase I on or in nanotubes with the free enzyme in the hydrolysis
of penicillin, however, showed a significant amount of the immobilised enzyme
remained catalytically active, implying that no drastic conformatiional change had taken
place. The carbon nanotube appears to act as a benign host in it~; ability to encapsulate
protein molecules within an environment which offers some prote:ction.
Nanodiamond was used for immobilization of antigens [60]. For this purpose,
diamond surface initially was treated with cellobiosis following adsorption of protein.
This system was used as antigens for prolong production of antibody.
Nanodiamond was also used for the purification of recombinant apoobelin from
Escherichia coli extracts. Apoobelin was selectively adsorbed onto nanodiamond and
then extracted with dithiothreitol [6\].
Reference [62] demonstrates a possibility of preparation of the nanodiamond-
protein-Ab03 complex, where ND and 8-Ab03 were used as building blocks and
protein as a cross-linking agents for design of new composites biomaterials. At the first
stage a recombinant apoobelin was adsorbed on nanodiamond particles. At the second
stage the nanodiamond-protein complex was adsorbed onto the particles.
198

Another perspective regions of NC concerns medical applications base on NC

derived devices. Thus new x-ray tube, in which radiation can be generated using a
carbon nanotube (CNT)-based field-emission cathode, was developed [63]. The device
can readily produce both continuous and pulsed x-rays with a programmable wave form
and repetition rate. A total emission current of 28 rnA was obtained from a 0.2 cm 2 area
CNT cathode. The x-ray intensity is sufficient to image a human extremity at 14 kVp
and 180 mAs. Pulsed x-ray with a repetition rate greater than 100 kHz was readily
achieved by programming the gate voltage. The CNT-based cold-cathode x-ray
technology can potentially lead to the portable and miniature x-ray sources for industrial
and medical applications.

5. Conclusion

Research at the NC frontier is unified by the need to share knowledge, tools and
techniques, and expertise on atomic and molecular interactions. The most emerging
problems related of this region concern the availability of well characterized NC
materials, with narrow fraction size distribution and controlled surface chemical
composition. However, powerful new concepts and capabilities, such as atomic-scale
imaging and manipulation, self-assembly, together with increasingly powerful
computing tools, are rapidly converging from different research fields. These will
enable based technology to progress and researchers to expand their expertise into new
application fields.

Acknowledgments

This work was supported by INTAS (Project No. 00-237, 01-254), the Russian
Foundation for Basic Research (Projects No. 02-03-32296) and SCOPES No.
7SUPJ062400.

References

1. Seigel, R.W. (1993) Nanophase Materials Assembled from Atom Clusters, Mater.
Sci. Eng. B-Solid State M 19,37-43.
2. H.S. Nalwa (ed.) (2000) Handbook ofNanostructured Materials and
Nanotechnology, vol. 1-5, Academic Press, Sandiego, San-Francisco, NY, Boston,
London,Sydney, Tokyo.
3. R.W.Siegel, E.Hu and M.C.Roco (eds.) (1999) WTECpanel on Nanostruture science
and Technology, Kluwer Academic Publishers, Dordrecht.
4. T.W.Ebbesen (ed.) (1997) Carbon Nanotubes: Preparation and properties, CRe
Press, Inc., Boca Raton, Florida.
 199

5. Harris, PJ.F. (1999) Carbon nanotubes and related structures, Cambridge University
Press, Cambridge.
6. G. Benedek, P. Milani, and V.G. Ralchenko, (eds.) (2001) Nanostructured Carbon
for Advanced Applications, Kluwer Academic Publishers, Dordrecht.
7. Proceedings ofthe International Workshop on the Science and Application of
Nanotubes, Potsdam (Germany), July 22-25, 2001.
8. Dolmatov, V.Yu. (2001) Ultradispersed diamond from detonation synthesis:
properties and applications, Uspekhi chimii 70(7), 687-708 (in Russian).
9. Vereschagin, AL. (2001) Detonation Nanodiamonds, Altai State Technical
University, Barnaul, Russian Federation.
10. Covan, J.R.,. Dunnington, B. W., Holzman, A.H. (1968) Process of diamond
synthesis, US Patent 3,401,019 (September 10, 1968).
II. Donnet, lB., Fousson, E., Wang, T.K., Samirant, M., and Barns, C. (2000) Dynamic
synthesis of diamonds, Diam. Relat. Mater. 9, 887-892.
12. Adadurov, G.A., Bavina, T.V., Breusov, D.N., Drobshev, V.N., Rogacheva, AI.,
Tatsy, V.F. (1976) The method of the diamond production, SSSR Patent 565474 (July
23, 1976).
13. Adadurov, G.A., Bavina, T.V., Breusov, D.N., Drobshev, V.N., Messinev, MJ.,
Rogacheva, AI., Ananiin, AV., Apollonov, V.N., Dremin, A.N., Doronin, V.N.,
Dubovitsky, F.I., Zemlyakov, L.G., and Pershin, S.V. (1984) Method of producing
diamond and/or diamond-like modilications of boron nitride, US. Patent 4,483,836
(November 20, 1984).
14. Volkov, K.V., Danilenko, V.V., and Elin, V.I. (1990) Diamond synthesis from the
carbon of detonation products, Phizika Gorenia i Vzriva (Rus) 26(3), 123-125.
15. Staver, A.M.,Gubareva, N.V., Liamkin, A.I., and Petrov, E.A. (1984) Ultradispersed
diamond powders prepared with the use of explosion energy, Phizika Gorenia i Vzriva
20(5),100-103 (in Russian).
16. Lia mkin, AI., Petrov, E.A., Ershov, A.P., Sakovich, G.V., Staver, AM., and Titov,
V.M. (1988) Diamond production from explosives, Doklady Academy of Sciences of
USSR 302(3),611-613 (in Russian).
17. Greiner N.R., Phillips D.S., and Johnson F.J.D. (1998) Diamonds in detonation soot,
Nature 333, 440-442
18. Petrov, E.A, Sakovich, G.V., and Bryliakov, P.M. (1990) Conditions for
stabilization of diamond in detonation production, Doklady Academy of Sciences of
USSR 313(4),862-864 (in Russian).
19. Kuznetsov, V.L., Malkov, I.Yu., Chuvilin, A.L., Moroz, E.M., Kolomiichuk, V.N.,
Shaichutdinov, Sh.K., and Butenko, Yu.V. (1994) Effect of explosions on the structure
of detonation soots: ultradisperse diamond and onion carbon, Carbon 32, 873-882.
20. Vereschagin, AL. Petrov, E.A, Sakovich, G.V., Komarov, V.F., Klimov, AV., and
Kozyrev, N. V. (1999) Synthetic diamond-containing material and method of obtaining
it, US Partent 5,861,349 (January 19, 1999).
21. Mironov, E., Koretz, A., and Petrov E. (2002) Detonation synthesis u1tradispersed
diamond structural properties investigation by infrared absorption, Diam. Relat. Mater.
n, 872-876.
200

22. Butenko, Yu.V., Kuznetsov, V.L., Chuvilin, A.L., Kolomiichuk, V.N., Stankus,
S.V., Khairulin, R.A., and Segall, B. (2000) The kinetics of the graphitization of
dispersed diamonds at 'low' temperatures, J. App. Phys. 88,4380-4388.
23. SchlOgl, R. (1997) Carbons, (1997) in G. Ertl, H. Knozinger, and J. Weitkamp
(eds.), Handbook ofHeterogeneous Catalysis, Vol. 1, Wiley-VCH, Weinheim, pp. 138-
19l.
24. Kuznetsov, V.L., Butenko, Y.V., Chuvilin, A.L., Boronin, A.I., Kvon, R.I., Kosheev
S.V., Stankus, S.V., and Khairulin, R.A. (1997) Self-assembled closed curved graphite-
like structures (CCGS) formation on diamond surface, CARBON'97, 23rd Biennial
Conference on Carbon, Pennsylvania, July 13-18, 1997, vol. 2, 326-327.
25. Kuznetsov, V.L., Aleksandrov, M.N., Zagoruiko, I.V, ChuviIin, A.L., Moroz, E.M.,
Kolomiichuk, V.N., and Sakovitch, VA (1991) Study of ultradispersed diamond
powders obtained using explosion energy, Carbon 29, 665.
26. Iijima S. (\980) Direct observation ofthe tetrahedral bonding in graphitized carbon
black by high resolution electron microscopy, J. Cryst. Growth 50, 675-683
27. Iijima S. (1994) Carbon Nanotubes, MRS Bulletin 19(11), 43-49.
28. Zhang, Q.L., O'Brien, S.c., Heath, J.R., Liu, Y., Curl, R.F., Kroto, H.W., and
Smalley, R.E. (1986) Reactivity of Large Carbon Clusters - Spheroidal Carbon Shells
and Their Possible Relevance to the Formation and Morphology of Soot, J. Phys. Chern.
90, 525-528.
29. Kroto, H.W., and McKay, K. (1988) The formation of quasi-icosahedral spheral
shell carbon particles, Nature 331, 328-330.
30. Cabioc'h, Th., Thune, E, and Jaouen, M. (2000) Carbon-onion thin film synthesis
onto silica substrates, Chern. Phys. Lett. 320,202-205.
31. Lambin, Ph., Henrard, L., Lucas, A.A., and Cabioc 'h, Th. (2001) Optical properties
of the carbon onions, in G. Benedek, P. Milani, and V.G. Ralchenko (eds.),
Nanostructured Carbon for Advanced Applications, Kluwer Academic Publishers,
Dordrecht, pp 273-284.
32. Sano, N., Wang, H., Chhowalla, M., Alexandrou, I., Amaratunga, G.A.J. (2001)
Synthesis of carbon 'onions' in water, Nature 414, 506-507.
33. Ugarte, D. (1992) Curling and closure of graphitic networks under electron beam
irradiation, Nature 359, 707-709.
34. Banhart, F. (\999) Irradiation effects in carbon nanostructures, Rep. Prog. Phys. 62
1181-122l.
35. Zaiser, M., Lyutovich, Y., and Banhart, F. (2000) Irradiation-induced transformation
of graphite to diamond: A quantitative study, Phys. Rev. B 62(5), 3058-3064.
36. Roddatis, V.V., Kuznetsov, V.L., Butenko, Yu.V., Su D.S., and SchlOgl, R. (2002)
Transformation of diamond nanoparticles into carbon onions under electron irradiation,
Phys. Chern., Chern. Phys. 4,1964-1967.
37. Kuznetsov, V.L., Chuvilin, A.L., Butenko, Yu.V., and Titov, V.M. (1994) Onion-
like carbon from ultra-disperse diamond, Chern. Phys. Lett. 222, 343-348.
38. Titov, V.M., Malkov, I.Yu., Kuznetsov, V.L., and Chuvilin, A.L., Method of
production of onion-like carbon particles, Russian Patent 209370 (October 19, 1993).
 201

39. Obraztsova, E.D., Fujii, M., Hayashi, S., Kuznetsov, V.L., Butenko, Yu.V., and
Chuvilin, A.L. (1998) Raman Identification of Onion-Like Carbon, Carbon 36, 821-
826.
40. Okotrub A.V., Bulusheva, L.G., Kuznetsov, V.L., Butenko, YIl.V., Chuvilin, AL.,
and Heggie, M.1. (2001) X-ray emission studies of valence band ofnanodiamonds
annealed at different temperatures, J. Chem. Phys. A 105,9781-9787.
41. Bulusheva, L.G., Okotrub, AV., Kuznetsov, V.L., Chuvilin, A.L., Butenko, Yu.V.,
and Heggie, M.1. (2002) Topology and Electronic Structure ofOniion-Like Carbon and
GraphitelDiamond Nanocomposites, in S. Komameni, J.-1. Matsushita, G.Q. Lu, J.C.
Parker, R.A. Vaia (eds.), Nanophase and nanocomposite materials, vol. 703, Mat. Res.
Sym. Proc., Material Research Society, Pittsburgh, pp. 381-386.
42. Tomita, S.; Fujii, M.; Hayashi, S.; and Yamamoto, K. (1999) Electron energy-loss
spectroscopy of carbon onions, Chem. Phys. Lett. 305, 225-229.
43. De Heer, W.A. and Ugarte, D. (1993) Carbon onions produced by heat treatment of
carbon soot and their relation to the 217.5 nm interstellar adsorption feature, Chem.
Phys. Lett. 207,480-486.
44. Kuznetsov, V.L., Chuvilin, A.L., Butenko, Yu.V., Malkov, I.Yu., Gutakovskii,
A.K., Stankus, S.V., and Khairulin, RA (1995) Study of onion-like carbon (OLC)
formation from ultra disperse diamond, in P. Bemuer, D.S. Bethune, L.Y. Chiang, T.W.
Ebbessen, R.M. Metzger, and J.W. Minmire (eds), Science and Technology ofFullerene
Materials, vol. 359, Mater. Res. Soc. Proc., Material Research Sodety, Pittsburgh, pp.
105-110.
45. Davies, G. and Evans, T. (1972) Graphitization of diamond at zero pressure and a
high pressure, Proc. R. Soc. Lond. A. 328, 413-427.
46. Evans, T. (1979) Changes produced by high temperature treatment of diamond, in
J .E. Field (ed.), The properties of diamond, Academic Press, London, New York, San
Francisco, pp. 403-424.
47. Banhart, F., Fuller, T., Redlich, Ph., and Ajayan, P.M. (1997) The formation,
annealing and self-compression of carbon onions under electron irradiation, Chem.
Phys. Lett. 269,349-355.
48. Terrones, M. and Terrones, H. (1996) The role of defects in graphitic structures,
Full. Sci. and Tech. 4,517-533.
49. Tomita, S., Sakurai, T., Ohta, H., Fujii, M., and Hayashi, S. (2001) Structure and
Electronic Properties of Carbon Onions, J. Chem. Phys. 114,7477-7482.
50. Kuznetsov, V.L., Butenko, Yu.V., Chuvilin, A.L., Romanenko, A.I., and Okotrub,
A V. (2001) Electrical resistivity of graphitized ultra-disperse diamond and onion-like
carbon, Chem. Phys. Lett. 336, 397-404.
51. Romanenko, A.I., Anikeeva, O.B., Okotrub, A.V., Kuznetsov,_V.L., Butenko,
Yu.V., Chuvilin, AL., Dong, e., and Ni, Y. (2002) Temperature Dependence of
Electroresistivity, Negative and Positive Magnetoresistivity of Carbon Nanopartic1es, in
in S. Komarneni, J.-1. Matsushita, G.Q. Lu, J.e. Parker, R.A Vaia (eds.), Nanophase
and nanocomposite materials, vol. 703, Mat. Res. Sym. Proc., M2lterial Research
Society, Pittsburgh, pp. 259-264.
202

52. Koudoumas, E., Kokkinaki, 0., Konstantaki, M., Couris, S., Korovin, S., Detkov,
P., Kuznetsov, V., Pimenov, S., and Pustovoi, V. (2002) Onion-like carbon and
diamond nanoparticles for optical limiting, Chem. Phys. Lett. 357,336-340.
53. Keller, N., Maksimova, N.l., Roddatis, V.V., Schur, M., Mesti, G., Kuznetsov, V.L.,
Butenko, Yu.V., and Schlogl, R. (2002) The catalytic use of onion-like carbon materials
for Styrene Synthesis by oxidative dehydrogenation of ethylbenzene, Angew. Chem. Int.
Ed. 41(11), 1885-1888.
54. Kovalenko, G.A., Kuznetsova, E.V., Mogilnykh, Yu.l., Andreeva, I.S., Kuvshinov,
D.G., and Rudina, N.A. (2001) Catalytic filamentous carbons for immobilization of
biologically active substances and non-growing bacterial cells, Carbon 39, 1033-1043.
55. Kovalenko, G.A., Komova O.V., Simakov, A.V., Khomov, V.V., and Rudina, N.A.
(2002) Macrostructured carbinized ceramics as adsorbents for immobilization of
glucoamylase, J Mol. Cat. A: Chemical 18Z/183, 73-80.
56. Tsang, S.C., Davis, J.1., Chen, Y. K., Green, M.L.H., Hill, H.A.O., Hambley, T.W.,
and Sadler, P.1. (1995) Immobilization of Small Proteins m Carbon Nanotubes - High-
Resolution Transmission Electron-Microscopy Study and Catalytic Activity, J Chem.
Soc. Chem. Commun. 12, 1803-1805.
57. Tsang, S.C., Guo, Z.1., Chen, Y.K., Green, M.L.H., Hill H.A.O., Hambley, T.W.,
and Sadler PJ. (1995) Immobilization of small proteins in carbon nanotubes, J Chem.
Soc. Chem. Commun. 18,2579-2581.
58. Tsang, S.C., Guo, Z,J., Chen, Y. K., Green, M.L.H., HilI.H.A.O., Hambley T.W.,
and Sadler P.1. (1997) Immobilization of platinated and iodinated oligonucleotides on
carbon nanotubes, Angew. Chem. Int. Ed. 36,2198-2200.
59. Davis J.J., Green M.L., Hill H.A., Leung Y.C., Sadler P.J., Sloan J., Xavier A.V.,
and Tsang S.c. (1998) The immobilization of proteins in carbon nanotubes, Inorganica
Chimica Acta. 272, 261-266.
60. Kossovsky N., Gelman A., Hnatyszyn J., Rajguru S., Garrell R., Torbati S., Freitas
S., and Chow G. (1995) Surface Modified Diamond Nanoparticles as Antigen Delivery
Vehicles, Biocorifugate Chemistry 6,507-509.
61. Bondar, V. S. and Puzyr, A. P. (2000) Application ofnanodiamond particles for
express extraction of recombinant apoobelin from Escherichia coli, Doklady Academy
of Sciences of USSR 373(2),251-253 (in Russian).
62. Puzyr, A. P.; Bondar, V. S.; Belobrov, P. I.; and Bukaemskii, A. A. (2000)
Preparation of nanodiamond-protein- 8-aluminum oxide complex, Doklady Academy of
Sciences of USSR 373(3), 408-410 (in Russian).
63. Yue, G.Z., Qiu, Q., Bo Gao, Cheng, Y., Zhang, J., Shimoda, H., Chang, S., Lu, J.P.,
and Zhou, O. (2002) Generation of continuous and pulsed diagnostic imaging x-ray
radiation using a carbon-nanotube-based field-emission cathode, Appl. Phys. Lett. 81,
355-357.
SYNTHESIS OF SPHERICAL, NON-AGGREGATED
SILICA NANOPARTICLES

o 0 0

A. VITAL, U. KLOTZ, T. GRAULE ,

§ § §
~. MUELLER, H. K. KAMMLER , S. E. PRATSINIS'
Swiss Federal Laboratories for Materials Testing and Research (EMPA),
High Performance Ceramics Section
Ueberlandstrasse 129, CH-8600 Duebendorf, Switzerland
§ Swiss Federal Institute of Technology (ETH),
Institute of Process Engineering, Particle Technology Laboratory
Sonneggstrasse 3, CH-8092 Zuerich, Switzerland

Commercially available Si02 nanopowders synthesised by flame aerosol processes

consist to a major part of large aggregates, formed by collision and sintering of primary
particles during the process. These aggregates impair the physical and chemical
benefits, which could be achieved from the properties of the primary, non-aggregated
nanoparticles.
In this work, Si02 nanoparticles were produced by high-temperature oxidation of
hexamethyldisiloxane (HMDSO) vapour in an aerosol co-flow diffusion flame reactor.
Oxygen and CH 4 were used as oxidant and fuel gas, respectively. Particle size and
morphology was controlled by flame configuration, O2 flow rate as well as CH4 and
HMDSO concentration. The mean particle diameter was determined from nitrogen
absorption measurement and particle morphology was examined by TEM.
SiOz nanopowders with mean primary particle diameters of 10 w 110 nm were
synthesised. The burner configuration and the O2 flow rate were found to be the main
parameters influencing morphology and size of the Si0 2 particles. Non-aggregated,
spherical Si0 2 nanoparticles formed at low O2 flow rates in a single diffusion flame as
well as in a premixed flame, while fractal-like structured aggregates were made at high
O2 flow rates.

1. Introduction

Fumed silica (SiO z) nanopowder is a chemical product of major industrial importance

[1]. Its wide field of applications extends from a filler and reinforcement material for
rubbers and polymers to a thickener for paints and pastes, a binder for concrete and
cement, an inert filler in pharmaceutics, an enhancement of powder flowability to an
additive in dental and medical products [2].
203
Y G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 203-210.
© 2003 Kluwer Academic Publishers.
204
Commercially available Si0 2 nanopowders produced by high-temperature aerosol
processes, such as in plasma and arc reactors or by flame hydrolysis/oxidation, consist
to a great extend of aggregates, formed by sintering of primary particles during the
synthesis process. These aggregates often have a fractal-like, highly branched structure
and reach a size of several 100 nm in several dimensions [2]. In contrast to particle
agglomerates, the primary particles of an aggregate cannot be dispersed individually in
a matrix, except by mechanical impact methods, such as comminution. As a result, these
aggregates impede the potential mechanical, optical and chemical benefits which could
be achieved from the properties of the non-aggregated primary nanoparticles, especially
when applied in nanostructured materials, like composites, catalysts, varnishes or
electronics [2]. For these reasons, the production of tailored, non-aggregated, nanosized
particles is very incentive.

2. Experimental Procedure

A schematic of the experimental set-up is depicted in Fig. 1. Hexamethyldisiloxane

(HMDSO) (Fluka Chemie AG, 99%) was used as silicon precursor and stored in a
1 litre bubbler flask, filled to a constant level. The flask was placed in a thermostated
silicon oil bath for control of the HMDSO temperature. Nitrogen (N 2) (PanGas,
~ 99.999%) was used as HMDSO carrier gas and was delivered to the bubbler flask
through a fritted cylinder. The Si02 powder production rate was determined by control
of the flask temperature and the carrier gas flow rate. Saturation of the entrained carrier
gas with HMDSO vapour was proofed by measuring the weight of the bubbler flask.
The tubing between the flask and the burner was heated by heating tapes to keep it at a
temperature at least 20°C higher than the flask temperature to prevent HMDSO
condensation [3]. Pure oxygen (02 ) (PanGas, ~ 99.999%) and methane (CH 4 ) (PanGas,
~ 99.5%) were supplied to the system as oxidant and fuel gas, respectively. All gas flow
rates were controlled by calibrated flow meters (Voegtlin Instruments, Typ VI00).
The burner consists of three concentric stainless-steel tubes (Fig. 1b, c) with the
following inner (Dn,i) and outer (D n,.) diameters: centre tube: Dl,i 4.8 mm I D 1,. 5.6 mm;
middle tube: D2 " 6.4 mm I D2 ,a 7.3 mm and outer tube D3 " 9.0 mm I D3 ,a 10 mm. The
gas supply system allows switching from a single diffusion flame (SDF) to a premixed
flame (PMF) configuration, where the gases enter the tubes according to Fig. 1b, c. The
N2 flow rate through the middle annulus was kept at 27 lIh for all experiments and lifts
the flame front from the burner, preventing particle deposition on the burner mouth [3].
Small samples (= 1 g) of product particles were collected by aid of a vacuum pump on a
glass fibre filter (Whatmann GF/A) located 30 cm above the burner and fixed in a steel
hood (Fig. la).
 205

~
HOOd
Filter

0
Collector

I I Chinney
----.:..< ~

Burner

(b)

(a) (c)

Figure J - a) Experimental set-up for SiO, nanopowder synthesis by a flame aerosol process and gas
inlet configurations of the co-flow burner for b) a single diffusion flame (SDF) and c) a premixed
flame (PM F).

3. Characterisation

The specific surface area As of the product powder was determined from a five-point N2
adsorption isotherm by the BET-method (Brunauer-Emmet-Teller) using a
Gemini III 2375 (Micromeritics Instr. Corp.). Prior to BET analysis, the powder
samples were degassed (flow prep station 060, Micromeritics Instr. Corp.) at 150°C for
2 h under N2 atmosphere to remove water trapped to the particle surface. Assuming
monodisperse, spherical primary particles, the BET-equivalent mean primary particle
diameter dp,so was calculated by

dp,50 = 6/(pp-AsJ (1)

where Pp is the density of Si0 2 (2.2 glcm\

Powder samples for transmission electron microscopy (TEM) were prepared by dipping
the carbon coated TEM grids (PLANO, W. Plannet GmbH) into the powder. The TEM
analysis was performed on a Hitachi H 600 electron microscope operated at 100 kV,
206

using magnifications of to'OOO to 30'000. TEM analysis focused on examination of the

particle morphology.

4. Results and Discussion

The influence of the flame detennining process parameters on mean particle size and
particle morphology was investigated by controlled variation of the O2 flow rate, the
flame configuration and the HMDSO concentration. All gas flow rates indicated in the
paper refer to the standard state (298 K, I atm).

4.1 INFLUENCE OF FLAME CONFIGURATION

In this series the influence of a SDF (Fig. Ib) and a PMF (Fig. Ic) configuration on
particle size was investigated. The major difference between the configurations derives
from the gas supply, especially the reactant gas mixing in the central tube. While in a
SDF fuel gases (CH4 and HMDSO vapour) and N2 carrier gas are supplied through the
central tube, in the PMF configuration the HMDSO vapour (plus N z carrier gas) is
mixed with O2 in the central tube and CH4 is provided through the outer ring. This leads
to different radial and axial temperature profiles in the flames as well as flame heights,
resulting in different temperature histories of the particles [4].

For both flame configurations and all experiments the N2 carrier gas flow rate was
176 IIh at a HMDSO temperature of 5°C, which gives a SiO z particle production rate of
17 glh. A constant amount of CH 4 (85 llh) and N z lift gas (27 llh) was supplied to the
burner throughout the series, while the O 2 flow rate was varied from 75 to 1437 IIh.

The mean particle diameter strongly depends on the O2 flow rate for both flame
configurations, as shown in Fig. 2. However, the PMF leads to smaller primary particles
than the SDF at the same Oz flow rate in the range> 250 llh, which is in agreement with
TEM observations (Fig. 3).
For the SDF, the particle size first increases with increasing O2 flow rate up to 280 IIh.
At higher O2 flow rates, the mean particle diameter decreases gradually with increasing
O 2 flow rate. The decrease of particle diameter with increase of O2 flow rate in a SDF is
in agreement with other studies [3-6]. However, the presence of the sharp maximum in
particle size at 250 IIh of O2 has not been reported yet and requires further research.
In the PMF, the diameter ofthe particles decreases continuously up to a O2 flow rate of
700 IIh. For higher flow rates, the particle diameter slightly increases with increasing O2
flow rate, which is again in correspondence with other studies of diffusion flame
reactors [3,4, 7].
At O 2 flow rates> 1200 llh Si0 2 particles with a similar mean size were synthesised in
both flame set-ups, indicating that particle fonnation is not sensitive to the flame
configuration when produced in flames with high O2 exit velocity.
Stoichiometric combustion conditions in both flames prevail at an O2 flow rate of
208 llh. Therefore higher O 2 flow rates do not contribute to oxidation of CH4 or
HMDSO and therewith heat generation, but cool the flame.
 207

120
110
100
-<>- "0,1. dI""M 'om. (SDF) I
E
..
.s 90 premixed flame (PMF)
~
.
E
'6
80
70
III

~.
60

a. 50
·S 40
~
I- 30
w
III
20 >
10
0
0 200 400 600 800 1000 1200 1400 1600

oxygen flow rate II/h]

Figure 2 - Influence of flame configuration and 0, flow rate on particle size at constant gas flow rates for: N,
carrier (176 11h), N, lift (27 JIh), ClL. (85 JIh) and a HMDSO evaporation temperature of 5°C resulting in a
SiO, production rate of 17 l!/h.

;lI h)

Figure 3 - TEM pictures of SiO, nanoparticles generated in a) a SDF and b) a PMF at an 0, flow rate of
510 llh with a mean particle diameter of 56 nm and 14 nm, respectively (gas flow rates see text of Fig. 2).

TEM investigations revealed that the particle diameter such as calculate'd by Eq. I is an
adequate indicator for the mean primary particle diameter in the powder ensemble. It
can be seen, that the powder with a small mean particle diameter (dp,50< 30 nm) consists
oflarge, fractal-like aggregates (Fig. 3b), whereas non-aggregated, spherical particles of
about 50-150 nm in size predominate in powders with dp,5o> 50 nm, beside a minor
fraction of fine particles forming aggregates and agglomerates (Fig. 3a).
208

For both flame configurations the fine particles were found to be highly aggregated
when produced at high O 2 flow rates. Non-aggregated, spherical particles only formed
at low O 2 flow rates in the range of 100 to 250 llh in both flame configurations. The low
O2 flow rate results in long, smooth flames, whereas a high O2 flow rate leads to short,
turbulent flames [3-8].

4.2 INFLUENCE OF PRECURSOR CONCENTRATION

Computation models for prediction of particle size, coalescence, collision and

morphology include temperature history of the particle formation process and
temperature dependence of the material properties as main parameters [9-11]. On an
experimental level, these parameters can be expressed by flame temperature and
precursor concentration. Flame temperature depends on the species and amount of fuel
supplied, namely CH 4 and the methyl groups of the HMDSO-precursor in the present
system. Complete combustion of the CH 4 and methyl groups at stoichiometric
conditions generates the maximum heat. Turbulent gas flow conditions at the burner
outlet lead to fast reactions and therefore short flames and high temperatures [3-8].

In this series the influence of the HMDSO concentration on the mean particle diameter
was investigated in a SDF. The HMDSO concentration was defined as:

~MDSO.g
C HMDSO = TilT TilT Tilr
"CH4 + Jj:j, + ~MDSO,g (2)

and varied between 1.2, 2.4 and 4.8 vol% in the gas delivered through the centre tube of
the burner. For all experiments the carrier gas flow rate was 86 IIh. Control of the
HMDSO concentration was achieved by increasing the flask temperature from 5 to 18
and 31°C. Particle production rate was 8.5, 17 and 34 gIh, respectively. A constant
amount ofCH4 (43llh) and N2lift gas (27llh) was supplied to the burner throughout the
series, while the O 2 flow rate was varied from 75 to 1437llh.

From Fig. 4 can be deduced, that the mean particle size increases with increasing
HMDSO concentration at a constant O 2 flow rate. This correlation is in agreement with
the observations made by Wegner [4] and Kammler et al. [6]. As the HMDSO
concentration is increased by a factor of 4, more fuel (methyl groups) is supplied to the
flame, leading to an increase of the flame height and a change in combustion
stoichiometry. The increased flame height results in a longer residence time of the
particles in the flame, as the change in exit velocity of the centre gas is not significant
with the change of HMDSO concentration. This increase of residence time at high
temperatures increases particle sintering rate and hence leads to formation of larger
particles. At the same time, the increased HMDSO concentration increases the Si0 2
particle concentration in the flame, consequently leading to more particle collisions and
thus formation of bigger particles [9-11].

TEM investigation on particle morphology (Fig. 5) was used to determine the O 2 flow
range at which spherical, non-aggregated Si0 2 particles are formed. It was found that,
 209

similar to the results from in the previous series, this specific particle shape can only be
achieved at low O2 flow rates in the range of 150 to 300 Ilh. This results indicate, that
spherical, non-aggregated Si0 2 particles predominately form in long, smooth flames.
120
110 -+-SDF - 1.2 vol% HMDSO
E 100
-e- SDF - 2.4 vol% HMDSO
.:."- 90
CD
1i BO --fr- SDF - 4.B vol% HMDSO
e
.!! 70
"CI
CD
U 60
'E
III
co. 50
·S 40
a
...
w
'I' 30 --t!.
---_.--§
III 20 ----.~---~
10
0
0 200 400 600 BOO 1000 1200 1400 1600
oxygen flow rate [Ithl

Figure 4 - Influence of HMDSO concentration (Eq. 2) and 0, flow rate on particle size in a SDF at constant
gas flow rates for: N, carrier (86 l/h), N, lift (27 11h), CH. (43 l/h) and a HMDSO evaporation temperture of 5,
18 and 31°C resulting in precursor concentration of 1.2, 2.4 and 4.8 vol% in the centre tube, and a SiO,
production rate of8.5, 17 and 34 gIh, respectively.

a) b)

Figure 5 - TEM pictures of SiO, particles generated in a SDF at a) precursor concentration of 2.4 vol% at
175 IIh of 0, resulting in a mean particle diameter of 102 nm and b) precursor concentration of 4.S vol% and
300 IIh of 0, leading to particles with a mean diameter of I OS nm.
210

5. Conclusions

Non-aggregated, spherical Si0 2 nanoparticles with a mean, BET-equivalent particle

diameter of 60 to 110 nm can be synthesised in a SDF as well as in a PMF, however
only in a narrow process window at low O2 flow rates (100 to 300 IIh). At this process
and gas flow conditions, the HMDSO precursor concentration can be varied within a
range of 1.2 to 4.8 vol%, without changing particle morphology, but the mean particle
diameter increases with increasing precursor concentration. In both flame
configurations, fractal-like aggregates with an equivalent particle size of 10 to 30 nm
formed at high O2 flow rates.

6. Acknowledgements

Financial support by KTI I Top Nano 21 is kindly acknowledged (project TNS 4954.1).

7. References

[I] Pratsinis, S.E. (1998) Flame Aerosol Synthesis of Ceramic Powders, Prog.
Energy Combust. Sci., vol 24, p 197-219
[2] Technical Bulletin Pigments, DegussaAG, no II, p 1-80
[3] Kammler, H.K., Pratsinis, S.E. (1999) Scaling-Up the Production of Nanosized
Si0 2-Particies in a Double Diffusion Flame Aerosol Reactor, J. of Nanoparticle
Res., vol I, p 467-477
[4] Wegner, K. (2002) Nanoparticle Synthesis in Gas-Phase Systems: Process
Design and Scale-Up for Metals and Metal Oxides, Ph D thesis, Swiss Federal
Institute of Technology, Zurich, ETH no 14568
[5] Kammler, H.K., Mlidler, L., Pratsinis, S.E. (2001) Flame Synthesis of
Nanoparticles, Chern. Eng. Tech., no 24, vol 6, P 583-596
[6] Kammler, H.K., Mueller, R., Senn, 0., Pratsinis, S.E. (2001) Synthesis of
Silica-Carbon Particles in a Turbulent H2-Air Flame Aerosol Reactor, AIChE
Journal, no 7, vol 47, P 1533-1543
[7] Wegner, K., Pratsinis, S.E. (2001) Aerosol Flame Reactors for Synthesis of
Nanoparticles, Kona Powder and Particle, vol 18, p 170-182
[8] Briesen, H., Fuhrmann, A., Pratsinis, S.E. (1998) The Effect of Precursor in
Flame Synthesis ofSi0 2, Chern. Eng. Sci., no 24, vol 53, p 4105-4112
[9] Lee, B.W, Jeong, J.I, et al. (2001) Analysis of Growth on Non-Spherical Silica
Particles in a Counterflow Diffusion Flame considering Chemical Reactions,
Coagulation and Coalescence, J. Aerosol Sci., issue 32, vol 32, p 165-185
[10] Koch, W., Friedlander, S.K., (1989) The Effect of Particle Coalescence on the
Surface Area of a Coagulating Aerosol, J. Aerosol Sci., no 8, vol 20, P 891-894
[11] Koch, W., (1990) Particle Growth By Coalescence and Agglomeration, J.
Aerosol Sci., suppl 1, vol 21, P S73-S76
FORMATION OF COMPOSITE SiC - SiB6 NANOPOWDERS

M. LASOVA(I), M. KAKAZEy(I), J.G. GONZALES-RODRIGUEZ(ll,

G. DOMINGUEZ(I), E. PRILUTSKII(2) T. TOMILA(2), L. ISAEVA (2), L.
DOMASEVICH(2), A. MYRA TOV A (2)
(1)The Autonomous University of State of Morelos, Cuernavaca, Mexico
(2)The Institute for Problems of Materials Science, National Academy of
Sciences of Ukraine, Kiev, Ukraine

Abstract
In a nanodisperse Si0 2 - B20 3 - C system, the reduction process was
investigated by X-ray phase analysis, IR spectroscopy, EPR and electron microscopy. It
was established that a temperature treatment of the mixture is accompanied by the
formation of a borosilicate glass-like system. The carbothermal reduction of borosilicate
melt at a temperature about 1600 K is accompanied by the formation of SiC, SiB 6, and
Si0 2 •

Key Words: borosilicate, reduction, silicon carbide, silicon boride

1. Introduction

Silicon carbide (SiC) belongs to the class of refractory nonmeta.llic compounds.

The multifunctionality of SiC-based materials, possessing a complex of qualities such
as high hardness and strength, chemical and temperature resistance, specific
electrophysical properties, etc., stimulates the development of special technologies
providing different combinations of properties in such materials. In recent years, the
synthesis of composite silicon carbide ceramics has attracted particular interest. New
properties of these ceramic materials extend significantly the field of application of a
monophase SiC material, that has been studied rather well [1-3].
The aim of the present work is to investigate the process of formation of composite
SiC-based powders with additions of boron silicides. It has been known [4] that alloys
on the base of B-SiC exhibit high chemical resistance, extremely high resistance to
thermal shocks, high strength at room and elevated temperatures, etc. It can be supposed
that SiC ceramics with additions of boron silicides will distinguish by increased strength
properties and possess conducting properties.
211
Y G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 211-218.
© 2003 Kluwer Academic Publishers.
212

2. Procedure of Sample Preparation

The procedure of preparation of mixtures included the following operations. Into

preliminarily synthesized gel of silicic acid (H2 SiO) saccharose was introduced in
amounts required for the formation of silicon carbide by the reaction Si0 2 + 3C = SiC +
2COt. Then boric acid and saccharose were added to the obtained mixture by the
solution method. In this stage, the amount of the saccharose introduced must be
sufficient for the formation of carbon, which is necessary for the reduction of boric
anhydride to boron. The taken total amount of boric acid (H)B03 ) and saccharose was
calculated for the formation of 4-23 mass % of B20) in mixture with SiC. Mixtures with
a two- and tree-fold excessive carbon content as compared to that required according to
the reaction Si02+C~SiC + cot were also prepared.
The mixtures were dried by moderate heating and subjected to pyrolysis in a
stepwise regime in the temperature range 273-723 K. The specific surface of the powder
mixtures was 200-250 m 2/g.
The temperature treatment of the mixtures was performed in a vacuum electric
resistor furnace with a tungsten heater. The ultimate vacuum was 2.7·10-) Pa. After an
exposure for 1 h under stable pressure, samples were taken at different stages of
temperature treatment.
The X-ray phase analysis of samples was performed using a DRON-I
diffractometer in Cu Ku radiation. The semiquantitative evaluation of changes in the
contents of phases was carried out using standards. IR absorption spectra were recorded
on a "Specord 75 IR" spectrometer. EPR spectra were recorded on a "SE/x 2547" unit
in vacuum and air using a standard. Electron microscopy studies were carried out on a
HU 200F unit.

3. Experimental Results

3.1. Starting Mixtures

In X-ray diffraction patterns of dry mixtures, a halo with traces of Si02 and 8 20 3
with a badly formed crystal structure is recorded. EPR spectra are characterized by an
intensive single line with g - 2.0027 and a width I1B - 2· W-4 T, which corresponds to
carbon, which is a product of thermal destruction of saccharose [5]. Thus, the mixtures
consist of Si0 2 , 8 20 3, and C particles. According to the electron microscopy data, the
particle size was -15 nm for carbon, -30 nm for 8 2 0 3 , and -50 nm for Si0 2 •

3.2. Heat Treated Mixtures

According to the X-ray analysis data, the main crystalline phases detected in
the heat treated mixtures are 8 2 0 3 , Si0 2 , and SiC of type I (hexagonal) (Fig. 1, 2). At
Ttre• t > 1100 K, in samples, an intensive halo is recorded in the range 26 - 14-32 grad
with a maximum at 26 - 21-22 grad. The appearance of the halo (of the X-ray
amorphous phase) may be caused by the formation of borosilicate glasses in the Si0 2 -
8 20) system [6]. An increase in the intensity of the halo after low-temperature treatment
 213

with increasing calculated boron content in the mixtures (actually, with increasing B20 3
content in the starting mixtures (Fig. 3, curve 3') and a set of IR absorption bands
characteristic of borosilicate glasses (Table 1) verify this assumption. Moreover,
according to the electron microscopy data, in the mixtures heat treated at 1273 K, glass-
like granules 1-3 nm in size are detected. However, in the high-temperature region of
synthesis (T = 1773 K), no increase in the content of the X-ray amorphous phase takes
place even at a high B20 3 content in the starting mixtures (see Fig. 3, curve 3).

3;:: 30 25 20 15

Figure 1. X-ray diffraction patterns of powders obtained by treatment of mixtures calculated for 8 mass %
B,O, (a, b, c) and 23 mass % B,O, contents with a doubled carbon content (d). 1273 K (a), 1473 K (b), 1773
K (c, d). Designations: 0 - B,O" v - SiO" x - SiC, • - SiB,.

160

C
120
.
::I

of!
80
~
•c
.! 40
.5
4
0
1273 1373 1473 1573 1673 1773
Temperature, K

Figure 2. Changes in the contents of crystalline (1, 2, 4) and amorphous (3) phases during heat treatment ofa
mixture containing 23 mass % B,O,. 1 - B,O" 2 - SiO" 3 - halo, 4 - SiC.
214

60

c: 50
:3
~ 40

.i
:>; 30

20
:5 10

O+---~----~--~~---r----'
o 5 10 15 20 25

Figure 3. Changes in the contents of crystalline phases (2, 4) and the intensity of halo (3,3') depending on the
B,03 content in starting mixtures. For 2 - 4 - T,~, = 1773 K; for 3' - T,,,,, = 1273; 2 - SiO" 3, 3' - halo, 4-
SiC.

Table I. IR absorption spectra of heat treated mixtures and standard samples

Sample type Frequency of IR absorption, v, cm"

Mixtures with 3 mass.% B,03 Mixtures with 20 mass.% B,03
Starting mixture 1650w.; 1450w.wd.; lloost.wd., 1200sh.; I 640av.; 1440st.wd.; 1200st.wd.,
920w.; 800av.n.; 470av.wd; lloosh.; 925w.; 880w.n.;
8IOst.wd.; 670w., 550av.;
470av.wd.,
Mixture, 1630w., 1390w.;109Ost.; 920w, 795av., 1640w.; 1460v.st.; 1190av.;
T,,~,=1273 K 7lOsh; 460st. I 100av., 1080sh.; 920w.; 880w.n.,
76Ost., 670av.; 540w., 460st.
Mixture, 1630w.; 1390w.;1090st., 920w.;795w.; 1640w., 1430st., 1370sh.; 12oow.,
T,,~,=1473 K 660v.w; 46Ost. 1150w., I 100w.; 1020w.; 920av.;
750w.; 640w., 460st.
Mixture, 1630v.w.; 1390w.; 1090st.; 920v.w., 800av., 1640v.w.; 1390w., 109Ost.; 920av.;
T,,,,, =1673 K 690v. w; 460st. 780w.; 660w.; 460st.
Mixture, T,,,,,-1773 K 163Ov.w.; 139Ow.;IIOOst.,l150sh., 1090sh.; 1640v.w.; 1390w.; 109Ost.; 920w.;
920w.; 8ooav.,780av; 46Ost. 8oow.,780w; 660w.; 460st.
B,03[7] 1680w.; 1462st., 1220sh.,1452sh., 1190sl.n.; 766w.wd
SiO, (a-quartz) [8] 1084st., II 72sh., 1152sh.; 798, 780n.; 697w.; 652w; 522w.; 465st.

SiO,(fused quartz) [8] 1114st., 1220sh., 80Iav.w., 630w.,580w.; 476sl.

SiO,-B,O, (glass) 1400v.w.; 1I00st., 925w.; 820av.; 670w.; 600v.w.;
[9,101

Subscripts: sl. - strong; avo - average; w. - weak; n. - narrow; wd. - wide; v. - very.

A decrease in the content of the 8 Z0 3 phase observed at Ttreat;e: 1273 K (see Fig. 2,
curve 1) indicates that the carbothennal reduction of 8 Z0 3, which, according to [11],
begins at 1200-1300 K, proceeds.
The fact that SiC is detected in the high-temperature region of synthesis can be
explained by the development of the process of carbothennal reduction of SiOz at a low
8 20 3 content in the starting mixtures and borosilicate glass in the mixtures with a high
8 20 3 content (see Fig. 2, 3, curve 4). Note that, after treatment at 1773 K, in the mixture
 215

containing 23 mass % of boron anhydride, a relatively insignificant amount of silicon

carbide is present (of no more than 10 mass %).
In the samples with a high B20 3 content synthesized at 1773 K, a decrease in the
intensity of the halo is accompanied by not only the formation of SiC, but by the
appearance of a-quartz as well (see Fig. 2, curves 2, 4). It can be supposed that, in this
stage of temperature treatment, the completion of the reduction of the borosilicate
system and carbide formation require a carbon content higher than that introduced into
the mixture with the following ratio of components: 39.4 mass % Si02 : 23 mass %
B2 0 3 : 37.6 mass % C. Actually, as the carbon content in the starting mixtures is
increased, the amount of silicon carbide rises, the halo disappears, SiB6 forms, and Si0 2
is retained (Fig. 4).

30
C
.e." 20

fc 10
~ 5
0
2 3
Content of C, arb. un.

Figure 4. Changes in the contents of crystalline phases with increasing carbon content in starting
mixtures. T..., = 1773 K; 2 - Si0 2, 4 - SiC, 5 - SiB6

In the IR spectra of the heat treated mixtures, absorption bands characteristic of

Si0 2 and B2 0 3 are present (Fig. 5, Table 1). In the samples with a low content of B2 0 3
synthesized at 1273 K, Si-O absorption bands of fused quartz are easily identified,
while, as the B20 3 content increases, the intensity of absorption bands of B-O bonds
increases essentially (Fig. 5 a, b, curves 1). The presence of additional absorption bands
at v - 670, 925, 1020, and 1400 em-I indicates the formation of borosilicate glass.

~\AbM
~2VV_~.
"\!\ ~'\
c

~
... , , ,
1!500 1000 «JO 1600 1000 4)J 1!500 1000 QJ
v~ em-I v, cm-I v,cm- l

Figure 5. IR absorption spectra in mixtures with 4 mass % B 20, (a) and 23 mass % B,O, (b) heat
treated at 1273 K (1),1473 K (2), and 1773 K (3).
216

As the heat treatment temperature is increased, the intensities of the absorption

bands of B-O bonds in B 20 3 decreases (see Fig. 5 a, b, curves 1-3). This corresponds to
a decrease in the content of free B20 3 in the mixture and coincides with the X-ray phase
analysis data. The absorption bands of borosilicate glass are more clearly defined (see
Fig. 5, curves 2). After treatment at 1773 K, in powders synthesized from the mixtures
with a low B2 0 3 content, the IR spectrum is close to that of fused quartz (Fig. 5a, curve
3), while, in powders obtained from mixtures with a high B 20 3 content, the IR spectrum
corresponds to the spectrum of borosilicate glass (Fig. 5 b, curve 3). Note that these
results, like the X-ray phase analysis data, indicate the incompleteness of the reduction
process of the oxide systems at the chosen ratio of components.
The formation of a doublet at v - 800 cm·] in samples with a high boron content at
Ttreat ;?: 1673 K (see Table 1) agrees with the X-ray phase analysis data on the formation
of a-Si02 as individual crystalline phase.
Though, according to the X-ray phase analysis data, in the high-temperature
samples synthesized from the mixture with 23 mass % B20 3 , silicon carbide forms, the
absorption band of SiC in the range 800-900 em·] was not recorded. The indicated
inconsistency is explained by the fact that silicon carbide particles are "shielded" by
borosilicate glass. However, in samples with an increased carbon content, this
absorption band is observed (see Fig. 5 c, curve 3).
For the EPR carbon signal with a width of about 2.10.4 T, "an oxygen effect" is
characteristic [10]. It can be observed in vacuum. However, the recording of a signal in
samples synthesized in air shows that carbon inclusions are isolated from environment.
Taking into account the low melting points of B2 0 3 and the Si0 2 - B2 0 3 mixture (T -
724-731 K) [11], it can be supposed that the carbon inclusions are "encapsulated" in the
glass phase. A change in the intensity of the EPR carbon signal shows (Fig. 6) that the
disappearance of free carbon inclusions is stepwise in character. An essential decrease
in the number of carbon paramagnetic centers takes place at temperatures -1473 and
-1773 K. At the indicated temperatures, a significant decrease in the B20 3 content and
then in the content of the glass phase and the formation of SiC are noted.

100

c: 00

.e." 60

ic 40
S
.5 20

0
1273 1373 1473 1573 1673 1773
T....,.........,K

Figure 6. Variations in the intensity of EPR carbon signal during temperature treatment of a mixture
with 8 mass % B,03.
 217

4. Discussion of Results

The starting mixtures are nanocompositions of Si02 , 8 20 3 , and C, whose

crystalline structure did not form completely. During temperature: treatment at a
temperature above 1100 K, a significant part of the Si0 2 - 8 20 3 mixture forms the glass
phase. The presence of the 8 20 3 phase in borosilicate glass is likely connected with the
chosen temperature treatment regime, which did not provide the homogeneity of the
melt due to its high viscosity [1].
The reduction properties of the used highly disperse carbon (the product of thermal
destruction of saccharose) manifest themselves beginning from T 2 II 00 K [12] for the
B20 3 - C system, from T 2 1500 K [5] for the Si02 - C system, and, according to the
results of the present work, from T 2 1700 K for the borosilicate system. Under the used
heating regime of the initial mixture, an eutectic Si0 2 - 8 20 3 melt, containing local
regions of a 8 20 3 melt and carbon inclusions, forms. That is why the reduction of boron
anhydride proceeds first, and then the borosilicate melt is reduced by carbon. In the first
stage, according to the reaction 8 20 3 + C ~ 80(B20 2)t + CO 2 (CO) t, the reduction
products are volatile boron oxides. As the treatment temperature is increased, in
mixtures with a small B20 3 content, the reaction Si0 2 + C ~ SiC + CO2 (CO) t
develops. At high temperatures, the reduction of borosilicate melt can realize both
through the carbothermal reduction (Si02 )n·(8 20 3 )m + C ~ SiC, 8 4C + 80(8 20 2)t +
CO 2 (CO) t + SiOt and the reduction of the melt by formed silicon and boron carbides.
It should be noted that reactions proceeding with the formation of SiB6 and Si02 ,
namely (Si0 2 )n·(8 20 3 )m + SiC ~ SiB6 + Si0 2 + cot are possible, which causes the
appearance of Si02 in the high-temperature samples. The participation of boron carbide
formed in reduction processes is most probably the reason why it is difficult to detect it
in synthesired powders.

5. Conclusions

The investigations performed showed that the use of the sol-gel and solution
method makes it possible to prepare a nanodisperse mixture Si02 -- 8 20 3 - C. The
temperature treatment of such a mixture is accompanied by the formation of a
borosilicate glass-like system. The carbothermal reduction of a borosilicate melt at a
temperature -1600 K results in the formation of SiC, SiB 6 , and Si0 2 _ The ratio of the
components in synthesized powders depends on the chosen ratio Si0 2 : 8 20 3 : C in the
initial mixture and heat treatment conditions.

References
I. Naslain R., Lamon J., and Doumeingts D. (1993) High Temperature Ceramic Matrix Composites,
WoodLead Pub!., Abington-Cambrige UK.
2. M.L. Scott (ed.), Proceedings of the 1/ th International Conference on Composites Materials (Woodhead,
London, 1997).
3. Evans A.G and Naslain R. (1995) High Temperature Ceramic-Matrix Composites If: Manufacturing and
Materials Development. American Ceramic Society, Westerville, OH.
4. Kosolapova T., Andreeva T., Bartnitskaya T. and others. (1985) Nonmetallic Refractory Compounds.
Metallurgia, Moscow.
218

5. VIas ova M.V, Domasevich L.T, Kakazey N.G, Pavlenko N.P., Tomila TV., Artyukh SYu., Ivchenko V.I.,
Kilimnik A.A., Lezhnenko Yu.I. (1993) Effect of the initial state of synthesis components in the reaction of
carbothermal reduction of silica on the structure of silicon carbide particles. II. Aerosil- sacharose system,
Powder Metallurgy and Metal Ceramics 32,606-613.
6. Abe T, (1952) Borosilicate Glasses, J Amer. Ceram.Soc. 35,284-299.
7. Galenner F., Lugovsky G., Mikkelsen J. (1980) Vibrational spectra of pure vitreous B,03, Phys. Rev. 22B,
3983-3990.
8. Vlasov A.G., Florinskaya V.A., Venedictov A.A. and others. (1972) Infrared Spectra of Inorganic Glasses
and Crystals, Khimiya, Leningrad.
9. Antonova S.L. and Dyakova V.V. (1979) Investigation of glass formation of high-silicon borosilicate
glasses synthesized by the sol-gel method, Fiz. Khim. Stekia 5,671-674.
10. Konijnendijk W. L. (1975) The Structure of Borosilicate Glasses, Phil. Res.Repts.Suppl. No.1, 158-159.
11 Toropov N.A., Barzakovskii V.P. Lapin V.v., and Kurtseva N.N. (1969) Phase Diagrams of silicate
systems I, Nauka, Leningrad.
12. Bartnitskaya T.S., Vlasova M.V, Kosolapova TYa., Kostyuk N.V. Timofeeva I.I., Tomila TV. (1991)
The role of structure ordering of source components in boron nitride formation, Soviet Powder Metallurgy and
Metal Ceramics 30, 491 -497.
A MODEL OF ELECTRICAL CONDUCTION IN XEROGRAPHIC
TONERS AS POLYMER - CONDUCTIVE FILLER PARTICULATE
COMPOSITES

VLADISLAV SKOROKHOD
Xerox Research Centre a/Canada
2660 Speakman Drive, Mississauga, Ontario L5K 2Ll, Canada

Abstract: An equivalent circuit model of electrical conduction in toner particles

with sub-mono layers of fme conductive additives on the surface was developed in
this study. The equivalent circuit was constructed and represented mathematically by
a linear system of nodal equations solved numerically by using Matlab software to
calculate composite conductance as a function of the amount of conductive filler and
filler-to-matrix size ratio. Additional capabilities of the present model include
calculation of the statistical distribution of composite particle conductance, and the
possibility to account for non-random configuration of filler particles.

1. Introduction

Conductive mixtures of polymers and conductive fillers such as carbon black, metal
or metal oxide powders fmd a variety of technological applications as antistatics,
materials for electromagnetic shielding, etc. Xerographic toner is another type of
polymer-filler composite wherein coloured polymer core particles with sizes of 5 -
20 Iilll are blended with nanostructured particulate surface additives. The primary
role of the additives is to modify various physical and electro-physical properties of
toners in order to facilitate toner transport in a xerographic printer [i, ii]. Because
toner transport is primarily controlled by electromagnetic fields, electrical properties
of toner materials such as conductivity, and dielectric pennittivity and loss are
usually required to meet tight specifications. Therefore, understanding the
mechanisms of electrical conduction in toners as insulator-filler particulate
composites is an important aspect of toner material design.
Typically, electrical conductivity of polymer-filler composites increases rapidly
at a characteristic filler concentration (usually referred to as the percolation
concentration), and then increases further with filler content. Classical statistical
219
Y.G. Gogutsi aruJ I. V. Uvaruva (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 219-227.
© 2003 Kluwer Academic Publishers.
220

percolation theory [iii, iv, v] describes the percolation effect in an insulative matrix
with uniformly distributed conductive filler as the formation of a continuous network
of contacting filler particles. Theoretical calculations in these models are based on
the assumption that the conductive particles are arranged into a two- or three-
dimensional periodic array with characteristic particle coordination. Comprehensive
reviews on percolation theory can be found in literature [vi, vii].
Since most previous studies on insulator-filler particulate composites were
driven by the needs of ceramic or polymer composite applications, they tended to
focus on the electrical properties of such materials in a compacted or consolidated
form. One group of percolation models describes and gives good approximations for
conductivity dependence on filler vol. % in particulate composites comprised of
individual insulative particles with a surface layer of fine conductive additives [viii,
ix, x]. However, although the composite configuration in these studies was
essentially similar to that of toners, their primary focus was bulk conductivity of
such composites in a compacted form, not the conductivity of an individual
composite particle.
The present study addresses several important aspects of electrical conduction
of matrix-filler particulate composites, which were left out of the focus of previous
studies but can be of great interest for toner applications. These aspects include the
method to estimate conductance of an individual toner particle with a monolayer or
sub-monolayer of fine conductive surface additives, the statistical distribution of
conductance, and possible effects of non-random structure of conductive surface
additives.

2. Experimental

Styrene-acrylate toner particles with a diameter of 5 !lm were prepared and

spherodized to the author's request. The particles were blended with a conductive
tin oxide powder with an average particle size of 50 nm at 1 - 4 vol. % by using a
standard blending process typically utilized in toner technology. The resulting
composite powders were compacted into circular pellets with a diameter of 10 mm
and a thickness of 1 mm in a steel die with insulative walls, under a pressure of 2000
psi. Current-voltage curves were measured in-situ by applying a voltage to the
plungers and measuring series current with a Keithley electrometer. Conductance of
each sample was evaluated at 10 Volts.
 221

3. Results and Discussion

Figure 1 shows a schematic of a bi-disperse composite particulate system comprised

of insulative matrix particles with a size D, and small conductive filler particles with
a size d (such that d« D) dispersed on the surface of the matrix particles. Further
in this paper the large insulative particles will be referred to as Ithe "matrix", the
small conductive particles as the "filler", and the matrix with filler dispersed on the
surface as the "composite". Finite-area electrical contacts occur between
neighbouring filler particles due to interparticle attractive forces (e.g., Van der
Waals force) [xi].
The typical approach used in past literature on electrical conduction modeling
in insulator-filler particulate systems was to approximate the random particulate
system by a regular periodic array [iii]. Electrical conduction in such systems is
affected by two major factors: (i) elementary conductance of an individual
composite particle between the points of contact with the neighbouring particles, and
(ii) extrinsic properties such as compaction density, porosity, intl~rparticle forces,
and particle coordination. The present study primarily focuses on the individual
conductance of composite particles. The extrinsic effects of compaction on the
electrical conductivity of particulate systems has been well studied in the past [iv,
xi].
To obtain conductance of an individual composite particle between the two
contacts with neighbouring particles (see Figure 1), it is convenient to represent
matrix particles as cylinders with two equipotential bases (ele,ctrodes), with a
diameter equal to its length, such that the surface area of the circumferential surface
will be equivalent to the surface area of a sphere with the same diameter (trD 2 ). The
filler particles are distributed on the circumferential surface of the cylinder and
arranged into a two-dimensional hexagonal close-packed surface array. The number
of filler particles required to form a hexagonal close-packed monolayer (n) can be
found as a ratio of the available surface area of the matrix to the area occupied by
one filler particle in an Rep lattice (;j3cfI2):

(1)

where D and d are the diameters of the matrix and filler, respectively. Ifthe number
of filler particles per one matrix particle is smaller than n in Eq. (1), the filler will
form a sub-monolayer where some of the Rep sites are occupied, and the other sites
are vacant (see Figure 1). From a simple geometrical calculation, the number of sites
222
in one circwnferential chain on the surface of the model cylinder (L) and the nwnber
of circwnferential chains between the electrode surfaces (K) can be found as:

L=27rD
d
(2)

K=~
-J3d

'3.0
V -- -'a-"---""
3,_1 __

OU

(a) (by

Figure 1. (a) schematic of a conductive Figure 2. Equivalent circuit

chain formed by composite
spherical particles with conductive
filler particles randomly dispersed
on the surfaces; (b) the equivalent
representation of an individual
composite particle by a cylinder
with a potential VA applied to its
bases and with a sub-monolayer of
filler particles arranged into a two-
dimensional hexagonal close-
packed (HCP) array.
representation of the model
cylindrical particle in Figure
1. The circuit nodes are
assigned to the HCP sites
and indexed according to
the number of
circumferential row (I st
index) and the node's
position in the row (2 nd
index). The filled and open
nodes correspond to the sites
that contain a filler particle
and a vacancy, respectively.
The arrows represent the
elementary currents jab<XI
from node a, b to node c, d
through the interparticle
contacts.

The equivalent electric circuit for the model cylinder with a sub-monolayer of
filler particles (Figure 1) is shown in Figure 2. The circuit consists of nodes assigned
 223

to the Hep array sites either occupied by filler particle (filled circles) or remaining
vacant (open circles). The arrows in Figure 2 correspond to the elementary currents
I:: flowing through an interparticle contact from node a, b to nodf: c, d which can be
represented in terms of the electrical potentials Vab and Vcd at nodes a,b and c,d,
respectively:

(3)

where g is the conductance of an interparticle contact. Because no current is flowing

in or out of the vacant nodes, i;:
= 0 if either a, b or c, d is a vacant node.
A current balance equation can be written for each node based on Kirchhoff's
rule:

·kl
1 + i (k-I)I
kl + i k! _ ik(l+I) _ i(k+I)1 _ i(k+I)(I-I) =0 (4)
k(l-I) (k-IX!+I) kl kl kl

where k is the index of the circumferential row of nodes, and I is the node index
within the row. The combination of Eqs. (3) and (4) leads to an equation for nodal
potentials written for any node k,l:

U(k_I)1 + U(k-I)(I+I) + Uk(l_l) - N klUkl + Uk(l+l) + U(k+I)(H) + U(k+I)1 = 0 (5)

where Ukl is the normalized potential (Ukl= VkyVA), VA is the potential drop across the
model particle (see Figure 2), Nkl is the coordination number of the node k,l, i.e., the
number of occupied neighbouring sites (Nkl = 6 if all the neighbouring sites are
filled). For the nodes adjacent to the electrodes (k=O and k=K), Eq. (5) is modified
by using the following boundary conditions: U (k.l)FO for k=O, and U(k+l)1 =1 for
k=K:

(6)

and
224

U ( K -I) I + U (K -I )(1 + I) + UK (I-I) - N kl U Kl + UK (/ +I) -1 (7)

Equations (5) - (7) written for all nodes form a system of linear equations which can
be solved for each u by using a standard solution technique such as by diagonalizing
the matrix of coefficients. The total conductance of the composite particle can be
found by using Kirchhoff's rule:
(8)
L
G gLu ol
1=0

Note that the nodal potential Uab = 0 for all vacant nodes a.h. Also note that there is
no need to write the nodal equations (5 - 7) for the vacant nodes.
The average conductance of a composite particle with a sub-monolayer of filler
particles was calculated through Eqs. (5 - 8) for different matrix-to-filler size ratios
(DId) and plotted in Figure 3 as a function of monolayer percentage. The term
"relative conductance" refers to the ratio of composite particle conductance to the
conductance of a contact between two filler particles. Random permutations of filled
and vacant nodes were generated for each monolayer percentage, and systems of
nodal equations (Eqs. 5 - 7) were built and solved by using Matlab 6.1. The
procedure was repeated 1000 times at each monolayer percentage to generate a
representative data set through random sampling. It is seen that relative conductance
changes with monolayer % in a broad range. The variation becomes more significant
at higher Did. For example, at a Did of 10, the change of monolayer percentage
from 80 to 100 % results in an increase in conductance logarithm by approximately
6 units, while at a Did of 100, the same change in monolayer percentage results in an
increase in conductance logarithm by approximately 70 units.
Figure 4 shows experimental data on toner conductivity plotted as a function of
vol. % of conductive tin oxide filler. The additional abscissa axis on the top of the
graph shows the corresponding percentage of the surface monolayer of conductive
particles calculated through Eq. (I). The onset of percolation is observed at
approximately 1.5 vol. % of filler, which corresponds to a 40 % monolayer. The
increases of conductivity logarithm with monolayer percentage plotted in Figure 4
appears to be somewhat shallower than it could be expected from the theoretical
plots in Figure 3. The experimental composite sample had a Did of 100. However,
the observed change of conductivity from 60 to 80 % of monolayer was only about
10 units, which corresponds to the theoretical curve in Figure 3 for Did of 20. One
of the possible causes for this discrepancy between the theoretical and experimental
data is filler agglomeration with an agglomerate size of approximately 5 times larger
than the primary particle size.
 225

% monolayer 20 40 60 80 100
2l 20
c Dld= -5
ti 0
:I

"c0 10 5 -10
en
f
-20
0 20
Gl
-40 -15
j -60 :::l
a::
c "c:0 -20
...J -80 t)
c: -25
40 60 80 100 ...J

% of Monolayer -30
0 2 3 4 5
Filler vol. %

Figure 3. Change of calculated relative Figure 4. Conductivity measured in pressed

conductance logarithm with toner pellets plotted against filler
monolayer percentage for volume percentage and
different matrix-to-filler monolayer percentage.
particle size ratios (DId)

In addition, two attractive and unique capabilities of the present model can be
demonstrated. First, if filler is randomly distributed on the matrix surface, the exact
configuration of the sub-monolayer varies from particle to particle at the same vol.
% of filler, and thus, conductance of an individual composite particle will also
somewhat vary depending on the actual position of filler particles on the matrix
surface. The distribution of conductance obtained through the random sampling
method is shown as a three- dimensional plot in Figure 5 for a Did of 20. It is seen
that the distribution of conductance is broader at lower monolayer p~!rcentages.
Second, conductive surface additives often tend to form a structured fractal-type
network (e.g., carbon black) rather than being randomly distributed within the sub-
monolayer. In the present model, such structured configurations of filler are
accounted for by combining the sub-monolayer's vacancies into clusters with a
characteristic minimum number of vacancies in each cluster. The change of
conductance plotted in Figure 6 with monolayer % for different sizes of vacancy
cluster indicates the expected increase in conductance as the conductive filler
network becomes more structured (i.e., with greater numbers of vacancies in a
cluster.)
226
0
I LLJ
-5

..
u
-10
Ii ~ ~+----------------,~~
ti
:I
'0
-15 ;
C g~O+i____________79~____~
..
0
u -20 'tI

..
>
i
Q:
c
-25
5
u ·15 +----'1"--:7'T7'7L-------;
~
-I
-30 i~r-~~----------~
i
0::
-35 5· 2 5 1 - - - - - - - - - - 1
-40 .JO +-----,---,-----,-------,---1
50 60 70 80 90 100
50 60 10 80 90 100
% or Monolayer
%of Monolayer

Figure 5. A three-dimensional plot showing the Figure 6. Change of relative

distribution of relative conductance conductance with
calculated by random sampling of sub- monolayer percentage
monolayer permutations at Did = 20. for different numbers
of vacancies in a
cluster (see numbers in
italic)

4. Conclusions

An equivalent circuit model was developed to describe electrical conduction in

particulate composites comprised of insulative matrix particles with sub-mono layers
of fine conductive surface additives. The equivalent circuit was represented
mathematically as a system of linear nodal equations. Conductance of an individual
composite particle was calculated as a function of matrix-to-filler size ratio (DId)
and monolayer percentage. General agreement was achieved between the theoretical
and experimental data. Unique capabilities of the present model demonstrated in the
present study include calculation of the statistical distribution of particle
 227

for the effects of non-random distribution of filler particles on composite particle

conductance.

References

i. Gutman, EJ. and Hartmann, G.c. (1997) The role of the electric field in triboelel~tric charging of two-
component xerographic developers, in (G. Marshall, ed.) Recent Progress in Toner Technology, The
Society for Imaging Science and technology, pp. 234 - 244.
ii. Jia, N.Y. and Fletcher, G.M. (1997) Measurements of the effective dielectric thidmess of toner layers
in transfer systems, in (G. Marshall, ed.) Recent Progress in Toner Technologies, The Society for
Imaging Science and technology, pp. 313 - 316.
iii. Kirkpatrick, S. (1973) Percolation and conduction I: Transport theory of percolation processes, Rev.
Mod. Phys. 45, 574 - 588.
iv. Janzen, J. (1975) On the critical conductive filler loading in antistatic composites, J Appl. Phys. 46,
966 - 969.
v. Scarisbrick, R.M. (1973) Electrically conducting mixtures, J Phys. D: Appl. Phys. 6,2098 - 21 10.
vi. Lux, F. (1993) Models Proposed to explain the electrical conductivity of mixtures made of
conductive and insulating materials, J Mater. Sci. 28,285 - 301.
vii. Clingerman, M.L., King, JA, Schulz, K.H., and Meyers, J.D. (2002) Evaluation of electrical
conductivity models for conductive polymer composites, J Appl. Polymer Sci. 8J, 1341-1356.
viii. Maliaris, A and Turner, D.T. (1971) Influence of particle size on the el,:ctrical resistivity of
compacted mixtures of polymeric and metallic powders, J Appl. Phys. 42,614 - 618.
ix. Rajagopal, C. and Satyam, M. (1978) Studies of electrical conductivity of insulator-conductor
composites, J Appl. Phys. 49, 5536 - 5542.
x. Bhattacharya, S.K. and Chaklader, AC.D. (1982) Review on metal-filled plastics, Poly.-Plasl.
Technol. Eng. 19,21- 51.
xi. Kendall, K. (1990) Electrical conductivity of ceramic powders and pigments, Powder Technology
62, 147 - 154.
COMPUTERIZED COMPLEX FOR SINTERING NANOCERAMICS
AT HIGH PRESSURES

V.S. URBANOVICH, and G.G. SHKATULO

Institute of Solid State and Semiconductor Physics
National Academy of Sciences of Belarus
17, P. Brovka St., Minsk, 220072, Belarus

1. Introduction

High-pressure sintering is one of the effective methods for obtaining high-density

nanoceramics based on high-melting-point compounds [1-4]. For example, TiN ceramics
with nearly full density and increased microhardness have been obtained by this method
[5,6]. It is evident that the structure and properties of nanornaterials sintered at high pres-
sure are defined by both the thermodynamic conditions of sintering (pressure, tempera-
ture) and kinetic parameters (time oftheir exposure, and also pattern of their change dur-
ing a sintering process). Sintering thermodynamic parameters are defmed to a great ex-
tent by characteristics of a high-pressure apparatus, while sintering kinetic parameters
are significantly defined by capabilities of the system controlling PT-c:onditions in a sin-
tering zone.
Conventional systems controlling temperature and pressure in a reaction cell of high-
pressure apparatus do not provide means for changing them by the dlesired law and do
not allow for their visual control during the sintering process [7-10]. Due to this, the po-
tentials of thermobaric treatment technique for forming the desired nanoceramics micro-
structure cannot be fully utilized. This paper reveals an automated complex for sintering
ceramic materials comprising a set of high-pressure and high temperature equipment for
the hydraulic press units with a force of 5-6.3 MN and a computer system to control and
operate by thermobaric sintering conditions.

2. Equipment for Higb-Pressure Nanoceramics Sioteriog

2.1. HIGH-PRESSURE AND HIGH-TEMPERATURE APPARATUS

Figure 1 shows a high-pressure and high-temperature anvil-type apparatus developed by

us for sintering high-melting compounds-based nanoceramics. This apparatus is the ret-
rofitted high-pressure apparatus described in work [6] allowing for sintering ceramic
samples of a larger size with a diameter up to 12 mm at pressures up to 4 GPa and tem-
peratures up to 2200° C. This Figure also shows components of a reaction cell for sinter-
ing of ceramics on this picture too. The apparatus consists of two hard-alloy WC-Co
molds 41 mm in diameter and 20 mm in height having central trapl~zoidal recesses 25
mm in diameter and 5 mm deep at their facing ends. The surface around central recesses
229
Y. G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured MateriaL, aruf Coatings for Biomedical and Sensor Applications, 229-237.
© 2003 Kluwer Academic Publishers.
230

is made to have a 5° cone angle. The molds are held together by a sandwich collar from
steel support rings. The container is made of pressed lithographic stone. An axial open-
ing of the container mounts molded graphite heaters consisting of two horizontally split
parts.

Figure 1. High·pressure and high-temperature apparatus for nanoceramics sintering.

2.2. SINTERING CONTROL COMPUTER SYSTEM

A number of devices in the field of computer-aided systems controlling sintering (syn-

thesis) temperature conditions at high pressures has been provided [7-11]. They can be
categorized as follows: apparatuses comprising an analog controller and an analog regu-
lator [7, 8]; apparatuses comprising a digital hardware controller and an analog regulator
[9]; and programmable microprocessor-controlled apparatuses provided with an ana-
logue regulator [10].
In principle, devices [7-9] cannot solve problems of controlled sintering. While the
KS-l synthesis controller [10] comprises a microprocessor, it is not suitable as a basis
for solving the problem set due to low efficiency and use of the analog regulator.
Of specific interest is designing the controller integrating assemblies of the program-
mable logical controller "S7-200 Simatic" [11]. The said controller is assembled from
finished modules, easily programmed and may be used in the diverse fields. The disad-
vantages of the S7-200 controller preventing form its application to address the above
problems are as follows:
• insufficient processor capacity (it takes 0.37 I!S to perform the Boolean opera-
tion);
• math coprocessor is not available;
• low speed of available analog-signal input modules (::;250 I!s), thereby condi-
tioning the need of using a unique block of functional converters of effective
current and voltage values;
• lack of a single case.
 231

The most optimum option is to build the controller using assemblit:s and units of the
IBM compatible industrial computer operating in the DDC mode (Direct Digital
Control). It would be possible to completely exclude the application of the analog com-
ponent in the control circuit and simplify and reduce the cost of the system to be devel-
oped, while maintaining high output parameters.

2.2.1. Block Diagram o/Complex

The controller integrated into the complex is built based on a compact IBM compatible
industrial workstation (personal computer application is possible) with an LCD display
and a keyboard. The electrical block diagram of connecting the controller to the power
unit of the press facility is shown in Figure 2.

I •I 1
3 4
~ ;:
5

G=
•• • 6
7
¢::::::::~

~
8 JI 9

Figure 2. Block diagram of heating circuits of press facility.

1- high-pressure and high-temperature apparatus, 2 - current transformer, 3 - heating transformer,
4 - thyristor regulator, 5 - electric power unit of press facility, 6 - sintering KS·3 controller,
7 - system block, 9 - keyboard.

A predetermined sintering temperature is reached by controlling a thyristor voltage

regulator connected in the primary winding of the heating transformer, the secondary
winding of which is connected to the high-pressure apparatus. Data on voltage (V)
across the heater of the sintering zone are recorded directly from the high-pressure appa-
ratus, while data on current (J) in the heater are read from the current sensor, with the
current transformer or low-voltage shunt serving as the sensor. These signals (u, J), as
well as a pressure signal (Pa) from the press facility compression system are supplied to
the controller. A number of other signals from the press facility is also supplied to it. Af-
ter processing them, the controller applies a corresponding control action on the thyristor
232
regulator and supplies a number of auxiliary signals to the press facility power unit. The
controller may be also connected to the LAN using RS-485 interface, while up to 25
controllers are allowed to be connected on one line. The developed system allows for
controlling the force of the press unit individually for different high-pressure apparatuses
and operating their cooling system.

2.2.2. Control and Operating System

The controller structure is shown in Figure 3. As it may be seen, the controller is based
on a full-function board of the «all in on Single Board Computer» class. The processor
and memory are embedded on the board, integrated LCD/CRT display interfaces, RS-
485, panel for DiskOnChip, etc are available. One more critical component of the con-
troller is a multifunction analog/discrete signal input/output board. The board is provided
with the multi-channel 12-bit AID converter with conversion time of 8 microseconds,
D/A converter, timer and digital input/output channels.

RS-485
3
1+--.c===::iJ4
5

6
to electroscheme of press

Figure 3. Electrical block diagram of sintering KS-3 controller:

1 - display, 2 - system bus, 3 - single-board computer, 4 - keyboard,
S - multifunction 110 board, 6 - interface board

Of significant importance is also an unique interface board. This board serves for rat-
ing and galvanically isolating input signals, timing operation of the controller with that
of thyristor block and controlling it, and interfacing the controller with the press facility
electric power unit. Since the high-speed AID converter jointly with the corresponding
software support allows the controller to be used in the DDC (Direct Digital Control)
mode, this completely eliminates utilization of analog circuits in the control loop,
thereby significantly simplifying the controller's design and improves its performance.
 233
2.2.3. Software and Technical Characteristics

The controller is provided with the unique basic software (SW), the general structure of
which is shown in Figure 4. As it may be seen, the basic SW comprises as follows:
• main software module loading the sintering program, a number of controller set-
tings, browsing the database, administrating, etc.;
• sintering control program controlling the sintering process and helping the con-
troller interact with the press facility in the "cycle" operation mode;
• program for setting control parameters.
In addition, the basic SW comprises the database itself and setting file. Since the control-
ler operates in real-time mode with tough requirements to its speed, tht: controller SW is
written on C++, Assembler languages using a number of hardware interrupts.

Software Main
for setup of programme
parametelS of
control

Software
of optjNBbng
bysintering

Figure 4. General software structure.

The main technical characteristics of the controller obtained through its software-and
hardware interaction are as follows:
1. The number of sintering program sections set by an operator, not more than 24 (100).
2. Duration of each sintering program section, s 1-60
3. The maximum value of controlled parameter:
• power 10.0 kV*A with discreteness 0.01 kV*A;
• voltage 10.0 V with discreteness 0.01 V;
• current 5.0 kA; with discreteness 0.01 A.
4. The mean control error, %, not more than ± 0.25
5. The controller provides an on-line digital control (DDC - Direct Digital Control) by
adaptive PID (PI) law.
6. Discreteness of processing and output of a control action to a controlled element, s,
not more than 0.02
234

7. The controller digitally displays all sintering parameters (P, U, /, R, current sintering
time and real time) to be set and controlled and also displays a curve of the controlled
parameter and reaction volume resistance during the sintering process.
8. The controller is provided with a memory storing a detailed sintering data and also
general data and allowing for data off-line browsing. The detailed sintering data may
be recorded on a floppy disk for exporting them to special-purpose utility programs
(for example "Origin").
9. The controller is provided with RS-485 interface to operate in the local area network.
10. The controller is protected from the unauthorized SW start.
The time dependencies of electrical sample heating power and cell electrical resis-
tance shown in Figure 5 illustrate the high speed of the controller. It is apparent from the
Figure that a stepwise decrease in electrical resistance at 6.8 second into the cycle has
caused only a momentary increase in electric power that demonstrates high speed of the
system.
r-------,;:;:::======iT 40
35
[
l 30 m
~: [
~ 25 ~
, / 0!:6,"'76-:~"'''':-~''''84''''''''6,..
::-::,,'' ' '-:'' '.' ':-7::!,1 20 ~
nrr.<-) ;:

15 @
10 ~
2.

10 12 14 16

Figure 5, Dependence of reaction cell resistance and electric power on sintering time,

Figure 6 shows as an example the controller display screen while setting the sintering
parameters in the form of a complicated electric power dependence on time. The behav-
ior of this dependence is simultaneously displayed in the lower part of the screen.
 235

--
I1Il""" ,1 IlnmJo.n>jl 'IlI!OOJ
.twllH1Z

t,
" c:p ,
t~ It II p, .~
t, h 1\ f~,
liM
" lfI,
"ti
~,
Jl'h ~ ~1I"i1 ~, ~t
t IL2l! . \.! 1,$0 l7 t,~ t
l IlAg til 111 2:$0 1~ fj ,,~ It
a 3 IL6i 11 ill 2,2l! ~9 1 L~ 3
4 1 B.lIIi! lit ~ 2,311 ;ro 1) Lilli '$
5 5 1000· 13 t5. 2,311 H ;m 1,19 S
~ 6 J,ziiI M z 2:2l! 12 :JII ILoo 1) ua
m
7
II
1
e i:: u.
::>
<1
2,00
COO
t3
t1
:JII
6i
IL'l'Il'
IUlil

Figure 6. Display screen with set sintering parameters.

Results provided in Table I and Fig. 7 demonstrate capabilities of the developed

equipment. Table I shows some physical and mechanical properties of nanoceramics
made of different high-melting-point compounds sintered at high pressures. The charac-
teristics given below exceed reported ones for bulk ceramic nanomaterials produced by
other techniques. Fig. 7 shows the view of fractured surface of said samples. As it can be
seen, all of them have a dense nanocrystalline structure.

TABLE I. Physical and mechanical properties of nanoceramics.

Sintering Density, glcm] Grain size, Microhardness, GPa

Compound
temperature, 'C nm (Load 0.5 N)

TiN 1200 5.10 -60 31.4±10

TiNffiB, 1200 4.72 -200 34.2±1.2
AI,O] 1000 3.87 -200 320+...1.4
236

a b c

Figure 7. Fracture surfaces ofnanocrystalline materials based on TiN (a), TiNffiB, (b) and AI,O, (c)
produced by high-pressure sintering.

The developed computerized complex opens up fresh and broader opportunities for
controlling the structure of nanoceramics during high-pressure sintering process and
producing novel nanocrystalline materials with desired properties, inclusive of applica-
tion of "rate-controlled sintering" technology [12].

3. Conclusion

The computerized complex of the apparatus for high-pressure sintering of nanoceramics

based on high-melting-point compounds has been developed. It allows sintering to be
proceeded according to the desired program and PT-parameters to be quickly and accu-
rately controlled. The sintering data are stored in the internal database and may be ex-
ported. The developed system may also be integrated into the shop floor local area net-
work.

4. Acknowledgements

The authors would like to extend thanks to Professors R.A. Andrievsky and E. Benko,
Dr. T. Yamasaki and Mrs. V.P. Lesnikova for fruitful cooperation. The work has been
performed with the fmancial support provided by the NATO's Science for Peace Pro-
gramme (Grant No. 973529).

5. References

I. Andrievski, R.A., Urbanovich, V.S., Kobelev, N.P., and Kuchinski, V.M. (1995)
Structure, Density and Properties Evolution of Titanium Nitride U1trafine Powders
under High Pressures and High Temperatures, in A. Bellosi (ed.), Fourth Euro Ce-
ramics, Basic Sciences - Trends in Emerging Materials and Applications, Gruppo
Edit. Faenza, Printed in Italy 4, pp. 307-312.
2. Andrievski, R. A., Urbanovich, V. S., Kobelev, N.P., Torbov, V.I. (1997) Reports
of Russian Academy ofSciences 356, 39-41 (in Russian).
 237
3. Andrievski, R.A., Kalinnikov, G.V., and Urbanovich, V.S. (1997) Consolidation
and Evolution of Physical and Mechanical Properties of Nanocomposite Materials
Based on High-Melting Compounds, in S. Komarneni, J.C.Parker and H.J. Wollen-
berger (eds.), Nanophase and Nanocomposite Materials II, 457, MRS, Pittsburgh,
pp.413-418.
4. Urbanovich, V.S. (1998) Consolidation of nanocrystalline materials at high pres-
sures, in G.M. Chow and N.!. Noskova (eds.), Nanostructured Materials. Science
and Technology, Kluwer Academic Publishers, Dordrecht, pp. 405-424.
5. Andrievski, R. A., Urbanovich, V. S., Ogino, Y. and Yamasaki, T. (1999) Consoli-
dation processes in nanostructured high melting point compound-based materials,
in British ceramic proceeding No 60. Book 718, vol.J, 10M Communications Ltd,
Cambridge, pp. 389-390.
6. Urbanovich V.S. (200 I) Properties of nanocrystalline titanium nitride-based mate-
rials prepared by high-pressure sintering in M.!.Baraton and I.V.Uvarova (eds.),
Functional Gradient Materials and Surface Layers Prepared by Fine Particles
Technology, Kluwer Academic Publishers, Dordrecht, pp. 169-176.
7. Shkatulo G.G. (1979) Thermoregulator of sintering TS -3. Passport, technical de-
scription and operating instruction, Report IFTTP NANB (in Russian).
8. Shkatulo G.G. (1980) Automatic regulator of power of electroheating. Sertification
of inventory No 752299, MKI G 05F, Bull. of inventories, No 28 (in Russian).
9. Shkatulo G.G. (1984) Thermoregulator of sintering TS-4. Passport, technical de-
scription and operating instruction, Report IFTTP NANB (in Russian).
10. Shkatulo G.G. (1987) Sintering controller KS-1. Passport, technical description and
operating instruction. Report IFTTP NANB (in Russian).
11. Siemens Simatic (2001) A system manual on programmable S7-200 controllers.
The user's guide (in Russian).
12. Ragulya A.V. and Skorokhod V.V. (1997) Validity of rate-controlled sintering
method of consolidation of dense nanocrystalline materials, 14 Plansee Seminar, 2,
735-744.
C 60 FULLERENE POLYMERS

R. A. WOOD', M. H. LEWIS', M. R. LEES', S. M. BENNINGTON 2,

M. G. CAIN3 and N. KITAMURA4.

1. Centre for Advanced Materials, Department of Physics, University of

Warwick, Coventry, CV4 7AL, UK.
2. ISIS Facility, Rutherford Appleton Laboratory, Chilton, Dicicot,
Oxfordshire, OX}} OQX, UK.
3. National Physical Laboratory, Queens Road, Teddington, Middlesex,
TW}} OLW, UK.
4. Osaka National Research Institute, Ikeda, Osaka, 563-8577, Japan.

Abstract. TEM investigations were performed on high temperature and pressure C60
fullerene. A mixed phase of alternating rhombohedral lamellae and tetragonal domains
is formed. At higher temperatures 900-1173 K graphitic 'hard' carbon phases formed,
via an inter-polymer interface instability and/or shear mechanism. This phase exhibited
a hardness of 10-15 GPa and a high degree of elastic recovery under indentation. The
kinetics of graphitisation were slow and proceeded via cross-linked intermediates,
which influenced the hardness markedly, up to 33.5 GPa. In-situ investigations into the
PIT field of C60 were performed using energy dispersive X-Ray diffraction techniques.
The PIT history and pressure isotropy were investigated to confinn and optimise
conditions for 3-D polymer formation. Formation of 3-D polymers has been carried out
in-situ, however the reclaimed sample failed to live up to hardness expectations, due to
depolymerisation problems on pressure release. Isotropy in the pressurl! field promoted
formation and retention of the 3-D FCC structure. The existence of a series of
ferromagnetic rhombohedral C60 polymers has been realised. This property has been
shown to occur over a range of preparation temperatures at 9 GPa. The structure is
shown to be crystalline in nature containing whole undamaged Buckyballs. Formation
of radicals is most likely due to thermally activated shearing of th.! bridging bond
resulting in dangling bond formation.

1. Introduction

Fullerene C60 was originally discovered in the middle nineteen eighties [1] where initial
interest was high. This however tapered off, until Kratschmer et al [2] discovered a
facile and economically viable synthesis route. With the advent of affordable fullerenes,
research resulted in the discovery of I-D and 2-D polymers [3], synthesised as a
phototransformed thin film. The polymer formation is through a [2+2] cycIoaddition of
239
Y G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 239-248.
© 2003 Kluwer Academic Publishers.
240

two adjacent '66' bonds, producing a four membered si covalently bonded ring
bridging two C60 molecules [4]. Bulk samples of these polymers were realised by
exposing these fullerenes to high temperature and pressure (HTHP) treatments [5-8].
This produced many new phases, at higher temperatures disordered si carbon [9-13] is
formed and with increasing pressure it is reported to form 3-D polymeric structures
[14,15] and a second disordered phase containing mainly si carbons.
Overall the reaction diagram [13,16] is well known and agreed, with
discrepancies due to variations in the shear and pressure uniaxiality of apparatus. To
investigate this and cover the rest of the PIT field we continued our in-situ X-Ray
studies, in similar fashion to beforehand [17]. This paper also tries to definitively
confirm the formation of 3-D polymers [14,15,18] in-situ as well as quenched, using
variations in the PIT history to optimise synthesis.
The recent discovery of ferromagnetism in rhombohedral C60 polymers [19]
has opened up the possibility of a whole new family of magnetic fullerenes. The
formation of magnetic carbon species is not entirely new. Identifications have been
made of ferromagnetic and superconducting hysteresis loops in highly oriented
pyrolytic graphite [20]. Magnetic properties have been found in super high surface area
carbons with micro graphitic structures [21,22]. In a recent study, ferromagnetic carbon
powder has been used in the production of ferromagnetic carbon electrodes [23]. In this
case we focused on the 2-D polymers especially the rhombohedral phase near its
stability limit, which has a structure analogous to pyrolytic graphite, but with
semiconductor behaviour.

2. Experimental

The high-pressure and high-temperature treated fullerene samples were produced at the
Osaka National Research Institute. The initial sample was 99.9% purity C60 purchased
from the Materials and Electrochemical Research (MER) Corporation. The thermobaric
conditions were achieved using an octupole press. Pressure was applied, before
subsequent heating and holding for periods ranging from half a minute to an hour.
Hardness and elastic modulus measurements were taken using a depth sensing micro-
indentation system within a SEM. This is able to account for the elastic recovery
exhibited by the sample, which is not allowed for with traditional indenting devices. For
each sample ten separate indents were performed using a load of 1000 mN, except for
the softer 773 K preparations where a lower load of 250 mN was used. TEM
experiments were performed on a JEOL 2000FX and high-resolution work on a
PhillipslFEI Technai F20 field emission unit. The in-situ diffraction was conducted at
the Spring-8 X-Ray synchrotron in Japan. The High Temperature Research BL04Bl
beamline utilises a SPEED 1500 high-pressure isostatic press with tungsten carbide
anvils, which gave access to pressures up to 13 GPa. Magnetisation versus magnetic
field (M vs. H) loops were collected using an Oxford Instruments vibrating sample
magnetometer (VSM).
 241

3. Results and Discussion

3.1. 2-D POLYMERS

These phases are produced at temperatures ranging from 650 to 900 K and pressures up
to approximately 9 GPa. Novel TEM investigations of the 2-D polymer samples
prepared at 773 K, 3-9 GPa all exhibited regular bands of contrast, which ran in three
distinct directions (figure 1). Also, the concentration of these bands increases with
treatment pressure. By varying the orientation of the sample it was noticeable that only
two directions could be clearly visible at anyone time. The two visible bands intersect
at an angle of approximately 70° indicating that a transformation has occurred parallel
to the original (llik planes. The electron diffraction pattern from the lighter mottled area
was indexed as a body-centred tetragonal phase with a (OIOtr electron beam direction.
The diffraction aperture was then moved over the contrasting areas without a change in
sample orientation. The diffraction pattern taken from this area bore a similarity to the
previous pattern, but subtle distortions were noticeable. The pattern was subsequently
indexed as the rhombohedral polymeric phase, with a [2T.O]R beam direction. Therefore,
it can be inferred that rhombohedral lamellar domains have formed along planes parallel
to the original (Ulk planes. The formation of microscopic domains in different variants
of the rhombohedral orientation occurs to minimise internal stresses produced in the
sample due to the rhombohedral or tetragonal lattice deformations.
The specific P-T conditions provide the thermodynamic potential and drive for
the phase transformation, the kinetics may be influenced by the enhanced (Ulk crystal
shear with increased non-hydrostatic pressure. This shear motion could enable the C60
cages to align themselves in a position where the double bonds on adjacent molecules
are parallel, as discussed in detail in [9]. This would facilitate a [2+2] <:ycloaddition and
hence the formation of the 2-D polymer.

3.2. GRAPHITIC 'HARD' CARBON

All samples synthesised at temperatures of 900 K and above behaved similarly under
XRD investigation, indicating C60 cage collapse and the formation of graphitic carbon.
The TEM images (figure 2) exhibit bands of contrast with similar frequency and
orientation to the mixed polymeric phase. The electron diffraction pattern had three
diffuse (00.2) reflections from turbostratic graphite in three layer-plane orientations
corresponding to the precursor rhombohedral lamellae. Additionally, there are diffuse
diffraction rings from a more disordered graphitic structure between the layered
graphitic bands, agreeing with previous neutron diffraction experiments [12].
The mechanism of cage collapse could be by two possible routes. Firstly, the
samples undergo vast pressures with a considerable magnitUde of uniaxiality present.
The consequence of this being a high probability of shear occurring in the system,
which could break the intermolecular bonds, initiating cage collapse. The shear bands
produced by this mechanism would be parallel to the (Ulk, which agrees with the
position and orientation of the contrasting areas visible in the TEM image. Also this
provides the first explanation for the graphitic structure that is truly consistent with the
three-fold diffuse electron diffraction pattern. The novel TEM observations, of the
242
tetragonal/rhombohedral mixed polymeric phase, offer an alternative mechanism
associated with the inherent instability of the interfacial region. With an increase in
temperature the molecules could acquire enough energy for bond dissociation
instigating the subsequent cage collapse.

Figure!. TEM images \Jffue a) mixed polynwrlc phase b) the rhomoohedrallamdlac.

Figure 2. From left to right: Low magnification image across an original Coo grain boundary and
more detailed HR-TEM oftbe graphitic 'hard' carbon phase.
 243

3.3. HARDNESS

Standard Vickers indentations, of the graphitised hard carbon phase, have been reported
in the literature to give calculated hardness up to a value two thirds that of diamond.
This is surprisingly large for a graphitic or disordered carbon phase. However, the
method used to determine the hardness value is related to the magnitude of the residual
indent left. This method cannot take into account the process of elastic recovery. The
depth sensing indentation techniques used in this work enabled an assessment of the
true hardness. The hardness values for the graphitic phase (973 K and 1173 K, one
hour) are considerably less than expected, 10-12 GPa. The reason for the recovery is
due to the ability of the corrugated graphitic layers, which have a variable spacing to
compress together without shearing, then subsequently expand back to the original
structure.
To discover fullerene derivatives with more advantageous hardness properties
a sequence of samples was produced at 9 GPa, over a temperature rangl! of 700-1200 K,
with shortened treatment times of one minute. The lower temperature hardness values of
1.3 to 2.9 GPa (figure 3) are typical of the different states of the polymerised phase. As
the temperature increases, there is a rapid increase in hardness to an initial peak at
900 K of 25.8 GPa with a decrease to -15 GPa, at 1100 and 1200 K. This variation in
the hardness indicates that the formation of the graphitic phase proceeds via an
intermediate state, with time-dependent structure. Therefore the relatively high hardness
indicates that the intermediate state must contain a substantial quantity of remanent
cross-linking between the graphitic layers. Further investigations, were: performed, into
the effects of treatment time at 900 K, 9 GPa, with variations from one half to thirty
minutes. The sample hardness increases through times of a half, one, five and up to a
maximum of 33.5 GPa at twenty minutes. The time/temperature influence on cage
collapse and subsequent reduction in remanent cross-linking has recently been
confirmed during in-situ diffraction experiments within the P-T field [17].
35
.t=20 mins:

30 .1=5 min • 2.8

• 1=30 min.
25

20

15

10

1.8

'===:::::=---------------------1
700 750 800 850 Q50 1000 1050 1100 1150
1.6
1200
Temperature (K)

Figure 3. Hardness and density graph for samples treated at 700-1200 K, 9 GPa.
244

3.4. IN-SITU XRD

The history dependence of 3-0 polymer formation was investigated, with MgO pressure
medium, by an isotherm at 800 K, directly from a 2-0 phase. The XRD evolution is
shown in figure 4. With pressure application, we lose the (III) reflection and the (220)
and (311) split, indicating the formation of a 2-0 polymer. As pressure increases the
shoulder on the (220) peak merges with the main peak as the rhombohedral fraction
increases up to 8 GPa. Once we go beyond 8 GPa all the previously split peaks remerge
resulting in a broad peaked highly strained FCC system. This would indicate that the
2-0 layers are being 'zipped' together forming the 3-0 FCC polymer. The sample was
then quenched and the FCC structure remained. The pressure was reduced and the FCC
structure held firm until the final gigapascal was removed. Then the (220) and (311) peaks
suddenly split, reverting back to a mixed phase of 2/3-0 or solely 2-0 polymers.
Evidently the strain in FCC system was too great for the intermolecular bonds to hold in
the third dimension without the assistance of applied pressure or without partially
depolymerising into a less strained mixed phase. Raman investigations, with the
assistance of vibrational spectral assignment, showed retention of a limited quantity of
the 3-0 polymer and reversion of the majority of the sample to 2-0 polymers.

750K Isotherm
Quenched POO

2 3 4 5 6 7 8 9
d-Value CA)
Figure 4. The influence of pressure on the XRD trace for an isotherm performed at 750 K.
 245

Hardness investigations of the sample returned disappointingly low values.

Their micro-hardness and plasticity were both characteristic of soft 2-D polymerics.
Therefore domains of the 3-D polymer are enclosed in a soft 2-D phase matrix. So, even
if the hard 3-D regions were indented a low hardness would be found because the
micro-domains would get pushed into the soft matrix, thus masking their hardness.
Therefore, the synthesis product is a mixed phase 2-D polymer, containing FCC 3-D
polymer micro domains.

3.5. FERROMAGNETIC FULLERENE

A similar 9 GPa isobar of samples were studied in this set of magnetic susceptibility
measurements. The evolution of the magnetization (M) vs. magnetic field (H) loops
collected at lO K for the various C60 samples can be seen in Figure 5. It is noteable that
the first two samples in the temperature series show a co-existence of paramagnetic and
ferromagnetic character, with the 800 K sample showing by far the strongest magnetic
behaviour. This indicates that the formation of carbon centres with non-zero spins,
analogous to radical molecules, is happening at an increasing rate up to -800 K. As the
preparation temperature increases above 800 K the ferromagnetic character is reduced
and eventually becomes dominated by diamagnetic behaviour, indicating a considerable
loss of magnetic centres during graphitisation.
O.2r-~~~--~--~--r-~---r--~--.~~

Figure 5. M versus H loops of high PfT treated C", at 9 GPa over the range 700-1200 K. Inset is an
enlargement of the most ferromagnetic 800 K sample.
246
We can postulate on the processes and mechanisms for magnetism in these
polymers. The radical sites are initially being generated at temperatures of about 700 K
and are formed more readily up to and probably a little over 800 K. Thus, the formation
of the magnetic species seems to be temperature dependent. This temperature
dependence would indicate that the process occurring is of a bond breaking nature, but
whether this is taking place inter or intra molecularly is not yet clear. If we compare the
temperatures of ferromagnetic formation and C60 cage collapse however, it is evident
that the generation of magnetic species is occurring at temperatures approximately
100 K less than the onset of cage collapse. Therefore it can be inferred that the radical
centres are being formed in the polymeric state before collapse, without damage to the
Buckyballs. For this to be the case we would expect the structure of the 700 K and
800 K samples to be polymeric and crystalline in nature. The confirmation of this is
given by TEM Figure 6. Ordered periodic lattice fringes and atomic columns are
evident in the TEM micrograph accompanied by a rhombohedral crystalline diffraction
pattern indicating a [OO.3]Rh zone axis. This verifies beyond doubt that the magnetism is
being generated in the rhombohedral polymeric phase with no damage to the C60
molecule. Therefore, it can be concluded that the magnetic character is produced when
the thermal energy is great enough for carbon bond cleavage to occur in the
intermolecular bridge and hence form a radical species in the form of a 'dangling' bond.

Figure 6. TEM images of the ferromagnetic phase illustrating its crystalline structure.
 247

4. Conclusions

At synthesis temperatures up to 773 K and heating times of one hour, 2-0 alternating
mixed polymeric phases are produced. TEM studies showed that at all pressures a
mixed state is produced with the proportion of rhombohedral and tetragonal polymer
being controlled by the treatment pressure. The subsequent cage ,collapse during
graphitisation results in the formation of a graphitic 'hard' carbon phase, with mainly
si bonding in regularly warped layers. The pressure uniaxiality, hence shear in the
system and the inherent instability present at the interface of the polymer domains
provide a mechanism for graphitisation. The kinetics of graphitisation are slow (up to
Ihr) and intermediates, with whole C60 cages and varying degrees cross-linking, are
formed. The magnitude of cross-linking influences the properties, reaching its peak after
treating for twenty minutes, resulting in a hardness of 33.5 GPa. With the evolution of
treatment time cross-linking becomes less dominant and the interlayer spacing increases
and approaches the (00.2) value of graphite.
The C60 reaction diagram has now been covered by in-situ X-ray diffraction
experiments below pressures of 13 GPa. It was confirmed that the boundary between
2-0 and 3-0 polymers lies at 9 GPa and cage collapse to form graphitic 'hard' carbon
commences at about 1000 K. With a different history the 3-0 polymer was synthesised
in-situ by 'zipping' together the 2-0 polymers layers. However on quenching the
extreme strain in the system could not be accommodated and some domains 'unzipped'
themselves forming lower order polymers for stability. The reclaimed sample was a
mixed phase 2-0 polymer, containing FCC 3-0 polymer micro domains.
Finally, it has been demonstrated that at preparation pressures of 9 GPa it is
possible to form a whole family of magnetic carbons with varying degrees of
ferromagnetic content. This unusual material is generated by the coupling of near
neighbour carbon radical centres, which are formed by thermally activated cleaving of
the strained tetrahedral s/ bridging bond. This process has been found to occur in the
polycrystalline rhombohedral polymeric phase, with the process initiating at a
temperature of approximately 700 K and increases until graphitisation occurs. Further
studies could be performed, using smaller increments in synthesis temperature enabling
production as near to the graphitic phase boundary as possible.

5. References
[I] Kroto. H. w., Heath, J. R, O'Brien, S. c., Curl, R. F, Smalley, R. E. (1985)
C-60-Buckminsterfullere, Nature 318,162-163.
[2] Kratschmer, W, Lamb, L. D., Fostiropoulos, K., Huffinan, D. R. (1990)
Solid C-60-a new form of carbon, Nature 347, 354-358.
[3] Rao, A. M., Zhou, A. P., Wang, K-A., Hager, G. T, Holden, J. M, Wang, Y., Lee, W-T.,
Bi, X-X, Eklund, P. c., Cornett, D. S., Duncan, M. A, Amster, I. J. (1993) Photoinduced
polymerisation of solid C-60 films, Science 259, 955-957.
[4] Xu, C. H., Scuseria, G. E. (1995) Theoretical predictions for a 2-dimensional rhombohedral phase
of solid C-60, Phys. Rev. Lett. 74,274-277.
[5] Persson, P-A., Edlund, U., Jacobsson, P., Johnels, D., Soldatov, A., Sundqvist. B. (1996) NMR
and raman characterisation of pressure polymerised C-60, Chem. Phys. Lett. 2,58,540-546.
[6] Sundqvist, 8., Edlund, 0., Jacobsson, P., Johnels, D., Jun, 1., Launois, P., Moret, R., Persson, P-A.,
248

Soldatov, A, Wagberg, T. (1998) Structural and physical properties of pressure polymerised C-60,
Carbon 36, 657~60.
[7] Rao, A M., Eklund, P. c., Venkateswaran, U. D., Tucker, J., Duncan, M. A, Bendele, G. M.,
Stephens, P. W, Hodeau, J-L., Marques, L., Nunez-Regueiro, M., Bashkin, I. 0.,
Ponyatovsky, E. G., Morovsky, A P. (1997) Properties of C-60 polymerised under high pressure
and temperature, Appl. Phys. A 64, 231-239.
[8] Davydov, V. A, Kashevarova, L. S., Rakhmanina, AV., Agafonov, V., Allouchi, H., Ceolin, R.,
Dzyabchenko, A V., Senyavin, V. M., Szwarc, H. (1998) Tetragonal polymerised phase ofC~O,
Phys. Rev. B 58, 14786-14790.
[9] Marques, L., Hodeau, J-L., Nunez-Regueiro, M., Perroux, M. (1996) Pressure and temperature
diagram of pol ymeri sed fullerite, Phys. Rev. B 54, 12633-12636.
[10] Kozlov, M. E., Tokumoto, M., Yakushi, K. (1997) Spectroscopic characterisation of pressure
modified C-60, Appl. Phys. A 64, 241-245.
[11] Blank, V D., Denisov, D. N., Ivlev, AN., Mavrin, B. N., Serebryanaya, N. R., Dubitsky, G. A,
Sulyanov, S. N., Popov, M. Y., Lvova, N. A, Buga, S. G., Kremkova, G. N. (1998) Hard
disordered phases produced at high-pressure-high-temperature treatment of C-60,
Carbon 36, 1263-1267.
[12] Bennington, S. M., Kitamura, N., Cain, M. G., Lewis, M. H., Arai, M. (1999) The structure and
dynamics of hard carbon formed from C~ fullerene, Physico B 263-264, 632~35.
[13] Blank, V. D., Buga, S. G., Serebryanaya, N. R., Dubitsky, G. A, Popov, M. Y., Sundqvist, B.
(1998) High-pressure polymerised phases of C-60, Carbon 36, 319-343.
[14] Brazhkin, V. V., Lyapin, A G., Popova, S. V. (1996) Mechanism of three-dimensional
polymerisation offullerite C6Il at high pressures, JETP Lett. 64, 802-807.
[15] Brazhkin, V. V, Lyapin, A. G., Popova, S. V, Voloshin, R. N., Antonov, Yu. V., Lyapin, S. G.,
K1uev, Yu. A, Naletov, AM., Mel'nik, N. N., (1997) Metastable crystalline and amorphous
carbon phases obtained from fullerite C6Il by high-pressure-high-temperature treatment,
Phys. Rev. B 56,11465-11471.
[16] Wood, R. A, Lewis, M. H., West, G., Bennington, S. M., Cain, M. G., Kitamura, N. (2000)
Transmission electron microscopy, electron diffraction and hardness studies of high-pressure and
high-temperature treated Coo, J. Phys.: Cont/ens. Matter 12, 10411-10421.
[17] Bennington, S. M., Kitamura, N., Cain, M. G., Lewis, M. H., Wood, R. A, Fukumi, K.,
Funakoshi, K. (2000) In-Situ Diffraction Measurement of the Polymerisation ofC 60 at High
Temperatures and Pressures, J. Phys.: Cont/ens. Matter 12, L451-L456.
[18] Blank, V. D., Buga, S. G., Serebryanaya, N. R., Dubitsky, G. A., Mavrin, B. N., Popov, M. Yu.,
Bagramov, R. H., Prokhorov, V. H., Sulyanov, S. N., Kulnitskiy, B. A, Tatyanin, Yeo V. (1998)
Structures and physical properties of superhard and ultrahard 3D polymerized fullerites created
from solid Coo by high pressure high temperature treatment, Carbon 36, 665~70.
[19] Makarova, T. L., Sundqvist, B., Hohne, R., Esquinazi, P., Kopelevich, Y., Scharff, P.,
Davydov, V A., Kashevarova, L. S., Rakhmanina, A V (2001) Magnetic carbon,
Nature 413,716-718.
[20] Kopelevich, Y., Esquinazi, P., Torres, 1. H. S., Moehlecke, S. (2000) Ferromagnetic- and
superconducting-like behavior of graphite,.!. Low Temp. Phys. 119,691-702.
[21] Ishii, C., Matsumura, Y., Kaneko, K. (1995) Ferromagnetic behavior of superhigh surface-area
carbon, J. Phys. Chern. 99, 5743-5745.
[22] Ishii, c., Shindo, N., Kaneko, K. (1995) Random magnetism of superhigh surface-area carbon
having minute graphitic structures, Chern. Phys. Lett. 242, 196-201.
[23] Ueda, H. (2001) Preparation offerromagnetic carbon electrode from charcoal blocks,
J. Mats. Science 36, 5955-5960.
A TIGHT-BINDING MOLECULAR-DYNAMICS APPROACH TO
STRUCTURAL AND ELECTRONIC PROPERTIES OF a-SiC

V.I. IV ASHCHENKO, P. E. A. TURCHI', V.I. SHEVCHENKO

Institute of Problems of Material Science, NAS of Ukraine,
Krzhyzhanovsky Str. 3, 03142 Kyiv, Ukraine
'Lawrence Livermore National Laboratory (L-353), P.o. Box 808,
Livermore, CA 94551, USA

1. Introduction

There have been a few theoretical studies addressing the topological structure,
electronic distribution in amorphous silicon carbide (a-SiC) [1-6]. These studies have
involved molecular dynamics (MD) simulations in combination with an ab initio
pseudo-potential approach (PA) in the local density approximation (LDA) for exchange
and correlation interactions among valence electrons [1-3], Monte Carlo calculations
within the Tersoff empirical potential (TEP) formalism [4,5] and MD simulations based
on the TEP [6]. The densities of states (DOS), computed for small super-cells using the
PA, do not show a distinct semiconducting band gap (BG), though thf: DOS of the 54-
atom sample [1] has the distinct dip demonstrating the trend towards gap formation.
Electronic states in large sized a-SiC samples were carefully investigated in the
framework of a spV tight-binding (TB) scheme [6]. However, the lattf:r investigation is
inconsistent, since by generating a-SiC samples the scheme was not involved. The first-
principles PA makes it possible to obtain accurately the atomic distribution in a-SiC but
electronic states are computed incorrectly, since the LDA is known to underestimate the
BG [2,7]. Therefore, it is rewarding to study both the atomic and electronic structures of
a-SiC using another procedures capable to provide the appropriate atomic distribution
and to describe the electronic states in the BG region in the framework of the same
approach.
In our investigation, the Sp3S* TB scheme was used in computing the atomic and
electronic structures of a-SiC during MD simulations. We have not used the
conventional Sp3 -TB scheme since it is not able to describe the indirect band gap
correctly [7]. Namely, the indirect transitions from XI to ~I form the BG in c-Si and 3C-
SiC. Only the introduction of the additional s' state in the Sp3S' TB scheme makes it
possible to reproduce the band gap in accordance with experiment. Besides, in contrast
to the PA-MD scheme, TB-MD simulations enable one to generate the amorphous
samples containing up to several hundreds atoms. In our variant. of the TB-MD
approach, the parameters of the repulsion contributions not included in the TB
calculation were determined from first principles pseudo-potential calculations. Using
the suggested scheme, we have generated as-quenched and annealed a-SiC samples of
249
Y. G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 249-255.
© 2003 Kluwer Academic Publishers.
250

128 atoms. By the inspection of the DOS we have put the accent on the localization of
electronic states in the BO region caused by homo-nuclear bonds and other coordination
defects, including three-fold coordinated (T3 ), five-fold coordinated (Ts) and strongly
distorted four-fold coordinated (T4) atoms. The spV TB-MD results are compared with
the data of other authors.

2. Computational framework

We have used MD simulations based on the Sp3S* TB orthogonal interpolation scheme

[7] to calculate with great accuracy both the atomic distribution and DOS of a-SiC. In
OUf variant of the TB-MD scheme, a pair-wise potential that accounts for the ion-ion
repulsive interactions and the correction to the double counting of the electron-electron
interactions was considered as follows [8]

V(r) = VI exp[-(r-fO)/rD] + V 2 rolr, (1)

where r is the distance between neighboring atoms, fo is its equilibrium value in the ZnS
phase, and VJ, V 2 and fD are the parameters defined for Si-Si, Si-C, and C-C pairs.
The parameters of the potential (1) are determined from total energy calculations
carried out for c-Si, c-C and 3C-SiC in the rock-salt and diamond-like structures using a
first-principles PA. The conditions of pseudo-potential calculations very careful
described by Bockstedte et af [9]. Therefore, here we shortly describe the main
approximations of these calculations. We have employed norm-conserving pseudo-
potentials constructed according to the scheme of Hamann. The pseudo-potentials are
represented in the fully separable form. The Brillouin zone (BZ) integral over k-points
is replaced by a sum over a set of 10 special k-points according to the Monkhorst-Pack
scheme. The cut-off energy fOf the Kohn-Sham plane-wave basis set was taken at 35
Ry. The exchange-correlation energy was considered in the LDA. The results of the
total energy pseudo-potential calculations of crystalline silicon, diamond and silicon
carbide in the two structural modifications are shown in Fig. 1.
The parameters of the Sp3S* -TB scheme for 3C-SiC, c-Si and diamond were
carefully determined in Ref. [6]. Therefore, we only note here that our scheme gives BO
of 1.17, 2.4 and 5.5 eV for Si, 3C-SiC and diamond respectively, that are close to the
experimental values. The scaling of the two-center hopping parameters with inter-
atomic distance was chosen according to the Harrison rule [10].
The constant volume ensemble was considered during MD simulations. The
equations of motion of the atoms are integrated using a velocity-Verlet algorithm.
Periodic boundary conditions were applied to super-cell of 128 atoms with a density,
which was equal to the one of the 3C-SiC crystal. We began from the dilute vapor phase
representing the initial cubic cell with a lattice parameter that is 1.5 times larger than
that of the 3C-SiC crystal, and containing randomly distributed atoms far apart from
each other. The cell is allowed to slowly shrink under applied pressure at 2000 K during
about 2 ps. The fully equilibrated high-temperature structures were slowly quenched
with an average cooling rate of about 10 1 Kls down to 300 K. We have applied
different time steps (h) depending on the simulation stage. Its value decreased when
 251

approaching the final amorphous configuration. The h-values changed in the range of
about 10-14 _10- 15 s. The generated a-SiC samples were then averaged up to 2 ps to get
the thermal equilibrium properties. The same conditions were applied by generating an
annealed sample, in which case the as-quenched sample was heated to 1500 K and
equilibrated for about 0.6 ps.
The DOS of a-SiC required for TB-MD simulations was computed at the center of
the BZ, and the resulting DOS was constructed by using the eigenvalm:s at 35 k-points
of the irreducible wedge of the BZ.
Following Fedders et al [11], to clarify the extent of the localization of electronic
states, we define a "localized charge" q(n,E) associated with each eigenvalue with
energy E and the atomic site centered on the atom number n. This charge is obtained
from eigenvectors of the TB matrix. The quantity q(n,E) summed over all atoms is 1,
therefore it is not suitable for determining the localization of the electronic states with
the energy E. For this purpose a mean-square charge (MSC) is introduced [11]

(2)
where N is the number of atoms in the sample. From the definition of the MSQ, we can
evaluate the localization of a given state by a comparative way _

>",-,
_ -107.75 ",----'-
,-----, -151,0

~/
-129,5
~ -151,5
E
.8 ·107,80 Si C
..: -130,0
·152.0

"
0-
>,
-152,5
~ ~107,85 ·130,5

""

~
;.L:
-153,0
a;

"
I-- -107,90
-153,5

5,0 5,2 5,4 4,0 4,2 4,4 4,6 3,2 3,4 3,6 3,8
a (IO"nm) a (IO-'nm) a (IO" nm )
Figure /, Total energy as a function of the lattice parameter (a) of crystalline Si, SiC and C in the NaCl (upper
curves) and ZnS (lower curves) phases computed with a first principles pseudo-potential method,

3. Structural properties

In Fig. 2, we show the total and partial pair-correlation functions (PC F) of the as-
quenched and annealed a-SiC samples together with the PA-MD analogous functions
computed by Finocchi et at [1]. The structural characteristics of the samples are given in
Table 1. We find a rather good agreement between the TH- and PA-I;urves, although
there is a difference in the peak position of the C-C correlation. This discrepancy is due
to an overestimation of the diamond lattice parameter in our pseudo-potential
calculations. The position of the peak associated with the Si-Si, Si-C and C-C nearest
252
neighbor distances is located around the mean value of 2.34 A, 1.89 A, and 1.56 A,
respectively. From Fig. 2 (b) it is seen that the annealing influences the structure of the
Si network making it more similar to that of the PA sample.

(c) Si-C

(b) Si-Si (d) C-C

2 4 6
R (lO·'nm)

Figure 2. Total and partial pair correlation function (PCF) of the as-quenched (thin line), annealed (solid line),
MC-TEP 216 atom [4] (dashed line (a» and PA-MD 54 atom [1] (dashed line (b)-(c)) samples.

TABLE 1. Structural characteristics of a-SiC samples. Rs, is the cut-off distance of Si-Si interactions. In our
calculations Rc~ and Rs•.(: correspond to the first minimum in the C-C and Si-C PCF, respectively; n; -
percentage of i-fold coordinated atoms; N; - average coordination number of species i; N;_j - percentage of i-J
bonds in the samples. TB I, TB2 is the as-quenched samples prepared at different cooling rates. TB3 is the
annealed sample. PA is the 54 atom sample generated by Finocchi et af [I].

Sam21e RSi,nm n3.0/0 14,% ns, 0/0 Ns• Nc NSi-Si, % N Si_C ,% Nc..c,%
TBI 0.247 11.7 82.0 6.3 3.94 3.95 25.4 49.2 25.4
TB2 0.248 6.3 84.4 9.3 4.09 3.97 24.0 53.5 22.5
TB3 0.250 7.0 86.0 7.0 4.06 3.94 25.0 52.0 23.0
PA 0.250 17 77 6 3.93 3.85 23 53 24

By inspection of Fig. 2 and Table 1, we deduce that the final amorphous structure
weakly depends on the cooling rate. One can see that both annealing and slow cooling
lead to enhancing chemical ordering. Since the percentage of homo-nuclear bonds in all
our samples fluctuates around 50%, we conclude that the suggested TB-MD scheme
generates a chemically disordered tetrahedral a-SiC structure consistent with those
generated with the ab initio PA-MD method [1] (Table 1). Fig 2 (a) demonstrates that
the amorphous network of our samples differs from that of the sample obtained by
Kelires [4]. In the Kelires model of a-SiC, both the silicon and carbon atoms mostly
exhibit the Sp2 bonding. Apparently, the latter is explained by the specificity of the
empirical potential for carbon used in the TEP-MC simulation, which was determined
from graphite.
 253

4. Electronic structure

The total densities of states of the TB 1 and TB3 samples are shown in Fig. 3. Both the
ionicity gap (around -11 eV) and BG contain deep levels associated with homo-nuclear
bonds, coordination defects, and distortions of tetrahedral coordination. If homo-nuclear
bonds essentially modify the ionicity gap region, other defects fonn the valence and
conduction band tails and deep levels inside the BG [6]. The DOS of both samples
exhibit a trend toward band gap fonnation around 0.0 eV. The band gap dip in the DOS
of the annealed sample is wider than in that of the as-quenched one. A reduction of c-c
homo-nuclear bonds in the annealed sample as compared with the as-quenched one
leads to a narrowing of the low energy carbon-like s-band.

,
15 15
~ ."-;:
(5 10 10
5

E(eV) E (eV)

Figure 3. Densities of states (DOS) and mean-squre charge distribution in the TBI (left panel) and TB3 (right
panel) samples. The Fermi energy is 0.0 eV (TBI) and 0.1 eV (TB3). Arabic numerals in the plots are the
heights of the main peaks of Q(E) in the band gap region divided by the number of the atoms in the groups
(Qn). The coordination of atoms inside a group (T" T, and T,) is given in parenthesis.
254

Since the local DOS of the TEP samples were studied earlier [6] and the general
features of these DOS are close to those of the local spectra obtained in the present
investigation, we put the accent on the analysis of electronic localization. Fig 3 plots the
MSC, see Eq. 2, versus energy for both the as-quenched and annealed samples. We are
particularly interested in those localized eigenvalues whose energies are in the band
gap. Since in our case it is difficult to determine the BG unambiguously because of gap
states we considered a band gap ranging from -1.2 to 1.2 eV, as in the spectrum of the
perfect 3C-SiC crystal. In favor of this selection one can consider the fact that the
distinct dip in the DOS of the amorphous samples is observed just in this energy region.
It is necessary to distinguish the energetic and spatial localization of the charge. The
peaks of Q(E) fall on the localized eigenvalues. Combining the atoms in clusters in the
amorphous systems depending on their local environment, we can compare the extent of
localization of the electronic states in the selected clusters. The main characteristic of
any localized eigenvalue in a cluster is its MSC divided by the number of atoms in the
cluster, Q", where n is the type of cluster (T3, T4 , T5 and others). From Fig. 3 one can see
that the states in the gap are quite localized, and that localization increases when there
are fewer states in the BG. Gap states in the TBI spectrum are more localized than in
the TB3 one. Ifin the TBI, localized gap eigenvalues are located at the band tails, in the
TB3 these eigenvalues are concentrated at the midgap. Such distribution of localized
states points to the fact that annealing leads to a removal of the large local distortions
around separate atoms. We note that the carbon states are more localized in the low
energy region, where the separate C-C bond s-states are predominant consistent with the
fact that deep carbon s-states in diamond and 3C-SiC are the most localized states.
The charges Qn can be sorted out by descending order according to the n-fold
environment of a given species as T3Si - T3C -T4 Si - T4C - T5Si. Hence, the dangling-
bond states are more localized than the states belonging to other groups of defects.
Concerning T5- atoms we note that the extra floating bond leads to an unpaired electron
spin on the neighboring atoms. Our results indicate that this unpaired electron can be
somewhat delocalized, and is able to resonate between three or four neighbors (in the
TBI case), or can be localized, in which case the floating bond consists of a single
dominant site involving an unpaired electron spin with one secondary atom (in the TB3
case). In the midgap, T4 Si atoms give strongly localized peaks in the Q(E) dependence.
No appreciable contribution from T4C atoms in the midgap region was revealed.

5. Conclusions

We have investigated the atomic and electronic structure of a-SiC using the Sp3S' TB-
MD scheme. The computed characteristics are in rather good agreement with ab initio
MD results, which proves the efficiency and the accuracy of our approach. The
generated samples have been found chemically disordered. The annealing process leads
to an improvement of the amorphous structure, and to a reduction of the extent of
localization of the gap states. Carbon and silicon dangling-bond states are more
localized than the states of other defects. Over-coordinated atoms give rise to both
localized and diffuse gap states. Strongly distorted four-fold coordinated silicon atoms
 255

cause the formation of strongly localized states in the midgap, while analogous carbon
atoms do not make an appreciable contribution in this region.

6. Acknowledgement

This work was supported in part by STCU Contract No. 1591. The work of P. T. was
performed under the auspices of U. S. Department of Energy by the University of
California Lawrence Livermore National Laboratory under Contract No. W-7405-ENG-
48.

7. References

I. Finocchi, E, Galli, G., Parrinello, M. and Bertoni, e. M. (\ 992) Microscopic structure of

Amorphous Covalent Alloys Probed by ab initio molecular dynamics: SiC, Phys. Rev. Letter 68,
3044-3047.
2. Kelires, P.C and Denteneer, P.J.H. (1993) Theo!)' of electronic properties of amorphous silicon-
carbon alloys: effects of short-range disordered, Solid State Commun. 87, 851-855.
3. Kelires, P.e. and Denteneer, PJ.H. (1998) Total-energy and entropy considerations as a probe of
chemical order in amorphous silicon carbide, 1. Non-Cryst. Solids 231, 200-204.
4. Kelires, P.C (\ 992) Short-range order and energetics of disordered silicon-carbon alloys, Phys.
Rev. B 46, 10048-1006\.
5. Tersoff, J. (\994) Chemical order in amorphous silicon carbide, P/rys. Rev. B 49,16349-16352.
6. Ivashchenko, V.I. and Shevchenko, V.1. (2001) Effects of short-range disorder upon electronic
properties of a-SiC alloys, Appl. Surf Science 184, 137-143.
7. Vogl, P., Hjalmarson, H.P. and Dow, J.D. (\ 983) A semi-empirical tight-binding theoty of the
electronic structure of semiconductors, 1. Phys. Chem. Solids 44,365-378.
8. Molteni, C, Colombo, L. and Miglio, L. (1994) Structure and properties ofam.orphous gallium
arsenide by tight-binding molecular dynamics, P/rys. Rev. B SO, 43-71.
9. Bockstedte, M., Kley, A., Neugebauer, J. and Scheffler, M. (\ 997) Density-functional theo!)'
calculations for poly-atomic systems: electronic structure, static and elastic properties and ab initio
molecular dynamics, Computer Phys. Commun. 107, 187-222.
10. Harrison, W.A. (\ 980) Electronic Structure and Properties ofSolids, Freeman, San Francisco.
II. Fedders, P.A. and Drabold, D.A. (\ 993) Hydrogen and defects in first-principles molecular-
dynamics-modeled a-Si:H, P/rys. Rev. B 47, 13277-13282.
THEORY OF NANOTUBE NANODEVICES

SLAVA V. ROTKIN
Beckman Institute, UIUC, 405 N.Mathews, Urbana, IL 61801,
USA

Abstract. The paper reviews quantum and classical effects which arise in physics of
nanotube devices. Knowledge of nanotube electronic structure has been used for a calcu-
lation of quantum capacitance and quantum terms in van der Waals energy. Combining
analytical theory and quantum mechanical micromodels I worked out a description for
nanoelectromechanical devices, for example, electromechanical switch. The theory takes
into account van der Waals forces which show up at the nanoscale and result in appearing
of a principal limitation for scaling down NEMS structures. A model, which has been
derived for a nanotube device electrostatics, includes an atomistic polarizability of the
nanotube in a selfconsistent way. This calculation yields a charge density distribution for
given external fields and specific device parameters. On the basis of these main elements
of continual compact modeling: quantum mechanical description of the nanotube elec-
tronic structure, theory of the van der Waals forces, quantum capacitance and continuum
mechanics,~ a general theory of nanodevices is proposed.

Key words: nanotube, theory, nanodevice, NEMS, MEMS, van der Waals, quantum
mechanics, electronic structure, quantum capacitance

1. Introduction

Applied physics of carbon nanotubes is an emerging new area ofnanoscience

and nanotechnology. Applied physics modeling and device theory at the
nanoscale require special techniques that are in between what have been
used in solid state physics and methods applicable to molecular systems
(Aluru et.al., 2002). I will identify in this paper several approaches that are
known to work for devices on the base of carbon nanotubes.
Various carbon, nitride and chakogenide nanotubes (NTs) have been
demonstrated recently (Tenne, 2001; Dai, 2002). A success of synthet-
ic technology immediately resulted in a number of applications. A few
to name are: ultrasharp and wear resistant tips for Scanning Probe Mi-
croscopy (SPM), Atomic Force Microscopy (AFM) and Scanning Thnneling
Microscopy (STM); electron guns for FPD technology and other electron
emitters; chemical sensors and gas storage; modified NT-AFM tips which
257
Y c. Cogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 257~277.
© 2003 Kluwer Academic Publishers.
258

are chemically or bio-sensitive; nanotube electromechanical systems, nan-

otweezers and nanoswitches and, last but not least, nanotube nanoelectron-
ics. I do not pretend to cover here all aspects of device physics of nanotubes
or even any substantial part of it. Instead, I will focus on fundamentals of
theory of nanodevices and will discuss how the classical description should
match the quantum one. Several theoretical methods will be illuminated.
Only analytical results will be presented for the sake of clarity. For specific
technical characteristics of concrete devices I will refer a reader to original
papers.
Graphite-like Systems and Materials, such as nanotubes, fullerenes,
onions, Graphite Polyhedral Crystals, nanographites, and many organic
macromolecules, are well known to have valence band system generated
by pi and sigma valence electrons (Dresselhaus, 1996). The latter ones are
localized and, normally, contribute only to mechanical properties of the
graphitic material. In contrast, pi electrons are mobile, highly polarizable
and define transport, electrical and electromechanical properties.
Many of graphites exhibit similar behavior at the nanoscale: the tenden-
cy for nanoscrolling. In the paper (Rotkin, 2002a) we predicted theoretically
and demonstrated experimentally formation of scrolls of carbon sp2-lattices
with a characteristic dimension about several nanometers. This scroll size
is typical for single-wall nanotubes (SWNTs) and fullerenes, as well as for
nanoarches at the edge of natural three-dimensional (3D) graphite. This
size was obtained within a Continuum Energetics theory which includes
continuum elasticity, microscopic model of van der Waals interactions and
surface energy theory (Rotkin, 2001).

Figure 1. Bandstructure of an armchair SWNT [10,10) (Right) and a zigzag SWNT

[17,0) (Left) within first Brillouin zone. Energy is in units of, ~ 2.5 eV, hopping integral.

While the sigma electrons form similar bond lattice in all graphite
like (sp2) substances, which reflects in a close similarity in their mor-
phology as discussed above, a versatility of their electronic properties is
due to sensitive electronic structure generated by pi electrons. The elec-
tronic structure of a monolayer of graphite (graphene) has a few (six)
 259

Fermi-points where an empty conduction band merges with an occupied

valence band. The graphene is a semi-metal (in planar morphology). The
pure fullerenes are mainly insulators (perfect spherical morphology). Car-
bon onions, nanohorns and other imperfect clusters are mainly conduct-
ing. The nanotubes may be either metals, or semi-metals (narrow gap
semiconductors), or insulators (wide gap semiconductors).
How does an electronic spectrum of a folded SWNT relate to the spec-
trum of a bare planar graphene? A simple but correct picture of the SWNT
bandstructure follows from a band folding argument: we imply extra quan-
tization for one of components of 2D wavevector of the electron in the layer
of graphene. This additional space quantization for the pi electrons appears
due to confinement in circumferential direction of the SWNT. As a result,
2D surface of the electron energy as a function of 2D wavevector is broken
into a number of ID curves: the nanotube electron subbands. The band
folding can be thought of as a mere cross sectioning of the bandstructure
of graphene along the nanotube symmetry direction (Fig.I). Depending
on the symmetry of the tube, three different situations can be realized:
(A) the armchair SWNT has a cross section passing through the Fermi-
point. In this case the SWNT is metallic and the conduction band merges
with the valence band. (B) The zigzag/chiral nanotube cross section is
distant from the Fermi-point. This tube has a nonzero gap and it is a
semiconductor tube. (C) One-third of zigzag and chiral nanotubes have a
very small gap, which follows from arguments other! than simple band-
folding. In our simplified picture these SWNTs have a zero band gap and
are semi-metals.
The bandstructure of the SWNT is highly sensitive to external fields.
Lattice distortions may cause changes in the bandstructure as well. A
proper lattice distortion moves the Fermi- point of graphite and results
in closing/opening of an energy gap, changing the electron density, and
charging the tube. Same is true for an external transverse electric field
or magnetic field. An impurity sitting on the nanotube or even placed
closely at a substrate surface may have similar action. These phenomena
open many possibilities for engineering of NT electronic bandstructure and
for application of nanotubes as nano-biosensors, mesoscopic devices and
nano-electromechanical systems.

1 Which is mixing of sigma and pi electrons at a finite curvature. This phenomenon is

more pronounced for small radius NT, R < 4A.
260
2. Modeling of Nanoscale Electromechanical Systems

Nano-electromechanical systems (NEMS) become an essential part of mod-

ern science and technology (Craighead, 2000). A number of applications is
already known: nanomanipulation, nanosensors, medical devices, nanofiu-
idic devices, to name a few. Even more applications are anticipated as a
result of the technological progress in this field.
I address in this section one of issues arising when one tries to under-
stand phenomena happened at the nanoscale with theoretical tools bor-
rowed from macroscopic physics. The latter has to reach its limits and new
micromodels are required for a quantitative description of a nano-device.
The object of study is a nano-electromechanical switch. Important changes
in its operation at the nanoscale are due to van der Waals forces. These
forces will change parameters, describing the equation of state of a NEMS.
All derivation will be performed analytically, which allows one to apply this
theory to a broad class of devices.

2.1. ANALYTICAL MODEL

I start here with a calculation of pull-in2 parameters of a general NEM

system which is an elastic media (elastic manifold) subjected to external
forces. The forces are changing during the NEMS operation and define a
dynamic shape of the NEMS. The specific forces, considered below, are (i)
the van der Waals force, (ii) the electrostatic force, and (iii) the elastic
force, which is able to restore the initial equilibrium shape of the NEMS.
In an earlier paper (Dequesnes, 2002a) an analytical derivation as well
as a numerical computation of the pull-in voltage, Va' have been presented
with account for the vdW correction. The other pull-in parameter, the
pull-in gap, X o , was treated as an independent quantity and taken from a
solution of a classical MEMS problem. Below I will extend the result of the
paper (Dequesnes, 2002a) and give an accurate derivation for both pUll-in
parameters.
The equilibrium dynamic shape of the NEMS satisfies the force balance
condition (A): the first derivative of the total NEMS energy is equal to zero.
In general, one has to calculate the energy gradient at every point of the
manifold and equate it to zero locally. This yields the equilibrium shape of
the system at given external forces applied to the NEMS. Main approxima-
tion, which allows one to obtain an analytical solution of the problem, is to

2 The pull-in is a phenomenon of loosing of NEMS stability at a certain (pull-in)

voltage. Then infinitely small increase of the voltage results in a sudden collapse of a
movable part of NEMS onto a ground plane.
 261

consider only one mechanical degree of freedom 3 . This approximation gives

an answer for the pull-in, which is correct up to a geometry dependent
numerical factor. The numerical factor is not altered by changing force
fields (e.g., by changing van der Waals to Casimir force) and has to be
calculated only once for a given geometry or used as a fitting parameter.
Within this one parameter model, I write the first equation of state as:
~~ = 0 where x is the single degree of freedom of a NEMS, for example,
the gap between the elastic movable part of the NEMS and the surface of
the ground plane, and E is the total energy given by E(x, <p) = T(x, hi k)-
V(x, <Pi C) - W(Xi E, 0:). Here, the energy depends on x and <p, the gap and
the voltage, which are the parameters governing the instability point. Three
energy components are the elastic strain energy, T, the electrostatic energy,
V, and the van der Waals energy (vdWE), W. All three terms depend on the
system geometry, shape, etc., as well as on material parameters (elasticity,
capacitance, etc.). The first type of dependence is expressed in terms of the
dynamic gap, 0 ::; x ::; hi in terms of the maximum separation between the
mobile and stationary parts of the NEMS (initial gap), hi and the voltage,
<po The material properties are collected in four constants: an elastic stiffness
with respect to the gap, k, a general capacitance with respect to the voltage,
C, and general vdW coefficients, 0: and E. The elastic energy component
reads as T = k(h - x)2 /2. The electrostatic energy term is V = C<p2/2,
where the capacitance has to be calculated with a specific micromodel, for
example, the model proposed in Ref.(Bulashevich, 2002).
The vdWE energy component is often approximated by a single attrac-
tion term (Dequesnes, 2002a; Barash, 1988), which depends on a distance
between interacting surfaces. Integrating out all system geometry (Barash,
1988; Girifalco, 2000), one obtains a simple dependence of the vdWE on
the gap: W ~ EX-a, where an exponent 0: defines the specific power law for
the specific dispersion force. For example, for the pure van der Waals inter-
action between small objects (atoms) 0: = 6, for the retarded Casimir force
between atoms 0: = 7, it can be fractional for the many-body terms in low
dimensional systems (Rotkin, 2002b; Barash, 1988). With these definitions
for the material constants I write the total energy of the NEMS as:

C(X)<p2
2 - W(x; E, 0:) (1)

where the dependence of the capacitance on the gap, C = C(x), has to be

defined separately.
To find an instability point of the NEMS I write the second pull-in
condition (B): the second derivative of the expression (1) must equal zero.

3 Which may be thought as a fundamental mechanical mode of a specific NEMS.

262
2.2. GENERAL EQUATIONS FOR THE PULL-IN

In what follows logarithmic derivatives of the energy components will be

used. When the dependence of the energy components on the gap is given by
a power law, the logarithmic derivatives are simply constants depending on
the material properties and the geometry of the NEMS: (31 = -xo log C /ox,
0:1 = -xOlogW/ox, (32 = x2(0l0gC/ox)2 + x 20 2 logC/ox 2 and 0:2 =
x 2 (Olog W/ox)2 + x 2 0 2 log W/ox 2 .
In terms of 0: and (3 (which are just numbers by this assumption) the
general physical 4 solution satisfying the pull-in conditions (A) and (B) is
as follows:

(2)

So far, this expression is still implicit because R.H.S. of the first equation
depends on the amount of vdWE at the pull-in gap, W(x o ). However,
the vdWE component is normally small at large distances and I propose
to substitute the bare value for the pull-in gap xo(O) = h /3!1rh into the
R.H.S. of the equations (2). This is allowed for large h because expanding
the expression in series in W, one gets the difference of this approximation
and an exact result only in the second order of W/kh 2 « 1. In the opposite
limit an explicit solution of the first of Eqs.(2) must be substituted in the
second one.

2.3. ROLE OF VAN DER WAALS ENERGY

I present here a specific case of the general equation of state (2) when the
electrostatic force can be described via a planar capacitor model: C = co/x
and the vdW contribution can be written as W = € / ha . For completeness,
I give here all logarithmic derivative coefficients: (31 = 1, (32 = 2, 0:1 = 0:

4 Bogus roots of this system of non-linear equations have to be discarded basing on

the physical meaning of the solution.

 263

1 2 3 4
h,nm

Figure 2. The pull-in gap as a function of the initial device gap. Red (solid) curve
represents the selfconsistent result. Green (dash~dotted) curve shows the first~order van
der Waals correction. Lower dashed line gives classical MEMS result.

and a2 = a(a + 1). Substituting these values into Eq.(2), I obtain:

.----------------,

1 + 3a(a - 1) w,.~o) )
)

(3)
v; _ 2Vkh 1 3 ( W(xo) 1~.
---aa+2)--+- )W(xo)
1+3a(a-1--.
0- 3JC(x o ) 2 2 kh 2 2 ' kh 2

These equations may be further simplified for the small vdW forces:
keeping only leading terms in W, I obtain:

(4)
.Jk- (2-h )3/2 ( 1- -
V; ~ - 9aW(xo)
- - + o(Wjkh 2))
o Fa 3 4 kh 2 .

The role of the vdW correction is to decrease the pull-~in voltage via in-
creasing the pull-in gap: the electrostatic term of the Eq.(1) becomes larger
because the vdW force brings the movable electrode closer to the ground
plane.
Numerical selfconsistent solutions of the equations (:3) are presented in
Fig.2 and Fig.3. The classical result (MEMS limit) is shown as a dashed line,
a dash-dotted line represents a result of first-order perturbation theory,
while solid line represents the selfconsistent solution for the pull-in gap,
264

10

Vo,V

---
--- --- 2
4

h,nm
2.5 3 3.5 4

Figure 3. The pull-in voltage as a function of the initial gap. Blue (solid) curve represents
the selfconsistent result as explained in the text. Red (dash-dotted) curve shows the
dependence in the first-order in van der Waals perturbation. Dashed line shows the
classical MEMS result.

x o, and the pull-in voltage, Vo. The material parameters taken for the
numerical estimate are k = c/nm 6 , Co = (2/3)VkV/nm. Typical value for
the c is similar for almost any solid substance and is about several eV A6 .
In contrast to the classical result, the pull-in voltage as a function of
the initial separation, decreases to zero at h = he (Fig.3). This is a critical
size of a smallest possible nano-electromechanical switch as discussed in
Refs. (Dequesnes, 2002a; Rotkin, 2002c).

CONCLUSIONS AND DISCUSSION

In summary, I presented in this section an analytical theory for simulation of

an electromechanical system. Using continuum model with a single mechan-
ical degree of freedom, I demonstrated the role of van der Waals forces for
nanoscale devices. A general equation of state and a closed form of solution
for pull-in parameters are derived for a planar capacitor NEMS. For the
NEMS operating at small gaps, the vdW correction is written explicitly. It
is discussed how the vdW interaction may restrict applicability of a classical
MEMS theory at the distances close to a vdW critical length. This length,
derived analytically, gives a principal physical limit for NEMS fabrication.
The theory presented in this section allows one to calculate the critical
gap as a function of material properties of the nanoswitch (to be found
elsewhere).

3. Van der Waals Energy for ID Systems

In this section I will address theory of the van der Waals interaction in
nanotube systems. The van der Waals terms were shown to be extremely
important for the NEMS operation in the last section.
 265

The van der Waals interactions have been studied over a considerable
period of time. Starting with the phenomenological work of van der Waals
(van der Waals, 1873), our understanding has developed from classical
models (Reinganum, 1912) to quantum mechanics (London, 1930) and to
full statistical quantum electrodynamics (Dzyaloshinskii, 1961). The earlier
semi-empirical approach is still considered accurate and adequate for de-
scription of many phenomena and it involves transparent physics even for
very complex systems.
I model the vdW cohesion5 following the dielectric function approach.
This method was shown to be useful for various solids (Dzyaloshinskii,
1961) and gives a simplest correction within many-body approximation and
beyond the 6-12 Lennard-Johns (LJ) potential (Lennard-Jones, 1930).

3.1. CALCULATION OF COLLECTIVE MODES

The theory starts with the calculation of the dielectric function of a single
SWNT in RPA neglecting all modes except collective plasmon modes that
have most of the oscillator strength. These modes contribute the major in-
put to the total vdWE. For standard semiconductors the dielectric function
in a high frequency limit reads as:

w2
c:(w) ~ 1 - --%'
w
(5)

where wp stays for a characteristic plasmon frequency of a material.

Instead of writing the NT dielectric function (refer to papers (Louie,
1995; Li, 2002)) and obtaining its high-frequency limit, I derive the answer
from equations of motion of a charge on a cylinder surface:

(6)

where n = 16/3V3b2 is the surface electron density in the graphene, m, e

are the electron mass and charge, <p is an acting potential on the surface
of the SWNT, which includes an induced potential of all charges on the
surface, a is the fluctuation of the charge density related to the plasmon
mode, j is its current, V' is 2D gradient operator along the surface.

5 Repulsion is due to the Pauli principle and will not be addressed here.
266

Figure 4. The ID systems considered in the section. A. Double-shell nanotube. B.

Nanotube on a substrate. C. Two nanotubes.

To close the set of equations I write the surface boundary condition for
the fluctuation density O"kJ1 using Gauss-Ostrogradskii theorem:
1
47rO"kJ1 = 'PkJ1 RKJ1 (kR)IJ1 (kR) , (7)

where KJ1(kR), IJ1(kR) are Bessel functions of imaginary argument and

of order jL taken at the SWNT radius R. k is the co-axial momentum
of the mode, jL is the circumferential quantum number of the mode. The
plasmon mode frequencies are given by the 1D series for each of allowed
jL = 0, .. N at -1, where Nat is the number of atoms along the circumference
in one real space unit cell of a SWNT. For example, for [10,10] armchair
nanotube it equals 20. The SWNT plasmon dispersion law is:

(8)

Here I use the notation wp = J 47rne 2 / mR for the 2D plasma frequency,

where a characteristic length scale is given by the SWNT radius, R. This
frequency sets the energy scale for the collective modes of the tube and,
hence, for the vdWE.

3.2. VAN DER WAALS COHESION ENERGY

The vdWE is given by the difference in the total energy of the modes of the
electromagnetic vacuum due to the interacting systems approaching each
other. The simplest way to calculate it is to find the collective mode fre-
quencies as a function of the distance between the systems. The frequencies
depend on the distance because of the Coulomb interaction mixes the modes
that are independent for infinite separation. In the zero temperature limit
(which is appropriate for distances smaller than 1000 nm), only the zero-
point oscillation term survives. The van der Waals force is the derivative
of the zero-point oscillation energy with respect to the distance which is
implicitly included in the plasmon frequency as it will be shown below.
 267
3.2.1. Intertube van der Waals attraction
In this section I calculate the attraction energy between two parallel nan-
otubes (Fig.4C). I derive the plasmon frequencies from the quantum me-
chanical Lagrangian within second quantization formalism. The Lagrangian
for a single shell is:

Ll -_ -1 L (w2
-2-
2 k,1-' wkl-'
- 1) O'kl-''Pkl-'
t + h.c., (9)

where I use Heisenberg operators of the classical potential, 'P, and the
classical charge density fluctuation, 0'. This form of the continuum La-
grangian is consistent with the equations of motion as given in Eqs. (6).
The generalization of the Lagrangian for the case of two tubes (two shells)
is straightforward:

L = ~ Lk,~ (-:;2~:(1)
kl'
-1) aL(l)'Pk~(l) + ~ K,ML (wKMw (2) -1') akM(2)'PKM(2)-
2
2

(10)
~ L V(l - 2) (O'L(1)'PKM(2) + O'kM (2)'Pkl-' (1) ) + h.c.
k,I-';K,M
where first two terms are the free Lagrangians of the plasmon subsys-
tems and last term represents the interaction between subsystems (1) and
(2), which I treat perturbatively. The matrix elements, that represent the
Coulomb interaction between the tubes in the continuum limit, are the com-
binations of the Bessel functions (compare with the exponents in the planar
problem (Chaplik, 1971) and the Legendre polynomials in the spherical
problem (Rotkin, 1996)). In order to proceed further analytically I restrict
the expression for the Coulomb matrix to the components with k = K
and J.L = M. This approximation has the same accuracy as before because
the dimensionless ratio of the interaction to the plasmon energy is small in
our problem. One has to keep only the first order of this parameter in the
secular equation for the plasmon mixing:

w2 ) ( w2 KI-'(kR2)II-'(kR2)) JI-'(kR2)K21-' (kD)

(
- wL(1) +1 - w~I-'(2) + KI-'(kR1 )II-'(kR1 ) = =t- KI-'(kRt}
(11)
here D is the distance between the tube centers and Rl and R2 are the NT
radiuses. The shifts of bare plasmon energy levels decrease with k and J.L as
well as the Coulomb matrix element itself (given by the RHS ofEq.(11)): in
the limit of kD » 1 it approximately equals 1/..j27r kD exp (-k(D - 2R)).
The specific van der Waals energy per atom is the sum of the plasmon
zero point oscillation energies divided by the total number of modes. The
268

integral over the first Brillouin zone of the SWNT has to be substituted for
the sum. I have considered the interaction between two tubes of the same
radius rv 71. The vdWE derived from the zero-point oscillation of those
mixed modes decays with D as D- 4 .5 (presented in Fig.5). In contrast to
one-body models this vdWE has a fractional exponent which reflects the
specific square root dispersion law of the bare plasmon frequency as given
by Eq.(8).

3.2.2. Cohesion to metal substrate

The solution obtained for the tube-tube interaction can be readily used to
yield the cohesion of the SWNT to the metallic substrate (Fig.4B) because
the electric field distribution near a flat conductor is given by an image
charge of the opposite sign. Hence, one can choose the odd solution (with
the minus sign) of the secular equation (11) while the even solution has
to be discarded. The same integration over the Brillouin zone gives now
a slower decay of the vdWE. It is because only one subsystem (SWNT)
possesses the ID plasmon modes in this case. The exponent is -3.5 as
compared to -4.5 for two SWNTs (Fig.5).

logW, eV
0.01

0.001

1 10-6

Figure 5. The calculated vdWE: (Upper) for nanotube on a substrate; (Lower) for two
identical nanotubes of R = 7A.

3.2.3. Inter-wall cohesion in DWNT

The energy of the inter-wall attraction in a multi wall nanotube seemed to
be an immeasurable parameter before the discovery of the peapod struc-
tures, C 6o @SWNT, and their transformation in double wall nanotubes
(Fig.4A). The process of creation of the second wall inside the initial nan-
otube is favorable owing to the inter-wall cohesion energy gain.
 269

The Coulomb interaction between shells is given by continuum electro-

statics in the same way as before. Because of the axial symmetry of the
problem only modes with the same quantum numbers (p = M, k = K) are
mixed and the interacting plasmon Lagrangian (10) is exactly diagonal in
Fourier space. The secular equation is as follows:

The plasmon frequencies depend on the radii of both shells. I plot here
the vdW cohesion energy as a function of the intershell distance (Fig.6).
This parameter can vary and can be measured experimentally(Iijima, 2001).

log W, eV
.. ......
1~:1'
11
I
0.1 t
0.01 !

~
0.0011
LL____ . ~_,_ ..

0.1 1 10 Ilog 0, nm
Figure 6. The dependence of the vdWE for a double wall nanotube on the distance
between shells, 8 (dots); and the linear fit (solid line) with the exponent 2.9,

CONCLUSIONS AND DISCUSSION

In this section a continuum theory has been presented which gives a fast and
accurate qualitative estimation of a many-body contribution to dispersion
attractive forces for ID tubular systems made from layered materials. I
have used a formalism of a dielectric function and have assumed that main
term in the many-body van der Waals cohesion is due to collective modes
(plasmons). The plasmon frequencies are explicitly calculated. As a result
of the plasmon mixing by the Coulomb interaction, the total system energy
is lowered by the van der Waals contribution, A distance dependence of
the new (many-body correction) term has a fractional exponent, 5/2 for
tube-metal cohesion and 7/2 for tube-tube interaction, unlike an one-body
270
energy given by LJ 6-12 potential. It was known that a direct summation of
the 6-12 atom-atom interactions for the carbon nanotubes gives exponents
of 3 and 4 for tube-substrate and intertube cohesion, respectively.
Our approach is almost independent on the structure of interacting
lattices, which is in contrast to the one-body LJ potential. It can be easily
applied for the cohesion of the tube to not graphitic substrate. This vdWE
contribution is especially important for the description of recently studied
friction properties of multi wall nanotubes and nanotubes on the graphite
(Yu, 2000; Cumings, 2000; Falvo, 2000).

4. Atomistic Electrostatics for Nanotube Devices

Now I switch to how the quantum properties of the nanotubes may reflect
in the macroscopic behavior of the nanotube based device. I have already
shown in the first section that the capacitance of the nanotube to the
backgate is one of important parameters in the modeling of the nanoelec-
tromechanical switch. Recent success in a fabrication of nanotube based
nanoelectromechanical devices (Akita, 2001; McEuen, 2002) confirms that
this theoretical research is very topical. Here I discuss how the classical
meaning of the electrostatic capacitance changes for a nanoscale system.
The transport properties of a SWNT device are also determined at some
extent by charge distribution along the nanotube channel (Rotkin, 2002d).
This charge density can be easily calculated on the base of the theory for
the quantum capacitance of the nanotube (Bulashevich, 2002).

4.1. FUNDAMENTALS OF SWNT ELECTROSTATICS

4.1.1. Classical vs. Quantum Modeling
The selfconsistent calculation of the equilibrium charge density for the
SWNT with a moderate mechanical deformation has been required to sup-
port a recent modeling of nanotube electromechanical systems (Dequesnes,
2002a; Dequesnes, 2002b; Rotkin, 2002c). Knowledge of the induced charge
allows us calculating an electrostatic energy of the system, which can be
rewritten in terms of a distributed capacitance. I demonstrated that this
atomistic capacitance has two contributions: purely geometrical term and
another one, specific for the nanotube. It is very natural to call the second
term "a quantum capacitance" as a similar definition was proposed for a
two-dimensional electron gas system in Ref. (Luryi, 1988).
It was found that a statistical description (similar to what was used
in Ref.(Odintsov, 2000)) is valid and gives a fairly good estimate for the
charge density as compared to the quantum mechanics (see also SecA.2.1).
The applicability of the macroscopic electrostatics modeling to an equilib-
rium charge distribution has been already discussed in Refs. (Bulashevich,
 271

Figure 7. Geometry of two single-wall nanotube devices studied in the section: (A)
cantilever NEMS, and (B) string NEMS.

2002; Rotkin, 2002e). Same arguments may hold for a system shifted from
equilibrium slightly, e.g., for description of a current carrying device (which
can be found elsewhere (Rotkin, 2002d)). The reason for this" classical"
behavior of so small quantum object is two-fold. Firstly, the selfconsistent
electrostatic energy level shift is the same for every subband of the SWNT in
a first approximation (Petrov, 2002). From the other side, the SWNT has a
very high depolarization factor as it was noted by Louie (Louie, 1995). The
higher the polarization of a molecular system, the closer is the selfconsistent
polarizability to its classical limit which is R 2 /2 for the SWNT (Li, 2002),
and R3 for a spherical fullerene (Rotkin, 1994). The polarization of very
mobile pi electrons in nanotubes is very high, hence, one is allowed to
apply a "classical" theory, taking into account the quantum mechanical
modification of the electronic structure perturbatively.
I studied a SWNT device in a "string" geometry: it comprises the
straight nanotube which is fixed (suspended without a slack) between two
metal side electrodes over a backgate electrode (Fig. 7B). The side electrodes
are kept at the same potential with respect to the backgate. This design is
standard for electromechanical systems, and the first experimental realiza-
tion of a single SWNT NEMS appeared recently (McEuen, 2002). I used
this and cantilever geometry (Fig. 7A) for theoretical study of the nanotube
electromechanical switch in Ref.(Dequesnes, 2002a).

4.1.2. Principles of Compact Model for Nanotube Devices

The continual compact modeling of the nanotube device bases on three
elements: (i) local statistical description of the charge density, (ii) perturba-
tion theory for changes in the charge density due to nanotube deformations
and(or) external fields, and (iii) external screening which results in a short
range Coulomb potential and allows obtaining analytical expressions. Hy-
pothesis (i) has been proved by comparing the result of the quantum
272

mechanics with the selfconsistent solution of the Poisson-Boltzmann equa-

tions (see Sec.4.2.1). Second supposition is valid until the deformation
or(and) axial component of the external field is not too large, which is true
for SWNT applications in nanoelectronics but may not fulfill for nano-
actuators. The change of the NT density of states with applied external
transverse field is studied in Ref. (Li, 2002). It was demonstrated that for
the nanotube, which is a one dimensional nanoscale system, it is not possible
to separate pure material properties and effects due to geometry/design
of the device. The electronic properties of nanotube-in-device differ from
what one obtains for a free tube in vacuum.

DOS,
arb.un.

E, arb.un.

Figure 8. Density of states of a metallic SWNT near the Fermi level: shaded area
represents an extra charge induced in the SWNT by shifting the Fermi level away from
a charge neutrality level.

4.1.3. Calculation of Atomistic Capacitance

In order to calculate the charge distribution of the straight SWNT as a
function of the total acting potential I represent the latter as a sum of the
external and induced potentials:

'Pact = 'Pxt + 'Pind . (13)

The statistical model supposes that the induced charge is an integral over
the nanotube density of states from a local charge neutrality level to a local
chemical potential which becomes a Fermi level at zero temperature (see
Fig.8). The local chemical potential is supposed to follow the local acting
potential. Great simplification is achieved in case of metallic nanotube op-
erating at low voltage when the Fermi level shifts within the first sub band.
Then, the electron dispersion is linear and the density of states is constant
and equals VM = 8/(3lry). Here b ~ 1.4 A is the interatomic distance and
I ~ 2.5 eV is the hopping integral. Within this approximation of the linear
energy dispersion in the lowest subband, the induced charge density reads
 273

as:
(14)
I note that Eq.(14) holds in one-dimensional (lD) case; while in 2D the
charge density is proportional to the electric field (first derivative of the
potential) by Gauss-Ostrogradskii theorem, and in 3D the charge density
is proportional to the Laplasian of the potential by Poisson equation.
In order to obtain a selfconsistent solution for the charge density I
calculate the induced potential with use of a Coulomb operator Green's
function, G(r, r'):

cpind(r) = 41T ! G(r, r')p(r')dr'. (15)

The Green's function of a 1D system is known to have a logarithmic

singularity until some external screening is considered. In case of a nanotube
device, this screening is due to the closest gates/contacts. An equation,
giving the nanotube charge density implicitly, follows from Eqs.(14,15) and
reads as:
- ~(r) _ 41T! G(r, r')p(r')dr' = cpxt(r). (16)
e VM

The equation can be inverted analytically in simple case. In general, it

allows only numerical solutions or may be expressed as a series.
An interesting result of our study is that the nanotube may be divided
in three parts: two contact regions and a "central" region. The side parts
are the regions near the contacts (NT ends) of a length about several h
(R) long, where h is the distance to the gate. Aspect ratio of devices of the
state-of-the-art of nanotube technology is very high, which means that
the length of the nanotube, L, is much larger than the h. Then, the central
region of the nanotube covers most of the device length.
The electrostatics of the central region is elementary and allows an
analytical solution for Eq.(16). Because of the screening of the Coulomb
interaction by the backgate and the valence electrons of the nanotube, the
corresponding Green's function is short-ranged. Therefore, (at the distance
about 2 - 3h from the contact) the selfconsistent charge density is given by
a simple expression:

here I used notations C;; 1 = 2 log (~) and C Q1 = 1/ (e 2 vM) for the in-
verse capacitance (potential coefficient) of a straight metal cylinder and
the quantum correction, respectively. Poo stands for an equilibrium charge
274

density of the SWNT, calculated at the distance from the side electrode
much larger than the distance to the backgate, h.

4.2. MODEL EXTENSIONS

4.2.1. Quantum Mechanical calculation

I derived the Green's functions for several realistic device geometries and
calculated the selfconsistent charge densities. These charge densities were
compared with the results of the quantum mechanical computation. I solved
joint Shroedinger and Poisson equations for the valence pi electrons of a
metallic armchair [10,10] SWNT in one subband approximation (in full
neglecting the intersubband or sigma-pi mixing which has been estimated
and is of minor importance for our problem). Aside such purely quantum
effects as quantum beatings at the ends of the finite length nanotube, the
statistical, semi-classical and quantum mechanical charge distributions are
almost identical (a cross check has been done with use of periodic boundary
conditions to exclude the finite length effects). Fig.9A shows the typical
charge density distributions calculated with use of a tight-binding theory
and the Boltzmann equation for the cantilever SWNT of 50 nm long. I must
conclude that a simple statistical description works fairly well for the case
of straight ideal single wall nanotube. A similar result has been obtained
for the string SWNT with two side contacts (Fig.9B).

2.S 0.8
E
--
E 2 c 0.6
c
Cil.L S <J.)
~0.4
0.. 1
O.S 0.2
o ~./~ -~-~~~~ __~J o ~10
o 10 20 30 40 50 20 30 40
z,nm z,nm
Figure 9. Specific charge density for two devices: (Right) string and (Left) cantilever
NEMS. The solid oscillating (red) curve is a result of the quantum mechanical calculation.
The solid (blue) line is a solution of joint Poisson and Boltzmann equations. The dashed
(green) line is by the analytical approximation.

4.2.2. Capacitance of Distorted SWNT

The equation for the equilibrium charge density is valid for a distorted
nanotube as well as for an ideal straight nanotube. In case of slightly bent
SWNT, one has to use in Eq. (17) the capacitance of the bent metallic
 275
cylinder, C;l(h(z)), instead of the logarithmic capacitance which is valid
only for a straight one. Thus, the atomistic capacitance of the distorted
SWNT depends on its shape:

This analytical form of solution for the device electrostatics is very useful
for calculating electrostatic forces in various NEMS devices.

CONCLUSIONS AND DISCUSSION

In this section NT electrostatics is studied and equilibrium charge density

calculation is performed, which are foundations for a continuum device
theory for nanotube electromechanical systems and nanotube electronics.
Concept of atomistic capacitance has been introduced. This model gives
a fast and accurate method for a simulation of the charge density for
the nanotube of an arbitrary shape displaced by a voltage applied to the
nanotube end(s). The one-dimensional charge density is given by the total
atomistic capacitance of the nanotube, which is not defined solely by ma-
terial properties of the nanotube itself. It depends also on the environment
because of the charge in a low-dimensional electronic system of the nan-
otube is screened by near placed electrodes. An analytical expression for
the atomistic capacitance of a nanotube subjected to moderate distortions
is found as well as for an ideal nanotube. The role of quantum effects is
evinced and an expression for a NT quantum capacitance is derived.

5. Summary

In this chapter I presented several examples of a state-of--the-art modeling

of nanotube device systems and focused on analytical models rather than
numerical approaches to give a clear qualitative physical picture of quan-
tum and classical phenomena at the nanoscale. I present the continuum
modeling approach which allows one to combine an atomistic computation
with a real device engineering, and gives a powerful tool for theoretical
study of nanodevices.

Acknowledgements

Author is indebted to Professor K. Hess, Professor N.R. Aluru, and Mr.

M. Dequesnes for introducing to the subject of MEMS and fruitful discus-
sions of obtained results, to Dr. A.G. Petrov, Ms. Y. Li and Mr. K.A. Bu-
lashevich for collaboration on the electronic structure computations, and
276

to Professor Y. Gogotsi, Dr. L. Rotkina and Dr. 1. Zharov for fruitful

discussions. Author acknowledges support through a CRI grant of UIUC,
DoE grant DE-FG02-0lER45932, NSF grant ECS-0210495 and Beckman
Fellowship from the Arnold and Mabel Beckman Foundation.

References

S. Akita et. al., Nanotweezers consisting of carbon nanotubes operating in an atomic

force microscope, Appl. Phys. Lett. 79(11): 1691-1693, 2001.
N.R. Aluru, J-P. Leburton, W. McMahon, U. Ravaioli, S.V. Rotkin, M. Staedele,
T. van der Straaten, B.R. Tuttle and K. Hess, Modeling Electronics on the Nanoscale,
in "Handbook of Nanoscience, Engineering and Technology", Eds.: W. Goddard, D.
Brenner, S. Lyshevski, G.J. Iafrate. CRC Press, 2002.
A. Bachtold, P. Hadley, T. Nakanishi, and C. Dekker, Logic Circuits with Carbon
Nanotube Transistors, Science, 294: 1317-1320, 2001.
S. Bandow, M. Takizawa , K. Hirahara, M. Yudasaka, S. Iijima, Raman scattering study
of double-wall carbon nanotubes derived from the chains of fullerenes in single-wall
carbon nanotubes, Chemical Physics Letters. 337(1-3): 48-54, 2001.
Barash, Iu. S., Sily Van-der-Vaal'sa, Moskva:"Nauka", 1988. 344 p.
L.X. Benedict, S.G. Louie, M.L. Cohen, Static Polarizabilities of Single Wall Carbon
Nanotubes, Phys. Rev. B. 52: 8541-8549, 1995.
K.A. Bulashevich, S.V. Rotkin, Nanotube Devices: Microscopic Model, JETP Letters
(Transl.: Pis'ma v ZhETF) 75(4): 205-209, 2002.
A.V. Chaplik, Zh.Eksp.Teor.Fiz. (Sov. Phys.-JETP), 60: 1845-1852, 1971.
H. G. Craighead, Nanoelectromechanical Systems, Science, 290: 1532-1535, 2000.
J. Cumings, A. Zettl, Low-Friction Nanoscale Linear Bearing Realized from Multiwall
Carbon Nanotubes, Science, 289(5479): 602-604, 28 July 2000.
Dai, H. J., Carbon nanotubes: opportunities and challenges, Surface Science, 500 (1-3):
218-241, 2002.
M. Dequesnes, S. V. Rotkin, and N. R. Aluru, Calculation of pull-in voltages for carbon
nanotube-based nanoelectromechanical switches, Nanotechnology, 13: 120-131, 2002.
M. Dequesnes, S. V. Rotkin, and N. R. Aluru, Parameterization of continuum the-
ories for single wall carbon nanotube switches by molecular dynamics simulations,
Int.J.Comp.Electronics, 1(3/4): (in press), 2002.
V. Derycke, R. Martel, J. Appenzeller, and P. Avouris, Carbon Nanotube Inter- and
Intramolecular Logic Gates, Nano Letters, 1: 453-456, 200l.
Dresselhaus, M. S., Dresselhaus, G., and Eklund, P. C., Science of Fullerenes and Carbon
Nanotubes. San Diego, CA: Academic Press, 1996.
1. E. Dzyaloshinskii, E. M. Lifshitz, and L. P. Pitaevskii, Adv. Phys., 10: 165, 1961.
M.R. Falvo, J. Steele, R. M. Taylor II, R. Superfine, Gearlike rolling motion mediated by
commensurate contact: Carbon nanotubes on HOPG, Phys Rev. B, 62: 10665-10667,
2000.
L.A. Girifalco, M. Hodak, and R. S. Lee, Carbon nanotubes, buckyballs, ropes, and a
universal graphitic potential, Phys.Rev.B-Condensed Matter, 62: 13104-13110, 2000.
P. Kim, C.M. Lieber, Nanotube Nanotweezers, Science. 286: 2148-2150, 1999.
J.E. Lennard-Jones, Perturbation problems in quantum mechanics, Proc. Roy. Soc
London, Ser.A, 129: 598-615, 1930.
Y.Li, S. V .Rotkin and U .Ravaioli, Electronic response of nanotubes in transverse electrical
field, submitted, 2002.
 277

F. London, Z. Physik, 63: 245, 1930.

S. Luryi, Quantum capacitance devices, Appl. Phys. Lett., 52(6): 501-503, 1988.
E. Minot, V. Sazonova, Y. Yaish, J-Y. Park, M. Brink, P. McEuen, unpublished, 2002.
A.A. Odintsov, Y. Tokura, Journal of Low Temperature Physics, 118: 509, 2000.
P.M. Osterberg, S.D. Senturia, M-test: A test chip for MEMS material property mea-
surement using electrostatically actuated test structures, J. of MEMS, 6(2): 107-118,
1997.
A.G. Petrov, S.V. Rotkin, unpublished, 2002.
M. Reinganum, Ann.d.Phys. Bd., 38: 649, 1912.
V.V. Rotkin, R.A. Suris, Calculation of electronic structure of fullerene within spherical
quantum well model, Sov.- Physics of the Solid State, 36 (12): 1899-1905, 1994.
V.V.Rotkin, R.A.Suris, C60 electron collective excitation nature, in Fullerenes. Recent
Advances in the Chemistry and Physics of Fullerenes and Related Materials. v.III. Eds:
R.S.Ruoff and K.M.Kadish. ECS Inc., Pennington, NJ, 1996,940-959.
S.V. Rotkin, I. Zharov and K. Hess, Zipping of graphene edge results in [10,10] tube
formation, in Electronic Properties of Molecular Nanostructures; (XVth International
WinterschooljEuroconference, Kirchberg, Tirol, Austria, 3-10 March 2001), Eds.: H.
Kuzmany, J. Fink, M. Mehring, S. Roth. AlP Conference Proceedings, 591: 454-457,
200l.
S.V. Rotkin, Y. Gogotsi, Analysis of non-planar graphitic structures: from arched edge
planes of graphite crystals to nanotubes, Materials Research Innovations. 5(5): 191-200,
2002.
S.V. Rotkin, K. Hess, Many-body terms in van der Waals cohesion energy of nanotubes,
Int.J.Comp.Electronics, 1(3/4): (in press), 2002.
S. V. Rotkin, Analytical Calculations For Nanoscale Electromechanical Systems, in
Microfabricated Systems and MEMS - VI, vol. PV 2002-6, Symposium - the Elec-
trochemical Society Proceedings, Eds.: P.J. Hesketh, S.S. Ang, J.L. Davidson, H.G.
Hughes, and D. Misra, ECS Inc., Pennington, NJ, USA., 2002, pp. 90-97.
S.V. Rotkin, A. Shik and H. Ruda, Nanotube and Nanowire Field-Effect Transistors,
unpublished, 2002.
S.V. Rotkin, V.Shrivastava, K.A. Bulashevich, and N.R. Aluru, Atomistic Capacitance of
a Nanotube Electromechanical Device, in the Special Issue "Recent Advances in Carbon
Nanotubes, Nanoscale Materials and Devices", Eds.: S. V. Rotkin and S. Subramoney.
International Journal of Nanoscience, 1 (6): (in press) 2002.
R. Tenne and A.K. Zettl, N anotubes from inorganic materials, in Carbon N anotubes: Syn-
thesis, Structure, Properties, and Applications, Eds.: Dresselhaus, M. S., Dresselhaus,
G., and Avouris, P.; Springer- Verlag, Berlin, 80: 81-112, 200l.
J.D. van der Waals, Over de Continuepteit van den Gasen Vloeistoftoestand (thesis,
Leiden, 1873).
M.F. Yu, B.I. Yakobson, R.S. Ruoff, Controlled sliding and pullout of nested shells
in individual multiwalled carbon nanotubes, Journal of Physical Chemistry B, 104:
8764-8767, 2000.
SOL-GEL PROCESSING OF LITHIA ALUMINO-SILICATES (LAS) SOLID SOLUTION
FOR FUTURE LOW TEMPERATURE CERAMIC SENSOR

WAFA I. ABDEL-FATTAH,* M.SH. FAYED**, SH. R. GOUDA** AND W.F.F.

MEKKY**
* Natiollal Research Celltre, Ceramics Departmelll Dokki, Cairo, Egypt.
** Techllical College, Kobri EI-Kobba, Egypt.

Abstract

Recent studies in the lithia alumino-silicate system proved its suitability as a dielectric for
low temperature sensors using conventional ceramic techniques. In the present work, lithia alumino-
silicate solid solution (LAS) with a molar ratio of I: 1:3 for Li 10:A l10,:Si01 was prepared through
chemical hydrolysis and condensation of colloidal precursors. Tetraethoxysdane (TEOS). lithium
chloride [LiC L 2H 10] and aluminium nitrate [Al(NO,)]. 9 H20] were used as precursors. The effect
of acid catalysts being acetic acid (CH,COOH), hydrochloric acid (HCl) . phosphoric acid (H,PO"
and nitric acid (HNO) as well as fonnamide (HCONH 2) as a drying control agent on the syntheses
process and its yield were studied. The complex fonnation. polymerization and crystallization
behavior of the solid precursors at various heat treatments were followed using Thenno-Gravimetric
Analysis (TGA). Differential Scanning Calorimetry DSC and Differential Thennal Analysis (DTA).
X-Ray Diffraction (XRD) analysis was supplemented by IR analysis for the phase identification of the
thennally treated powders. Scanning electron microscopy (SEM) was used to assess the
morphological features for samples sintered at 900°C.
Results proved that the maximum yield of the LAS is in the presence of HNO) whose DSC
thenno-gram recorded a minimum Tg temperature of 313°C compared to other acids denoting the
lowest degree of cross-linking in parallel with the lowest percent weight loss. The DT A confinned
the Tg peak recorded on the DSC thenno-gram by an exo-peak at 316°C which was followed by two
peaks at 763 and 945°C due to the crystallization and sintering of the LAS solid solution respectively.
XRD proved initial crystallization for the samples calcined at 750°C following, the transfonnation of

279
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 279-289.
© 2003 Kluwer Academic Publishers.
 280

amorphous hallows recorded in powders calcined at 650°C. The sintering of the powders calcined at
900°C is confirmed by its DT A peak at 945°C and the highly crystalline structure especially in the
presence of HNO, as a catalyst compared to others. Advanced sintering was proven through a
minimum presence of hydroxyl groups for HNO) as proved by IR analysis along with the fused edges
characterizing the grains.

1. Introduction

Porous B-eucryptite, lithium alumino-silicate (LAS) has the potential application as a

humidity sensor [11. Recently its application as a low temperature sensor was explored [2]. Among
the factors that are crucial for sensor performance are internal stress state, uniformity and phase
stability [3]. Previous work by the present authors assessed the preparation of LAS (1: 1:3) 13-
eucryptite solid solution via the sol-gel technique using several concentrations of H 2S0 4 catalyst as
well as 0.1 M of either HCI, HNO, or Hl P0 4 [4]. For obtaining highest yield only a 0.05M ratio of
H2S04 was sufficient. The present work is extended to assess the calcination routes of the differently
catalyzed gels.
The sol-gel technology was first applied to prepare spherical particles of oxide nuclear fuel to
eliminate the grinding and sieving steps, as well as to prepare ceramic fibers and abrasive grains
[5].The use of solution mixing followed by gelation to make glasses and ceramics of various oxide
compositions, first with organic and later inorganic precursors, was developed almost four decades
ago [6].
At present. an atomic scale mixing of constituents of high purity allows the manufacturing of
a variety of new materials with high homogeneity, and well-controlled stoichiometry, microstructure.
and porosity.
The sol-gel process is well known to obtain solid polymerized phases trom organo-metallic
raw materials that lead to gels of high reactivity at relatively low temperatures [7]. However, with the
use of inorganic chemical compounds such as nitrates and chlorides and with an adequate percent. it is
possible to obtain sols with two or more hydrated cations. The gelation control leads to a glassy gel
without tree water, cracks and with high strength [7]. The gelation reaction was designated by the
additive type (acid) and Rh (molar ratio of water to alkoxide). Gelation of metal alkoxide solution is a
consequence of hydrolysis and the condensation reactions which are greatly affected by the molar
ratio of water to alkoxide (Rh ). The formation of drying cracks may be due to large capillary stresses
during drying [8] and an-isotropic shrinkage [9, 10]. Therefore, the use of drying control chemical
additive (OCCA) such as glycerol or formam ide can eliminate cracking by increasing the flexibility of
the dried gels [I I].
 281

The generation of a series of intermediate states of the host gel during its thermal evolution
to the corresponding ceramic is important to control. Such gels are known to have large internal pore/
surface area covered with hydroxyl ions. therefore. attracting ambient water molecules. The thermal
treatment of the gels, leading to decreased surface area and hydroxyl content. is expected to lead to
variety in the nature of coordination. No investigations have apparently been made to understand the
nature of these intermediate stages of sol·gel produced LAS I: 1:3 solid solution undergoing thermal
densification.
The present work aims at the preparation of p·eucryptite (LAS) solid solution (I: I :3) using
alcoholic solutions of aluminum nitrate and lithium chloride with silicium alkoxide. The influence of
sol parameters, acid peptizing and % solid yield were followed by TGA and DSC. Crystallinity of
calcined powders were assessed through DT A and XRD supplemented by IR and SEM.

2. Experimental:

Alcoholic solutions of LiCl'H,O (Koch·light) and AI(N0 3),-9 H,O (Fisher) were mixed and
added to TEOS [Si(OC,H,).] (Aldrich). The effect of catalyst at a ratio of (0.1: 10TEOS) was studied
by introducing H* represented by CH,COOH, H3 PO, HNO] and HCI at a ratio of 0.1 M. The
precursors were refluxed for I hour at 85°C. After cooling, the hydrolyzates were added to TEOS
along with the drying agent. Formamide (Prolabo) being the Lewis base like ammonium catalyzer was
added equal to alcohol to control drying. The TEOS/water/ethanol molar ratio was 1110/5. Stable gels
were obtained at ambient conditions within 3·5 days, aged for one week, then, dried at 120°C, till
constant weight [7].
TGA (Netzch, TG 209) and DSC (Netzch, DSC 200) were adopted at 10°C/min to follow the
decomposition of the dried gels. Phase analysis of the gels calcined at 550°. 650°. 750 0 and 900°C was

detected using X·ray diffractometer (Philips. PW 1390, 28 = 2°/min & eu Kcx radiation). Semi
quantitative infrared spectral analysis with KBr disc technique (Philips PU 9712 spectrophotometer)
and SEM (lEOL JSM T20) were used to supplement XRD data.

3. Results and Discussion

3.1. Thermal Analysis:
3./.1. Differential Scanning Odorimetry (DSC):

The sample with HCI catalyst exhibits a different behavior where two adjacent exothermic
peaks with different intensities are detected. The small peak is recorded at 269°C while the other
appeared at 272.7"C with similar glass transition temperature and with a total energy of 686.95 J/g
282

(Figl). In the sample with the HNO, catalyst, the exothennic peak is at 264.4°C with an intennediate
energy value of 662.72 Jig and holds for 55°C with a shoulder at 264°C, and a lower glass transition
temperature of 309.3°C compared to others.
The sample with H)PO, has only one peak of large exothermal activity at 273.7"C where its
maximum holds steady for about 55°C and has an energy of 634.63 Jig. As the temperature increases
a glass transition was reached which started at 329°C for another S°e. The sample with CH,COOH
catalyst has also one exothennic peak at 273.6 °c with an energy of 60 Jig which is similar to that of
the HNO, catalyst sample but it holds for about 75°C with a little shoulder at 260"C and glass
transition at 327°e.
The registered exothennic peaks are associated with the removal of all organic residues
incorporated in the dried gel structure as well as removal of physically adsorbed water. The
represented peaks have significant temperature shifts as the shifts range in about 2 min, but have
different energies. This difference means different amounts of water and alcohol in the samples as
consequence of difference in degrees of hydrolysis and condensation of the prepared samples.
From the above data and combining the effect of different acids. the order of heat evolved is
coinciding with temperature loss. The values of 6.C p followed the order H,PO,> HCI>HNO, denoting
the complexity of polymer length and relative ease of protonation of the negatively charged alkoxide
group. Acid catalyzed condensation is directed towards the ends rather than middle of the chain,
resulting in a less highly branched polymer giving monolithic gels.

3.1.2.Thermo-gravimetric analysis TGA :

The HCI sample recorded a steady decrease in weight up to a temperature of 273°C at which
a sharp decrease in weight occurs followed by another steady decrease giving a total loss of 82.8%. A
constant weight is then attained up to 1000°e. A different behavior is shown for the 0.1 M HNO,
sample (Fig. I b) where weight loss started at a lower temperature of 90°e. This may be due to the
adsorption of moisture by the sample during preparation and gives an indication that the prepared
sample is very hygroscopic. At about 230°C, weight loss increased with a faster rate up to a
temperature of 250°C where a sharp decrease in weight occurs and lasts for S°e. Loss in weight
continued with the increase in temperature up to S20"C giving a total loss of about 77.7%. The H,PO,
sample exhibits a weight loss of88 91% at a temperature of 273°C with no other losses recorded until
1000°e. The CH, COOH sample has no weight loss up to a temperature of 130°C at which a steady
decrease in weight starts. At approximately 273°C a sharp decrease in weight is recorded with a total
percent of loss 92%. The observed loss continues for about 5°C and stops.
 283

-2
E
~ -4

-!'
~-6

(a) -B

-10

-12

-14
200

100
-~, ~,

,
\ \
BO CH3COOH \ HO
\

,
\
60
_91.7 ~/.
I -88.9'" \ -830'"

(b) ";: 40 ,
I
I
,
I

20

Temperature;C

Fig.l. (a) Differential Scanning Calorimetry. (blThermo-Gravimetric Analysis

Therefore. the present data reveals that the weight loss is a single reaction step and is almost
complete at 273:+: 2°C in all samples and is attributed to the dissociation of organics. The order of
highest % of solid yield is HNO) ::: HCI > CH,COOH >H 3 PO,. The % yield value of the acid free
sample is only 45.6% which is nearly one half of the others.
The lower % yield implies that hydrolysis of OR groups to form Si(OH)4 is small and so
condensation reactions start before the complete hydrolysis of Si(OH),. The order of % yield in the
presence of HCl and HNO) catalyzers is attributed to the common ion effect with the precursors and
to the differences in their dissociation constants.
284

3.1.3.Diffirential thermal analysis DT A :

The experiments were only conducted at temperatures above 300"C due to large peaks

corresponding to decomposition and removal of organic residues which will decrease the sensitivity
for differentiating small peaks [12].
The first exo-peak above 300°C corresponds to de-carbonization while the second exo-peaks
are recorded between 752"C and 763"C having the same phenomena and corresponding to
crystallization. The third exo-peak is above 900"C revealing lower crystallization temperatures than

that corresponding to HNO, (Fig. lc).

Comparing these data with those obtained by TGA. it has to be notified that differences in
peak temperatures are related to differences in the extent of cross-linking between fonned species
during the preparation sequence as a result of different types of acid catalysts employed. Therefore. a
calcination temperature of 875°C is expected to give the stan of crystallinity.

Peak !l-l5.06 ·C
~. rtB3.911 ·C I •

Paak i63.H 'e IINO,

AJ"I'~ ·43.5~ 'e ~ ,
Put 3i5.96 ·C
(c) Anli -1.005 ·C I !

;---~~I------~ t
600.0
,
100.0 8OG.O
I I
900.0
1
1000.0 HOO.a
~o.o -400.0 500.0

Fig I.(c) Differential thennal analysis DT A

3.2. Structural Analysis

3.2.1. X-ray Dif/raction (XRD):

Powders for sintering should have to be isometric in shape. small in size «1 ~). have narrow
size distribution and do not form hard agglomerates to attain high sinterability. All the prepared
samples that were calcined at 450°C did not show any definitive diffraction peaks in their X-ray
diffraction patterns indicating that the powders remained amorphous up to this temperature. At 550°C

only the HCI sample gives a hallow in the 28 rang from 18 to 32°, and that of HNO) gives a similar
 285

trend having the 28 range extended to 38° This range is also recorded for the catalyst free sample. At
650°C, although a higher calcination temperature, the HCI sample and that of HNO, give only the
hallow fonll, while the catalyst free sample starts a single peak (d=3 4797 A) which is characteristic
for LAS, ~-eucryptite. At a higher annealing temperature of 750°C, the pattern of the HCI sample
shows a sharp peak (d=3.5203A) together with two small peaks (d=1.9028A and 1.652IA). The
sample of HNO, shows only two small peaks (d=3.5618A and 34931A) without any other peaks
denoting lower crystallinity. The catalyst free-sample has the highest peak (d=3.5203A) together with
three peaks (d=3.26667 A, 1.8954A and 1.4968A) of similar low intensity. When comparing the data

with those of standard ~-eucryptite and ~-spodumene (JCPDS file), the structure of the present
samples is characteristic for the solid solution.

Ie .. flm Cu-K..

Fig. 2 X-ray Diffraction pattern of samples calcined at 900"C

It is clear that the highest produced crystallinity is due to the HNO, catalyst, while the
CH,COOH sample exhibits a unique peak at d=3.847A which is not resolved well in other samples
except for the H,PO, sample. This peak may be due to the strong chelating power of CH 3 COOH that
affects the structure developed during gelation leading to some sort of different produced structure
due to different ways of cross-linking. The present data reveals the effect HNO, on inducing better
crystallinity compared to others.
286

3.2.2. Infrared Analysis

3.2.2.1. Calcination at 650°C

a. Group ji-equency region: The absorption band at 3450cm' for the CH,COOH and HCI samples is
very strong and broad, while for HNO, and H,PO. it is of lower intensity denoting lower water
capture in the structure. Greater absorption of the HN0 3 sample in contrast to other samples is for the
band at 2920 cm" pointing to various sites ofOH' species. Here the reappearance of the 1650 em"
band for the OH' stretching is clear. This gives an indication to the redistribution of these sites and
inclusion of water species in the lanice.
b. Finger print region The broad band is located in the range from 1200-900 cm". All samples
recorded a characteristic absorption at 1120 and 1020 cm" with variable intensities and a different
behavior especially for the H,PO, sample where a very broad shoulder appears at 1120 cm". For the
CH,COOH sample it is only an inflection. For the HCI sample, two weak shoulders at 1120cm"and
1160 em" were resolved. while for HNO, a very weak shoulder is recorded at 1120cm'. The band at
880 cm" assigned for AI 6 _OH disappeared from all the spectra accompanying the reorientation of the
structure due to calcination at such high temperatures. The HNO, sample differs from the others with
a very sharp set of peaks in the region from 700-400 em' (ie. amorphous silica and alumina). For the
HCI and CH,COOH samples they are only resolved. while for the H,PO, catalyst sample they are
absent. Therefore. it could be noted that the order of effectiveness of acid catalyzer is HNO, > HCI ~

CH,COOH > H,PO,

3.2.2. 2. Calcination at 750°C:

Xerogels obtained at 750°C have an IR spectrum very similar to those obtained at 650°(,
although it was expected to reflect the start of crystallinity.

3.2.2. 3.Calcination at 900°C:

(1- In group ji-equenc)' region. Only one band at 3430cm" assigned for OH' bridging appeared in all
spectra. Other bands due to other OH' species previously recorded at 2920. 2850 and 1650 cm"
disappear completely as a consequence of higher calcination temperatures and crystallinity (Fig. 3).
b- In/inger print region' The 1120cm,1 band assigned for Si<-+O AI. Si-O-Si and Si-O str is shifted to
1160 cm" (i.e. to a lower wave length). A new band at 1060 cm" assigned for Si(AI)-O which beside
1020 cm" assigned for Si-O str. AI" -O<-+AI' revealing reorientation and stability of the structure. The
880cm' band is shifted to 820 em" to tetrahedral symmetry Si'-O<-+AI'. Also the 720 cm" band
 287

shifted to 750 cm- l (i.e a lower wave length) i.e. from AI· ..... O-Si(AI0 6 ) od to Si-AI(Si) str, Si-Si str
AI6 Od. The 670-650 cm- l for (AI)" and 550-415 cm- l for (AI-O) are recorded at constant wave
numbers_ The IR spectra confirmed the XRD findings as all bands are recorded in the samples with
great similarities for their band intensities and wave number positions except for the H,PO" catalyst
sample that has a lower intensity.

~
g
1~
c .
§
~
.':( ~
..!.

§
~

E
§ "-
~

~
~ ~
~

Fig. 3. Infrared spectra of samples calcined at 900°C

3.3. Scanning Electron Microscopy (SEM):

During thermal treatment of the precursor. the starting particles develop a texture

characterized by a cluster consisting of a large number of small particles «0.5 11m in diameter)

agglomerated in a large size particle (20-30 11m in diameter). These features point to random
nucleation mechanism. A common phenomenon is that the size of the particles increases with
increasing calcination temperature due to partial sintering.
The HNO, sample a shows fused surface which is confirmed in the focused picture showing
fused grains (Fig 4a). The symmetric two crystals formed indicate a very high degree of crystallinity
at this temperature (Fig 4b).
The SEM with the various acids has inhomogeneous particle size distribution with some
neck areas inbetween. Few long laths are observed with the HN0 3 catalyzer.
The HCI sample shows pores of comparable size together with fused rounded pores which
gives an indication of an early stage of sintering. The formed crystal inside the pore confirms the high
crystallinity of the sample (Fig 4c &d).
288

(a) (b}

Fi", 4, Scanning: dectron mk'fI);;copic analysis of I [NO, catalyzcd samples calcined at 900"C

( c) (J)

Fig, 4, Scanning electron microscopic analysis of HCI catalyzed samples calcined at 900°C

4- Conclusions:

It was possible to obtain maximum yield of LAS solid solution via sol-gel route in the
presence of nitric acid with minimum glass transition, Gelation was perfonned in open aiL The
cystallization of the fonned gels is proved by strong exo-thennic reaction, Morphologies of the
calcined xerogels presented various degrees of sintering,
 289

The effect of either nitric or hydrochloric acid catalysts is attributed to the common ion effect
and strong chelating of the nitrate group. The role of acetate group as mono-dentate, chelate bi-dentate
or bridging bi-dentate warrants further research.

5-References:

1. Susumn, N., Hiroshi, K., Yoshiko, S. and Yoshiro, S. (1992). 1. Ceram. Soc. Japan, 100 (1163)
968-971.
2. Abdel-Fattah, Wafa, I., EI-Amir ,G. and Abdellah, R.A. (2002) Dopants and dielectric
properties of ~-eucryptite and ~-spodumene, Key Eng. Mat.206-213.13 73-13 76.
3. Lisa, K.1. (1989) Am. Ceram. Soc. Bull. 68, 4, 860-865.
4. Abdel-Fattah. Wafa, I., Fayed, M.Sh., Gooda. Sh. R. and Mekky. W.F. (1998) J. Sol-Gel.
Sci. and Tech. 13,981-985.
5. Brinker, C.J. and Scherer, G. W, (1990) Sol-Gel Science, The physics and chemistry of sol
gel processing, Academic Press Inc., New York.
6. Roy. R. (1987) Science. 238.1664.
7. Zarzycki. 1. (1984 ) Ultra-structural processing for ceramic. glasses and composites Ed. LL
Hench and D.R. Ulrich, 27 42, New York.
8. Scherer, G.W., (1986) in Better ceramics through chemistry, II ed. C.J.Brinker, D.E. Clark and
D.R. Ulrich Materials Res. Soc., Pittsburgh, PA, 225.
9. Leenaars, A.F.M., Keizer, Land Burggraaf, A.J. (984) J. Mat. Sci. 9,1077.
10. Pierre, A.C. and DR. Wlmanin, D.R. (1986) in Better ceramics through chemistry, II ed.C.
11. Brinker. DE Clark and D.R. Ulrich. Materials Research .Soc.Pittsbergh.PA,48.
12. Lannutti J.1. and Clark. DE. (1985) Ceramics 1m. II. 91.
13. Huling, S.c. and Messing, G.L (1992) J. Non-Crystalline Solids, 147-148,213.
STRUCTURAL AND OPTICAL CHARACTERIZATION OF InN THIN
FILMS - NOVEL PHOTONIC MATERIALS FOR PHOTOVOLTAIC AND
SENSOR APPLICATIONS

V.Ya. MALAKHOV
Institute for Problems of Materials Science, National Academy of
Sciences of Ukraine
3, Kr=hi=hanovsky St., 03142 Kiev, Ukraine

The basic structural and optical properties of low temperature plasma enhanced
reactively sputtered (L TPERS) InN thin films are presented. The X-ray diffraction,
RHEED, Auger, SEM and AFM studies of a surface morphology plus a cross-
section of InN polycrystalline layers deposited on various substrates were performed
to evaluate of an inner granular structure and cluster's properties as well. The optical
absorption and reflectance spectra of the InN textured films were taken to determine
some optical and electronic parameters including the direct band gap energy (2.03
eV). Moreover the TO (480 cm") and LO (585 cm") phonon features of the indium
nitride polycrystalline films in the NIR and Raman spectra are observed and
discussed as well. These results both obtained now and a few years ago from
identical film samples are in a good agreement and thus demonstratl! a long time
stability of this compound as to its electronic parameters. The attractive possibilities
of InN layers especially as top coatings of InN lSi tandem heterojunctions for
potential applications both in sensor engineering and PV devices including high
efficiency solar cells are confirmed.

Keywords: InN film, microstructure, Raman spectra, photonic devices

1. Introduction

In the last ten years, indium nitride has become a focus a growing interest as a III-V
semiconductor. It possesses about a 2.0 eV direct band gap and therefore is a
potential application in photonic devices such as LEDs, lasers, full color displays,
and especially in high efficiency solar cells [1]. Physical properties of InN thin films
obtained by different methods have been studied in numerous works [2-7] including
ones ofthe author. However some optical and electric parameters, namely: dielectric
and optical constants, energy gap, effective mass of carriers as well as phonon
wavenumbers await for further more accurate definition. A lack of single crystalline
samples explains the situation around mentioned above data. Any structural and
thermal properies of InN have not been studied for epitaxial films on lattice-matched
291
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 291-295.
© 2003 Kluwer Academic Publishers.
292
substrates. As was reported in [3] the TO and LO phonon features which observed in
transmittance spectra of reactively sputtered InN thin films are close to determined
earlier from reflectance of Gal.xInxN solid solution [4]. But Kwon with co-worker
[5] did not reveal the LO phonon peak in the Raman spectra ofInN monocrystalline
films. For this reason, the presented paper demonstrates original structural and
optical data obtained for InN polycrystalline films synthesized earlier and nowadays
by LTPERS [6-7] and offers to use potential possibilities of InNISi heterojunctions
in sensor and PV devices including high efficiency solar cells.

2. Experimental details

Inasmuch as a dissociation temperature of InN films is about 650°C [2] therefore

low temperature growth technique is required. In our case the LTPERS equipment to
synthesize InN thin films was used. Intensive Ti-wire evaporation applied as getter
was carried out during all time of deposition process to reduce oxygen
contamination inside of reactor and in growing films as well. High quality smooth
surfaces of Si, quartz, fluorite and compound ceramics were used as substrates. In
time of films growth substrates temperature was about 350°C due to intensive ion
bombardment of a top electrode (anode) during sputtering process. The films
thickness was changing in the range 100-2000 nm. To determine chemical
composition of sputtered InN films Auger spectrometer JAMP-1O was used. The
surface morphology including a microstructure of cross-section of the films were
investigated by Philips SEM5V scanning electron microscope and also by AFM
standard equipment. The crystalline structure parameters of sputtered films were
determined using the X-ray diffractometer DRON-3 employing Cu(ka) radiation
and also by means of the electronograph EG-IOOa. Reflectance and transmittance
measurements in visible and NIR regions (25000-200 cm- I ) have been carried out
with help of Bruker IFS66 Fourier transform spectrometer (FTIR) and Carl Zeiss
M40 grating spectrometer respectively. The Raman spectrometer Dilor XY
equipped with microscope was used for study of phonon spectra of nitride films.

3. Results and discussion

X-ray diffraction patterns of a-InN (wurtzite) layers deposited on ceramic substrates

demonstrate most strong diffraction peaks corresponding to the InN (002). It
suggests some textured crystalline structure of deposited films where the c axis is
perpendicular to the plane of substrate. The same result was obtained from a study
of the RHEED pattern of InN film deposited on ceramics (Fig. 1). The AFM image
of natural surface morphology of InN film on polished ceramics is shown on Fig. 2.
These results and also Auger-investigation testify that any outsider phases except
InN were found inside films. Some off-stoichiometric InIN ratios with In abundance
inside nitride films is caused by nitrogen vacancies. A perceptible oxygen
concentration has been revealed in the films leads probably due to amorphous
indium oxide forming partialy binded oxygen.
 293

vl"'_I~
"'../-<I¥I<1'_h

Figure 1. RHEED pattern from InN layer Figure 2. AFM image of natural surface ofInN film
deposited on ceramics synthesized on ceramics

Moreover, the other parts of unactive oxygen molecules both incorporate to voids
between InN grains orland places on film's surface.
Further, both transmittance and reflectance spectra in the visible spectral range
were taken to determine the band gap energy of InN thin films on CaF2 (fluorite)
substrates. At this procedure the band gap energy was derived from absorption
coefficient a of InN layer and yielded E g := 2.03 eV which is very close to the value
reported earlier [6]. Moreover the PDS method was used to explore some
peculiarities in free carriers absorption spectra for InN polycrystalline films
deposited on different substrates. In order to obtain necessary information about
phonon features of InN films as well as to precisely determine some optical
parameters the Drude-Lorentz formalism was used for dielectric function modeling
procedure to reflectance spectra from both InN films surface and bare fluorite
substrate. Fig. 3 shows a good agreement between experimental data and calculated
curve. Moreover as shown there the reflectance peaks at 485 and 590 cm'!
respectively in our opinion are connected with TO and LO modes vibration of
indium nitride [3-5]. Identical result were obtained from study of Raman spectra for
InN textured films > 1000 nm of thickness on ceramics (Fig. 4). The
broadening peak at 485 cm'! in Raman spectra of nitride films is close to one
observed in the reflectance spectra. But because of imperfect crystalline structure of
prepared samples we can observe only two apparent optical phonon modes there:
E!(TO) at 485 cm'! and A!(LO) at 585 cm'!. For this reason it was difficult to find
here also the main LO phonon mode at 694 cm'! as was predicted earlier by
Osamura et al. [4].
294

10.3 -me.Burament
model calculation
2(b<10'

"
t!. 0.2 , .
J! •
.c l.!lo::lO

: 0.1

"
OO~--~~~~--~2~OO~O--~25~O~O--~300·0
Wlly"number (em -') 'lJ,CM·1

Figure 3. NIR reflectance spectra ofIoN

Figure 4. Typical Raman spectra ofInN textured
polycrystallioe layer on CaF,
film deposited on ceramics

4. Theoretical and practical considerations

To achieve optimum power conversion efficiency 1) for solar cells based InN n-Iayer
(emitter) and p-Si (base), large aCE), minority carrier lifetime T, diffusion length L
(at least aL>3 for front side illumination) as well as small surface recombination
velosity S have to be combined. It also can be achieved under compromise condition
of optimum band gaps Eg (-1.0 and 2.0 eV) for an efficient tandem system
consisting of several stacked single cells between large Isc or large Voc [I]. An
appropriate solar cell base material should exhibit a proper energy gap and a strong
absorbtion (alnN > 104 cm· l ) adjusted to the solar spectrum, a dopability by carriers
which exhibit high mobilities and long lifetimes. Existing problems for
heteroepitaxy of InN films on Si will be fulfilled by option and research of
appropriate buffer layer (e.g. AIN thin layer) on InN-Si interface in order to match
lattice periods and improve running heterojunction parameters. A fabrication of high
quality InN/Si heterojunctions and its further characterization are also of great
importance for the manufacturing of high efficiency solar cells.

5. Conclusion

The polycrystalline a-InN (wurtzite) thin films synthesized by LTPERS were

textured and therefore demonstrated preferable <002> orientation. A perceptible
oxygen concentration revealed in the films probably leads to amorphous indium
oxide forming partialy (binded oxygen). Other parts of unactive oxygen molecules
were incorporated to void between InN grains or places on a surface of films. A
direct energy gap 2.03 eV of InN polycrystalline films was confirmed. Moreover
E!(TO) at 485 cm'! and A!(LO) at 585 cm'! phonon features observed from Raman
and NIR reflectance spectra are in a good agreement with those obtained by other
authors. These studies show a long term stability of this compound as to its optical
and electrical characteristics. Thus the promising possibilities of InN stable
polycrystalline layers for potential applications in sensor engineering, high power
photonics and PV devices including high efficiency solar cells based on InN lSi
heterojunctions are confirmed.
 295
6. References

I. Yamamoto, A, Tsujino, M., Ohkubo, M., Hashimoto, A (1994) Metalorganic chemical vapor
deposition growth ofinN for InN/Si tandem solar cell, Sol. Energy Mater. and Sol. Cells 35, 53-60.
2. Ambacher, 0 (1998) Growth and applications of group III nitrides, 1. Phys. D: Appl. Phys. 31, 2653
2710.
3. Tansley, T., Egan, R., Horrigan, E. (\988) Properties of sputtered nitride semiconductors, Thin Solid
Films 164,441-448.
4. Osamura, K., Naka, S., Murakami, Y. (\ 975) Preparation and optical properties of Gal.' In,N thin
films,1. Appl. Phys. 46, 3432-3437.
5. Kwon, H-J., Lee, Y-H. (1996) Raman spectra ofinN monocrystalline films, Appl. Phys. Lett. 69, 937-
940.
6. Tyagai, V., Evstigneev, A., Krasiko, A, Andreeva, A and Malakhov, V. (1977) Optical properties of
indium nitride films, J. SOY. Phys. Semicond. (USA) 11, 1257-1261.
7. Malakhov, V. (\999) Growth and optical characterization ofinN thin films synthesi~:d LTPERS,Proc.
of Euromat'99 9,75-79.
AN OVERVIEW TO MAGNETIC BEADS USED IN ELECTROCHEMICAL
DNA BIOSENSORS

JOSEPH WANG 1, ARZUM ERDEM 1

1 Dept. of Chemistry and Biochemistry,
MSC 3C, New Mexico State University
Las Cruces, NM 88003, USA
joewang@nmsu.edu
2 Dept. of Analytical Chemistry,
Faculty ofPhannacy, Ege University,
35100 Bomova-IZMIR. TURKEY
erdema@phann.ege.edu.tr

ABSTRACT

Nucleic acid layers combined with electrochemical or optical tr~msducers produce a

new kind of affmity biosensors such as "DNA Biosensor" or "Ge:nosensor" for small
molecular weight molecules. Electrochemical DNA biosensors are attractive devices
for converting the hybridization event into an analytical signal for obtaining sequence-
specific information in connection with clinical, environmental or forensic
investigations. DNA hybridization biosensors, based on electrochemical transduction
of hybridization, couple the high specificity of hybridization reactions with the
excellent sensitivity and portability of electrochemical transducers. The ultimate goal
in all research is to design DNA biosensors for preparing a basis for the future DNA
microarray system. Recently, there has been an interest in magnetic beads for
designing of new electrochemical DNA biosensor approaches resulting in efficient
magnetic separation. Attractive features of this technology include simple approach,
rapid results, multi-analyte detection, low-cost per measurument, stable, and non-
hazardous reagents, and reduced waste handling. Some of these new approaches and
applications of the electrochemical DNA biosensors based on magnetic beads are
described and discussed.

1. INTRODUCTION

After the discovery of electroactivity in nucleic acids at the beginning of the sixties
[1], many electrochemical approaches have been tried for analyzing or quantification
of nucleic acids. There has been an increase in the use of nucleic acids as tools in the
recognition and monitoring of hybridization of analytical interest [2-8]. Recent
advances in automated DNA synthesis and the convenient site··spesific labeling of
synthetic oligonucletides with advances in microelectronics, have generated a number
of attempts at the development of novel biosensor devices for the: analysis of spesific
gene sequences and for the study of nucleic acid-ligand interactions. A biosensor is a
297
Y.C. Cagatsi and I. V Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 297-303.
© 2003 Kluwer Academic Publishers.
298

device that incorporates a biologically active layer at the surface as recognition

element and converts the physical parameters of a spesific biological interaction into a
measurable analytical signal. Various combinations of biological material associated
with different types of transducers are an attractive subject of research. A biosensor
associates a bioactive sensing layer with any suitable transducer and gives an output
signal. Biomolecular sensing can be defmed as the possibility of detecting analytes of
biological interest, like metabolites, drugs and toxins, by using an affmity layer like
enzymes, receptors, antibody or nucleic acids. The affmity layer can be a natural
system or an artificial one, able to recognize a target molecule in a complex medium
among thousands of others.
A genosensor, or gene-based biosensor / DNA biosensor, normally employs
immobilized DNA probes as the recognition element and measures spesific binding
processes such as the formation of DNA-DNA and DNA- RNA hybrids, and the
interactions between proteins or ligand molecules with DNA at the sensor surface [5].
Typically, the design of a genosensor involves the following steps [8]: 1)
modification of the sensor surface to create an activated layer for the attachment of the
DNA probe; 2) immobilization of the probe molecules onto the surface, preferably
with controlled packing density and orientation; and 3) detection of target gene
sequence by DNA hybridization at the sensor-liquid interface.
What makes biosensors the most difficult and the most appealing study is the
interdisciplinary aspect of this field and the competency associated with designing
these new analytical tools. Biosensors are able to work in real and practical conditions
and thus challenge well-established methods in molecular diagnosis.
The detection of specific DNA sequences provides the basis for detecting a
wide variety of microbial and viral pathogens. Traditional methods for DNA
sequencing, based on the coupling of electrophoretic separations and radioisotopic
e 2 p) detection, are labor intensive and time consuming, and are thus not well suited for

routine and rapid medical analysis, particularly for point-of-care tasks.

Electrochemical hybridization biosensors for the detection of DNA sequences
may greatly reduce the assay time and simplify its protocol. Such fast on-site
monitoring schemes are required for quick preventive action and early diagnosis. The
development of DNA hybridization biosensors holds great promise for obtaining
sequence-specific information in connection with clinical, environmental or forensic
investigations. Nucleic acid hybridization is a process in which inconsonant nucleic
acid strands with specific organization of nucleotide bases exhibiting complementary
pairing with each other under specific given reaction conditions, thus forms a stable
duplex molecule. This phenomenon is possible because of the biochemical property of
base-pairing, which allows fragments of known sequences to find complementary
matching sequences in an unknown DNA sample [9]. DNA hybridization biosensors
can be employed for determining early and precise diagnoses of infectious agents in
various environments [10,11] and these devices can be exploited for monitoring
sequence-specific hybridization events directly [12,13] based on the oxidation signal
of guanine or by DNA intercalators (metal coordination complexes, antibiotics, etc.)
which form complexes with the nitrogenous bases of DNA [14-22]. Electrochemical
transducers offer a very attractive route for converting the hybridization event into a
useful analytical signal [23-26].
 299

2. ELECTROCHEMICAL DNA HYBRIDIZATION BIOSENSORS

BASED ON MAGNETIC BEAD ASSAY

One of the newest approaches in designing electrochemical DNA biosensors is based

on particles called magnetic beads. These magnetic particles are strongly affected
when exposed to external magnetic fields. Their size is typically between I and 10 J..lm.
These materials normally contain metals such as iron, nickel or cobalt [27]. The use of
monosized magnetic beads gives the convenience of magnetic separation. These
particles are superparamagnetic, meaning that these microparticles can be easily
separated from the liquid phase with a small magnet, but can be redispersed
immediately after the magnet is removed [28]. The use of magnetic particles can bring
novel capabilities to bioaffmity assays and sensors. Bioanalysis has benefited from the
use of magnetic beads in electrochemical immunosensors [29] or for fluorescence
DNA hybridization approaches [30]. The new biomagnetic process combines efficient
magnetic mixing and separation into a single mechanism [31].
Magnetic bead capture recently has been used for eliminating non-spesific
adsorption effects hampering label-free detection of DNA.
Wang et al. [32] reported a novel biomagnetic assay of DNA sequences
related to BRCAI breast-cancer gene based on the coupling of the label-free guanine
detection route with an efficient magnetic isolation (by procedure represented in
Scheme A) of the hybrid for addressing the non-spesific adsorption problem. The
results have showed that this proposed magnetic separation has been extremely useful
for discriminating against unwanted constituents, including a large excess of co-
existing mismatched and non-complementary oligomers, chromosomal DNA, tRNA
and proteins such as bovine albumin. The new protocol involves the attachment of
biotinylated oligonucleotide probes onto streptavidin-coated magnetic beads, followed
by the hybridization event, dissociation of the DNA hybrid from the beads, and PSA
measuruments at a renewable pencil graphite electrode (PGE). For optimization of the
experimental parameters on PSA, the effect of the deposition potential, stripping
current and accumulation time was studied and most subsequent work employed on
the optimum conditions as + 0.5 V, +5 J..lA and 1 min, respectively. The effect of
amount of beads and hybridization time was studied and the optimum hybridization
time was found as 15 min. All experiments was done as using 100 J..lg beads which
corresponded to l.9 J..lg probe. An excess amount of captured probe as 4 J..lg was used
for a full coverage of beads. It was reported that the detection limit corresponds to 5 ng
in the 50 ilL samples (i.e., 60 pM) at hybridization time as 10 min.
300

SCHEME B ... Streptavidin coated SCHEME A

!
. , magnetic beads
Biotinylated probe
* Inosine labelled one for Scheme A
~
~ ______-vv.
Biotinylated ~
~ target ~
I
'" streptavidin.Enzyme +I NaOH
~ e.-w
,'\IV
! s- IX· noltilfo_
".. ar-nllltol
Pencil
__
8 p
...::::~-="""'~ adropl.toftcrwhgl. Graphite
./ , pracess In Scheme B
Screen-printed electrode Electrode

~
ELECTROCHEMICAL
DETECTION FOR BReA 1
Scheme: Electrochemical detection of DNA hybridization related on BRCAI gene. (A) (a) introduction of
the streptavidin-coated beads; (b) magnetic capture of the biotinylated inosine labelled probe; (c)
hybridization event; (d) release the hybrid DNA from the beads using 0.05 M NaOH solution, (e) adsorptive
stripping chronopotentiometric detection of the target guanine oxidation peak at a renewable graphite
electrode. (B) (a) introduction of the oligomer-coated beads; (b) addition of the biotinylated target (T)
oligomer-hybridization event; (c) addition of the streptavidin-enzyme (E=a1kaline phosphatase) and its
conjugation with the biotinylated target; (d) addition and enzymatic reaction of the substrate (S=a-naphthyl
phosphate); (e) placement of a droplet of the supernatant onto the thick-film electrode for the voltammetric
meaurument.

hybridization based on electrochemical techniques; such as, differential puis

voltammetry (DPV), or potentiometric stripping analysis (PSA). In electrochemical
DNA biosensor approaches based on magnetic beads [32-35], the spesific detection of
DNA hybridization observed in exceedingly low detection limits resulting in efficient
magnetic separation.
Another study of Wang et al. [33] represents the coupling a magnetic
isolation with electrochemical detection of DNA segments related to the breast-cancer
BReAl gene. This new protocol employs an enzyme-linked sandwich solution
hybridization isolation (by procedure represented in Scheme B), with a magnetic-
particle labeled probe hybridizing to a biotinylated DNA target that captures a
streptavidin-alkaline phosphatase (AP). The a-naphthol product of the enzymatic
reaction is quantitated through its well-defined, low-potential DPV peak at the
 301

disposable screen-printed electrode. The effect of the hybridization time, the enzyme
association time, ex-naphthyl phosphate reaction time was studied ,md most subsequent
work employed on the optimum conditions as 20, 25 and 20 min hybridization,
enzyme association and reaction times, respectively The effect of amount of beads was
also studied. All experiments was done as using 50 Ilg beads which corresponded to
950 ng captured probe. A larger amount of biotinylated probe as 4 Ilg was employed
during coating step to assure a full surface coverage. It was report1ed that the detection
limit corresponds to 500 pg target in the 50llL samples at hybridization time as 20 min.
The reported and on-going genomagnetic electrical assays yield a very
attractive bioanalytical performance. The attractive features of this technology include
simple approach, rapid results, multi-analyte detection, low-cost per measurument,
stable, and non-hazardous reagents, and reduced waste handling. Further
improvements, particularly towards the detection of point mutations, are expected
from combining the new challenging protocol with peptide nucleic acid (PNA) probes.
Such coupling of magnetic hybridization surfaces with renewable transducers and
label-free electrical detection eliminates the needs for external indicators and advanced
surface modification or regeneration schemes, and hence results in a greatly simplified
protocol. The use of magnetic beads for detection of DNA hybridization has been
developed in flow detection systems combined with the advantages of solid state
chemistry, for designing of DNA microarray in diagnosis or drug discovery.

3. FUTURE PERSPECTIVES

The completion of the first draft of the human genome has made it possible to foresee
major steps forward in our understanding of the molecular basis of disease, both from
attack by external pathogens and internally from variations within the human genome
resulting in a plethora of new molecular therapeutic targets fiJr drug design and
discovery.
It is hoped that continued development through combined efforts in
microelectronics, surface/ interface chemistry, molecular biology, and analytical
chemistry will lead to the establishment of genosensor technology as a major
component of analytical biochemistry [8].
Immediate applications will include directly quantifying DNA samples for
use in sequencing or polymerase chain reactions (peR), or pharmaceutical testing and
quality control. Eventually, they could be applied to producing credit card-sized sensor
arrays for clinical applications such as detection of pathogenic bacteria, tumors, and
genetic disease, or for forensics.
A quantitative understanding of such factors that determine recognition of
DNA sites would be valuable in the rational design of new DNA targeted molecules
for application in chemotheraphy and in the development of new tools for the point-of-
care tests and diagnosis based on DNA.

4. ACKNOWLEDGEMENTS

A.E acknowledges the scientific scholarship from Highly Skilled Young Scientist
Programme of The Turkish Academy of Sciences (TUBA-GEBIP). A.E wishes to
express her gratitute for scientific and technical contributions from the members of
302

DNA biosensor teams at Senso-chip lab in New Mexico State University (NM, USA)
during the course of these studies.

5. REFERENCES

1. Palecek E. (1960) Oscillographic polarography of highly polymerized deoxyribonucleic acid, Nature,

188,656-657.
2. Wang J. (1999) Towards Genoelectronics: Electrochemical Biosensing of DNA Hybridization, Chem.
Eur. J., 6, 1681-1685.
3. Wang J. (2000) Survey and Summary: From DNA biosensors to gene chips, Nuc!. Acids Res., 28, 3011-
3016.
4. Palecek E., Fojta M. (2000) DNA hybridization and Damage, Anal. Chem., 73, 75A-83A.
5. Pividori M. I., Merkoci A., Alegret S. (2000) Electrochemical genosensor design: immobilisation of
oligonucleotides onto transducer surfaces and detection methods, Biosens. Bioelectron., 15,291-303.
6. Chiti G., Marrazza G., Mascini M. (2001) Electrochemical DNA biosensor for environmental monitoring,
Anal. Chim. Acta, 427,155-164.
7. Erdem A., Ozsoz M. (2002) Electrochemical DNA Biosensors based on DNA-Drug Interactions,
Electroanal., in press.
8. Yang M., McGovern M. E. , Thompson M. (1997) Genosensor technology and the detection of
interfacial nucleic acid chemistry, Anal. Chim. Acta, 346, 259-275.
9. A. K. Bej, Nucleic Acid Analysis: Principles and Bioapplications, Wiley-Liss Press, New York, 1996,
Chapter I, pages 1-29.
10. Hodgson J. (1998) Shrinking DNA diagnostics to fill the markets of the future, Nature Biotech., 16,
725-727.
II. Millan K., Saraulo A., Mikkelsen S. R. (1994) Voltamrnetric DNA biosensor for DNA based on
electroactive hybridization indicators, Anal. Chem., 66, 2943-2948.
12. Wang J., Cai x., Rivas G., Shiraishi H. (1996) Stripping potentiometric transduction of DNA
hybridization processes, Anal. Chim. Acta, 326,141-147.
13. Wang J., Rivas G., Cai x., Dontha N., Shiraishi H., Luo D., Valera F. S. (1997) Sequence-spesific
electrochemical biosensing ofM. tuberculosis, Anal. Chim. Acta, 337, 41-48.
14. Wang J., Rivas G., eai x. (1997) Screen-printed electrochemical hybridization biosensor for the
detection of DNA sequences from the Escherichia coli pathogen, Electroanal., 9,395-398.
15. Palecek E., Tomschik. M., Stankova V., Havran L. (1997) Chronopotentiometric stripping of DNA at
mercury electrodes, Electroanal., 9,990-997.
16. Liu J., Li J., Dong S. (1996) Interaction of brilliant cresyl blue and methylene green with DNA studied
by spectrophotometric and voltammetric methods, Electroanal., 8, 803-807.
17. Fojta M., Doftkova R., Palecek E. (1996) Determination of traces of RNA in submicrogram amounts of
single- or double-stranded DNAs by means of nucleic acid-modified electrodes, Electroanal., 8, 420-426.
18. Erdem A., Kerman K., Meric 8., Akarca U. S., Ozsoz M. (1999) DNA electrochemical biosensor for the
detection of short DNA sequences related to the Hepatitus B virus, Electroanal., 10,586-588.
19. Erdem A., Ozsoz M. (2001) Interaction of anticancer drug, Epirubicin with DNA, Anal. Chim. Acta, 437,
107-114.
20. Erdem A., Kerman K., Meric B., Akarca U. S., Ozsoz M. (2000) A novel hybridization indicator
methylene blue for the electochemical detection of short DNA sequences related to the Hepatitis B virus,
Anal. Chim. Acta, 422,139-149.
21. Jelen F., Erdem A.. Palecek E. (2002) Cyclic voltammetry of Echinomycin and its interaction with
double-stranded and single-stranded DNA adsorbed at the electrode, Bioelectrochem., 55,165-167.
22. Marrazza G., Chiti G., Mascini M., Anichini M. (2000) Detection of human apolipoprotein E genotypes
with DNA electrochemical biosensor coupled with PCR, Clin. Chem., 46, 31-37.
23. Millan K. M., Mikkelsen S. R. (1993) Sequence-selective biosensor for DNA based on electroactive
hybridization indicators, Anal. Chem., 65, 2317-2323.
 303
24. Millan K. M., Spurmanis A. J. ,Mikkelsen S. R. (1992) Covalent immobilization of DNA onto glassy
carbon electrodes, Electroanal., 4, 929-932.
25. Wang J., eai x., Rivas G., Shiraishi H., Farias P. AM., Dontha N. (1996) DNA electrochemical
biosensor for the detection of short DNA sequences related to the human immunodeficiency virus, Anal.
Chern., 68, 2629-2634.
26. Wang J., Fernandes J. R., Kubota L. T. (1998) Polishable and renewable DNA hybridization biosensors,
Anal. Chern., 70, 3699-3702.
27. Larsson K., Kriz K., Kriz D. (1999) Magnetic transducers in biosensors and bioassays, Analusis, 27,
617-621.
28. Sole S., Merkoci A, Alegret S. (2001) New materials for electrochemical sensing III. Beads, Trends in
Anal. Chern., 20,102-110.
29. Gehring AG., Brewster J.D., Irwin P.L., Tu S.I., Van Houten LJ. (1999) I-Naphthyl phosphate as an
enzymatic substrate for enzyme-linked immunomagnetic electrochemistry. J. Electroanal. Chern .. 469, 27-
33.
30. Dequaire M., Degrand C, Limoges B. (1999) An immunomagnetic electrochemical sensor based on a
perfluorosulfonate-coated screen-printed electrode for the determination of 2,4-dichlorophenoxyacetic acid,
Anal. Chern., 71, 2571-2577.
31. MCB 1200 MixSep Technology, http://www.sigris.comlmcb_technology.html.
32. Wang J., Kawde A-N., Erdem A, Salazar M. (2001) Magnetic bead-based label-free electrochemical
detection of DNA hybridization, Analyst, 126. 2020-2024.
33. Wang 1., Xu D., Erdem A, Polsky R., Salazar M. (2002), Genomagnetic electrochemical assays of
DNA hybridization, Talanta, 56, 931-938.
34. Palecek E., Billova S., Havran L., Kizek R., Miculkova A, Jelen F. (2002) DNA hybridization at
microbeads with cathodic stripping voltammetric detection, Talanta, 56, 919- 930.
35. Wang J., Kawde A-N. (2002) Amplified label-free electrical detection of DNA hybridization, Analyst.
127,383-386.
MECHANICAL PROPERTIES OF OXIDE COATINGS HAVING SENSOR
APPLICATION

D.GRABCO, N. PALISTRANT, R. JITARU, E. RUSU

Institute 0/Applied Physics 0/ Moldavian Academy o/Sciences, 5 Academy

str., Chisinau. MD-2028, Moldova, e-mail: daria.grabco@phys.asm.md

Abstract

An experimental investigation of the mechanical properties of planar structures of

ZnO/Si for different indentor loads has been described. The influence of intensive solar-
like light and temperature on hardness and plasticity parameters has been also
investigated. It is obtained that the light exposure and temperature tn:atment cause not
great changes of the microhardness and plasticity properties.

Keywords: nanoscale ZnO/Si planar structure, fabrication, microstructure,

mechanical properties

1. Introduction

Planar structures have a wide application in modem engineering. In particular, planar

structures like ZnO/Si are of interest as potential materials for optically-transparent
electrodes in photonic devices such as solar batteries, photodetectors and luminescent
diodes.
The successful use of planar structures is impossible without the study of their
mechanical characteristics. Indentation hardness testing of solids provides a simple and
quick method to obtain information about the mechanical properties of materials. These
properties mainly include microhardness and microbrittleness. A particular advantage of
the indentation hardness testing is that the investigation can be carried out on small
specimens only a few micrometers of size. During the last two decades, significant
advances have been made in applying surface engineering technique:s to materials for
improving their mechanical and chemical properties. These techniques include coating,
ion implantation, multilayer superlattices and others. To test the mechanical properties
of engineered surfaces and thin coatings deposited on substrates, nanoindentation
hardness machines and techniques have been elaborated which allow the investigations
to be carried out using low indentor loads over a range from !-tN to mN [1,2].
305
Y. G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 305-310.
© 2003 Kluwer Academic Publishers.
306

The proposed work involves the study of the mechanical characteristics of

planar structures consisting of composites of the nanoscale films deposited on
crystalline substrates (ZnO/Si) and the influence of light irradiation and temperature
treatment on them. The results of mechanical behaviour investigations can be used for
the optimisation of the manufacturing process of the nanocomposites in order to obtain
the high-quality structures with the properties required for the practical applications.

2. Experimental details

The fabrication process of the ZnO/Si planar structures included three consecutive
stages. At first, a conducting n+-ZnO layer was formed on a p-Si substrate by
organometallic deposition; zinc acetylacetonate Zn(AcAc) was used as a source of Zn.
The ZnO layer was deposited in a vertical reactor; the substrate temperature was
maintained within 300-350"C. In the capacity of supporting gas, Ar saturated with Zn
(AcAc) vapours and O2 were utilised. During n+-ZnO layer deposition, the temperature
of the Zn (AcAc) source was 120'C; a flux of Ar mixed with O2 in the proportion of
[Ar]/[Ar] + [02])=0.5 was used. The obtained n+-ZnO layer thickness was within the
limits from 150 nm to 700 nm; the electron concentration in these layers was (5-
8)xl0 19cm·3, and electron mobility was ~ 50cm2 y-1S·I. An additional n-ZnO layer of
I 00-300A 0 thickness was then deposited using electron-beam evaporation of n-ZnO
crystals (p= 102Qcm).
The microindentation hardness testing was made using a PMT-3 device and Vickers
diamond square pyramid as indentor. The mechanical parameters were calculated using
the usual formulae [3,4] with an indentor load P between 0,2 ~ 2N.
The brittleness of ZnO/Si samples was determined by measuring the summary
amount (N) of acoustic emission (AE) signals excited in the tested material under a
concentrated loading. The application of this method is based on the concept that the
deformation and fracture processes are the sources of the acoustic signals bearing the
profound information about the internal structure transformation. In the case of quasi-
static indentation the acoustic emission signals were registered at indentor penetration
into material (N I) and as a summary account of AE signals accumulated under
loading/unloading process (N2). The value tJ.N2= N2 - NI indicates the AE signal sum
appearing at the unloading stage.
Taking into account that the ZnO/Si planar structures, in the case of a practical
application, can undergo an intensive light exposure, be influenced of high or low
temperatures, it was of interest to carry out the respective experiments. To study the
influence of an intensive light exposure the specimens were irradiated during 30
minutes using a quartz-mercury lamp PRK. Heat treatment of samples was made by two
means: annealing over 6 hours at T=573 K and exposure during 1 hour in liquid
nitrogen (T=77 K). The optical microscopes Neophot, Amplival, microinterferometer
MII-4 and scanning electron microscope Tesla BS-350 were utilised for the surface
microstructural investigations.
 307

3. Results and discussion

The ZnO/Si planar structures have a composite structure, in which one can
distinguish several substructures: a thin nanostructure coating, the substrate and the
interface zone between the film and substrate. The study of microstructure has shown
that ZnO layers possess a smooth surface with a homogeneous microstructure. The
layer thickness ( t ) varied within the limits of 150 nm to 700 nm. The thickness of the
film is maximal in a central part of the sample and decreases to edges. The
microindentation was carried out in various parts of the compound: in the center, middle
and close to the margins. It was established that the composite zim: oxide layer
deposited on the surface of a single crystal silicon has a higher micro hardness than the
bulk ZnO and Si single crystals separately. So, for the composite with t ,= 700 nm, the
hardness of zinc oxide is 2.3 GPa for a load of P = 0.03 N, that is more than
microhardness of the ZnO bulk crystals (H = 1.8 GPa) and is constant in all of the
investigated regions. The microhardness value increases with applied load increase (Fig.
I). This result can be attributed to the substrate influence; the substrate contribution
increases for more heavy loads because of indentation depth increase. The
microhardness value becomes equal to 9.5 GPa for P= 0.5 N that corresponds to the
microhardness value of Si substrate and exceeds it when the load is equal to 1.0 N.
The obtained result is indicative that the ZnO coating improves the strength properties
of the studied planar structure.

h, nm
700 1200 1800
12

10 ... ........f:!.SJ . .

ro
Q.
<.?
I 6

~?""
0.0 0.2 0,4 0,6 0,8 1,0
P, N

Figure 1. The microhardness dependence (H) of the planar structure ZnO/Si vs applied load;
h is the indentation depth for the respective load.

In addition to that it was found that the deposition of thin nanoscale films on the
surface of hard and brittle crystals reduces the brittleness of the latter [5, 6]. The results
of AE signal measuring confirm this conclusion. So, it was found that the AE signal
number (N), N2, ~N), appearing during the micro indentation of the ZnO/Si structures, is
less than for the identical deformation of the Si bulk crystals: the AE imensity for the
former objects is approximately two times smaller than that for second one:s.
308

Simultaneously it was established that the ZnO layers possess a good adhesion to the
substrate. This fact was confirmed by the relief topography around indentation: layer
integrity was saved for all applied loads. Only at the high load (P=1.0 N) thin cracks
appear near the indentation. This crack appearance is connected with the Si support
brittleness appearing at these loads.
N
H,GPa 6000 • 2
14 5000
13 4000
12 ~ 3000 ~

11
--i--l
-j
2000 ~
10 ~ "
'------.. 1 1000
-------------- ------.,.. 2
9 :+~-'--r P,N o i
o 0.4 0.8 1.2 1,6 o 0.4 0.8 1.2 1.6 2

a b

Figure 2, The dependence of microhardness (a) and the summary account of acoustic
emission signals (b) vs indentor loads for ZnO/Si planar structures .
Curves: I - as-grown sample; 2 - sample which suffered the strong light irradiation,

Results somewhat different from those presented in Fig. I were found for the ZnO
thin layers (t '" 150 nm) (Fig. 2 a, curve I). In this case, the curve has a hand-downing
shape that can be also explained by growing contribution of the Si crystal brittleness at
high loads. This suggestion was confirmed by studying the indentation microstructure
and by measuring the AE signals (Fig. 2, b, curve I). At the same time it can be noted
that all microhardness values on curve I (Fig. 2 a) exceed the H values of the Si support
(cf. Fig.1 with Fig.2 a, curve I). This illustrates an improvement of the mechanical
properties of the substrate material when it is covered by a more plastic layer.
The intensive light irradiation lowers the micro hardness of the composite structure
(cf Fig. 2, curves I and 2) at the expense of additional brittleness (P>0.5 N). This as
well as the above mentioned was confirmed by a microstructural investigation and
acoustic emission signal estimation (cf. Fig. 2 b, curve I and 2).
Note one more peculiarity observed on Fig. 2 a. The shape of both curves possesses
a nonmonotonic nature and their arrangement is different in the small load range. This
effect is especially pronounced at P<0.5 N. Here, the microhardness of as-grown planar
structure increases with load decrease, but H of sample which undergone an intensive
light exposure, decreases together with load diminution. Herewith, how it is seen from
Fig. 2 b, the microhardness decrease can not be connected to the brittleness increase as
the AE number at these loads has practically not changed. On the one hand, such an
effect can result from the change of the ZnO layer microstructure state softening the
layer in comparison with the hardness increase of initial material. On the other hand, the
curve shape difference can be related to the increase of the experimental error when
indentation dimensions become too small (P<0.5 N). For this effect clarification it is of
 309

interest to carry out investigations using a nanohardness tester. Such experiments are
planned for the near future.
The heat treatment of ZnO/Si planar structures for 6 hours at 573 K did not bring any
essential changes of both the hardness and AE signals account. However the cooling in
liquid nitrogen for I hour caused the micro hardness decrease and the AE signals to
increase (Table).

TABLE . The change of mechanical parameters of the ZnO/Si planar structure after the high and low
temperature treatment. P~.5 N.
Type of
temperature H,GPa N, number of AE
treatment signals

The initial sample 11.0 840

6 hours annealing 11.0 1000

at 573 K

1 hour cooling in 7 .. 5 1400

liquid nitrogen

Microstructural investigation of the surface relief around the indentations allows to

attribute this effect to the change of the layer state and durability of their adhesion to the
substrate. We can see the different image of indentations made on sampks which have
undergone a heat and cooling treatment ( Fig. 3). The material pile-up is seen in the

c
Figure 3. Optical images of micro indentations (a,c) and their interferograms (b, c) made on the ZnO/Si
planar structures which have undergone to the high (a, b) and low (c, d) temperature treatment.
p= 0.5 N.
310

neighbourhood of the indentations made on the annealed sample (Fig.3 b) and the
wrinkled surface can be found in the indentation region on the cooled specimen (Fig.3
c, d). The interferogram lines rise near the indentation in the first case and their smooth
course is seen on the Fig. 3 d. This results show that the behaviour of the ZnO/Si
structures after the temperature treatments have a different microdeformation. However
it is required to indicate that low temperature treatment did not break the layer integrity
and caused no microstructural modification that indicates the good quality of the
fabricated planar structure.

4. Conclusion

The study of the mechanical characteristics of planar structures consisting of nanoscale

films deposited on crystalline substrates (ZnO/Si) and the influence of light irradiation
and temperature treatment on it is carried out. It is shown that the obtained material
possesses good microstructural and mechanical properties. The results of mechanical
behaviour investigations can be used for the optimisation of the manufacturing process
of the nanocomposites in order to obtain high-quality structures with properties required
for practical applications.

References
[I]. PETHICA, J., HUTCHINGS, R., and OLIVER W. (1983). Hardness measurement at penetration
depths as small as 20 nm. ,Phil. Mag. A, vol. 48, no. 4, .593-606.
[2]. CASTELL, M., SHAFIRSTEIN, G., and RITCHI, D. (l996). The indentation response ofGaAS-AIAs
heterostructures. Phil. Mag. A, vol. 74, No.5, 1185-1194.
[3]. BOYARSKAYA, Yu.S., GRABCO, D.Z., KATS, M.S. (l986). Physics of Microindentation Processes,
edited by "Stiinta", Kishinev, 294 p.
[4} BERKOVICH, V.F., BOGOMOLOV, N.I., BABANIN, A.V. et. al. (1974). In News in
Microindentation Tests, edited by "Nauka", Moscow, 119 - 225.
[5]. GRABCO D., RUSU E., GROSESCU I. (2000) Microstructure and Strength Properties of ZnO/Si Planar
Structures. In Proceedings of National Conference ofPhysics, Constanta-2000, Romania.
[6]. GRABCO D.Z., MEDINSCHI M.I. (l995), Micromechanical Properties of the Epitaxial Structures on
the Base of InP Crystals. izvestia AS RM. ser. Phys.& Techn., N I, pp.28-31 (in Rom.)
APPROACHES FOR STRUCTURED IMMOBILISATION OF RECOGNISING
ELEMENTS ON THE TRANSDUCER SURFACE OF BIOSENSORS

V.M. STARODUB1,1.J, A.V. NABOK4, N.F. STARODUB\ W.

TORBICZ1
IA.V.Palladin Institute of Biochemistry, National Academy
of Sciences, Kiev, Ukraine
2Institute of Biocybemetics and Biomedical Engineering,
Polish Academy of Sciences, Warsaw, Poland
3N.V. Organon, Oss, the Netherlands
4Sheffield Hallam University, School of Engineering, Sheffield, England.

Abstract In this article a short analysis of the existing approaches on the

immobilisation is given as well as attention is paid to the experimental
data obtained by the authors at the development of different types of
immune and enzyme sensors. Among the big set of examined approaches
the use of polyelectrolytes (PE), lectins, protein A from Staphylococcus
aureus, different thiol groups and aminosilanes (AS) as well as the
combination of these substances as an intermediate components between
recognising structures and surface is considered in detail. Immobilisation
of biological molecules on the silicon surface through its preliminary
activation by the AS with glutaraldehyde (GA) or alone with GA is
separately discussed.
Keywords: Biosensors, sensitive elements, immobilisation, structuring.

1. Introduction

The successful development of analytical tools for bio-sensing, which allows to meet
practical demands for bio-sensors, namely, their high sensitivity and selectivity, fast
response, cost effectiveness and suitability for field analysis, is tightly connected to
solving a number of tasks related to the immobilization of biological materials on the
transducer surface. It is very important for this complex problem to provide not only the
reliable integration and preservation of a native structure of biological material but also
to achieve the situation, when recognizing elements are presented in its maximal density
on the transducer surface, and their active centers are exposed to solution for the
interaction with analytes. These are the main conditions to provide maximal sensitivity
of bio-analysis.
At the present time extended information about the structure of biological molecules
on different surfaces was obtained. It is a background for a new generation of
biotechnology, which should materialize the results of theoretical investigations into
practical achievements. The following conditions have to be fulfilled in order to design
311
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 311-325.
© 2003 Kluwer Academic Publishers.
312

a right combination of biological molecules with the transducer surface: (I) maximal
preservation of the activity of bio-molecules as recognizing structures; (2) high density
of these structures on the transducer surface; (3) exposition of the majority of active
centers into solution; (4) high level of adhesion of bio-molecules to transducer surface;
(5) optimal conditions for diffusion of substances to be analyzed; (6) simplicity of the
procedure of sensitive layer formation; (7) combination of the formation of sensitive
layer with the transducer manufacturing; (8) long term activity of bio-sensors in both
working and storage conditions; (9) the formation of reusable or easily replaceable
sensitive layers on the transducer surface. These demands can be fulfilled by
combination of a number of approaches: (1) physical approach, which includes
adsorption on porous materials, such as porous silicon or the use of polyelectrolytes; (2)
chemical approach, which comprises: (a) direct covalent binding of biological material
to surface via BrCN or bifunctional agents, e.g. glutaraldehyde (GA); (b) the use of
intermediates such as 3-aminopropyltriethoxysilane, or thiols; (c) entrapment of
biological molecules into gels, e.g. ion-, photo-, electro- and beam polymerization; (3)
physical-chemical approach, which is based on the Use of membranes with chemically
attached biomaterial (e.g. cellulose, Biodine C, N and others).
For directed structuring of the immobilized material it is possible to use some other
methods. In case of immobilization of the immune components, the most widespread
and simple method is introducing intermediate layers of lectins or protein A from
Staphylococcus aureus. In this article we demonstrate the application of existing
approaches in practice for the directed functionalization of transducer surfaces as well
as for the structuring of biological molecules through the use of physical adsorption,
covalent binding, polyelectrolytes self-assembly, as well as, the use of lectins and
protein A from St. aureus as intermediate layers between surface and receptor
molecules.

2. Structuring biological molecules with the direct and indirect

functionalization of the transducer surface.

The immobilization of biological molecules by direct physical adsorption is often used

in practice for the development of biosensors. We have realized this approach at the
development of the immune sensor based upon the photoluminescence of porous silicon
(PhL ofPS) [1-3). Because of well developed surface area ofPS and its high adsorption
activity, this approach provides the required density of immobilised biological
molecules. Prior deposition PS transducers were rinsed several times in distilled water
and then in alcohol and finally dried at room temperature. For spontaneous sorption of
antibodies or antigens, PS samples were immersed into solution of biological material
of a certain concentration solutions in 20 mmoVI (pH 7.3) phosphate buffer, containing
0,14 M sodium chloride (PhBS) and retained there for 15-60 min. After that PS surface
was washed with PhBS buffer. According to our previous results [4], the time of 60 min
was chosen as the most suitable for our experiments. Such conclusion was made,
because the periods of 15 and 30 min of the immobilization of biological material were
rather short to have a quantity of antibodies the surface sufficient for the sensitive
detection of the analyte. At the same time, periods of 60 and 90 min were sufficient for
an efficient antibody immobilization and showed similar results (Fig. 1).
 313

The approach of physical sorption of one of the immune components can be very
simple realized by using the sheets or strips of nitrocellulose (NC) as a solid phase [5,
6]. In this case, it is enough to immerse NC in bio-material solution of the concentration
of 0.1 Ilglml. The sorption lasts for about 1 hour at room temperature. After that the
remaining free functional groups should be blocked. The level of immobilized material
depends on the capacity of NC. Moreover, NC surface can be special activated for
covalent binding of biological molecules. In spite of simplicity of the immobilization
procedure, some problems still exist with providing a good adhesion between the solid
phase and transducer surface. Unfortunately, not in all cases it can be successfully
solved.
Lx. %
lOO~~~----~------~

90

80

70

60

50

40 ~"+,-.........~.,......................~,...........,.........j
o w ~ W ~ m ~ •
~min

Fig. I Changes ofChL intensity at the immune complex fonnation on the PS surface in dependence on
different time of immobilization of specific antibodies. 1- 4 - time of immobilization: 15,30,60 and 90 min
respectively.

We tried to use the procedure of physical sorption of biological molecules when

developed immune sensors were based on the ion-selective field-effect transistors
(lSFETs) [7] and surface plasmon resonance (SPR) [8]. Unfortunately, immobilization
of biological molecules by physical adsorption in those cases was ineffective. In any
case it is necessary to pay attention to the fact that the sorption of some biological
molecules from complex mixtures is not a simple procedure. It was revealed in so-called
Worman effect [9], that the amount of immobilized proteins on the surface is
determined by both the duration of the procedure and composition of mixture, from
which individual molecules are immobilized. This effect was checked up in course of
immobilization of albumin, fibrinogen and immunoglubulin (IgG) from mixtures
containing all these components [10]. It turned out that albumin immobilizes on the
solid phase more quickly than other components, but IgG and, particularly, fibrinogen
were able to displace albumin molecules. Generally speaking, the process of primary
immobilization of individual proteins from the mixture is very complicated and depends
on number of factors, including molecular weight of participated components, the
duration of the process, hydrophobicity of the surface and other components [11, 12].
At the same time, it was established in [13] that the treatment of the surface in
triethylene-glycol-dimethyl-ether prevents the sorption of proteins.
For the reliable obtaining oflayers of biological molecules structured and assembled
with a high regularity it is often used the Langmuir-Blodgett technique (LB) [14], which
allows the transfer of the preliminary formed monolayers of bio-molecules on the solid
phase. The classical execution of this technique was recently modified by using
314

amphiphilic polymers or cellulose derivatives. With the help of the LB technique it is

possible to form biological layers of nanometer thickness on the transducer surface. The
latter is very important for enzyme sensors as well as for different affinic-sensors,
including immune sensors.
The procedure of covalent immobilization of antibodies on the surface of LB layers
on the silicon transducer with the help of cyannamoylthrimethylsilylether cellulose
(CTMCE) includes the number steps [15]. The fist step consists of a thorough
preparation of the surface by subsequent treatments in chloroform, ethanol and in the
solution of octadecyltrichlorsylane for surface hydrofobization followed by washing in
polytrimethylsilylether and ethylacetate. CTMCE-cellulose was obtained by substitution
of lateral groups at the trimethylsilylether of cellulose. As a second step, monolayer of
CTMCE-cellulose was formed by stratification of its chloroform solution on the water
surface and than the obtained layer was transferred on the silicon surface by method of
vertical immersion. Then the layer of CTMCE-cellulose was stabilized in the presence
of osmium oxide and sodium peroxidate to form of aldehyde groups. Finally, antigens
were covalently bound to the aldehyde groups through their amino residues. The
sensitivity of such optrode sensor prepared with the above recipe and based on the
principle of evanescent wave with the fluorescent label (Cys5) was found to be of about
1 ng/mL.
The classical version of the surface modification of the SPR immune biosensor is
the use of carboxymethylated dextran (CMD) treated by N-hydroxysuccinimide and N-
ethyl-N/(dimethylaminopropyl)carbodiimide to form N-hydroxysuccinimide esters for
the covalent fixation of biological material [16, 17]. This approach was exploited for the
fabrication of special SPR cartridges by BIA-core firm (Uppsala, Sweden). The protein
solution with low ionic strength and pH values lower than isoelectric point was slowly
passed through the cartridge with activated CMD. At these conditions protein is
concentrated by electrostatic strength and reacts with active groups of CMD. Non-
coupled binding sites are blocked by ethanolamine hydrochloride. Usually, the
procedure takes 30 min to allow immobilization of about 70 fmol/mm 2 of proteins. The
quantity of immobilized material can be varied by altering ionic strength, pH, the
concentration of reagents and the duration time of individual steps. This approach was
used in [18] for the detection of ouabain, ouabagenin and strofantidin, and the
concentrations of 5xlO- 1O , 7xlO- IO and 2xlO-8 M, respectively, were registered for the
above compounds. A large number of data were collected on the application of these
types of membranes for different investigations and, particularly, for the determination
offolding-defolding ofluciferase at the various conditions [19].
Some other versions of surface modification were exploited for the formation of
sensitive layers for affinic biosensors, namely, preliminary silanization of silicon
surface with the help of [3-(2-aminoethyl) aminopropyl] trimethoxysilan (AEAPTMS),
aminobutyldimethylethoxysilan (ABDMES), n-octadecyltrichlorsilan (ODTClS) or
other similar compounds with their subsequent activation by glutaraldehyde (GA) or by
other bifunctional reagents. It was studied in detail the application of silanization, which
was accomplished in two steps: the sorption on the surface and the formation of
appropriate covalent binding [20). An optimal regime at 28° C was established for the
above process. The thickness of 0,1-100 !-lm was achieved for the ODTClS according to
atomic force microscopy (AFM) data.
 315

It is well known that the thiolalcanes allow to carry out effective and stable
modification of metallic surfaces. They were used for the development of bioaffinic
sensors based on SPR and piezo-crystals [21-25]. Detailed investigations of the bonding
type between the sulfur of alcanthiols and metallic surfaces showed that these bonds are
covalent, and sulfohydryl proton is admittedly realized in forms of H+, 1I2H or H20 2.
The use of alcanthiols is tightly connected with the problem of their oxidization during
the exposure of prepared surface to the air. As a rule, organothiols under such influence
are tend to form sulphynates and sulphonates [26].
One of the possibilities to solve problem of immobilization of immune components
on the metallic surface lies on the use of biotin-avidin pairs (immobilized biotin and
biotinilated antibodies were connected through avidin) [27], the second one - by the
application of two different conjugates (biotin with long-chain alcanthiols and
antibodies labeled by streptavidin) [28]. Both methods mentioned above provide a
simple solution because Ig does not have free SH-groups. Disulphide bonds, which
stabilize the molecular conformation, are situated between heavy and light chains, and
therefore they are masked. It requires conjugation of the immune components with the
additional substances, which have these groups. In this case, RNase was chosen as one
of suitable component [29]. At first, metallic surface was treated by gelatin to form
mercapto-bounds and then conjugated RNase-antibodies were immobilized in the
reaction of thiol-disulphyde exchange. To avoid conjugation of antibodies with SH-
contented components, an attempt of their immobilization was made after splitting
F(abirfragments with subsequent light reduction of SH-groups to provide a possibility
of mercaptane formation [30]. The sensitivity of SPR immune sensor with the selective
layer prepared with the above technique was in the range of several microgram per I m!.
The surface has undergone regeneration three times in the buffer solution with low pH.
Some combined versions of immobilization of biological molecules were proposed
on the basis of self-assembled layer of thioalcanes and immune components modified
by hexahystidine sequence [31]. This approach was taken from the well developed
metal-affinic chromatography [32]. The realization of this approach promises good
results with reversible immobilization and without any conformational changes of
biological molecules according to the data of infrared spectroscopy [33]. Although the
formation of mercapto-bonds of biological molecules with metals is a very promising
way, there is a large probability that the sensitive surface may quickly lose its activity
due to the washing out conjugated Ig during the analysis of real solutions, which may
contain substances with free SH-groups and thus will be able to take part in the reaction
of thiol-disulphide exchange. Additionally, the reduction of SH-groups on F(ab)'2-
fragments does not exclude destroying of structure of antigen-binding site as result of
the destabilization of bounds between variable domains of light and heavy chains of
antibodies.
At the development of the SPR immune sensors we examined the efficiency of the
immobilization of immune components on the bare surface as well as on the surfaces
treated by thiols, in particular, dodecanthiols [8]. It was stated that the physical sorption
of immune components was increased up to 1,8 times after the treatment of gold surface
by dodecanthio!. Probably, this is a result of the optimization of a
hydrophilic/hydrophobic ratio of the surface. Nevertheless, in that case recognizing
elements were not structured, and a number of binding sites were in an inactive state
(see the explanation below).
316

In our experiments (Fig. 2) on the direct immobilization of one of the immune

components on the silicon nitride surface, a following routine was undertaken [7, 33-
35]. A drop of the 2,5% water solution of the GA was placed on the prepared surface.
After drying the sample for 90 min at room temperature (25°C), the GA was placed
again following the same procedure. Then, after another 90 min drying the surface was
washed with the distilled water. After this, the transducer surface was activated, and it
was able to bind any proteins. In that procedure, it is necessary to use protein solution in
concentration of 20 )lglml in 10 mM phosphate buffer (PhB) and to keep it on the
surface for 1 hour at humid conditions. The treatment of the chip with the GA was
carried out at room temperature. Finally, the surface was kept overnight at 4°C, and then
after vigorous washing in PhB was immersed into glycine solution (10 mglml in the
PhB and the room temperature) in order to block non-reacted groups of the GA. After
30 min at room temperature the surface was washed with the PhB [7]. So, it was
demonstrated [7, 36, 37] that the silicon nitride surface could be directly, e.g. without its
preliminary modification, activated by GA in special conditions.
Another method of direct immobilization of one of the components of the immune
reaction at the silicon surface or on the inner walls of a special capillary tube was
fulfilled as follows. At first, before the immobilization, the surface was activated with
cyanogen bromide at 150° C in the presence of triethylamine. Then, the sample was
immersed into the antigen or the antibody solution for about 30 min and then into 0.1 M
glycine solution for 40 min to block the remaining free functional groups [33-35]. It is
necessary to mention that the efficiency of both processes is approximately the same,
but the first procedure is more simple in comparison to the second one, and the former
does not demand any special conditions.
In the past the sol-gel technique was widely spread for the fabrication of biosensors
[38]. Typically, a low-molecular weight metal alcoxide precursors molecule (usually
tetramethoxysilane: TMOS; Si(OCH3)4 or tetraethoxysilane (TEOS; Si(OC2Hs)4) was
hydrolyzed first in the presence of water, acid catalyst, and mutual solvent, generally
ethanol. Hydrolysis caused the formation of Si-OH group on the solid surface. On the
next step, Si-O-Si polymers were formed in a condensation reaction. As a result a
colloidal suspension (sol) was formed on the solid substrate and a gel are in solution.
Then solvents were removed from the interconnected porous solid network (Fig. 3).
 317

Silicon oxide surface

Steps:
I. Fonnation of silanol bonds on the surface by the treatment in the water
vapor phase;

2. BrCN- activation of hydroxyl groups of silicon:

-OH + CNBr - O,C=NH

3. Covalent binding of proteins:

O,C=NH + H,N-R - 0,= C-NH-R + -OH
Silicon nitride surface
Steps:
I. Activation of imino groups:

(Si-NHh + OCH-(CH,),-CHO _ (Si-N)=CH-(CH,),-CHO

2. Covalent binding of proteins:

(Si-N)=CH-(CH,h-CHO + H,N-R-(Si-N)=CH-(CH,h-HC=N-R

Fig. 2 Approaches for direct covalent immobilization of recognizing biological molecules on silicon
transducer surfaces

The selectivity of silicon-organic polymers with respect to the sorption of biological

molecules can be achieved in two ways [39, 40]. One of them is related to the geometric
modification by alternation the parameters of adsorbent structure under appropriate
conditions of synthesis, when the chemical structure of the polymers remains
unchanged. The second way is based on chemical modification of silicon-organic
polymers either (a) by alternation of the character of the group linked with silicon atoms
(-Si-CH3, -Si-OH, -Si-CH=CH2' Si-(CH2)-CH-NH2 etc.; or (b) by inclusion of hetero-
atoms in the silicon chain (AI, Zn, Cu, Co, Mn, Fe). We have obtained following
modifications of silicon-organic polymers: (1) polymethylsiloxane (PMS), a
hydrophobic adsorbent; (2) silicapolymethylsiloxane (S-PMS), a polymer with a given
ratio of hydrophobic (CH 3) and hydrophilic (OH) groups on the surface; (3)
polymetalmethylsiloxane (PolyMet-PMS), with metal included in the structure of the
siloxan chain; (4) polyheterometalmethylsilixane (polyheteroMet-PMS), with the
structure of the polymer chain containing two different atoms of metals instead of some
silicon groups; (5) PMS with included aminogroups or modified by aminosilanes for
covalent immobilisation of biological molecules with the help of GA.
It was shown in [41] that PolyMet-PMS containing Co-ions is very effective as a
membrane for immobilization of urease on the gate surface of ISFETs. Such membranes
increase the sensor signal in comparison with those where proteins were bound via GA.
Most likely, NH3 ions, formed during enzymatic reaction, bound selectively with Co-
contained organosilane and cause more intensive changed in local pH. The same effect
was observed for ISFET immune sensor with the use of urease labeled conjugates.
In another case, the silicon surface was initially silanized and then activated with
GA. The process of silanization was performed as follows. Aminopropyltriethoxysilane
(APTES) was frozen and placed in a vacuum chamber together with the solid substrate.
The vacuum container was filled with APTES vapor in which the surface was held for
no less then 12 hours. Then it was placed for 3 hours in GA vapors. Finally, the
activated surface was immersed into the solution of one of the immune components for
20-30 min. After washing the surface in PhB the remaining free functional groups were
318

blocked by the treatment in 0.1 M glycine solution. It was stated that the biological
material was bound 1.8 times more effectively when BrCN was used instead of APTES
and GA [33-35].

Hydrolysis
M(OR). + xH,O --+ M (OH). + xROH
M+ Si, Ti, AI, B, Zr, Ce

Condensation
-M-OH + RO-M- --+ - M-O-M- +ROH

Polycondensation
x(- M-O-M-) --+ (- M-O-M-)x

Fig. 3. Steps of sol-gel process [17]

Two possibilities, namely, silanization and thiolation can be combined.

Aminosilanisation of the surface allows covalent immobilization of protein molecules
and their subsequent treatment by pyridyl, which opens a possibility of thiol-disulphide
exchange reaction to happen[42].
Unfortunately silanization does not provide the required density of surface
modification. It is stipulated by sterical restriction and existing lateral reactions which
lead to olygomerization and multi-layer formation [43]. Moreover, very often the
silanized surface becomes strongly hydrophobic, and it promotes to increasing non-
specific effects [44]. Therefore, pure silicon surface, which initially possesses a
negligible non-specific binding, after silanization has become capable of sorption of
biological molecules in concentrations of up to 1,5 nglmm 2 [45]. Such effect can be
decreased by subsequent deposition of hydrophilic polymer, which reduces level of non-
specific interaction but at the same time it decreases the specific reactions too [43, 44).
To avoid of these restrictions, it was proposed [45-47] to modify the surface by
polyethylene glycol (PEG) at certain conditions. For that, the surface was preliminary
treated by pyranha (3 M ofNaOH, 60% of H2 S04 and 40% ofH 20 2) and then activated
by glycidyloxipropyltrimethoxisilane. The level of non-specific adsorption of
ovalbumine on the prepared surface was about 100 pglmm2, and the value of 2 nglmm 2
was achieved by increasing the protein concentration by only 10%. For covalent
fixation of biological molecules, PEG in forms of diamino- and dicarboxyl derivatives
and appropriate cross-linking agents were used. It should be noted that the level of non-
specific adsorption on the surface prepared with the help of PEG does not change even
at the chemical modification of its functional groups. In addition, the stability of the
formed layer is characterized by a hundred cycles of regeneration of the surface after
accomplishing selective interactions. Since immobilized molecules are superficially
situated, their reactions with ligands do not have those restrictions, which took place in
the case of the application of developed dextran. At the same time, since the surface
contains a thin, dense and stable layer of polymer, the sensitivity of the analysis does
not reach the level, which was achieved in experiments with developed dextran [45].
Sometimes the proposed way of surface modification reminds the situation with the
application of thiolalcanes mentioned above. Nevertheless, the later interact with the
surface selectively but spontaneously, and their orientation is stabilized by the inter-
chain interactions. The density of layers based on PEG, in contrary, is forced at the
 319

extreme conditions (high temperature, concentration of biological molecules) and the

orientation of polymer chains in the layer in water solutions is determined by inter
chains repulsion due to their large volume. All these factors lead to low level of non-
specific interaction. It becomes very important, because the elimination of non-specific
effects is one of the main tasks at the development of any methods for biochemical
analysis.
In case of immobilization of antibodies, the use of protein A from St. aureus as
intermediate layer is more preferable in comparison with their direct fixation on the
surface preliminary modified by aminosilanes and activated by GA [7]. Protein A
prevents undesirable conformational changes of antibodies' structure at their interaction
with solid phase as well as provides directed exposition of antigen-binding sites to the
solution and allows to accomplish immobilization of immunoglobulin (Ig) G straight
from the serum blood. The latter allowed us using total antiserum to some herbicides to
develop very sensitive electrochemical immune sensor based on ISFETs and electrolyte-
insulator-semiconductor structures. Unfortunately, not all specific classes and
subclasses of Ig can interact with protein A [48]. In that case lectins are more suitable,
since they bind carbohydrate compounds associated with Ig in the region of second
constant domain. Although it is necessary to choose appropriate lectins for certain Ig
because their individual varieties have the appointed species of carbohydrate
compounds with different types of end residue.
We have investigated the interaction between human IgG and lectins by SPR
immune sensor [49]. It was found that human IgG reveals maximal affinity to wheat
germ (WGA)-Iectin in comparison with phaseolus vulgaris (PLA)- and helix pomatia
(HPA) lectins. The density of the IgG attached to the surface treated with WGA-lectin
was higher than that for the bare gold surface (Fig. 4). There is necessary to notice that
the density of IgG immobilised on the bare gold is higher than that on the surfaces
treated by HPA- and PLA-Iectins. Probably, they modified the surface in such way that
it stands very hydrophilic from one side, and from other side its affinity to
carbohydrates of IgG is not very strong. To reveal the possibility of non-specific
binding of IgG on the surface modified by lectins, the solution of F(ab)2 fragments of
human IgG was introduced in the measuring cell. It was shown that the value of a
resonance angle of the sensor did not change. Moreover, after next introduction of IgG
solution this parameter was remained the same, as it was in case of the use of the
surface treated by lectin alone. If the gold surface was treated by the solution of bovine
serum albumin, it does not bind any IgG. Thus, the obtained results confirm that the
surface with the immobilised lectin has specific interaction with IgG.
Subsequent binding of antigen has produced a maximal response on the surface
treated with the WGA lectin in comparison with that one for both bare gold surface and
that treated by dodecanthiol. It was shown, that lectin treatment of the gold surface not
only increases the density of IgG immobilisation on the surface and also provides
proper orientation of IgG molecules on the surface. Lectin specificity to the
monocarbohydrates located at the Fc-fragment of IgG molecules provides the
orientation of IgG molecules with their variable sites oriented towards the solution and
thus take part in the immune reactions between specific antibodies and antigens.
320

3000

2500

2000

1500

1000

500

35 70 105 140 215 250 285

t, sec

The same experiments were carried out for other pair of antigen-antibody: mouse
monoclonal antibody to myoglobin and human myoglobin. It was stated that mouse
monoclonal antibody reveals maximal affinity to HPL-lectin. The surface treated by
HPA- and PLA-lectins exhibited larger shift of the resonance angle than that of bare
gold surface. In this case the value of the resonant angle shift for myoglobin was in the
range of 1400 sec; and at same time, for bare gold surface, it was in the range of 1000
sec. Similar results on the relationship of the efficiencies of antibody immobilisation on
the surfaces treated by dodecanthiol and lectins were demonstrated. Specific lectins
provided proper orientation of antibodies on the surface and thus increased immune
sensor signal. It was stated that features of carbohydrate component, which is located in
second constant domain of IgG molecules, provide their different affinity to lectins.
According to our experiments, rabbit antibody reveals maximal affinity to HPA-lectin.
At the same time, pig antibody, similar to human IgG, shows the greatest affinity to
WGA-lectin.
Therefore, some lectins are able to increase the response of the SPR based immune
sensor by increasing the density of the immobilisation of IgG molecules and improving
their orientation on the surface. Depending on the carbohydrate contents and the
specificity of IgG, specific conditions were selected to increase the immune sensor
response as compared to that for bare gold surface or gold surface preliminary treated
with dodecanthiol.

3. Self-assembled polyelectrolytes and structuring of biological sensitive

elements.

In the present time, self-assembled polyelectrolytes are widely used for biological
material immobilisation or for the separation of initial components from the products of
immune reaction. The latter can be achieved in a simple way, however the chemical
synthesis of desirable conjugates of polyelectrolytes with biological molecules is
sometimes very complicated [50].
Obtaining multi-layered films with the help of small organic molecules found a wide
dispersion. The matter is that these molecules form non-solving polymer during the
sorption on the charged surface. At that time this polymer is able to electrostatic
 321

sorption of other molecules, which have an opposite charge and are able for the
formation of non-solving film too. In such way a number of self-assembled layers, with
the thickness from 5 to 500 nm can be deposited [51-55]. The increment of film
thickness at the self-assembly was named as a growth step. The concentration of
poly ions at this process should be more than 0,2-0,5 mglmL [52-56]. It is often used
polyalylamine hydrochloride (PAA), polyethylenemin (PEl), polyvinylsulphonate
potassium (PVS) and polysterensulphonate (PSS). PSS and PVS are used as water
solution at pH 2 and concentration of 3-5 mglmL. PAA and PEl are used at pH 8 [51,
52]. It is possible with the help of PAAIPSS-polyelectrolytes to form films of about 200
layers with the typical growth step of 1,2-1,4 nm (Fig. 5).

~n (a)
\ ~N~I'CI-
Fig. 5. Chemical fonnula ofPAA (a), PSS (b)

The value of growth step can be varied by changing of pH, ionic strength of
solution, solvents, time of adsorption, charge density on the surface of the solid
substrate and number of other factors. Inclusions of NaCI in the concentration of 2
mollL to the solution of PSSIPAA pair increases the growth step from 1,09 to 2,5 nm,
and it does not effect the film quality. Similar results were obtained for other mono- and
bivalent ions, for example glycerol [51]. The increasing of growth steps at these
conditions may be explained by the formation of developed chains of polyelectrolites
[56]. At the same time, increasing of growth steps was not observed at the small salt
concentrations (less than 0,1 mol/L). It is an important difference from the situation
with the LB-layers, where the high level of solution purity is absolutely vital for
successful film formation [51, 57]. It is very important to mention that PSSIPAA films
are able to withstand the heating up to 100°C and are not changed at the storage during
20 days in the dry state. Polyelectrolitethiol films may include layers, which have some
electroconductivity [51].
Cyclo-tetrachromotropylene (CTChTP) can be included in PSSIPAA layers and it
may serve as pH indicator. The level of pH can be altered as a result of interaction of
urease with the substrate [58]. Other authors [59] have obtained polyelectrolyte layers
using thiocyanine dye and polydialylmethylammonia chloride. A possibility of
incorporation of such dyes as eosin and phenol red in the polymer layers formed
sulphonated polyvinyl alcohol as source of anionic groups and poly-NN-
dialkylaminoalkylacrylamide as target of cationic groups was recently investigated [60].
Still, due to the phenol red molecules staying in form of stable salt in positive charged
polyelectrolytes, they are not able to react to local changes of pH. It was demonstrated
in [52] the formation of films with the use of PAA and polylisine, polyuracil (PU) and
polyadenine. The efficiency ofporphirines with attached RuII(bipyhCl to meso-tetra(4-
pyridil)porphyrine as well as diethylaminorthyldecstrane for polyelectrolyte deposition
was examined in [61,62]. For the formation of films containing DNA, which is a strong
polyanion, PAA and PEl it were used. In this work, the concentration of DNA was
322

about 1 mglmL in solution of 0,02 M NaCI at pH 5,5 [52]. The similar approach was
realized by including viral particles in polyelectrolyte layers [58].
The use of polyelectrolytes was found very effective for simultaneous incorporation
of urease or cholinesterase with dyes: CTChTP or thymol blue at the development of
optical enzymatic biosensors targeted the determination of phosphororganic pesticides
and heavy metal ions [63, 64).
The method of polyelectrolyte self-assembly was successfully exploited for the
immobilisation of the protein A as well as the immune pair of human IgG /goat on
human anti-IgG [65,66]. These biocomponents, which are negatively charged in
TrismaBase-HCI buffer with pH 8.3, can be alternated with polycations
(polyalylamine). In this investigation, monolayers of the above biological components
were deposited on the top (PSSIPAA) multi-layer. Series of SPR experiments showed
that pre-adsorption of the protein A molecules improves the orientation of the immune
components (either IgG or anti IgG), and thus increases the sensitivity of the SPR
immune sensor (see the Table). Using polyelectrolyte self-assembled technique with an
intermediate protein A layer allowed us to register. both immune components in very
low concentrations starting from 30 nglml, which is remarkable for the SPR method.
Further improving the sensitivity of the immune analysis could be achieved by
employing interferometry techniques, such as Mach-Zhender interferometer or planar
waveguide. Our preliminary experiments with planar interferometer have shown much
higher sensitivity in the range of 1 nglml.

Table. The experimental SPR data for the effective thicknesses of the adsorbed bio-Iayers.

The thickness of respective layers (nm)

The adsorption sequence with the accuracy of 0.05 nm
protein A BSA IgG anti-IgG IgG
Aufprot.AlBSAlIgG/anti-IgG 0.45 0.7 1.75 3.52
Auf IgGI anti-IgG 1.84 0.35
P ANprot.AIBSAlIgG/anti-IgG 1.0 0.4 3.2 5.1
PANprot.AlBSA/anti-IgGlIgG 0.55 0.3 4.39 1.76
PAN IgGI anti-!gG 2.09 1.81
PAN anti-IgGI IgG 2.33 0.5

As to conformational changes in proteins at their sorption in the solid phase in

general, and on polyelectrolytes, in particular, it depends on their previous state as has
been shown in [67]. According to the results of this investigation, proteins with small
dimensions undergo any considerable changes, but large and conformational labile
proteins may experience substantial changes. To estimate intensity of conformational
changes of proteins adsorbed on some solid phase or introduced into molecular layers it
is recommended [68] to measure tryptophane fluorescence. This approach was realized
at the control of conformational reconstruction in apoazurine, alcohol dehydrogenase,
glyceraldehyd-3-phosphate dehydrogenase and glutamate dehydrogenase. At the same
time it was demonstrated with high resolution of spectroscopy [69] the loss of electron
energy for the characterization of silicon and platinum surfaces covered by monolayers
of silane and thiol as well as for some self-assembled molecular layers. Information
about the use of different physical approaches for the characterization of surfaces was
summarized in a special review [70]. It contains the detailed data obtained for a number
 323

of different surfaces, and the application of these methods for the study of protein
immobilisation is currently of a great importance.
4. AckDowledgemeDts
This work has been supported by the grant from INCO-Copemicus Programs (Contracts N: ICA2-CT-2001-
10007 and ICA2-CT-20oo-10033) as well as INTAS Program (project N 99 - 0995).

References
I. Starodub N.F., Fedorenko L.L., Starodub V.M., Dikij S.P. and Svetchnikov S.V. (1996) Use of the
silicon crystals photoluminescence to control immunocomplex formation, Sensors and Actuators B. 35-
36,44-47.
2. Starodub V.M., Fedorenko L.L., Sisetskiy AP and Starodub N.F. (1999) Control of myoglobin level in
a solution by an immune sensor based on the photoluminescence of porous silicon, Sensors and
Actuators B. 58 (1-3), 409-414.
3. Starodub Y.M., Fedorenko L.L., Starodub N.F. (2000) Optical immune sensors for the monitoring
protein substances in the air, Sensors and Actuators B 68 (1-3), 40-47.
4. Starodub V.M., Starodub N.F. (2001) Electrochemical and Optical Biosensors: Origin of Development,
Achievements and Perspectives of Practical Application. In: Proceedings of NATO ASI "Novel
Processes and Control Technologies in the Food Industry", 7-17 Sept. 2000. Eds. Bozoglu F. et al.
v.338, lOS Press, P. 63-94.
5. Kukla A, Starodub N.F., Kanjuk M.J., Shirshov JU.M. (1999) Multienzyme electrochemical sensor array
for determination of heavy metal ions, Sensors and Actuators 57 (1-3),213-218.
6. Starodub N.F., Kanjuk MJ., Kukla AL., Shirshov Y.M. (1999) Multi-enzymatic electrochemical sensor:
field measurements and their optimization, Analitica Chimica Acta 385,461-466.
7 Starodub N.F., Dzantiev B.B., Starodub Y.M., Zherdev AV. (2000) Immunosensor for the
determination of the herbicide simazine based on an ion-selective field effect transistor, Analytica
Chimica Acta 424,37-43.
8. Starodub N.F., Dibrova T.M., Shirshov Y.M., KoSIyukevych K. V. (1999) Development of the
myoglobin immune sensor based on the surface plasmon resonance, Ukrainian Biochemical Journal 41
(2),33-37.
9. Vroman L., Adams A.L. (\969) Findings with the recordings ellipsometer suggesting rapid exange of
specific plasma proteins at liquid/solid interfaces, Surf Sci. 16, 438-446.
10. Green RJ., Davies M.e., Roberts C.J., Tendler S.JB. (1999) Competitive protein adsorbtion as
abserved by surface plasmon resonsnce, Biomaterials 20, 385-391.
II. Malmsten M., Lassen B. 91995) Ellipsometry studies of protein adsorption at hydrophobic surfaces.
ACS Symp. Ser. 602, 228-238.
12. Green R.J., Davies J., Davies M.C., Roberts e.J, Tendler S.J.B. (1997) Surface plasmon resonance for
real time in situ analysis of protein adsorption to polymer surfaces, Biomaterials 18, 405-413.
13. Mar M.N, Ratner B.D., Vee S.S. (\999) An intrinsically protein-resistent surface plasmon resonance
biosensor based upon a RF-plasrna-deposited thin film, Sensors and Actuarors B54, 125-131.
14. Erbach R., Hoffinan B., Vogel A. (1991) Impedance spectroscopy on Si-based chemical sensors with
Langmuir-Blodgett films, Sensors and Actuators B4, 379-384.
15. Loscher F., Ruckstuhl Th., Seeger S. (\ 998) Ultrathin cellulose based layers for detection of single
antigen molecules,,Adv. Mater. 10, NI3, 1005-1009.
16. Sjolander S., Urbaniczky e. (\991) Integrated fluid handeling system for biomolecular interaction
analysis. Anal. Chem. 63, 2338-2345.
17. Johnsson B., Lofas S., Lindquist G. (\991) Immobilization of Proteins to a carboxymethydextran-
modified gold surface for biospecific interaction analysis in surface plasmon resonance sensors, Anal.
Biochem. 198,268-277.
18. Laricchia-Robbio L., Balzan S., Ghione S., Montah u., Revoltella R.P. (\ 998) Detection of digitalis
compounds using a surface plasmon resonance-based biosensor, Biosensors and Bioelectronics 13,
1055-1060.
19. Zako T., Harada K., Mannen T., Yamaguchi S., Kitayama A, Ueda H., Nagamune T. (2001)
Monitoring ofthe refolding process for immobilixed firefly luciferase with a biosensor based on surface
plasmon resonance, 1. Biochem. 129, 1-4.
20. Loscher E, Ruckstuhl Th., Seeger S. (\998) Ultrathin cellulose based layers for detection of single
antigen molecules, Adv. Mater. 10, NI3, 1005-1009.
324

21. Boitieux J.L., Groshemy R., Thomas D., Ergan F. (19987) Reversible immobilization of an antibody
with a thiol-substituted sorbent. Application to enzyme immuneassays, Anal. Chim. Acta 197, 229-237.
22. Duschl C., Sevin Landais AF., Vogel H. (19%) Surface engineering: optimization of antigen
presentation in self-assembled monolayers, Biophys. J. 70, 1985-1995.
23. Rickert J., Weiss T., Kraas W., Jung G., Gopel W. (19%) A new affinity biosensor: self-assembled
thiols as selective monolayer coating of quartz crystaJ microbalances, Biosensors and Bioelectronics 11,
591-598.
24. Sigal G.B., Bamdad c., Barberis A., Strominger J., Whitesides G.M. (1996) A self-assembled
monolayer for the binding and study of histidine-tagged proteins by surface plasmon resonance, Anal.
Chem. 68,490497.
25. Kin W.B., Fafferic-Rebault N., Chovelon J.M., Lacoroix (2001) Optical fiber as a whole surface probe
for chemical and biological applications, Sensors and Actuators B74, 207-211.
26. Zhang J., Zhao J., Zhang H.L., Li H.L., Liu Z.F. (1998) Effect of the molecular interaction on molecular
packing and orientation in azobenzene-functionalized self-assembled monolayers on gold, Thin Solid
Films 327-329, 195-198.
27. Morgan H., Taylor D.M. (1992) A surface plasmon resonance immunosensor based on the streptavidin-
biotin complex, Biosensors and Bioelectronics 7, 405410.
28. Knichel M., Heiduschka P., Beck W., Jung G., Gopel W. (1995) Utilization ofa self-assembled peptide
monolayer for an impedimetric immunosensor, Sensors and Actuators B28, 85-94.
29. Nuzzo R.G., Allara D.L. (1983) Adsorption of bifunctional organic disulphides on gold surfaces, JAm.
Chem. Soc. 105,44814483.
30 Suzuki M., Nakashima Y., Mori Y. (1999) SPR immunosensor integrated two miniature enzyme
sensors, Sensors and Actuators B54, 176-181.
31. Kroger D., Liley M., Schiweck W., Skerra A, Vogel H. (1999) Immobilization of histidine-tagged
proteins on gold surfaces using chelator thioalkanes, Biosensors and Bioelectronics 14, 155-161.
32. Zachariou M., Hearn M.T. (19%) Application of immobilized metal ion chelate complexes as
pseudocation exachnage adsorbents for protein separation, Biochemistry 35, 202-211.
33. Starodub N.F., Arenkov PJ., Starodub AN., Berezin VA (1992) Optoimmunosensors for analysis of
specific and non-specific classes of immunoglobulins, Sensors and Actuators B7 (1992) 371-375.
34. Starodub N.F., Arenkov Pl., Starodub A.N., Berezin VA (1993) Fiber optic immunosensors for
detection of some drugs, Sensors and Actuators B 13-14, 728-731.
35. Starodub N.F., Arenkov Pl., Starodub AN., Berezin VA (1994) Fiber optic immunosensors based on
enhanced chemiluminescence and their application to determine different antigens, Sensors and
Actuators B 18-19, 161-165.
36. Harame D. L., Bousse L. J., Shott J. D., Meindl J. D. (1987) Ion-sensing devices with silicon nitride and
borosilicate glass insulators, lEE Transaction on Electron Devices, ED-34, N8, 1700-1707.
37. Pijanowska D. G., Torbich W. (1997) pH-ISFET based urea biosensor, Sensors and Actuators B44, 370-
376.
38. Wang J. (1999) Sol-gel materials for electrochemical biosensors, Anal. Chim. Acta 399, 21-27.
39. Samodumova LM., Slinjkova I.B. and Kisileva L.L (1976) Influence of some factors on the porous
structure ofxerogels ofpolyaluminomethylsiloxanes, Col/oidal J (USSR) 37, 502-507.
40. Samodumova LM. (1983) Adsorptional properties of new group ofelementorganic-silicon sorbents with
the bivalent metals, Theor. Exp. Chem. (USSR) 748-751.
41 Starodub N.F., Samodumova I.M. and Starodub V.M. (1995) Usage of organosilanes for integration of
enzymes and immunocomponents with electrochemical and optical transducers, Sensors and Actuators B
24-25, 173-176.
42. Starodub N.F. (1990) Electrochemical biosensors and biochemical diagnostics. In: Lecture notes of the
ICB Seminar. Biomeasurements. Ed. Pungor E., Torbicz W. v.3, Jablonna, Poland, pp. 173-201.
43. Stevens M.J. (1999) Thoughts on the structure of alkyl silane monolyers, Langmuir, 15,2773-2778.
44. Piehler J., Brecht A., Hehl K., Gauglitz G. (19%) Surface modification for direct immunoprobes,
Biosensors and Bioelectronics 11,579-590.
45. Piehler J., Brecht A., VaJiokas R., Liedberg B., Gauglitz G. (2000) A high-density poly(ethylene glycol)
polymer brush for immobilization on glass-type surfaces, Biosensors and Bioelectronics 15,473481.
46. Piehler J., Brecht A., Gecke1er G. (1999) Protein interaction in covalently attached dextran layers,
Col/oids Surfaces B: Bioimerfaces 13, 325-336.
47. Lee S.W., Laibinis P.E. (1998) Protein-resistant coating for glass and metal oxide surface derived from
oligo(ethylene glycol)-terminated alkyltrichlorsilanes, Biomaterials, 19, 1669-1675.
48. Inununology. U. Pol (Ed.), 1987,1. pp. 255-313.
 325

49. Starodub N.P., Pirogova L. V. (2002) Development of immune component immobilization on the
optical transducers by use of some lectins. In: Proc. of Intern. Conf. "Optoelectronics, Optical Sensors
& Measuring Techniques", Exhibition Centre Erfurt, 14-16 May, Germany, AMA, 175-180.
50. yazynina E.V., Zherdev A.V., D2antiev B.B., Izumrudov VA, Gee Sh.I., Hammock B.D.(1999)
Immunoassay techniques for detection of the herbicide simazine based on use of oppositely charged
water-soluble polyelectrolytes, Anal. Chern., 71, 3538-3543.
51. Ulman A. (1991) An introduction to ultrathin films, from Langmuir- Blodgett to self-assembly. Academic
Press: Boston, New York, Toronto, 440.
52. Lvov Y. M. and Decher G. (1994) Assembly of multilayer ordered films by alternating adsorption of
oppositely charged macromolecules, Crystallography Reports, 39, N 4, 628 - 647.
53. Lvov Y., Ariga K., Ichinose 1., Kunitake T. (1995) Assembly of multi component protein films by means
of electrostatic layer-by-layer adsorption,J Am. Chern. Soc. 117, N 22, 611 7-6123.
54 Caruso F., Niikura K., Furlong D. N., Okahata Y. (1997) Ultrathin multilayer polyelectrolyte films on
gold: construction and thickness determination, Langmuir 13, 3422-3426.
55. Caruso F., Niikura K., Furlong D. N., Okahata Y. (1997) Assembly of alternating polyelectrolyte and
protein multilayer films for immunosensing, Langmuir, 13,3427-3433.
56. Hesselink E. (1983) Adsorption from solution at the solidlliquid interface. London: Acad. Press, 377.
57. Lvov Y., Haas H., Decher G., Mohwald H., Mikhailov A., Mtchedlishvily 8, Morgunova E.,
Vainshtein B. (1994) Successive deposition of alternate layers of polyelectrolytes and a charged virus,
Langmuir, 10,4232-4236.
58. Nabok A.V., Davis F., Hassan A.K., Ray A.K., Majneed R., Ghassemlooy Z. (1999) Polyelectrolyte
self-assembled thin films containing cyclo-tetrachromotropylene for chemical and bio-sensing, Mat. Sci.
and Enginnering, C309, N8-9, 123-126.
59. Fukumoto H., Yonezawa Y. (1998) Layer-by-Iayer self-assembly of polyelectrolyte and water soluble
cyanine dye, Thin Solid Films 327-329,748-751.
60. Fabianowski W., Roszko M., Brodzinska W. (1998) Optical sensor with active matrix built from
polyelectrolytes-smart molecules mixture, Thin Solid Films, 327-329,743-747.
61. Simone de Souza Lima K., Winnischofer H. (2000) Thin molecular films of supramolecular porphyrins,
A. Acad. Bras. Ci. 72, NI, 27-32.
62. Gavalas Y.G., Chaniotakis NA, Gibson T.D. (1998) Improved operational stability ofbiosensors based
on enzyme-polyelectrolyte complex adsorbed into a porous carbon electrode, Biosensors and
Bioelectronics 13, 1205-1211.
63. Starodub N.F., Nabok A.V., Starodub V.M., Ray AX., Hassan A.K. (2001) Immobilisation of
biocornponents for immune optical sensors, Vkr. Biochem. Jorna173, No 4, 55-64.
64. Nabok A.Y., Ray A.K., Starodub N.F., Dowker K.P. (2000) Enzymelindicator optrodes for detection of
heavy metal ions and pesticides. Proc. of SPIE "Biochemical and Biomolecular Sensing", Ed.
Lieberman RA, v. 4200, Boston, USA, pp. 32-41
65. Starodub N.F., Nabok A.V., Ray A.K. (2002) Visualization of optical enzyme sensor responses at the
detection of heavy metal ions and pesticides. In: Abstracts of International symposium on
Visualization and imaging in transport phenomena. May. 5-10,2002, Antalya, Turkey,pp. 174-176.
66. Nabok A.V., Starodub N.F., Ray A.K., Hassan A.K. (2000) Registration of immunoglobuline Ab/Ag
reaction with planar polarization interferometer. Ibid, pp. 1-9.
67. Blomberg E., Claesson Per M., Froberg J.C. (1998) Surfaces coated with protein layers: a surface force
and ESCA study, Biomaterials, 19,371-386.
68. Starmbini E.G., Strambini G.B. (2000) Tryptophan phosphorescence as a monitor of protein
conformation in molecular films, Biosensors and Bioelectronics, 15,483-490.
69. Vilar M.R., Bouali Y., Kitakatsu N., Lang Ph., Michalitsch R., Garnier F., Dubot P. (1998) HREELS
characterization of surfaces and interfaces in self-assembled molecular mono layers, Thin Solid Film
327-329,236-240.
70. Fulghum Jx., McGuire G.E., Musselman I.H., Nemanich R.I., White J.M., Chopra D.R., Chourasia A.R.
(1989) Surface characterization, Anal. Chern., 61, 243R-269R.
PERFORMANCE OF AMPEROMETRIC ALCOHOL ELECTRODES
PREPARED BY PLASMA POLYMERIZATION TECHNIQUE

D. C;OKELiLER and M. MUTLU

Hacettepe University, Engineering Faculty, Food Engineering Department
Beytepe Campus, 06532, Ankara, Turkey

1. Introduction

Ethyl alcohol assays are necessary in many industries and biotechnological

applications. The content of ethanol should be monitored in its production from
mono- and oligosaccharides, cellulose, oil hydrocarbons and natural gas as well as
when it is utilized as a substrate in different biotechnologies and food processing.
Biosensor analyzers for ethanol would simplify the procedure without loosing
important characteristic of the chromatographic method such as accuracy,
reproducibility and also providing fast, cost-efficient assays [1].

In this study, we describe the preparation and use of an amperometric alcohol

electrode for sensing the analytically significant substrate ethyl alcohol based on
alcohol oxidase. Alcohol oxidase has been incorporated on ethylenediamine cold-
plasma modified polycarbonate membrane. The performance of the electrode was
tested by measuring the alcohol content of commercially available beers [2].

2. Experimental

Polycarbonate (PC) track etched membranes with nominal pore size 0.03 ~m was
obtained from Poretics (California, USA). Alcohol oxidase was obtained from the
Institute of Biochemistry, Grodno, Belarus. Glutaraldehyde, monomer
ethylenediamine, ethyl alcohol, hydrogen peroxide and the other buffer chemicals
were obtained commercially (Sigma, Steinheim and Merck, Darmstadt, Germany).

Electrode System: An oxidase based probe type enzyme electrode was used. The
electrode consists of Pt working electrode and Ag/AgCI reference electrode. The Pt
electrode was polarized at +650 mV for H20 2 detection and the amperometer was
linked to a data acquisition system and a computer.
327
Y. G. Gogotsi and I. V Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 327-332.
© 2003 Kluwer Academic Publishers.
328

2.1. MEMBRANE PREPARATION

The PC membranes were treated in a glow discharge reactor by plasma

polymerization technique to incorporate amino groups on the surface [3].

Plasma Polymerization System: The reactor was a glass tube with two external cupper
electrodes. A 13.6 MHz radio frequency generator was used to sustain plasma in the
reactor. Power losses were minimized by means of a matching network. In a typical
glow discharge treatment the membrane was placed in the middle of the reactor where
only one side of the membrane is exposed to the glow (Fig. I). Using this process, PC
membranes were modified in different plasma polymerization conditions (discharged
power: 10 to 30 W; monomer flow rate: 20 to 60 mUmin; exposure time: 15 and 30
min).

The effect of different plasma polymerization parameters on the PC membrane

permeability was tested by a diffusion cell system.
The diffosion cell system: It consisted of a cylindirical plexiglass chamber with two
compartments each of them with a volume of 170 ml and divided by a membrane.
There is an intensive stirring close to the membrane surface, which eliminates the film
mass transfer resistance [4). A Solution ofH202 in distilled water (2% v/v) was placed
in one compartment while distilled water was placed in other. The concentration of
H20 2 in the first compartment was followed by immersing a probe type amperometric
H202 electrode into the solution. The concentration change in each experiment was
evaluated to calculate the permeability of plain and modified polycarbonate
membrane for hydrogen peroxide [5].

Monomer Tank
Electrode Needle Valve

Plasma Reactor

Vacuum Pump

Figure I. Plasma Polymerisation System

 329

The plasma polymerization modified membrane with the highest resistance to H20 2
diffusion was selected as the optimum membrane for alcohol electrode preparation.
So, that membrane with incorporated amino groups was incubated in 25 mL of
glutaraldehyde solution (% 2.5 v/v) at 25°C for 2 h for activation. Then it was washed
with distilled water to remove unbound glutaraldehyde.

2.2. ALCOHOL OXIDASE IMMOBILIZATION

Alcohol oxidase solution was prepared in pH 7 phosphate buffer at three

concentrations (0.20%, 0.36% and 0.60% w/v), then activated membranes were
placed into enzyme solution and left to interact for 24 h at 25°C. This single
recognition layer was constructed on to probe by using O-ring. In our studies the
membrane was placed to the electrode surface where the enzyme layer facing to the
electrode for both to protect the enzyme from the medium that could be affect the
stability and to improve the sensitivity. So, this probe utilizes the catalytic oxidation
of alcohol, by alcohol oxidase to H202, which then is oxidized at an anodic polarizing
voltage (650 mY).

2.3. ELECTRODE PERFORMANCE

The alcohol electrodes were first allowed to stabilize for 5-10 min in the isotonic
buffer solution (pH: 7). Then ethyl alcohol was added to achieve 2.5 mM alcohol
concentration in the medium. The medium was stirred continuously. The response
was measured in terms of current. The time to reach the 95% of the maximum signal
obtained was taken as response time. This procedure was repeated several times until
the deviation from linearity was observed. These data were evaluated for each single
layer alcohol electrode with different enzyme content, to obtain the calibration curves.

The single-layer amperometric alcohol electrode with highest performance parameters

was selected to measure the alcohol content and by using an appropriate calibration
curve, the ethyl alcohol concentration of samples were found and compared with the
manufacturers' rating marked on the can.

3.Results and Discussion

In this study, we report the effect of different plasma polymerization parameters on

the PC membrane permeability by evaluating the data from diffusion cell system
experiments [4]. The dilution curves of hydrogen peroxide through the plain and
plasma modified membranes were evaluated, and the slope at t=O is considered to be
the permeability. The highest slope value, that is for the plain polycarbonate
membrane was considered to be 100 and the results are presented in Table 1.
330

TABLE 1. The effect of plasma polimerisation parameters on permeability

of H20 2

Power Pressure Exposure Time Flow Rate Relative

Permeability
(W) (mbar) (min) (ml/min) (dimensionless, %)
100
15 10 ·4 15 60 60
30 10 ·4 30 20 44
30 10 ·4 15 20 32
30 10 ·4 15 20 28

Table 1 shows that plasma modification of the PC membranes in general decreased

the permeability of hydrogen peroxide due to the deposition and polymerization of the
monomers on the surfaces, which has been shown in our previous studies [4]. The
maximum deposition or minimum permeability was observed at discharged power of
30 W, exposure time of 15 min and monomer flow rate of 20 mLimin. This particular
membrane with highest resistance against hydrogen peroxide permeation was selected
as the carrier membrane to obtain high linearity alcohol electrode.

The recognition layers which were prepared by immobilizing different amounts of

AOx were used as the alcohol electrodes. The performance parameters of these
alcohol electrodes were summarized in Table 2. Table 2 clearly shows that the drastic
increase in sensitivity after the critical enzyme concentration on the membrane
surface. The response time of the electrode is decreasing while the amount of enzyme
bound to the surface increases.

TABLE 2. The Performance Parameters of Alcohol Electrodes

Alcohol Linearity Sensivity Response Correlation

Electrode Time Coefficient (R2)
(AOx (mM) (nA/mM» (s)
content)
0.20% 6.0 0.075 240 0.995
0.36% 5.0 0.127 120 0.998
0.6% 2.0 5.6 50 0.992

The alcohol electrode that is made up of 0.60 % w/v AOx, which has the highest
sensitivity and minimum response time was selected (Fig.3.). It was used to measure
the alcohol level in various types of commercially available beers with different
alcohol content
 331

o 2 3
Concentration, C (mM)

Figure 2. The Calibration Curve Of Alcohole Electrode

The results are to compare the values rated by the manufacturer and printed on the
can. The results show that, there is no significant difference were also between the
measured values and the manufacturers' rating. The label values are checked by
Refractive Index method (Table 3).

TABLE 3. Test of The Alcohol Electrode in Beer Samples

Beer Concentration Concentration Refractive Index

Type Values Values Results
(enzyme electrode) (manifacturer)
(%VN) (%VN) (%VN)
Light 3.3±O.! 3 2.9±O.3
Normal 4.6±O.2 5 4.8±O.2
Extra 7.4±O.5 8 7.8±O.!

4. Conclusion

The plasma polymerization technique introduces a new area for biosensor researchers
for preparing single layer enzyme electrodes with many tunable properties [6]. In this
study, we prepared an alcohol electrode by EDA-plasma modification on
polycarbonate membrane. The alcohol electrode with highest sensitivity was
preferred because the analyte, ethyl alcohol was diluted before measurement. The
performance of the electrode against commercial data was very satisfactory.
332

Acknowledgement
This research has been partly supported by Turkish Scientific and Technical Research
Council (TUBITAK) Project (TARP 2057) and EUREKA Project (E! 2080,
PLASMA//BIOSENSE).

REFERENCES
[I] Boujtita M. and. El Murr N (1995), Journal of Food Sciences, 60 pp. 201-205.
[2] Cokeliler D. and Mutlu M., Analytica Chimica Acta, (2002) (in press).
[3] Mutlu M., Mutlu S., Boyacl LH., Alp B., Piskin E. (1998), "Polymers in Sensors: Theory &
Practice", A.M. Usmani and N. Akmal, eds., American Chemical Society (ACS) Symposium Series,
690 pp. 57-65.
[4] Turhan M., Desai MA., Vadgama P., Mutlu, M. (1995) J. Biotechnology Techniques, 9 pp.413-416.
[5] C;;iikeliler D., (2000) M Sc. Dissertation, Hacettepe University, Ankara.
[6] Mutlu S., Alp B., Onnelles R.S., Mutlu M. (1997) J. Food Engineering. 33 pp. 81-86.
CHEMICAL ACTIVE CENTERS AT SURFACES OF
Si-BASED MATERIALS

A.E. KlV\ V.G. LITOVCHENK02, D. FUKS',

V.V.GOLOVANOV 3, I.P. LISOVS Kyy 2, T. 1. MAXIMOVA'
1Depart. 0/ Mat. Eng., Ben Gurion University,
Beer-Sheva, 84105 Israel
2Institute o/Semiconductor Physics, Kyev-28, 252028, Ukraine
3South-Ukrainian Pedagogical University, Odessa, 65008, Ukraine

1. Introduction

Recently, the new types of the Si-based sensors have been

developed for medical and industrial applications. The further improvement
of these devices demands a deeper study of silicon surface in real
conditions.
Last time it was found that the relaxation of real Si surface depends
strongly on the environmental conditions and may be changed at elevated
temperatures and under irradiation [1,2].
Using Molecular Dynamics (MD) simulation we observed
appearance of quasi-disordered phase (QDP) at temperatures above 650 0 C,
which is spread to four atomic layers [2]. The (001) Si surface was studied.
QDP contains dangling bonds. Moreover, the sites with two-three dangling
bonds are typical for QDP.
The atomic radial distribution functions in QDP layers are similar to
those in a-Si. Each atomic layer in QDP has peculiar characteristics. The
third layer almost does not contain dangling bonds, but a hard distortion of
this layer and the observed directions of bonds argue that this layer plays a
connecting role between the QDP and the bulk. The QDP contains atomic
rings with the number of atoms, which is larger than 6 and less than 5. The
3-atomic rings contain the most stressed bonds and they are quite labile.
The silicon cluster configurations analogous to the QDP structures were
studied by ab initio calculations [3].
The features of chemical bonds and the near-surface structure
determine the chemical activity of silicon surface. That is why the
simulations of the temperature dependence of the QDP properties is
important for applications in the sensor physics.
The possibility to vary the chemical activity of silicon surface by
the low-energy irradiation is shown in [4,5].
The calculated characteristics of the QDP are consistent with
experimental results [6].
333
YG. Gogotsi and 1. V. Uvarova (eds.),
Nanostructured Materials and Coatings/or Biomedical and Sensor Applications, 333-341.
© 2003 Kluwer Acudemic Publishers.
334

We perfonned the MD simulation of structural changes in Si

surface as a function of the temperature and the parameters of the ion
bombardment. The standard MD approach was modified in two directions.
First, the new calculation algorithm allows to correct the interatomic
potential at each step of the simulation procedure. The second innovation is
a possibility to model re-bonding and re-hybridization of chemical bonds
during the relaxation processes.

2. Modification of the MD approach and the features of the

model

MD method was applied in its standard fonn [7] i.e. the equations
of motion were solved by using the central difference scheme. The time-
step was 10. 14 s. But the usual MD algorithm was changed to account the
specific features of the dynamics of covalent structures. At each step of the
calculations we estimated the changes of the hybridization coefficients
(CH) and obtained new Sp3 -hybrid wave functions. This allows to correct
the potential at each step of calculations and to find new coordinates using
the potential improved at the previous step. The widening of the MD
algorithm is shown schematically below:

A scheme of the widening of the MD algorithm

Xi (0), Vi (0), Fi (0) -> Xi (1'), Vi (1'),

Xi (1'), Fi (1') -> new CH and \jI (t),
\jI (1') -> ~ U[,
U (0) +~UI -> Fi (1'),
Xi (t), Vi (t), Fi (t) -> Xi (2t), Vi (2t), etc.

Here Xi (t), Vi (t), Fi (t) are coordinates, velocities and forces of i-

th atom at the moment 1', \jI (t) is the sp3-hybrid wave function at the
moment t and U (0), ~UI are the initial potential and the first correction
to the potential.
The next modification of the MD algorithm is linked to the necessity to
describe disordering of the covalent structure. In this case we have to
account the re-bonding processes. If the relaxation leads to large distortion
of the covalent structure, the stressed bonds appear, which are subjected to
predissociation with a large probability. New bonds are fonned in
accordance with new atomic configuration, arising as a result of relaxation.
Two conditions should be satisfied in the modeling of the re-bonding
process:
Each atom in the distorted covalent structure is connected
with its nearest neighbors.
 335

The angles aspire to have values inherent to the tetrahedral

structure (~I 09°).
From the above conditions it follows that in the re-bonding processes not
all bonds are restored, and thus dangling bonds arise. The situation may be
illustrated by the following example (See Table 1).

Table 1. Illustration of situation leading to formation of dangling bond

The central atom The nearest neighbors

a) 5 41 72 100 201
b) 100 5 23 75 130
c) 100 11 23 75 130

In Table 1 we demonstrated the initial situation (a) when in the ideal lattice
the atom 5 is surrounded by atoms 41, 72, 100 and 20 I. As a result of the
relaxation two new atomic configurations are possible. In the first case (b)
the atom 100 has the atom 5 as a nearest neighbor as it was in the ideal
structure. In the second case (c) the atom 100 changed the atom 5 for the
atom II. In the last situation the atom 5 remains without the forth neighbor.
The dangling bond appears.
In the MD simulation the starting configuration was taken as a
parallelepiped containing 864 atoms: 12 layers with 72 atoms in each one.
Periodic boundary conditions were used in two dimensions. The atoms
were in their normal positions inherent to silicon lattice. In the simulation
process we studied a free relaxation of the system at different temperatures
of the lattice and in conditions of irradiation. Simulations were performed
with Stillinger-Weber (SW) potential [8], with the quantum corrections, as
it was described above. The experience with SW potential indicates that it
is a reasonable presentation of silicon for the study of ion beam processing,
accurately describing many properties of small silicon clusters, bulk and
surfaces.
We use the following the scenario of the simulation ofthe pulse ion
irradiation. After equilibration ofthe system, one atom is given the velocity
corresponding to the chosen beam energy and the beam angle of the
incidence. Then we wait when the system reaches the equilibrium with the
thermal bath. In the case of the continuous ion irradiation the next ion pulse
follows immediately after the previous one.
Parameters of ion beams were varied in wide region. We
investigated the spatial distribution of electronic and structural defects in
the QDP as a function offluencies and energies ofthe ion beam.
336

3. Description of results

Surface bonds are chamcterized by a wide spectrum of bonding

energies [9]. This is linked to specific features of silicon surface
reconstruction at elevated temperatures. In Fig. I we demonstrate a spatial
orientation of atomic dimers at silicon surface layer, appearing at
tempemture larger than 600K. On the vertical axis is the percentage of the
total number of dimers in the first silicon layer, corresponding to the given
inclination angle to the surface plane (00l). P is the inclination angle.

o 10 20 30 40 50 60 70 8
Fig. 1. Spatial orientation of atomic dimers in the surface layer.

One can see that the essential part of dimers in the first surface layer
do not dispose in the surface plane. This part of dimers is characterized by
stressed bonds, which are unstable and can be broken as a result of small
pertubations. We studied the dimers structure in the temperature interval
from 200 to 1000K. It was found that the distribution of angles P, which is
shown in Fig. 1, depends on the temperature, as we demonstmte in Table 2.
The upper limit of the angle pfor 50% of the surface dimers at different
temperatures is shown in Table 2.

Table 2. The temperature dependence ofthe angle ~ distribution in QDP

T,K 300 600 1000

{3< 100 15° 25°

Formation of dangling bonds in QDP is revealed in the of oxygen

agglomeration in polycrystalline silicon films [10]. In [10] the semi-
insulating polycrystalline oxygen-doped silicon (SIPOS) layers were
studied. It was found that the high-temperature treatment of SIPOS films
causes the increase of the concentration of SiO x clusters with a large
 337

content of oxygen (X= 3, 4). An analysis of the oxygen agglomeration

effect in SIPOS films leads to result, which is consistent with the computer
simulation of the relative content of double and triple dangling bonds (See
Fig. 2). In Fig. 2 one can see the distribution of atoms in the QDP with N=
1, 2, 3 and 4 dangling bonds. Atoms with four dangling bonds are
interstitials.

%
40

30

20

10
~ ____ ~ ____ ~ ____ ~ N
2 3 4

Fig. 2. Distribution of nodes with 1,2,3 and 4 dangling bonds in QDP.

The distribution of oxygen over the Si-O clusters was obtained

experimentally by using a method of deconvolution of Si-O IR spectra
stretching band into the elementary Gauss profiles. This approach allows to
find the angles between chemical bonds in different Si-O clusters, which
are formed in the SIPOS films in the process of high-temperature oxygen
agglomeration [10]. The typical angles are represented in Table 3.

Table 3. The experimental values of angles between chemical bonds in

SIPOS films.

Si-O-Sb

In Fig. 3 we represented the simulation results on the distribution of the

angles between chemical bonds in QDP of silicon surface. It is seen that the
deviations from the tetrahedral angle are an attractive fit to the
experimental data
338

9(1J)

20

10

Fig. 3. Distribution of angles between chemical bonds in QDP.

In Fig. 3 the dotted line corresponds to a-Si. Thus we can conclude

that structural features of the QDP in the silicon surface layers at elevated
temperatures or in the conditions of irradiation are close to those of a-Si.
The increase of the temperature in the interval 600-1000K leads to
essential changes of structural and electronic characteristics of the QDP in
the silicon surface. Moreover these uneven changes of parameters of the
QDP are observed every 50-100°. As a result of transition to higher
temperatures the following changes in QDP take place:
- The increase of the concentration of dangling bonds.
The changes of the statistics of atomic polygons in silicon
surface layers [6].
The changes of the distribution of angles between bonds
shown in Fig. 3.
Accordingly to these temperature effects the equilibrium state in the
interaction of the silicon surface with environmental atoms becomes
broken. The diffusion conditions are also changed. The new chemical
active centers appear, and a redistribution of the bonded impurity atoms
takes place.
At the temperature 1000K the disordering of silicon surface layers
is characterized by the increased concentration of dangling bonds. The rows
of dimers disappear. The dimers are distributed in space randomly. The
concentration of interstitial atoms is larger than it is estimated for this
temperature by thermodynamic approach.
There are wide possibilities to control the chemical activity of Si
surface by irradiation. For example, the low energy electron irradiation
leads to formation of extremely strong adhesion bonding in the Si-AI
system [11]. The radiation changes of the silicon surface chemical activity
depend significantly on the parameters of irradiation. Computer simulation
of the reconstruction of silicon surface layers shows that the low energy ion
bombardment can decrease the concentration of dangling bonds. At the
same time we observe in our simulations the essential increase of chemical
active centers in silicon surface at comparatively large energies of ions. In
Fig. 4 we demonstrate the results of simulation of the pulse ion
bombardment. We modeled the ion pulse, corresponding to the flux 107
ion/cm 2 sec-to On the vertical axis is the change of the total energy of the
 339

system induced by the ion bombardment. For ~E>O we have the annealing
of the broken bonds and the decrease of the number of dangling bonds. For
~E<O the number of dangling bonds increases. The analysis of Fig. 3
shows that the third layer is the most sensitive to the ion bombardment. We
see that the transition from the transferred energy of 30 eV to 40 eV
changed the radiation effect radically.

AE,eV/Atom

0,2 ... 0 .. 1 eV

_ ..... lOeV
0,15

0,1

0,05

o Layer

-0,05

Fig. 4. Changes of energies for different layers in QDP as a result of

bombardment ofSi surface by ions in the energy interval 1-40 eV.

At the transferred energy 30 eV the number of non-pair electrons is

minimal. For 40 eV the situation is opposite. In the last case we have the
maximal number of dangling bonds and accordingly the large chemical
activity of the surface. The environmental atoms can easy penetrate inside
of QDP and interact with chemical active centers located in all layers of
QDP.
It follows from the described results that the combination of the
temperature effect and the irradiation opens the possibility to control the
chemical activity of the silicon surface layers. For example, by increasing
the temperature in the given interval we can increase the total number of
dangling bonds in the QDP. Then using the low energy ion irradiation we
can change the distribution of dangling bonds between atomic layers of the
QDP,
We describe silicon surface atoms implying their belonging to the
corresponding layers of the ideal lattice [9],
340

4. Conclusion

The modified method of MD was used to study the relaxation

processes in the silicon surface layers at elevated temperatures and in
conditions of the low energy irradiation. We improved the traditional
approach by adding the new procedures of correction ofthe potential in the
process of calculations and by account the re-bonding of atoms. Thus we
were able to study the full relaxation of the silicon surface layers at
different temperatures and under the irradiation.
It was found that QDP with dangling bonds is formed in the near-
surface layers of silicon. There are the node atoms in the QDP with more
than one dangling bond. The non-uniform distribution of dangling bonds in
the silicon near-surface layers is confirmed by the experimental study of
SIPOS films. Though the bulk-like hexatomic polygons predominate in
QDP, abnormal polygons with 3, 4, 5 and so on atoms and with stressed
bonds are inherent to the QDP. Electronic and structural features of QDP
determine the chemical activity of the silicon surface at the elevated
temperatures.
New dangling bonds appear with the temperature increase. The
structural parameters of the QDP also change. This leads to changes of the
surface diffusion coefficients and to the redistribution of the bonded
impurity atoms in the surface layers. Thus we observe the essential
temperature influence on the sensor properties of silicon surface.
The low energy ion irradiation of the silicon surface enables to
perform the precise changes of the distribution of dangling bonds in the
silicon near-surface region. The combined affect of the temperature and the
low-energy irradiation open possibilities to control the chemical activity of
silicon surface in a wide region.

Acknowledgments

This work was supported by EC in the frame of INTAS Program (Grant

INTAS -01-0009)
 341

References

1. Hibino, H., Sumitomo, K. (1999) High-temperature disordering of

(Ill) Si surface, Phys. Rev. B58, 12587 - 12589.
2. Kiv, A.E., Soloviev, V.N., Maximova, T.I. (2000) Microstructure of the
relaxed (001) Si surface, Semicond Phys., Quantum 200.
3. Eguchi, H., Tsumuraya, K., Nagano, T. and Kihara, S. (1999) Bonding
nature of Sb bicapped pentagon c1uster:ab initio study, Mater. Trans.,
JIM 40, 1198-1205.
4. Jacobs, P.W., Kiv, A.Soloviev, V.N., Maximova, T.I.(1999) Radiation-
stimulated processes in Si surface layers, CM&NT3, 15 -20.
5. Kiv, A.E., Maximova, T.I. and Soloviev, V.N. (2001) MD simulation of
the ion-stimulated processes in Si surface layers, in M.-I. Baraton and
LV. Uvarova (eds), Functional Gradient Materials and Surface Layers
Preparated by Fine Particles Technology, Kluwer Academic
Publushers, Dordrecht, pp.297 - 303.
6. Lisovskyy, LP., Litovchenko, V.G., Gnennyy, B.M., Mazunov, D.O.,
Fussel, W., Kiv, A.E., Maximova, T.I. and Soloviev, V.N. (2001)
Oxygen-Assisted Structural Transformations in Silicon Polycrystalline
Films, Phys. ofLow-Dimensional Structures 7/8, Il3 -126.
7. Frenkel Daan and Smit Berend (1966) Understanding Molecular
Simulation, Academic Press, New-York.
8. Stillinger, F.H., Weber, T.A. (1984) New interatomic potential for
silicon, Phys. Rev. B31, 5262 - 5267.
9. Jacobs, P.W., Kiv, A.E., Balabay, R.M., Grischenko, N.V. (1998)
Atomic configurations in a-Si obtained by ion implantation, CM&NT 2,
15-20.
10. Lisovskyy, I.P., Litovchenko, V.G., Gnennyy, B.M., Fusse1, W., Kiv,
A.E., Maximova, T.I. and Soloviev, V.N. (2002) Efect of oxygen
agglomeration in polycrystalline Si (SIPOS) films, J Mater. Sci.:
Mater. Electronics 13, 167 -171.
DEPOSITION OF SN0 2 -BASED THIN FILMS BY REACTIVE
DC MAGNETRON SPUTTERING FOR GAS SENSING
APPLICATIONS

J. MIGUEL CORREIA-PIRES, VASCO TEIXEIRA and J. B.

ALMEIDA
Physics Department, Universidade do Minho, Campus de Gual-
tar, PT-4710-057 Bmga, Portugal

Abstract. Tin dioxide is a semiconductor with good sensitivity to a wide variety of gases.
It has been produced by several thin and thick film techniques, and successully used in
gas sensor devices. Some of the parameters affecting gas sensing response are already
known and may be used to guide the layer production process to improve gas sensing
performance of the fabricated devices.
Magnetron sputtering is a reliable and environmentally clean technique, suitable to
produce thin film oxide materials, like Sn02. Layers with widely different characteristics
may be obtained by varying the deposition parameters. Namely, is has been verified that
composition, grain size, porosity and surface roughness may be varied within the ranges
reported to correspond to best gas sensitivity.
The influence of the deposition temperature, total pressure and magnetic flux density
on the thin film characteristics have been studied and it was verified that it is possible
to get layers with characteristics similar to the ones produced using other techiques
and showing high sensitivity. Deposition rates and film density have been determined
using data obtained from mass difference measurements, optical transmission spectra
and scanning electron microscopy (SEM). Structural properties - crystal structure and
orientation, surface morphology and grain size - have been studied using data from atomic
force microscopy (AFM), SEM and X-ray diffraction (XRD). The coatings bulk and
surface composition was investigated using energy dispersive X-ray analysis (EDX) and
X-ray photoelectron spectroscopy (XPS).

1. Introduction

The effect of the ambient atmosphere upon the electrical conductance of

semiconductors is known for a long time [4, 11, 20]. The gas sensing prop-
erties of tin oxide layers result mainly from chemisorption of oxygen and
water vapour. Chemisorption takes place on the 8n02 surface and results
in the trapping of free electrons, building a negatively charged layer at the
surface. In some polycrystalline materials, the negative barrier develops at
343
Y G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 343-355.
© 2003 Kluwer Academic Publishers.
344

the surface of each grain, forming a barrier to electron conduction at each

grain boundary.
Tin oxide has been produced by several thin and thick film techniques,
and successully used in gas sensor devices. Sn02 sensitivity is, however,
highly dependent on the technique used to produce the layers. The depen-
dence of the tin oxide conductance on the gas concentration fits usually to
a power law of the type

G ex P;as (1)
where G is the layer conductance, P gas is the partial pressure of the reducing
gas and n is an exponential factor lower than 1. Some models have been
developed to explain gas sensitivity response, but much has yet to be done
to understand the full process, particularly, the factors that explain the
different power laws observed and the conduction processes through the
polycrystalline layers.
Gas sensitive Sn02 layers have been produced by several techniques,
such as: sol-gel [1, 24, 14], magnetron sputtering [28, 18, 8], powder sin-
tering [3, 2], spray pyrolysis [16, 23], laser ablation [31] and RGTO [25].
Although physical vapour deposition technology has shown good possi-
bilites in the preparation of gas sensing layers, the majority of the reports
deal with materials produced by thick film processes.
In undoped form, Sn02 has shown sensitivity to: carbon monoxide [13,
28, 7], ethanol [28, 3, 12], H2 [1] and NO x [30, 25], among others. Metal
additives are sometimes introduced in the tin oxide layer to change its gas
sensing [33, 17, 27, 19], electrical [22] or structural [9, 34] properties.
Although tin oxide gas sensitive layers have already been produced by
DC magnetron sputtering, the full possibilities of the techique have not
been studied yet. The ability to produce films with different composition,
porosity and grain size is important to the development of gas sensor tin
oxide technology. In this work it is shown that it is possible to obtain layers
with characteristics suitable to gas sensor applications.

2. Experimental details

The layers were produced in a home made sputtering chamber equipped

with a Hiittinger PFG 2500 DC generator. The system is evacuated using
baffled mechanical and diffusion pumps [5]. The target used was a high
purity tin disk (Goodfellow, 99.9 %) with 10 rum diameter. The deposition
atmosphere was a mixture of argon and oxygen whose flow was maintained
by two independent Bronkhorst High-Tech B. V. mass flow controllers. Be-
fore each deposition the chamber was evacuated to at least 3.0 x 10-3 Pa
and chamber walls heated during some hours to reduce degassing during
 345

deposition. Most of the samples were deposited on glass substrates, either

with 1 mm or with 0.1 mm thickness. A few were also deposited on steel and
alumina. The substrates were weighted using a Sartorius MP5 electronic
microbalance, before and after each deposition to determine deposited mass.
The DC sputtering system was operated at 20 W power and a current
of 80 mAo The oxygen partial pressure was preset to a value between
1.3 x 10- 1 Pa to 1.0 Pa and then argon flow was added until the desired pres-
sure was achieved. The deposition pressure was varied from 2.0 x 10- 1 Pa
to 3.8 Pa. The target to substrate distance was set to 65 mm and the
substrate temperature was varied between 1000 C and 400 0 C. The position
of the samples in the substrate holder was recorded in order to evaluate
film homogeneity.
After removing the films from the chamber its optical transmittance
was measured in a range of wavelengths from 200 nm to 2600 nm and from
the obtained spectra film thickness and refractive index using Swanepoel's
method [26] were determined. Those spectra were obtained using a Shi-
madzu UV-3101PC spectrophotometer. The crystalline structure of some
samples was determined using the XRD spectra and from this set a few
were analyzed by EDX and XPS. The surface morphology was analyzed
by atomic force microscopy and SEM. The cross-section of some layers
with thickness higher than 500 nm was observed in the scanning electron
microscope to determine film growth structure.
Some samples were then tested in a specially designed chamber with
controlled atmosphere, detailed elsewhere [5]. This test chamber has a
volume of 50 cm 3 and its pressure is controlled by a Setra 204E variable
capacity pressure sensor. The gas lines are connected to a mechanical pump
that is used to evacuate the system down to 6.0 x 102 Pa before starting the
test. The reference gas was dry synthetic air and the test gas was a high
purity mixture of carbon monoxide in nitrogen. The electric contacts on the
layer surface are made with two gold-coated tips. The sample is mounted
on a metallic disk heated with an electrical resistance and measured using
a thermocouple placed on the holder. Both the resistance and the ther-
mocouple are connected to a Shimaden SR53 temperature controller that
maintains the holder temperature within 0.10 C of the preset value. While
the reference gas is flowing, the system is heated during several hours to
stabilize temperature gradient. Both the temperature and the resistance
are registered in a computer at regular time intervals previously specified.

3. Results

The transmittance spectra of the samples permitted to determine the thick-

ness and the refractive index of produced layers. The thickness values were
346

then used to compute the deposition rate, and the layer density. Fig. 1 shows
the dependence of both the deposition rate and the layer density on the total
deposition pressure during layer growth. It is shown that the deposition rate
has a maximum when the total deposition pressure is about 1.4 x 10- 2 mbar.
At this pressure the deposition rate is 29 nm/min. On the other hand, the
density is shown to decrease with increasing deposition pressure, showing
a steep variation when the pressure is reduced from 6.2 x 10-3 mbar to
2.9 x 10- 2 mbar. When the deposition pressure is below 6.2 x 10- 3 mbar
the layer density is very close to the crystal density of Sn02. The refractive
index relative to a wavelength of 600 nm, has a dependence on the total
deposition pressure similar to the one observed for the density, as shown
by Fig. 2.

30

c
·E
E 5.5
s
.Sl 20
~
c
g
.0;
0
4.5
Co
CD
"C

10 deposition rate
density
'--~-=----~----'-----~-~---' 3.5
10
deposition pressure (x10-3 mbar)
Figure 1. Effect of total deposition pressure on deposition rate and layer density, using
oxygen partial pressure equal to 0.15 Pa.

The transmittance spectra of four different samples deposited using an

oxygen deposition pressure of 1.5 x 10- 3 mbar are shown in Fig. 3. The
ones shown in Fig. 3(a) were deposited with a substrate temperature of
100° C and have different thickness. The thicker sample is the one produced
with a deposition pressure of 4.8 x 10-3 mbar. Higher thickness is revealed
by the higher number of interference fringes in the spectrum. It is also
clear the decrease in the refractive index, shown by the lowering of the
difference between the maxima and minima of the interference fringes. In
Fig. 3(b) it should be noted a small increase of the refractive index when the
substrate temperature is increased at a total deposition pressure of 3.8 x
10-2 mbar. Simultaneously it is observed at higher substrate temperature
 347

1.9
x
Q)
"0
.£:
Q)
.~ 1.8
"0
ttl
~
~
1.7

1.6
10 20 30 40
deposition pressure (x10-3 mbar)
Figure 2. Effect of total deposition pressure on tin oxide refractive index at a wavelength
of 600 nID. The oxygen partial pressure was kept at 0.15 Pa during deposition.

a small increase of absorption in the infrared region. Calculation of the

deposition rate permitted to conclude, in the case of the samples from
which spectra of Fig. 3(b) were taken, that the deposition rate slightly
decreases with substrate temperature.
Films with high density have been analysed by XRD and it was shown
that when the deposition pressure is increased, maintaining the oxygen
to argon ratio in the deposition atmosphere at 1.0, there are significant
variation in the dominant crystal orientation of the nanocrystals. Namely
it was shown that the dominant orientation was in the (101) direction at
a total pressure of 1.36 x 10-3 mbar, changed then to (211) at 1.50 x
10- 5 mbar and finally at 1.89 x 10- 3 mbar the dominant orientation was
in the (110) direction. This dependence is shown in Fig 4(a). Unlike in
the above case, when the substrate temperature is varied between 100° C
and 400 0 C, maintaining the total pressure at 3.2 x 10- 3 mbar, there are
no significant changes in the crystallite orientation of the layer as shown
in Fig. 4(b). Small changes in peak intensity may however be observed in
spectra shown in Fig. 4(b).
SEM micrographs showed the growth structure of some layers and also
film failure modes. As expected films produced with low total deposition
pressure were highly dense, as shown in Fig. 5(a) for a sample deposited at
5.0 x 10-3 mbar, and had very good adhesion to the substrate. On the other
hand, when the pressure was increased various forms of layer deformation
were observed. The most typical were cracks, has shown in Fig. 5(b). The
348

3.6x10-2 mbar

~
'"
u
c:
E 100
11l

E
r/) IVVV\J~"---
c:
~ I 4.8x1O-3 mbar

I
I
I

0
I
500 1000 1500 2000
f...(nm)

(a) Substrate temperature: 100° C

~~~~c

~
'"
u
c:
.;: 100 ---
11l

E ~~------
r/)
c: !
I 100' C
~ I
I
I
I
i

0 i
500 1000 1500 2000
f...(nm)

(b) Deposition pressure: 3.8 Pa

Figure 3. Effect of deposition pressure and substrate temperature on transmittance
spectra of DC sputtered tin dioxide films.

form and dimension of the cracks varied with deposition pressure, but no
clear trend Wa.<l found yet. In some cases, high stresses caused adhesion
failure of the layer, as shown in Fig. 6. The sample shown, was produced
at 1.4 x 10- 2 mbar, with a substrate temperature of 100° C.
AFM graphs of the more dense films showed a grain like structure,
as observed in Fig. 7 for a sample grown using a substrate temperature
of 210° C and a deposition pressure of 0.65 Pa. Samples deposited using
lower pressure and similar thickness, showed a small decrease of the surface
 349

<110> <211>
<101:>

10 20 30 40 50 60 70 80
2e (0)
(a) Substrate temperature: 300 0 C

<110>

:i
~
~~;...;;.-~
"iii
ji _ _ _
.E ~
<321>

10 20 30 40 50 60 70 80
2en
(b) Deposition pressure: 3.2 x 10- 1 Pa
Figure 4. Effect of deposition pressure and substrate temperature on polycrystalline
structure of deposited tin dioxide layers

roughness and the grain size [6J. The substrate temperature was observed
to change considerably the grain size above 200 0 C [5J.
The gas sensing response of the analysed samples showed the usual
exponential dependence on the gas concentration, as shown in Fig. 8. The
resistance of the sample decreases when exposed to CO, but the change
is proportionally higher for small concentration. In Fig. 9 is shown the
normalized conductance change, observed upon exposure to 200 ppm of
350

(a) 0.50 Pa

(b) 3.8 Pa
Figure 5. SEM imaS'BS of SuCh films produced at different pressures

Figure 6. SEM images of Sn02 films produced at a total deposition pressure of 1.0 Pa,
showing adhesion failure.

carbon monoxide with an operating temperature of 3000 C, against the

deposition pressure. Analyses performed until now seem to show a small
increase of the conductance change with deposition pressure. However all
the undoped films show significant conductance drifts. Addition of molyb-
denum during layer deposition was tried. Some of these samples had better
electrical stability. Studies are now under way to reveal the effect of this
 351

Figure 7. AFM graphs of a sample produced at a total deposition pressure of 0.65 Pa

and substrate temperature of 210° C.

element on the layer structure, in order explain the change in the electrical
behaviour.

Figure 8. Sample response to concentrations of carbon monoxide between 100 and

1200 ppm.

4. Discussion

Films with decreasing density have been produced, by increasing total de-
position pressure. This density variation seems to be caused by the cracking
phenomena on the layer surface. The refractive index variation shown in
Fig 2, points also to an increase in layer porosity, that could be caused by
352

5r-------------------~--~~_.

OL-----------------~--~--~
2 5 6 7 8
deposition pressure (X10·3 mbar)
Figure 9. Conductance variation upon exposure to 200 ppm of CO, observed on samples
produced with different total sputtering pressures.

the cracks. If the cracks do not cause failure of the thin film they would
be important in gas sensitivity application, since the sensitivity mechanism
involves the interaction of the gas with the layer surface, thus an increase
of porosity would increase the exposed surface of the sensing material.
Grain size is also important to reducing gas sensitivity of semiconductor
metal oxide materials. The model most widely accepted to explain the
grain size effect on the Sn02 sensitivity assumes that an electron depleted
surface layer (also called space-charge layer) develops in the material when
its surface is exposed to oxygen. This space charge layer has high resistivity,
therefore if its extension is large it will reduce significantly the conduc-
tance of the layer. When carbon monoxide (or other reducing particle)
collides with the surface it may react with adsorbed oxygen species and
desorb in CO 2 form. During the reaction, the oxygen species will loose the
captured electron onto the sensing layer, increasing its conductance. The
conductance change will depend on the reaction rates taking place at the
surface, however, providing that the gas reaches the surface of the grains,
smaller grains could (and in some cases have been verified [21, 32, 10])
lead to higher conductance changes upon reducing gas exposure. Since it is
known that substrate temperature during sputtering increases significantly
the grain size of the produced layers [5], in the majority of the depositions
low temperatures were selected.
Another factor than have been reported to influence sensitivity is the
oxygen content of the oxide film [29, 18, 28, 15]. The oxygen content can be
altered by changing oxygen partial pressure in the deposition atmosphere.
When O 2 partial pressure decreases, oxygen content in the deposited film
also decreases. It is demonstrated that there are three main deposition
regimes: the one with lower oxygen partial pressure that produces metallic
 353

tin films; a middle regime in which carbon monoxide amorphous films are
produced and a third regime using higher oxygen partial pressures that
results in the deposition of polycrystalline tin dioxide films. Tin monox-
ide films have a characteristic black colour and are easily oxidized when
annealed in an oxygen containing atmosphere. Since these are unstable at
the working temperature used to test the sensitivity, careful selection of the
oxygen partial pressure during deposition was done in order to obtain poly-
crystalline tin dioxide layers. Choosing a low deposition current permits to
experiment with lower oxygen pressures, since higher currents need higher
oxygen pressure in order to obtain poly crystalline films.

5. Conclusions

DC magnetron sputtering offers very good possibilities in the development

of gas sensitive metal oxide thin films. Namely, it permits to obtain layers
with high porosity, different grain size and oxygen content. Total deposition
pressure has to be carefully selected and controlled during deposition inas-
much as it affects significantly the structural properties of the produced
layers. High pressures seem to enhance gas sensing characteristics, but
adhesion failure problems may arise if proper substrate is not chosen.

Acknowledgements

This work has been supported by the Fund~ao para a Ciencia e a Tec-
nologia, through research project: PRAXIS/2/2.1/TPAR/2044/95. One of
the authors (J.M. Correia-Pires) gratefully acknowledges a grant from the
Fund~ao para a Ciencia e a Tecnologia.

References

1. Ansari, S. G., P. Boroojerdian, S. R. Sainkar, R. N. Karekar, R. C. Aiyer, and S. K.

Kulkarni: 1997, 'Grain size effects on H2 gas sensitivity of thick film resistor using
Sn02 nanoparticles'. Thin Solid Films 295, 271-276.
2. Blaustein, G., M. S. Castro, and C. M. Aldao: 1999, 'Influence of frozen distributions
of oxygen vacancies on tin oxide conductance'. Sensors and Actuators B {Chemical]
55,33-37.
3. Botter, R., T. Aste, and D. Beruto: 1994, 'Influence of microstructures on the func-
tional properties of tin oxide-based gas sensors'. Sensors and Actuators B {Chemical]
22, 27-35.
4. Brattain, W. H. and J. Bardeen: 1953. Bell Syst. Tech. J. 32, 1.
5. Correia-Pires, J. M., J. B. Almeida, and V. Teixeira: 2001, 'Characterization of
sputtered Sn02 thin coatings for gas sensing applications'. Vacuum. submitted.
354

6. Correia-Pires, J. M., J. B. Almeida, and V. Teixeira: 2002, 'Gas sensitive response

of Sn02 thin film sensors produced by reactive DC magnetron sputtering'. Key
Eng. Mat. 230-232, 388-391.
7. Cricenti, A., R Generosi, M. A. Scarselli, P. Perfetti, P. Siciliano, A. Serra, A.
Tepore, J. Almeida, C. Coluzza, and G. Margaritondo: 1996a, 'Pt:Sn02 thin films
for gas sensor characterized by atomic force microscopy and x-ray photoemission
spectromicroscopy'. J. Vac. Sci. Technol. B 14(2), 1527-1530.
8. Cricenti, A., R Generosi, M. A. Scarselli, P. Perfetti, P. Siciliano, A. Serra, A.
Tepore, C. Coluzza, J. Almeida, and G. Margaritondo: 1996b, 'Morphological,
chemical and electrical characterization of Pt-Sn02 thin film grown on rough and
mechanically polished Ab03 substrates'. J. Phys. D: Appl. Phys. 29, 2235-2239.
9. Ghiotti, G., A. Chiorino, W. X. Pan, and L. Marchese: 1992, 'A comparative study of
Sn02 and CrO x /Sn02 powders: their microstructures and their different response
t.o CO, NO, H2 and CH4 atmospheres'. Sensors and Actuators B [Chemical] 7,
691--699.
10. Gopel, W. and K.-D. Schierbaum: 1991, Electronic conductance and capacitance
sensors, Vol. 2 of Sensors: A comprehensive survey, Chapt. 9. VCH.
11. Heiland, G.: 1954. Z. Phys. 138, 459. H03.
12. Ho, J.-J., Y. K. Fang, K. H. Wu, W. T. Hsieh, C. H. Chen, G. S. Chen, M. S. Ju, J.-
J. Lin, and S. B. Hwang: 1998, 'High sensitivity ethanol gas sensor integrated with a
solid-state heater and thermal isolation improvement structure for legal drink-drive
limit detecting'. Sensors and Actuators B [Chemical} 50,227-233.
13. Hiibner, H.-P. and S. Drost: 1991, 'Tin oxide gas sensors: an analytical comparison of
gas-sensitive and non-gas-sensitive thin films'. Sensors and Actuators B [Chemical}
4,463-466.
14. Jin, Z., H.-J. Zhou, Z. L. Jin, R. F. Savinell, and C.-C. Liu: 1998, 'Application of
nano-crystalline porous tin oxide thin film for CO sensing'. Sensors and Actuators
B [Chemical] 52, 188-194.
15. Kissine, V. V., S. A. Voroshilov, and V. V. Sysoev: 1999, 'Oxygen flow effect on
gas sensitivity properties of tin oxide film prepared by r.f. sputtering'. Sensors and
Actuators B [Chemical] 55, 55-59.
16. Labeau, M., B. Gautheron, G. Delabouglise, J. Pena, V. Ragel, A. Varela, J. Roman,
J. Martinez, J. M. Gonzruez-Calbet, and M. Vallet-Regi: 1993, 'Synthesis, structure
and gas sensitivity properties of pure and doped SnO,'. Sensors and Actuators B
[Chemical} 15-16, 379-383.
17. Maekawa, T., J. Tamaki, N. Miura, N. Yamazoe, and S. Matsushima: 1992, 'Devel-
opment of SnO,-based ethanol gas sensor'. Sensors and Actuators B [Chemical] 9,
63-69.
18. Micocci, G., A. Serra, P. Siciliano, A. Tepore, and Z. Ali-Adib: 1996, 'CO sens-
ing characteristics of reactively sputtered Sn02 thin films prepared under different
oxygen partial pressure values'. Vacuum 47(10),1175-1177.
19. Morimitsu, M., Y. Ozaki, S. Suzuki, and M. Matsunaga: 2000, 'Effects of surface
modification with platinum and ruthenium on temperature and humidity depen-
dence of Sn02-based CO gas sensors'. Sensors and Actuators B [Chemical] 67,
184-188.
20. Morrison, S. R: 1955. Adv. Catal. 7, 259.
21. Ogawa, H., M. Nishikawa, and A. Abe: 1982, 'Hall measurement studies and an
electrical conduction model of tin oxide ultrafine particle films'. J. Appl. Phys.
53(6),4448-4455.
22. Osaki, Y., S. Suzuki, M. Morimitsu, and M. Matsunaga: 2000, 'Enhanced long-term
 355
stability of Sn02-based CO gas sensors modified by sulfuric acid treatment'. Sensors
and Actuators B [Chemical} 62, 22(}-225.
23. Ratcheva, T., 1. Stambolova, and K. Konstantinov: 1994, 'PH3 detection by Sn02-
Zr02 thin films'. Sensors and Actuators B [Chemical} 21, 199-204.
24. Rella, R., A. Serra, P. Siciliano, L. Vasanelli, G. De, and A. Licciulli: 1997, 'CO
sensing properties of Sn02 thin films prepared by the sol-gel process'. Thin Solid
Films 304, 339-343.
25. Ruhland, B., T. Becker, and G. Muller: 1998, 'Gas-kinetic interactions of nitrous
oxides with Sn02 surfaces'. Sensors and Actuators B [Chemical} 50, 85-94.
26. Swanepoel, R.: 1983, 'Determination of the thickness and optical constants of
amorphous silicon'. J. Phys. B: Sci. Instrum. 16, 1214-22.
27. Tamaki, J., T. Maekawa, N. Miura, and N. Yamazoe: 1992, 'CuO-Sn02 element for
highly sensitive and selective detection ofH2S'. Sensors and Actuators B [Chemical}
9, 197-203.
28. Vlachos, D. S., C. A. Papadopoulos, and J. N. Avaritsiotis: 1996, 'Dependence of
sensitivity of SnO z thin-film gas sensors on vacancy defects'. J. Appl. Phys. 80(10),
605<Hi054.
29. Voshchilova, R. M., D. P. Dimitrov, N. I. Dolotov, A. R. Kuz'min, A. V. Makhin,
V. A. Moshnikov, and Y. M. Tairov: 1995, 'Forming the structure of gas-sensitive
layers of tin dioxide produced by reactive magnetron sputtering'. Semiconductors
29(11), 1036-1039.
30. Williams, G. and G. S. V. Coles: 1993, 'NO z response of tin dioxide based gas
sensors'. Sensors and Actuators B [Chemical} 15-16,349-353.
31. Williams, G. and G. S. V. Coles: 1999, 'The gas-sensing potential of nanocrystalline
tin dioxide produced by a laser ablation technique'. MRS Bulletin 24(6),25-29.
32. Xu, C., J. Tamaki, N. Miura, and N. Yamazoe: 1991, 'Grain size effects on gas
sensitivity of porous Sn02-based elements'. Sensors and Actuators B [Chemical} 3,
147-155.
33. Yamazoe, N., Y. Kurokawa, and T. Seiyama: 1983, 'Effects of additives on
semiconductor gas sensors'. Sens. Actuators 4, 283-289.
34. Zhang, G. and M. Liu: 2000, 'Effect of particle size and dopant on properties of
Sn02-based gas sensors'. Sensors and Actuators B [Chemical} 69,144-152.
REHYBRIDIZATION AT (110) FACES OF Sn02-BASED GAS
SENSOR

V. V.GOLOVANOV, TAPIOT.RANTALA 1, TUOMOS.RANTALA 1 AND

A. KIV2

South-Ukrainian Pedagogical University, Odessa, 65008, Ukraine

iTampere University a/Technology, FIN-33101, Tampere, Finland
2Ben Gurian University a/the Negev, Beer Sheva, 84105 Israel

1. Introduction

Tin dioxide (Sn02) is a wide-gap semiconductor with applications as gas sensor material and
catalyst. The structure and electronic properties of its surface have been experimentally studied
in detail [1, 2]. Also theoretical investigations of oxide surfaces based on an accurate first-
principles calculations have been carried out [3-6]. Such approaches allow one to create an
atomistic picture of sensing mechanisms and methods for prognosis of gas response properties
of the semiconducting materials.
However, ab initio calculations of Sn02 (110) face based on utilizing the different
computational methods reveal diversity of bare surface relaxation data [3-6]. Therefore,
calculations based on different methods are welcome, to establish reliable structure of relaxed
Sn02 (110) atomic plane and to understand the related differences in rehybridization occurring
at stoichiometric and reduced surfaces.
Recently, we have investigated the sttu:ture of the (110) face of rutile structure Sn02
(cassiterite) with ab initio - DFT calculations, the local-density approximation (LDA) and the
generaIized-gradient approximation (GGA) [5]. There are also semi-empirical approaches
available for the surface relaxation and electronic structure of the (110) face of Sn02 to
compare with [10].
In the present study, by comparing with the ab initio data, we search for the best
fitting Universal force field (UFF) of Rappe et al. [8], where the hybridization is explicitly
assigned to the atoms. Such comparison of the data from ab initio calculations and results of
molecular mechanics (MM) minimization allows one to judge between the various presupposed
hybridizations.
A more detailed analysis of the hybridizations based on our ab initio electronic
structure results [5] has been included here, too. The Mulliken population analysis of the

357
Y. G. Gogotsi and I. V. Uvarova (eds.).
Nanostructured Materials and Coatings Jar Biomedical and Sensor Applications. 357-364.
© 2003 Kluwer Academic Publishers.
358

LCAO electronic structure has been used to project the molecular or crystal one-electron
wave functions onto the atomic orbitals, which were used as the basis set. Together with the
energy eigenvalues this defines the hybridization of original atomic orbitals and also reveals the
remaining lone pair electrons.

2. Slab Model and Computational Method

The rutile structure of Sn02 is 6:3 coordinated and the Sn4+ ions in the structure are centered in
slightly distorted oxygen octahedra. The (110) layers of the structure consist of neutral groups
of three parallel planes O-SnzOrO and the cleavage cut between these groups of planes
breaks the least number of cation-anion bonds. This results in (110) stoichiometric face, which
is the most dominant surface of the Sn02 material. Removal of the surfacemost bridging oxygen
layer of the stoichiometric surface results in the reduced (110) surface, see Fig. 1.
In the present study we examine the rehybridization of the surface atoms. This is
coupled with consideration of the force field, as many details of the potential functions are
hybridization dependent. We employ the Universal force field (UFF) which is suitable for our
purposes. It is a general purpose force field that has been parametrized from a set of rules
based on element, hybridization and connectivity. The UFF has been carefully validated for
many structure types. In UFF the hybridization is explicitly assigned to the atoms, defining the
nearest neighbor coordination attributes, such as equilibrium bond lengths and angles. The UFF
has been developed in conjunction with the Charge Equilibration method. This approach
predicting charge distribution in system makes it applicable for materials with high ionicity [9].
For MM calculations we have adopted the same periodic slab that was used for the ab
initio calculations [5]. We adopt the slab model of the (110) surface with an orthorhombic
supercell with 16 atoms (reduced, SIlr;OIO) or 18 atoms (stoichiometric, (Sn02)6) oriented as
shown in Fig.!. Vacuum between the surfaces was chosen to

Figure 1. (a) Primitive tetragonal unit cell of the bulk Sno,. The experimental lattice parameters are
a ~ 4.737 A, c ~ 3.186 A and u ~ 0.306. (b) Orthorhombic supercell of the slab model. The top face
of the supercell illustrates the stoichiometric and bottom face the reduced (110) surface.
 359

occupy half of the slab supercell. The upper and lower surfaces were treated identically.

3. Results and Discussion

Total energy minimization of the bulk Sn02 using ab initio calculations and starting from the
experimental lattice constants with the LCAO method leads to about 2% expansion, whereas
the expansion with PWPP is somewhat less [5]. We were able to find a force field for the
molecular mechanics calculations that exhibits negligible shrinking of the lattice constants
(0.001%), see Table 1. These bulk relaxed structures were used as reference to the surface
relaxation.

TABLE I: Bulk relaxation of cassiterite, Sno,.

Experimental LeAO pwpp PWPP[3) MM

a 4.737 4.832 4.730 4.637 4.721
c 3.186 3.266 3.212 3.060 3.185
u 0.307 0.307 0.306 0.307 0.307

The bonding between atoms in Sn02 has a relatively strong ionic character. At the
stoichiometric surface the tin atoms appear also fivefold coordinated with less ionic character.
Removal of the bridging oxygen changes the coordination of tin atoms from sixfold to fourfold,
which in turn, change ionic character of the bonding. Reduction of the surface leads to
rehybridization of the surface tin atoms and oxygen atoms that couples to the surface
reconstruction or relaxation. Establishing the correct electronic structure of the reduced and
stoichiometric Sn02 surfaces is important for the further investigations of adsorption/desorption
and catalytic properties connected with construction of adsorbate-surface site conformations
and understanding of the different reaction paths at the tin oxide surfaces.
The most prominent feature of surface relaxation confirmed by the two different ab
initio methods is outward displacement of in-plane oxygens of 0.4 A relative to the perfect-
lattice (or truncated bulk) positions. On one hand, such surface deformation at the reduced
surface may be connected with rehybridization of in-plane oxygens environment. On the other
hand, strong outward displacement of in-plane oxygens not connected directly with
rehybridization of Sn and/or 0 surface ions may be attributed to the changes in electrostatic
interactions between atoms, coupled with changes in the ionicity of the surface bonding.
Possible rehybridizations at the (110) surface of Sn02, leading to the outward
displacement of in-plane oxygen ions may be assigned to subbridging tin atoms, D.-plane
oxygen itself, or both of those at the same time.
Most of the surface Sn4+ ions usually have the octahedral valence state of the hybrid
sp 3d 2 orbitals. At the fully oxidized surface Sn4+ ions carry two oxygen monodentately
coordinated on it. If these oxygens are eliminated then the cations becomes four -coordinated.
360

This arrangement is energetically less favorable owing to the relatively high degree of
coordinate unsaturation of the cation and could lead to the surface reconstruction. A possible
mechanism of the reconstruction is the change in the rehybridization of the cation into some
inactive fonn, for example, into distorted sp3. Usually [10,11], the surface reconstruction
resulted in such rehybridization displace the superficial cations down and cause the oxygens
raise above the surface plane (Fig. Ib).
Another possible reason for the relaxation leading to the outward displacement of in-
plane oxygen atoms at the reduced surface is rehybridization of in-plane oxygen from the bulk
configuration, which is sp2 plus a non-bonding, lone-electron pair. At the reduced surface, the
oxygen atoms are in effort to lower their electronic energies, attempt to move out from the
surface into more tetrahedral Sp3 configuration.
It must be noted, that change in the coordination of subbridging tin from sixfold to
fourfold at the reduced surface, generates unsaturated dangling bonds, making rehybridization
highly probable. In-plane oxygen at reduced surface remains three-fold coordinated and its
rehybridization can take place only if it is energetically more favorable.
The surface geometry of Sn02 (110) face from ab initio calculations and molecular
mechanics (MM) minimization with various hybridizations is shown in Table 2. For the reduced
surface both octahedra1ly and tetrahedrally hybridized tin and oxygen are considered. For the
stoichiometric surface octahedrally hybridized tin leads to a better description of relaxation.
According to previous predictions [10,11] tetrahedral hybridization of subbridging tin atoms at
reduced surface accompanied by outward displacement of in-plane oxygens. At the same time,
subbridging tin atoms move down significantly, that displays main difference between the MM
calculations and ab initio results, where an outward displacement i; found, instead. Such
strong difference between MM and ab initio results makes this choice ofFF rather unreliable.
How it follows from the MM calculations (Table 2), assignment of the tetrahedral
hybridization to in-plane oxygen ions is not enough to provide an outward displacement of 0.4
A.
TABLE 2: Vertical displacements of the surface atoms from their original "bulk cut" lattice sites (in ?) from
the ab initio calculations (LCAO and PWPP) and molecular mechanics (zero temperature molecular
dynamics). Tin has been considered in octahedral and tetrahedral, and oxygen in trigonal and tetrahedral
hybridizations.

LCAO PWPP Sn-oct Sn-oct Sn-tet Sn-tet

.(GGA) [5] (GGA) [51 O-tri O-tet O-tri O-tet
Stoichiometric
Bridging tin (6 fold) +{l.B +{l.21 +{l.()43 +{l.033
Bridging oxygen +{l.05 +{l.l3 -0.038 -0.050
In-plane oxygen +{l.18 +{l.22 +{l.IOO +{l.124
Reduced
Bridging tin (4 fold) +{l.l4 +{l.20 -0.125 -0.147 -0.26 -0.33
In-plane oxygen +{l.39 +{l.49 +{l.034 +{l.073 +{l.38 +{l.50
 361

Moreover, minimization of the flexible outennost surface layer of (110) Sn02 face,
consisted of the local environment of the superficial Sn and 0 ions, reveals antisymmetrical
(110) surface for oxygen in tetrahedral state (Fig. 2b), and makes this setting less probable.
The surface symmetry was restored, however, for more rigid slab model (Fig. Ib).
Nevertheless, applying the tetrahedral hybridization for tin and oxygen surface atoms leads to
increase of the slab total energy (Table 3), which demonstrates that such choice is energetically
less favorable. Such results are also in good agreement with [10], where it was concluded that
the oxygen atoms cannot fully rehybridize, because of the large local strain field (i.e., the large
energy cost associated with significant accompanying changes in bond lengths).

a) b)

Figure 2. Model of the outermost surface layer of(IIO) SnD, face for the a) Sn-tet+O-tri and b) Sn-
oct+O-tet hybridization states.

TABLE 3: Comparison of the total energies of the stoichiometric and reduced slabs with different
hybridizations.

Sn-oct Sn-oct Sn-tet Sn-tet

O-tri O-tet O-tri O-tet
Reduced -151.12 -153.11 -165.15 -166.38

Mulliken population analysis of LeAO electronic structure gives information on the

rehybridization at the tin oxide surface directly. Computed energy levels from the stoichiometric
and reduced surfaces of the reduced, SIl60 IO or stoichiometric, (Sn02)6 slabs are presented in
the Fig. 3.
Mulliken population analysis performed for stoichiometric and reduced surfaces of
Sn02 displays presence of the lone-electron pair at in-plane oxygen atoms in both surface
states (Fig.3) indicating that these ions remain in a trigonal hybridization state - Sp2 plus a non-
bonding, lone-electron pair. Such conclusion is in agreement with results of our MM
calculations.
At the stoichiometric surface d-orbitals are involved in the bonding of subbridging tin
ions, supporting conclusion made from MM calculations about the octahedral valence state of
362

the hybrid sp3 d 2 orbitals. It is extremely significant that dorbitals are still involved in the
bonding of subbridging tin atoms at the reduced surface. Such results exclude capability for
tetrahedral Sp3 state of the superficial tin orbitals and also go along with data of MM
calculations. Participation of the d-orbitals in the bonding of the subbridging tin atoms allows
one to assume, for instance, distorted trigonal bipyramidal sp3 d hybridization state for such
cations, where 5s-5p hybrid lone pair occupying stereochemical position on the vacuum side of
the ions.
PA Cox et at [12] have suggested that after removing of the bridging oxygen atoms
at (110) face two electrons remains at subbridging tin making it sri+. The defect states near the
valence-band maximum (VBM) are due to a rehybridization of Sn 5s-5p states at reduced
Sn2+ sites adjacent to oxygen vacancies or in local SnO-like environments at the surface. An
alternative suggestion explaining appearance of Sri+-like ions is that the inert pair of 5s
electrons occupies the apex of the tetragonal pyramid, how it takes place in tetragonal SnO
structure.
The Mulliken population analysis perfonned on Sn02 structure allowed us to attribute
an occupied electron energy levels related to the subbridging tin ions with the 5s- 5p hybrid lone
pair located near the VBM (-4.33 eV), see Fig. 3. Removing of the bridging oxygen atoms
results in strong dispersion at the VBM, which has four-fold tin 5s and 5p character. These
levels do not relate to the bonds to the neighboring in-plane oxygen atoms, however. Thus,
these are loosely bound electrons from the bridging oxygen vacancies, the levels which are
experimentally seen in the band gap [12].

s 111I1I111111111 ~ 11111111 II I I III 111111

R III ~ 11II1I1111111111 I IIII I IIII1 111111111 III

B

-4 -2 o 2 4 6 8
eV

Figure 3. Computed LCAO one-electron levels for the stoichiometric (SnO,)6 (S) and reduced,
Sn60,O (R) surfaces, and for the bulk (B). A scissor operator of2.2 eV has been applied to make the
band gap close to its bulk value. The fractional occupation numbers in the case of reduced surface
are illustrated by the barlength. For the bulk and stoichiometric cases the VBM is adjusted to 0 eV
and for the reduced case it is adjusted to fit the bands in Ref. 5b.
 363

The Mulliken population analysis allows one to understand the ionic character of the
bonding in crystal. Localized orbitals may be constructed from hybrids on both Sn and 0 and
these overlap, forming occupied bonding and unoccupied antibonding orbitals. Because of the
very different electronegativities of Sn and 0, the bonding orbitals are concentrated on 0
atoms and antibonding - on Sn. The ionic crystal can be therefore regarded as the case where
bonding and antibonding orbitals become completely localized on oxygen and metal,
respectively. It was found, that at the reduced surface the bonding orbitals related to the
subbridging tin atoms are localized on the metal ions rather then on oxygens, displaying
covalent character of the bonding of such tin atoms. This is in agreement with well-known
covalent character of the bonding in SnO crystal.
The results of the ab initio and MM calculations indicates that the two electrons left
behind after removing of the bridging oxygen ions occupy orbitals (a mixture of 5s and 5p) on
surface Sn ions, converting them to sd+ [12,13] and abnormal stereochemistry with inert
electron pair at the reduced surface of Sn02 is due to this sop hybridizations. Therefore
observed strong outward displacement of in-plane oxygen ions cannot be connected directly
with tetrahedral rehybridization neither superficial tin or oxygen atoms and may be attributed to
the changes in electrostatic interactions between atoms, coupled with changes in the ionicity of
the surface bonding.
As it was mentioned Sn02 must be regarded as essentially ionic crystal. On the other
hand, appearance of the Sn2+ atoms at the reduced surface of Sn02 changes character of the
bonding to the essentially covalent. Therefore, bound between subbridging tin and in-plane
oxygen became weaker at the reduced Sn02 surface. Atomic configuration of the surface is
determined by superposition of the stretching surface bondings and repulsion electrostatic
interaction with neighboring oxygen atoms. Weakening of the superficial Sn-O bonding, due to
the reduction of the tin ions from Sn4+ to Sn2+ leads to the domination of the electrostatic
repulsion interaction and resulted in the observed outward displacement of the in-plane oxygen
atoms.

4. Conclusion

Relaxation of the (110) surface of Sn02 occurs essentially in the perpendicular direction, only.
In-plane relaxation is negligible. Symmetry breaking reconstruction was found only in case
where the in-plane oxygen has been allowed to rehybridize. The most prominent feature is the
surface layer anion relaxation outwards with respect to their bulk positions at the reduced
surface. This appears to be strong at the reduced surface, about 0.4 A. The mechanism
responsible for such relaxation is supposed to be rehybridization of the subbridging tin cation
coupled with changes in the ionicity of the surface bonding. Some arguments were found to
assume that in-plane oxygen remains in trigonal hybridization state.
364

Acknowledgements

This work was supported by EC in the frame ofINTAS Program (Grant INTAS - 01- 0009).

References

1. V. A Henrich and, P. A Cox. (1994) The Surface Science of Metal Oxides,

University Press, Cambridge, p.432.
2. R.G. Egdell. (1994) in The Science of Ceramics Interfaces II, ed. 1. Nowotny,
Elsevier, N.Y. p.527.
3. 1. Manassidis and M.J. Gillan. (1994) First-principles study ofSn02 (110), Surf Rev.
and Lett., 1, No.4, 491-494.
4. 1. Manassidis, 1. Goniakowski, L.N. Kantorovich and M.J. Gillan. (1955) The
structure of th: stoichiometric and reduced Sn02 surface, Sur/Sci, 339,258,381-
394.
5. Tapio Rantala, Tuomo Rantala, Vilho Lantto. (1999) Surface relaxation of the (110)
face of rutile Sn02, Surf Sci., 420, 103-109; Tapio T. Rantala, Tuomo S. Rantala
and Vilho Lantto. (2000) Electronic structure of Sn02 (110) surface, Materials
Science in Semiconductor Processing, 3, 103-107.
6. 1. Goniakowski, M.J. Gillan. (1996) The adsorption of H.z0 on Ti0 2 and Sn02 (110)
studied by first-principles calculations, Surf Sci., 350,145-158.
7. M. Prutton. (1975) Surface physics, Oxford Univ. press, London, p.245.
8. AK. Rappe, C.J. Casewit, K.S. Colwell, W.A. Goddard III and W.M. Skiff. (1992)
UFF, a full periodic table force field for molecular mechanics and molecular dynamics
simulation,J. Am., Chern. Soc., 114, 10024-10035.
9. AK. Rappe and W.A. Goddard III. (1991) Charge equilibration for molecular
dynamics simuiations,J. Phys. Chern., 95, 3358-3363.
10. T.J Godin and 1. P. LaFemina. (1993) Surface atomic and electronic structure of
cassiterite Sn02 (110), Phys. Rev. B, 47, No. 11, 6518--{)523.
11. M. Egashira, M. Nakashima, S. Kawasumi. (1981) Temperature programmed
desorptiion study of water adsorbed on metal oxides. 2. Tin oxide surfaces, J. Phys.
Chern., 85, 4125-4130.
12. D.F. Cox, T.B. Fryberger, S. Semancik. (1988) Oxygen vacancies and defect
electronic states on the Sn02(l1O)-lxl surface, Phys. Rev.B, 38, No.3, 2072-2083.
13. Tapio T.Rantaia. (1999) Ab initio studies of compound semiconductor surfaces in
Computational Studies afNew Materials, Eds. D.A Jelski and T.F. George, World
Scientific, Singapore, 6--26.
BEHAVIOUR OF MICRO-FABRICATED COMPOSITE MEMBRANE AS
AMPEROMETRIC GLUCOSE BIOSENSOR

C. M. PEREIRA, J. M. OLIVEIRA AND F. SILVA

Departamento de Quimica
Faculdade de Ciencias da Universidade do Porto
R. do Campo Alegre, 687
4169-007 Porto - Portugal
e-mail: cmpereir@fc.up.pt

Abstract

An electrochemical glucose biosensor based on the assisted transfer of It across a

gel/liquid interface was developed. The transducer element consisted of a composite
membrane prepared by assembling a micromachined polyester film with a DBS-NPOE
gel layer. The biocomponent, glucose oxidase (GOx), was attached to the composite
membrane by adsorption. The glucose analysis was made following the gluconic acid
production by the enzymatic reaction between glucose and glucose oxidase. The
change in pH was monitored by measuring the current intensity due to the transfer of
H+ across the interface between the organic gel and aqueous solution assisted by an
ionophore present in the organic phase.

1. Introduction

Selectivity is a problem of extreme importance in analytical chemistry and it is

seldom achieved due to the lack of chemical reactions that satisfies the appropriate
requirements. However extraordinarily selective and versatile reagents are available in
nature in various forms such as enzymes, antibodies and receptors.
The use of enzyme technology to build sensors has some advantages over
ordinary chemical sensors primarily because it offers a wide combination of selectivity
and sensitivity and allows a wide range of transduction technology to be used.
Biosensors can be defined as devices that use immobilized biomolecules
combined with transducers to detect or respond to specific interactions with the analyte.
The biosensor must have some particular properties, namely the biological component
must be specific and stable as well as the response should be accurate, precise and
reproducible. The reaction should be as independent as possible from physical
parameters such as pH, temperature and stirring and furthermore the sensing element
should be small and biocompatible. At last the complete unit should be cheap and
portable.
365
Y. G. Gogotsi and l. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications. 365-370.
© 2003 Kluwer Academic Publishers.
366

Over the last two decades there has been a growing interest to use interfaces
between two immiscible electrolyte solutions (ITIES) to follow enzyme reactions [1]
and to build amperometric biosensor devices [2]. The first work using ion transfer
reactions through ITIES was reported by Osakai et al. [2] and described the
development of a biosensor based on the ammonium produced during the urease
decomposition of urea. The facilitated transfer of ammonium ion through the ITIES
was the detection method used [3]. In a similar manner creatinine was assayed at a
microiTIES coupling the enzyme reaction with ion transfer process [4].
Glucose oxidase catalyses the oxidation of B-D-glucose in a highly specific way
and is the recognition element in most glucose electrochemical biosensors [5].
The gluconic acid generated in the enzyme cycle can be followed by the use of a
protonable ligand in the organic phase opening the way to a vast field of new
application for ITIES. The transfer of a charged specie (H+) from water (w) to an
organic phase (0) facilitated by a ligand is based in the method first described by
Koryta r61. Figure 1 shows a simplified model of the detection process.

gluconate

organic phase aqueous phase

Figure 1. Schematic representation of the assisted transfer of Ir through the organic gel/aqueous
solution used in the detection process.

2. Experimental

The aqueous and organic phase solvents were respectively ultrapure water, Milli-Q,
Millipore purification system (p =18 MO) and 2-nitrophenyloctylether, o-NPOE
(Fluka, Selectophore). The organic solvent was used without any further purification
however the bottle was always kept in the refrigerator. All solvent manipulations were
carried out in a glove box under a nitrogen atmosphere.
 367

The supporting electrolytes used were 10 mM NaCl (Merk, p.a. grade) in the
aqueous phase and 10 mM tetraoctylammonium tetraphenylborate, TOATPB in the
organic phase, the latter was prepared as described elsewhere [7].
The organic gelifying agent was 1,3:2,4-dibenzylidene sorbitol, DBS (Milliken),
and agar (DlFCO) as aqueous gelifying agent. The complexing agent used in this work
was the 3-(2-pyridil)-5,6-diphenyl-I,2,4-triazine, PDT (Aldrich). D(+)-Glucose
monohydrate and glucose oxidase (EC 1.1.3.4, 23U/mg, from Aspergillus niger) were
supplied by Fluka. GOx solutions were prepared by dissolving analytical-grade reagent
in ultrapure water. All glassware was thoroughly cleaned by rinsing first with acetone
(Riedel-de Haen) and then with ultrapure water.
The microhole array was prepared by UV excimer laser photoablation of 12 /lm
thick Melinex (S grade) film at the Laboratoire d'Electrochimie in EPFL - Lausanne. A
detailed procedure for microhole array fabrication is given by Seddon et al. [8]. The
array used throughout this work, consists of a hexagonal pattern of 66 holes of 10 /lm
diameter that are 110 /lm apart. The film was glued to a glass cylinder, using a solvent
resistant fluorosilicone sealant 730 RTV (Dow Corning Corp - USA).
The details of preparation and properties of the gels used as aqueous reference
and organic phases can be found in the literature [9, 10]. In order to assure that the
gelification process is completed the gel was allowed to set overnight.
The electrochemical measurements were carried on a SI 1287 Electrochemical
Interface (Solatron Instruments) and the data acquired using home developed MS-DOS
software. The cell used throughout this work can be schematically represented as:

Ag IAgClllOmMNaTPBllOmMTOATPB
IOmMNaCl yMPDT
r zMglucose
xMGOx
I AgCI lAg'

agar 1.5% DBS 1.8 % LiCI 10 mM

(w) (o-npoe) (w)

where • indicates the polarized interface and y, z and x represent the concentration of PDT, glucose and
GOx, respectively.

Cyclic voltammograms (CVs) were recorded with a scan rate of 50 mV/s. The
amperometric experiments were carried out applying to the electrochemical cell a
potential difference in such a way that the assisted proton transfer through the interface
can be monitored (circa 0 V).
All experiments were performed at room temperature of 20±2°C.

3. Results and Discussion

In order to work as a transducer ITIES relies on the transfer of charged species through
the liquid-liquid interfaces. Since the glucose molecule is neutral it is not suitable to be
detected voltammetrically at ITIES.
When both the ionophore and enzyme are present respectively in the organic
and in the aqueous phase, a net increase of current intensity is observed in the cyclic
368
voltammograms as shown in figure 2. When the ionophore is present the proton formed
during GOx cycle can be transferred from aqueous phase to the organic phase resulting
in the increase of current intensity at more positive potentials.

9
--baseline
6 - - 7 mM glucose
- - -- - 56 mM glucose
3

i. 0

---..... -3

-6

-9
-800 -600 -400 -200 o 200

E/mV

Figure 2. Cyclic voltammograrns obtained when 10 nM GOx is present in the aqueous phase and
20 mM PDT in the organic phase. Effect of glucose concentration on the peak current
intensity for the assisted transfer onr.

The optimization of sensor performance should consider the effect of GOx

ionophore concentrations. Figure 3 shows that there is an increase of the peak current
when the glucose concentration increases. For higher concentration of glucose the peak
current tends to a maximum and this value depends both on the ligand concentration
and GOx concentration. The shape of this calibration curve is typical of an enzymatic
amperometric detection.
Figure 3 also shows a decrease of the peak current when the concentration of
GOx is increased from 10 nM to 500 nM. This behavior may be related to different
states of adsorbed GOx at higher concentrations [11].
 369

7
• • •
•
• •
6

5 • • 10 nM GOx
• •
--
500 nM GOx
~ 4
• 10 11M GOx

3
• • • •• • • ••
•
•••• • • • • • • •
2

• • •

0 1 2 3 4 5 6

CglucoseImM

Figure 3. Peak current as a function of the concentration of glucose. Effect of glucose oxidase
concentration in the aqueous phase. (5 mM PDT in the organic phase).

4. Conclusions

The simultaneous presence of two elements (ionophore/GOx) is needed for operability

of a glucose amperometric biosensor based on !TIES.
The biosensor's performance is optimized using GOx solutions with
concentrations close to IO nM. The better performance of the biosensor for lower GOx
concentrations may be related to a change in the adsorbed state of the enzyme.
The sensitivity of proton assisted transfer is dependent on the ionophore
concentration and a higher sensitivity is obtained for lower concentration of ligand.
This model biosensor satisfies some needed properties for biosensor's
development such as specificity, stability, precision and reproducibility (data not
shown).
The effect of some known interferents for the glucose sensor is under study.
370

Acknowledgements The authors wish to acknowledge financial support given by FCT

(CIQ-L4).

REFERENCES
I Vanysek, P. and Behrendt, M. (1981) Investigation of acetylcholine, choline and acetylcholinesterase at
the interface of the two immiscible electrolyte solutions J. Electroanal. Chem. 130,287-292.
2. Osakai, T" Kakuatani, T and Senda, M. (1988) A novel amperometric urea sensor Analytical Sciences,
4,529-531,
3. Osakai, T, Kakuatani, T. and Senda, M, (1987) A novel amperometric ammonia sensor Analytical
Sciences, 3, 521-527.
4. Osborne, M.D. and Girault, RH, (1995) The micro waterll,2-dichloroethane interface as a transducer for
creatinine assay Mickrochim, Acta, 117, 175-185,
5, Memoli, A, Annesini, M,e" Mascini, M., Papale, S. and Petralito, S. (2002) A comparison between
different immobilized glucose oxidase-based electrodes Journal of Pharmaceutical and Biomedical
Analysis 29, 1045-1052.
6. Koryta, J. (1979) Electrochemical polarization phenomena at the interface of two immiscible electrolyte
solutions Electrochim, Acta 24, 293-300
7, Pereira, eM., Schmickler, W., Silva, F, Sousa, M.D. (1997) Ion association at liquidlliquid interfacesJ,
Electroanal. Chem. 436, 9-15.
8. Seddon, BJ, Shao, Y., Fost, 1. and Girault, RR(l994) The application ofexcimer laser micromachining
for the fabrication of disc microelectrodes Electrochim. Acta 39, 783-791.
9, Silva, F" Sousa, M,D" Pereira, eM, (1997) Electrochemical study of aqueous-organic gel micro-
interfaces, Electrochim, Acta 42,3095-3103,
10 Pereira, eM., Tirilly, N., Martins, M.e and Silva, F, (2001) Development of Zn(II) sensors based on the
assisted transfer of metal ions by hydrophobic ligands through gel-supported microinterfaces Fresenius
1. Anal, Chem, 369,609-{j12.
II. Georganopoulou, D,G, Williams, DE, Pereira, eM., Silva, F" Su, T-1. and Lu, J.R. in press
Adsorption Of Glucose Oxidase At Organic-Aqueous And Air-Aqueous Interfaces
EQCM STUDY ON THE POLYTYRAMINE MODIFIED ELECTRODES FOR
THE PREPARATION OF BIOSENSORS

C. M. CORDAS', A. TENREIROb, L. M. ABRANTES'

a Departamento de Quimica e Bioquimica, Faculdade de Ciencias da
Universidade de Lisboa, Campo Grande, 1749 - 016 Lisboa
b Laboratorio de Bioelectroquimica, [CAT, Faculdade de Ciencias da
Universidade de Lisboa, Campo Grande, 1749 - 016 Lisboa

1. Introduction

Modified electrodes prepared by electropolymerisation of different monomers aiming

the immobilisation of biomolecules have known a growing interest in the recent years.
Among the advantages concerning this approach for biosensor preparation are the
control of the thickness and structure of the electrode coatings and its consequent
properties like permeability, stability and reproducibility [I].
Several methods are used to immobilise the biomolecules on the films such as their
entrapment during the polymerisation [2] or posterior attachment onto the polymer
surfaces by chemical grafting [3]. The polymerisation of functionalized monomers
allowing a more specific interaction by chemical affinity between the species of interest
and the film is another route to achieve a biosensor [4]. This method has been used
with monomers bearing specific groups, as tyramine (Ty), which can enhance the
interaction between polymers and electroactive species through the formation of
covalent bonds either between the amino and aldehyde or carboxyl groups [5, 6]. In the
literature the tyramine electropolymerisation mechanism is associated with the phenol
group oxidation and subsequent radical formation [7] as represented in figure I.

tyramine polytyraminc

Figure I. Schematic representation of the tyramine polymerisation.

The potentiodynamic electropolymerization of tyramine is often described as an

effective route to prepare the polymer modified electrodes and has also been employed
in this study [8].
371
YG. Gogorsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 371-376.
© 2003 Kluwer Academic Publishers.
372

The attachment of nucleotides, such as the deoxyguanosine triphosphate (dGTP)

can be accomplished by covalent coupling with l-ethyl-3-(3-dimethylaminopropyl)-
carbodiimide (EDC) in the presence of N-hidroxysuccinimide (NHS) [9]. The
immobilisation process can be evaluated by electrochemical quartz crystal microbalance
(EQCM) [10]. The advantage of the EQCM is that this technique offers very sensitive
detection of mass variation which can be use to identify the biomolecules grafting on
the polymer and the hybridization of the oligonucleotides without the use of another
type of transducers like optical or redox indicators [11].
In the present work, the biomolecule immobilisation in polytyramine films has
been investigated by open circuit potential measurements and by EQCM technique. The
data allows to discuss the role of the polymer matrix structure I morphology imparted
by the electropolymerisation conditions, on the modified electrodes performance.

2. Experimental

Tyramine (4-hydroxyphenethylamine) and the coupling agents N-hydroxysuccinimide

(NHS) and 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC) were obtained from
Sigma. Characterisation solutions were phosphate buffer, (Na2HP04 and NaH 2P04)
and ferro/ferricyanide, K3[FeCN6] and ~[FeCN6] in KN03, which were p.a. grade and
obtained from Merck. Deoxyguanosine triphosphate (dGTP), PCR grade, was
purchased from Invitrogen Life Technologies. All reagents were used as received and
Millipore water used to prepare the solutions.
The electrochemical and gravimetric measurements were performed in a 420
Model CH Instruments Electrochemical Quartz Crystal Microbalance. One
r
compartment EQCM cell was used. The workin electrode was a 8 MHz AT-Cut quartz
crystal coated with 1000 A Pt disk with 0.2 cm geometrical area, the counter electrode
a Pt wire and the reference a saturated calomel electrode (SCE).
Polytyramine (PTy) films were obtained by consecutive cycling in the potential
domain -0.1 V and different anodic potential limits, using 50 mV s'\ potential scan rate,
from a solution containing 0.5 mol dm·3 ofTy in phosphate buffer, pH 6.8 (0.1 mol dm,3
Na2HP04and 0.1 mol dm,3 NaH2P04).
The immobilisation of the dGTP onto the potentiodinamically synthesised PTy
films was performed at room temperature from solutions containing 0.001 mol dm,3 of
the nucleotide, 0.03 mol dm,3 NHS and 0.01 mol dm,3 EDC in water.

3. Results and Discussion

3.1 PREPARAnON OF POLYTYRAMINE MODIFIED ELECTRODES

The current and frequency responses of a Pt electrode subjected to a potential scan

in Ty containing solution are illustrated in figure 2.
The monomer oxidation starts at 0.46 V (figure 2 A) and the anodic current at
more positive potentials suggests the occurrence of a sequence of oxidation processes
which are probably related to different reactions taking place during the polymerisation
of tyramine. Notwithstanding, the polymer deposition on the electrode is clearly
 373

indicated by the decrease in the frequency observed for E ~ 0.51 V (figure 2B). During
the cathodic potential sweep, there is no indication of polymer reduction (no cathodic
wave in the cyclic voltammogram) or removal from the electrode (no change in
frequency). The second potential cycle indicates a non-conductive character of the
polymeric matrix, which is responsible for the small observed current (figure IA) and
for the low frequency decrease (figure IB) .
.
'.
;::

,,,
;~.ij.
,.,
'"
~~

i, ,. =!! '#)

C> ;;: '"

'"
'"
:~~==~~~=====:::J
U '0 ta
""
~ ~ ~

E/V.,SCE E1Yv.SCE

Figure 2. 1" and 2"" cyclic voltammograms (A) and simultaneously recorded frequency variation with the
applied potential (B) of a PI electrode in 05 mol dm" Ty I phosphate buffer; v=50 m V 51,

Taking into account the above mentioned features, lower anodic potential limits,
E,.., should be used in the potentiodynamic preparation of PTy modified electrodes.
Aiming the adequate definition of that potential limit, several PTy films of similar mass,
(evaluated by EQCM data) have been prepared and their behaviour analysed in solution
containing the Fe(CN 6)3+/Fe(CN 6)2+ redox couple and compared to the bare Pt electrode
response. The respective cyclic voltammograms, displayed in figure 3, confirm the
passive nature of PTy layers which dramatically increases with the EI.a used for their
synthesis.
Films grown with E'.a higher than 0.5 V hinder the electroactive species fluxes to
and from the electrode surface; the observed signoidal current shape (inlet of figure 3) is
very likely due to very compact structures of the polymer layers prepared under those
conditions.
so
~o
40
3Q
.n

1
to

-
..., ·'0 ~, f

I
20
-30 3 and 4
-,hi
4(,

.'"
-60
0):1) ""to Ot~ (,21) O~ 0_40 000

E/V •• SCE

Figure 3. Cyclic voltammograms obtained in 0.1 mol dm" K,[FeCN.] I 0.1 mol dm" K.[FeCN.] in 1 mol dm"
KNO, at PI (1) and PTy synthesised using different E 1•a : 0.5 V (2); 0.6 V (3) and 0.7 V (4); ); v=50 mV 51.
Inlet: magnification of current responses for 3 and 4.
374

Figures 4 and 5 present the first cycles and the respective change in the frequency
for the potentiodynamic growth of PTy, keeping the anodic potential limit at 0.5 V and
0.6 V, respectively. Although the voltammetric response suggests a self-limiting growth
after a few cycles, the gravimetric data analysis shows a continuous mass increase. It is
worthwhile to point out that a similar increase in mass is observed after 25 cycles for
the first case whereas only 14 are required if higher anodic potential limit is chosen.
Therefore, under these conditions, polymer layers with dissimilar structures shall also
be expected.

1.... -.IS..

EIV".SCE E/VnSCE

Figure 4. Cyclic voltammograms (25 cycles) (A) and simultaneously recorded frequency variation (8) for the
potentiodynamic growth ofPTy on Pt from 0.5 mol dm~3 Ty/ phosphate buffer; v~50 mV 51; E1.a ~.5 V.
n.---~~-~-~-~-~-,

14'
21' A
I.e i
15 '

t---.~,~,,,,""'"-,...._:..~~ ,~ ... __.... _ p.~

·24 a ~._~"~'¥''''''''''''''''¥'''.-r''H'''''''''''~'.'<O.''¥'~''t •••.••• ~_H!,··,.··-.·_·~·,.··,.··.·····t
0'0 OiG O~ 000 il.'j) coo- .o1O 0 0'0 020 0:1) OM) OID cro
E/VvoSCE E/VvoSCE

Figure 5. Cyclic voltammograms (14 cycles) (A) and simultaneously recorded frequency variation (8) for the
potentiodynamic growth ofPTy on Pt from 0.5 mol dm· 3 Ty/ phosphate buffer; v=50 mV s'; E1.a ~.6 V.

3.2 NUCLEOTIDE IMMOBILISATION

The performances of potentiodynamically grown PTy films (100 cycles and anodic
limits of 0.5 V and 0.6 V, respectively) towards the biomolecule immobilisation
process have been investigated. For that purpose the polymers were subjected to a pre-
treatment conslstmg on a polarisation at --{).l V for 900 s for the expulsion of
incorporated anions, and then rinsed with Millipore water in order to remove any
 375

monomers left in the film surface and porous. The so-treated modified electrodes were
then immersed in the solution containing the dGTP and the coupling agents to
accomplish the immobilisation.
The frequency variation, ~f, with time and the evolution of the open circuit
potential ,E oe , have been recorded for several hours.

tis
o 1000 2000 3000 4000 5000 6000 7000
0.45

~~
0.40

. ~ 0.35 B

~:~:
~ 0.30
:r:
;. 0.25 ~

1
iil 0.20 ~

0.15
0.10
0 2000 4000 6000 8000
lis

Figure 6. Frequency variation (A) and evolution of open circuit potential (8) during the dOTP immobilisation
in PTy film, grown with 100 cycles between -0.1 and 0.5 V (in 0.5 mol dm" Ty I phosphate buffer on Pt);
v=50mVs·'.

tl s
o 1(XX) 200) 300l 4<XXl 5<XXl trol 'iIDl
0.45,--------------,
0.40

"' 0.35
~ 0.30
; 0.25
Oil 0.20
0.15
0.1O+---~--~--~--__1
o 2000 4000 6000 8000
tis

Figure 7. Frequency varIation (A) and evolution of open circuit potential (8) during the dOTP
immobilisation in PTy film, grown with 100 cycles between -D.I and 0.6 V (in 0.5 mol dm" Ty I phosphate
buffer on Pt); v=50 mV s·'.

As expected and shown in figures 6 and 7, providing evidence of the dGTP

insertion into the polymer, a continuous decrease of both ~f and Eoe with time is
observed; after about 4 hours the parameters approach constant values.
In the presence of the dGTP, the film prepared with the lowest anodic potential limit
present a total frequency variation much higher (at least 2 fold) than observed for the
polymer synthesised with anodic potential limit of 0.6 V; the correspondent amount of
immobilized dGTP (1.6 nmol) is a far reasonable value considering biosensing
purposes.
The observed behaviour justifies further research efforts for the optimisation of the
conditions for Ty electropolymerisation.
376

4. Conclusions

The morphology/structure and relative electroactivity of electrochemically prepared

PTy films strongly depend on the conditions of polymer growth.
The EQCM has proven to be a powerful tool for the characterisation of polymer
synthesis; it has been successfully employed to provide evidence that the polymerisation
of tyramine is not a self-limiting process, in spite of the low current densities observed
in the cyclic voltammetry. It also elucidates about the influence of
electropolymerisation conditions on the films structure.
Tests on the performance of PTy films towards the immobilisation of dGTP
revealed promising characteristics: up to 1.6 nmol ofbiomolecule / 490 ng polymer.
Under strict control, the potentiodynamic electropolymerisation of Ty originates
polymer films displaying interesting properties for biosensors preparation.

Acknowledgements

C. M. Cordas gratefully acknowledge to the Fund"l'ao para a Ciencia e Tecnologia for the scholarship (SFRH
/ BD / 2917 / 2000) and to the NATO and the INVOTAN in Portugal for the financial support to participate in
the NATO ARW (Kiev, 2002).

References

1. Cosnier, S. (1999) Biomolecule immobilization on electrode surfaces by entrapment or attachment to

electrochemically polymerized films. A review., Biosensors & Bioelectronics 14, 443-456.
2. Wang, J. and Jiang, M. (2000) Towards genoelectronics: nuclei acid doped conducting polymers,
Langmuir 16, 2269-2274.
3. Palmisano, F., De Benedetto, G. E. and Zambonin, e. G. (1997) Lactate amperometric biosensor based
on an electrosynthesized bilayer film with covalently immobilized enzyme, Analyst 122, 365-369.
4. Lassalle, N., Vieil, E., Correia, J. P. and Abrantes, L. M. (2001) Study of DNA hybridization grafted
with oligonucleotides by photocutrent spectroscopy, Biosensors & Bioelectronics 16,295-303.
5. Dubois, J-E., Lacaze, p-e., Pham, M. e. (1981) Obtaining thin films of "reactive polymers" on metal
surfaces by electrochemical polymerization. Part Ill. Amino-substituted polyphenylene oxide films.
Application to preparation offerrocene electroactive films, J. Electroanal. Chern. 117,233-241.
6. Situmorang, M, Gooding, J. J., Hibbert, D. B. and Barnett, D. (1998) Electrodeposited polytyramine as
an immobilisation matrix for enzyme biosensors, Biosensors & Bioelectronics 13, 953-962.
7. Torii, S. (1985) Electroorganic Syntheses - Methods and Applications. Part I:Oxidations, Kodansha,
Tokyo.
8. De Benedetto, G. E., Palmisano, F. and Zambonin P. G. (1996) Flow-through tyrosinase enzyme reactor
based on reticulated vitreous carbon functionalized by na electrochemically synthesized film, Anal.
Chim. Acta 326, 149-154.
9. Liu, S., Ye, J., He, P. and Fang, Y. (1996) Voltarnmetric determination of sequence-specific DNA by
electroactive intercalator on graphite electrode, Anal. Chim. Acta 335, 239-243.
10. Zhou, X. C., Huang, L. Q. and Li, S. F. Y. (2001) Microgravimetric DNA sensor based on quartz crystal
microbalance: comparison of oligonucleotide immobilization methods and the application in genetic
diagnosis, Biosensors & Bioelectronics 16, (2001) 85-95
11. Wang, J., Nielsen, P. E., Jiang, M., Cai, X., Fernandes, J. R., Grant, D. H., 02S0Z, M., Beglieter, A. and
Mowat, M. (1997) Mismatch-sensitive hybridization detection by peptide nuclei acids immobilized on a
quartz crystal microbalance, Anal. Chern. 69,5200-5202.
NANOCOATING OF PARTICULATE SURFACE IN COLLOIDAL
PROCESSING FOR PIEZOELECTRIC SENSORS APPLICATIONS

WEI-HENG SHIH, WAN Y. SHIH, CHIA-YI YANG, HUIMING GU,

AND JEONG W. YI
Department of Materials Engineering, Drexel University, Philadelphia, PA
19104

1. Abstract

Colloidal processing offers the advantage of controlling the interactions between particles.
Nanometer is the critical length scale for interparticle interactions. Modification of
interparticle separation in the nanometer range plays a critical role in the properties of
particulate suspensions. We have developed a nanocoating method that modifies the surface
chemistry of ceramic powders. The nanocoating serves multiple purposes. It enhances the
chemical and thermal stability of the powders, improves the consolidation and rheological
properties of slurries, increases the homogeneous distribution of processing additives, and
lowers the sintering temperatures of the green compacts. For example, we showed that
boehmite (AIOOH)-coated silicon nitride and silicon carbide powders have a significantly
higher solids loading than the uncoated powders in water. Coated suspensions have lower
viscosity and wider linear viscoelastic region, and lower shear modulus than that of
uncoated suspensions. With an increasing coating thickness, the shear modulus of coated
suspensions decreases indicating the coating prevents close contact of core particles thereby
reducing the van der Waals attraction interaction between the core particles. Boehmite
coating is shown to enhance the thermal stability of catalytic oxide zr0 2 • Coating of
Mg(OH)2 layer on Nb20 s particles facilitates the synthesis of pyrochlore-free perovskite
lead manganese niobate (PMN) powders by a single heat-treatment step. Normally this must
be done in two heat-treatment steps as in the columbite process. More recently, the coating
technique has been used to fabricate piezoelectric PMN-PT thick-film sensors.
Piezoelectric-metal bimorphs have been fabricated to detect yeast cells and biotinylated
polystyrene spheres with a sensitivity of 10-9 glHz.

377
Y.G. Gogotsi and I. V. Uvarova (eds.),
Nanostructured Materials and Coatings for Biomedical and Sensor Applications, 377-394.
© 2003 Kluwer Academic Publishers.
378

2. Introduction

Particulate suspensions have a wide range of applications including ceramic

processing, drug delivery, paint, ink, waste treatment and others. In recent years,
there has been a renewed interest in particulate suspensions and gels due to
increasing activities in the areas of nanostructured materials and nanotechnology.
One of the major nanotechnology developments is biosensor. In this paper, we will
summarize our research activities in the synthesis of nanostructured materials and the
development of piezoelectric cantilever biosensors.
Nanoscale materials have been shown to possess unique properties that
materials of larger scales lack. While nanocomposites and nanomaterials promise
exceptional properties, processing nanoparticles is challenging. Nanoparticles are
more reactive and less stable due to its large surface area and high solubility in
solution. Also, nanoparticles tend to agglomerate in suspensions, making it difficult
to process nanoparticles. The reason that nanoparticles tend to agglomerate is the
dominance of attraction over repulsion in nanoscale. This is illustrated by the
calculated DLVO potential for alumina particles of various radius at pH=3 as shown
in Fig.l. I All colloidal particles have a primary attraction minimum at close contact,
i.e. at s=O. At pH=3, alumina particles have a high zeta potential, s=55 m V, that
gives rise to a large repulsion barrier
for large Ah03 particles (r > 160 nm 25C
with r being the radius). The
repulsion barrier prevents particles
20C '\60"m
from getting into the primary energy ~ ISC

~
minimum at close contact. Thus, ~
suspensions of large alumina ~ JOC
particles are dispersed at this pH. :> '"
so
However, with the same suspension .~nm
pH and the same zeta potential, the 0
repulsion barrier decreases with a IS 20
0 S 10
decreasing particle radius. The S (nm)
repulsion barrier almost diminished Fig.l. Calculated DL VO potential between
at r = 8 nm even with a high zeta AI 20 3 particles where C; =55 mV, A= 4.Sx I 0. 20 J,
potential, dramatically increasing the radius of the AbO, particles ranges from 8
nm to 160 nm as labeled, and the electrolyte
the aggregation probability of these concentration is 0.003M as typical near pH=3.
nanoparticles. Thus, to make
 379

nanomaterials that involve processing of nanoparticles, control of the aggregation

behavior and hence the rheological properties of particulate gels is critical.
The formation of particulate gels is due to the attraction among particles.
Aggregation of particles often leads to fractal clusters that have self-similar structural
properties at a length scale smaller than the cluster size. The fractal aggregates can
further aggregate to form a continuous network throughout the system creating a
particulate gel. The elastic properties of fractal aggregates are unique, determined by
the fractal dimension, df, and the backbone dimension, db. The backbone dimension
characterizes the scaling of the contour length within the cluster. The relationship
between the storage modulus, G', and the limit of linearity, Yo, of a gel of volume
fraction ~, have been obtained by us previously as,2
3+d b
G' ~ qJ3-d,
for the strong-link regime, (1)
_I+d b

Yo ~ rp 3 d, o

and
I
G' ~ rp3-d,
I
for the weak-link regime. (2)
Yo ~ rp3 d r
0

In addition to boehmite and silica gels, a variety of protein gels and fat crystal
networks have also been shown to obey the relationships described by Eqs. (1) and
(2).3
Colloidal processing of ceramics requires a homogeneous microstructure with
high density. For micron-size powders, this is normally achieved with a dispersed
suspension obtained by adjusting the interparticle interaction to be repulsive.
However, dispersed suspensions do not produce cohesive green bodies that are
necessary for shaping and microstructure homogeneity. On the other hand, attractive
interaction between particles .leads to agglomerated structures and highly viscous
suspensions that are difficult to deform and shape into complex geometry. Between
highly dispersed suspensions and strongly flocculated suspensions, controlled
gelation of suspensions presents an opportunity to shape suspensions into complex
shapes. The final gelled green body provides the cohesiveness required for
microstructure homogeneity and keeping in shape.
To address the challenges of controlled rheological properties, gelation, and
chemical stability, one needs to control the interparticle interaction in suspensions.
Only when control of the interaction between particles is accomplished can we fully
control the behavior of particles in suspensions and gels. Traditionally, to modify the
380

surface chemistry of micron-size particles, one would change the pH of the aqueous
suspension to move away from the isoelectric point (IEP). Alternatively, one can
modify the surface by adsorbing charged species to alter the IEP of the particles.
These methods do not work effectively in
the case of nanoparticles since the dominant
factor is the attraction rather than the
repulsion. As shown in our earlier study on
the Ab03-coated ShN4, nanosize Ab03
adsorbed on micron-size Si3N4 even when
they have the same sign of charge. To
disperse nanoparticles, new approaches are
needed that address the issue of attraction.
Spalla and Kekicheff have also raised a
similar view that the stability of nanosize
dispersions is better achieved by controlling
the adhesion at contact rather than affecting FigJ. TEM irnagf: ofhodu"ite-cmltoo SiC partidt,
the long-range electrostatic repulsion.
We used a nanolayer coating on particles to modify the interaction between
particles. Nanometer coating thickness is just the right scale to significantly alter the
interaction between particles, as seen in Fig.l. The nanolayer coating can prevent the
close contact of particles and reduce the interparticle attraction. The nanolayer
coating can also improve the chemical stability of the core particles making it
redispersible, recyclable, and less
likely to dissolve. 60

In the following review,

40
we will describe how we utilize
the coating approach in
modifying the properties of
particulate suspensions and gels
and its applications in biosensors. + Uncoated SiC
!! x. Pure Boehmite
Our results indicate that the N -20 --0_1\,,% AlOOH-SrC
control of rheological properties ........... - 3",1% A100H-SiC
-A- 5wt% AlOOH-SiC
of precursor solutions and -40
-'t'-IO\'lt%AJOOH-SiC
reactivity of powders play an --O-15\Vt%A100H-SiC
-60 __
essential role in the fabrication of
L-~----L_~-'-~ '----~-'-~_-'-~-----'

o 2 4 6 10 12
piezoelectric films that will be pH
Fig3. Zeta potential of SiC for different amounts of boehmite
used in a biosensor.
 381

3. Boehmite Coating on SiC Particles

We have studied the coating of boehmite on SiC and the effect of coating on the
rheological and consolidation behavior of the boehmite-coated suspensions. A
boehmite coating on SiC was fonned by adding acid after the initial precipitation of
boehmite in the solution. As an example, Fig. 2 shows that boehmite fonned a
coating layer on SiC. The
coating of boehmite has 16
:

r 1r
been confinned by 15
S
measuring the zeta potential
of the coated SiC. Figure 3
,s
~
14 I
"c
1: 1
.
shows the zeta potential of ~

rj
:§ 13
SiC as a function of pH for
.: :
various amounts of ;; 12

I
0
boehmite added. The IEP U

I
r 1
increased and approached 11
:
that of boehmite as the :

11 1
10
boehmite coating amount
was increased. 9
The coating layer
has been shown to act as a
barrier to prevent particles
8

7
1 Coating Regime Excess Regime

from close contact, thus 0 2 4 6 8 10 12 14 16

wt% of Aloon to SiC
reducing the attractive Fig.4. Boehmite coating thickness versus boehmite concentration.
interaction between
neighboring particles. Thus, a plausible way of controlling the interaction between
particles is to modify the van der Waals attraction by coating particles with a
nanolayer to provide a barrier to prohibit close contact between particles. In addition,
if the coating material is chemically more stable, the dissolution of the core particles
can also be minimized.
As seen in Fig.l, the critical range of interaction is in the nanometer range. A
nanolayer coating can provide a barrier that can significantly modify the interaction
between particles. Figure 4 shows the coating thickness as obtained from the TEM
pictures versus the boehmite amount. When the amount of boehmite was less than 10
wt%, the coating was observed to be complete and very few individual boehmite
particles were observed in the TEM. This regime is identified as the coating regime
and labeled as Regime 1. The coating thickness increased with an increasing coating
concentration in this regime. When the coating concentration exceeded 10 wt%, the
382

coating layer thickness

70
saturated at about 14 run and ~
"0
a significant number of -t.
l~ =:=::~ ~~~::

t/ \\
individual boehmite particles .~ 60
c / -A- 5 wt% A100H-
were observed in the TEM. Q" / 0-0 \ -~- \0 wt% A100H-
This regime is identified as c 50
~ -0-- 15 wt% AIOOH-
the excess regime and labeled
~
.sc ! & \\ \
as Regime II. While earlier
we have shown that boehmite i
E 40 1/ III \ \
/! 1../ "I \" \

//;·0 \\\ \
/ I //
'" 30
II I,
coating improved thc
rheology and consolidation of
the SiC suspensions, here we
further show that the
20
~~D/ "D~\U~
rheology and consolidation
density of the suspension can 2 4 6 10 12
pH
change significantly with the
Fig.S. Sedimentation density ora boehmite-coated SiC suspension versus pH.
coating thickness. Figure 5
shows the consolidation
density of the boehmite-coated SiC suspension versus pH with various boehmite
amounts. In Regime I, the maximum consolidation density increased from 25 vol%
with I wt% boehmite coating to about 68 vol% with 10 wt% boehmite coating.
Further increasing the coating amount in Regime II decreased the maximum
consolidation density to 52 vol% with 15 wt% boehmite coating. The different
behavior in the consolidation
density versus the coating amount :
(I ) ( II )
is closely related to the different
behavior in the suspension ME Ixl0 5
• •
rheology versus the coating ~ •
amount in the two regimes. Figure :::,
c
.... • • •
6 shows the G' O.max versus the 0
amount of boehmite added in the
•
•
SiC suspension where G' O,max is
the maximum in the storage
modulus of the coated SiC
suspension with respect to pH. Ixl04
0 2 4 6 8 10 12 14 16
Note that the pH value at which wt% of AIOOH to SiC
G' O,max occurs differed among Fig.6. G' versus boehmite concentration where G' is the storage
modulus of a boehmite-coated SiC suspension at its IEP.
suspensions with different coating
 383

amounts since the IEP varied with the coating amount. In the coating regime (I),
G' O,max decreased with an increasing amount of boehmite, In contrast, in the excess
regime (II), G' O,max increased with the amount of boehmite, The increase of G' O,max
with an increasing boehmite amount in regime II can be interpreted as the result of
the excess individual boehmite particles, The individual boehmite particles can
participate in the gel formation process thus contribute to the strength of the gel, in
tum lowers the consolidation density, Due to their small particle size, a small amount
of boehmite can contribute significantly to the storage modulus of the suspension,
This can be seen from the fact that a suspension of I vol% of small boehmite
particles alone can gel over time.
In the coating regime, the relationship between the G' O,max and the coating-layer
thickness has been examined in our previous theoretical study.s In the theory, the porous
coating layer acts as a barrier to prevent SiC particles from close contact. By considering the
sliding between particles separated by a distance twice the coating-layer thickness, the
theory predicted that G' O,amx decreases with the coating-layer thickness, t, as

(3)

To compare with the Eq. 3, the storage modulus, G'O,max, in the coating regime shown
in Fig. 6 is replotted as a IxlO r-~-.~-.-~--'-'---'~-,-,
function of C 2 in Fig. 7. .//
That G' O,max is linear in C 2 .//
indicates that the boehmite 8xl0
,.-,'"

l
coating acts as a barrier NS /

preventing two particles 6xl6 ......

//
from close contact, thus ~
lowering the storage .../ / '
modulus of the suspension, 4

which in tum increased the 4xl0 •.,l

consolidation density. • //
.,/
The nanolayer coating 2xl0 L-~--l._~.....L.~_...l-~_'--~......J.----'
0.004 0.006 0.008 0.010 0.012 0.014
on particles may also
modify the attraction
between the particles Fig.?: G' versus r2 where G' IS the storage modulus of a boehmite-coated
SiC suspension at its lEP and t the boehmite coating thickness.
because the Hamaker
constant of the particles may be modified. Boehmite has a smaller Hamaker constant
(-5xlO-21 J) than SiC (_25xlO- 21 J). As a result, the attraction between SiC particles
could be reduced by the boehmite coating. However, the reduction of G' O,max, was
more than two orders of magnitude from particles with no coating to particles with
384

10 wt% coating, which cannot be accounted for by the change of the Hamaker
constant alone. Moreover, this argument could not explain the thickness dependence
that we observed in the coating regime.

4. Numerical Simulations of Coating Process

Numerical simulations on the coating process have been carried out. The simulation
was performed to provide a fundamental understanding for the boehmite coating on
SiC surface (or in general, any ceramic coating on another ceramic surface). The
model used was a modified SAK (Shih, Aksay, and Kikuchi) modd- Io that allowed
both cluster-cluster aggregation and dissolution (or unbinding) to occur.
The coating simulations were done in 2d on a 60 x 90 simulation cell with
two walls located at the right and left-hand sides of the simulation cell representing

Fig. 8: Snapshots taken at 500,000 Monte Carlo steps with total input monomer concentration
(from top to bottom) Ct = 0.019, 0.056, 0.111, 0.148, and 0.167 for (a) E=3 kaT, (b) E=4 kaT. and
(c) E=5 kaT For all snapshots, Ew=5 kaT.

the flat core particle as shown in the snapshots in Fig. 8(a)-(c). The individual dots
represent monomers of the coating material and clusters the particles formed by
monomers. The monomers may be thought of as clusters of atoms of the precipitating
species. The core particles were assumed to be flat and the curvature effect was not
considered. We considered nearest-neighbor attraction energy, -E, between the
precipitating monomers, and nearest-neighbor attraction energy, -Ew, between the
 385

monomers and the core surface. In Fig.8, Ew = 5 kB T and the snapshots taken at
500,000 Monte Carlo steps are shown in Fig. 8(a), (b), and (c) for E = 3 kBT, 4 kBT,
and 5 kBT, at total input monomer concentration ct=0.019, 0.056, 0.111, 0.148,
0.0167, respectively. The snapshots were

I
taken long after the coating thickness
saturated. The coating thickness
eventually saturated and individual (a)
particles formed in the suspension, as
similar to the experimental observation
described above. The success of the model unbinding,
dissolution
can be attributed to (1) that the model
allows both aggregation and dissolution as (b)
shown in Figs. 9(a) and 9(b), and (2) that
the model considered a size-dependent Fig. 9: Schematic of (a) the aggregation and (b)
aggregation probability, Pagg , between the dissolution processes. The aggregation
probability, P au' d~pends on the partide SlZe. Fig.
particles formed by the precipitating 10 illustrates the stepwise Pagg used in the
species as shown in Fig.l0. The nearest- previous simulations.
neighbor monomer-monomer attraction
energy, -E, controlled the dissolution or
unbinding process as schematically shown in Fig. 9(b). As a result, the solubility
roughly decreases exponentially with an
increasing E (not shown). As shown in Fig. I,
under repulsive conditions, i.e., with high zeta
1;--
potentials, the repulsion barrier in the DLVO
interaction energy increases with an increasing
particle size, leading to a diminishing Pagg. In the
simulations, we approximated P agg to be unity "-,1
below a critical size Nc and zero above Nc in the ~-­

simulations as schematically shown in Fig. 10 N

N
where N is the number of monomers in a particle.
Nc can vary with the zeta potential and the salt Fig. 10: A stepwise P 'g~ versus N used
in the previous simulations where N is
concentration. Generally, a smaller Nc represents the number of monomers in a particle.
a more repulsive condition or a high zeta
potential and N c= 00 represents the attractive
conditions at and near the isoelectric point where electrostatic repulsion is negligible.
In the simulations, Nc was taken as 15 for the repulsive conditions.
As a result of the interparticle repulsion, particles were formed with a well-
386

defined size distribution and the particle size increases with an increasing solubility
(decreasing E), in agreement with the experimental observations. The particle size
distribution, P(N) versus N is plotted in Fig.II where N is the number of monomers
in a particle.
As shown in Figs. 8-11, the model captured the correct physics of the coating
process: (1) that the saturation of
coating thickness coincided with
the appearance of individual 0.3
--.-- E = 3
o E=4
particles in the solution, (2) that E=5
particles size increased with
increasing solubility (smaller E),
0.2
o
00 •
(3) that increasing solubility
P (N)
• •
(decreasing E) increased coverage 0.1 A 0 ·0
and hence the likelihood of layer- •••
o
like coating, (4) decreasing the 0.0

wall attraction increased the

100 1000
likelihood of island-like coating,
N
comparing well with the
Fig. II: Particle size distribution for E=3, 4, and 5 kB T.
experimental observations.
Moreover, the onset of coating-
thickness saturation was accompanied by the formation of individual particles. Also
note that the size of the particles formed in the solution decreases with an increasing
E (or reduced solubility) as can be seen from Fig. 11. This is also in good agreement

RxE =5

.-- E = 3 ••••••
---0-E=4
,.....5

•
E=5 ell
~ en
~ 4 .0- ·0 f 0.6

• ~
III O· RxE =5
:::::'3 0
.... .0- 0
U 0.4 A
w
.0- 0" A A. A A --.- E =3
~. Z -0-E=4
1 a 0.2 A At. E=5
0.00 0.03 0.06 0.09 0.12 0.15 0.18 0.00 0.03 0.06 0.09 0.12 0.15 0.18
C C
t t
Fig. 12: (a) Coating thickness, L, and (b) coverage versus c, where coverage is the area fraction
covered by coating. c, the total input monomer concentration.
 387

with the experimental observation that boehmite particles formed at pH=3 where the
solubility is high are much larger than boehmite particles formed at higher pH.
The coating thickness, L, and the coverage versus Ct are plotted in Fig. 12 (a)
and (b) where the coverage is defined as the area fraction that is covered by the
coating and Ct is the total input monomer concentration. Note that in all three cases,
as the monomer concentration increases, both the coating thickness and the coverage
initially increased and eventually saturated at higher concentrations. Moreover, the
saturated coating thickness is smaller with an increasing E (decreasing solubility).
Notice that for a high E, i.e., E=5kB T, the coverage has never reached 100% (i.e., the
coating was always patchy) even at high monomer concentrations. It turned out that
the ratio of Ew/E is crucial in
determining whether the coating is
•• ,(3,3)
layer-like or patchy. In Fig.13 we plot
the coverage versus E/E w. For E/Ew<l,
..... 100
~90
t (4,4)

+ (5,5)
f80
GI
coverage=l, and the coating is layer-
like. For E/Ew> I, coverage decreases
rapidly with increasing E/Ew (i.e.,
GI
8 70
I ..
decreasing solubility), and the coating til 60

i
C
is patchy or island-like. For ElEw=l, :;:I
III 50
coverage can be layer-like if E is small 8 ! •
or island-like ifE is large. These results 416.0 0.5 1.0 1.5 2.0

compare well with the experimental E lEw

observations described above. II . ls Fig. 13: Coverage versus EIE w for various E and
Ew where blue diamonds, red squares and green
squares denote E1Ew<l, EIEw> I, and ElEw~l,
5. Synthesis of Piezoelectric PMN-PT respectively.

Relaxor Pb(Mgl/3Nb 2l3 )03 (PMN) has been studied extensively because of its high
dielectric constant and large e1ectrostrictive coefficients. 16 For example, 0.9PMN-
0.1 PT (lead titanate) relaxor ferroelectrics are a good candidate for multi-layer
ceramic capacitor applications due to its high dielectric constant at room temperature.
Furthermore, the lower sintering temperature of 0.9PMN-0.l PT also makes it more
advantageous than BaTi0 3 . However single-phase perovskite PMN could not be
obtained by the conventional solid oxide method because of the presence of the
pyrochlore phase, a reaction product between Nb 2 0 s and PbO. Swartz and Shrout!7
first succeeded in eliminating the pyrochlore phase by developing the columbite
method that involved two calcination steps:

(4)
388

In the first calcination step, mixtures ofNb 20 s and MgO were heated near lOOO°C to
form the columbite phase,
MgNb 20 6 . In the second
calcination step, MgNb 20 6
was mixed with PbO and
heat-treated. The
perovskite phase began to
appear near 700°C and
complete perovskite
conversion occurred near
900°C. The two
calcination temperatures
may vary with a number of
parameters such as
reactivity of MgO, degree
of mixing, and control of
the PbO volatility.18
Nevertheless, two Fig.14. Optical micrograph of Mg(OH)z-coated NbzOs partick'S.
The white core is Nb,Os and the dark layer surrounding the core
calcination steps were is Mg(OHh The identification has been verified by uplkal
needed to prevent the micrographs of pure Nb2 0 S and Mg(OHh particles.
direct contact between
Nb 20 s and PbO and, thus, the formation of the pyrochlore.
The solid solution, PMN-PT is a piezoelectric material with piezoelectric
coefficients higher than that of PZT (lead zirconate titanate). In the development of a
piezoelectric biosensor, to be discussed later, we need to fabricate piezoelectric
films. The main challenge in the synthesis of PMN-PT piezoelectric is the
elimination of pyrochlore phase. In our approach, the pyrochlore phase was
prevented by coating Mg(OHh on Nb 20 S particles. Coating of Mg(OH)2 on Nb 20 S
was done by precipitating Mg(OHh in an aqueous Nb2 0 S suspension at pH=lO. The
coating was confirmed both by the optical micrographs and the zeta potential
measurements. 19
An optical micrograph of the coated Nb20 S particles is shown in Fig.14 with
light-colored particles surrounded by dark-colored coating layers. Comparison with
micrograph of Nb20 S particles indicates that the light-colored core particles in Fig.14
were Nb20 S particles. The dark-colored coating layers surrounding the Nb2 0 S
particles were designated as Mg(OHh because the XRD study showed the powder
 389

obtained by the same

10000
precipitation procedure Perovskite phase
without the Nb20 s core 9000
I (110) peak
particles, were Mg(OH)2. 8000
!i
,
In addition, the XRD 7000 "

pattern of the coated

powder after
6000
Z. 5000
heat '00
Pyrochlore phase !I\
(222) Peak
c:
treatment at 500 0 e .& 4000

showed MgO peaks. c: 3000

Therefore we expect the 2000
coating layer to be 1000 >' ~ 800°C calcinate 2 h
---
Mg(OHh The measured 900 oC calcinate 2 h
0
coating thickness was
29 30 31 32 33 34
consistent with the
amount of 28
Fig.15. XRD pattern of mixed powders of Mg(OH)rcoated
Mg(N03)2·6H20 added in Nb 20 5 and PbO calcined at 800°C and 900°C for 2 h.
the system. In addition, Pyrochlore peak disappeared after 900°C heat treatment.
the measurements of the
zeta potential of the coated particles showed that the Nb 20 s particles were coated
with Mg(OHh
The diffraction pattern of the coated powder mixed with PbO and 10%PT
calcined at 800 0 e for 2 h is shown in Fig.15, indicating that the powder was not
100% perovskite. Some amount (25 wt%) of pyrochlore phase was present in the
powder after being calcined at 800 0 e for 2 hr. The results in Fig.15 indicate that
complete conversion to the perovskite phase (or complete elimination of the
pyrochlore phase) occurred between - - .. -.--.-------------,
800 0 e and 900 o e. A DT A analysis was
carried out on the Mg(OH)2-coated
Nb 20 s powder mixed with PbO. It
showed a peak at around 848°e,
PZT
consistent with the result inferred from
Fig.I5 that complete perovskite
formation occurred between 800 0 e and steel
900 o e. Am
The elimination of the
pyrochlore phase is attributed to the
separation of PbO and Nb20 s by the Fig.l6. Schematic of a piezoelectric
Mg(OH)2 coating. Also, the Mg(OH)z cantilever sensor.
390

coating on Nb 20 s improved the mixing of Mg(OH)z and Nb 20 s and reduced the

temperature for complete columbite conversion to around 850°C. The pyrochlore-
free perovskite 0.9PMN-0.I PT powders were sintered to 97% density at 1150°C. The
sintered 0.9PMN-0.IPT ceramics exhibited a dielectric constant maximum of
approximately 25,000 at 45°C for a frequency of 1 kHz.

6. Piezoelectric Cantilever Sensors

When an ac voltage is applied to a unimorph cantilever, a piezoelectric piece bonded

to a nonpiezoelectric piece, shown in Fig.16, vibration occurs. The nth -mode flexural
resonance frequency is related to the bending modulus per unit width, the length, and
the mass per unit area, m, of the cantilever as

(6)

where v~ is the dimensionless nth_mode eigen value.

I,'-~ '2~ (7)

n - 27r Me + flm .

For Lim «Me, the resonance frequency shift, Llf", due to the added mass, Lim, is thus

flif,n -- I,n' -! .,n 1 I, flm

- -- (8)
= 2 n-Me .

The resonance frequency shift per unit loaded mass is therefore

 391

The PZT/stainless steel cantilever was rigidly clamped at one end of the cantilever with an
aluminum fixture. Measurements were carried out with cantilevers of 0.4 cm and 0.2 cm in
width. The length of the cantilevers was varied from 1.37 cm to 0.44 cm by changing the
clamp position. The experimental results are shown in Fig.I7. Note that off resonance,
the cantilever was a capacitor exhibiting a phase angle around _90 0 • At resonance, the
flexural motion gave rise to a peak in the real part of the impedance, and hence a
peak in the phase angle due to the direct piezoelectric effect. As an example, the
first-mode resonance frequency spectrum, i.e., phase angle versus frequency, of a

-40
---air
-50 ~R"'. -0-0.35 mg
-c;
II)
M'
r+?1:'1'16 \,~ -+-0.59 mg
~ -x-0.78 mg
-60 f'J-'i .I 'f.~ ~
II)
c, rfr J ~o\ '~
C rtf J >t o\ } .
c(
II)
-70
~d
~(t ~
"'.
Wb \ b,-.i.,
VI
CIS o .. a ill
..c -80 000 • °0 •
a..
-90
1.20
~~--'---'---1.~~~
1.22 1.24 1.26 1.28 1.30 1.32 1.34
Frequency (KHz)
Fig.I? Phase angle of a piezoelectric cantilever as a function of frequency. As
the mass on the cantilever increases, the resonance frequency shifts to lower
values.

cantilever of L = 1.37 cm and w = 0.4 cm without loading (solid squares), loaded

with an aluminum foil of 3.Sx 10-4 g (open circles), S.9x 10-4 g (pluses), and 7.8x 10-4
g (crosses) at the cantilever tip is shown in Fig.I7. Clearly, the resonance peak
shifted to a lower frequency as the mass of the aluminum foil is increased. Currently,
a cantilever of 1 cm length has a llflllm = 3.2 xl 05 Hz/g. When the cantilever is
reduced to 10 11m in length, llflllm will approach 10 17 Hz/g. More detailed
description of the application of piezoelectric cantilever as a mass sensor and how
the sensitivity change with the length, width, and height of the cantilever can be
found in Yi et al. 20
392

7. Yeast Cell Detection

As an application of the piezoelectric

cantilever, it was used to detect the
adsorption of yeast cells on a substrate. The
cantilever consisted of a PZT layer 3 mm
long, 3.7 mm wide, and 0.26 mm thick
bonded to a 0.1 mm thick nonpiezoelectric
stainless steel (SS 304, Shop-Aid Inc.
Woburn, MA) layer that was as wide as the
PZT layer. The cantilever had a 3.5 mm long
stainless steel tip at the free end (see the
schematic of the cantilever shown' in
Fig.18). The stainless steel tip at the free end
was steeped in 4 N nitric acid for 30 minutes
and was subsequently washed in distilled fig.llL The sensor design j('r the
water, and then thinly coated with a poly-L- detection of yeast teHs in sllspensions
lysine (Sigma P8920, 0.1 % w/v). Poly-L-
lysine offers a positively charged environment to which negatively charged yeast
cells are attracted and
immobilized. Yeast
suspension was prepared 150 ,-~--

• •
by suspending bakers

-- DOD 0
yeast (Saccharomyces
cerevisae, Fleischmann's 100

• •
Yeast, Fenton, MO N
:::I:
DO •
obtained from a local
store) in deionized water ~ fO
at room temperature. In I 50 l
.1 rrg/ni
the experiment, the I
02 rrg/ni
stainless steel tip was
immersed in a yeast t
0 __
suspension of 1mg/ml or
0 20 40 60 80
2 mg/ml concentration.
The dipping depth was t (min)
3.0 mm. At 60 min, the Fig.19. The frequency shift due to the adsorption of yeast
cantilever had a -11/ of cells as a function of time for two concentrations of yeast
120 Hz. The sensitivity
 393

of the cantilever is -t'1m/ t'1J= 4 xlO- 7 g/Hz, -t'1j= 120 indicated a t'1m = 4.8 xlO- 5 g.
The concentration effect on the adsorption kinetics is shown in Fig.19 where -t'1J
versus t of cantilever in both 1 mg/ml and 2 mg/ml yeast suspensions is shown.
Clearly, -t'1J increased faster with time in the 2 mg/ml suspension than in the I
mg/ml suspension but -t'1J eventually saturated at about 120 Hz. More detailed
results on yeast detection are presented in a recent paper by Yi et al. 2l

8. References

I W.-H. Shih, D. Kisailus, W. Y. Shih, Y.-H. Hu, and J. Hughes, "Rheology and Consolidation of
Colloidal Alumina-Coated Silicon Nitride Suspensions," JAm. Ceram. Soc., 79[5]1155-62 (1996).
2 W.-H. Shih, W. Y. Shih, S. 1. Kim, J. Liu, and 1. A. Aksay, "Scaling Behavior of the Elastic
Properties of Colloidal Gels," Phys. Rev. A. 42,4772-4779 (1990).
3 S. S. Narine and A. G. Marangoni, "Relating Structure of Fat Crystal Networks to Mechanical
Properties: a Review," Food Research International 32, 227-248 (1999).
4 O. Spalla and P. Kekicheff, "Adhesion between Oxide Nanoparticles: Influence of Surface
Complexation," J Colloid & Inter! Sci., 192, 43-65 (1997).
5 W. Y. Shih, W.-H. Shih, and 1. A. Aksay, "Elastic and Yield Behavior of Strongly Flocculated
Colloids," JAm. Ceram. Soc., 82[3], 616-24 (1999).
6 W. Y. Shih, 1. A. Aksay and R. Kikuchi, "Reversible Growth Model: Cluster-Cluster Aggregation

With Finite Binding Energies," Phys. Rev. A 36, SOlS (1987).

7 1. Liu, M. Sarikaya, W. Y. Shih and 1. A. Aksay, "Fractal Colloidal Aggregates With Finite

Interparticle Interactions: Energy Dependence of the Fractal Dimension", Phys. Rev. A 41, 3206
(1990).
8 W. Y. Shih, W.-H. Shih, and 1. A. Aksay, "Mechanical Properties of Colloidal gels subject to particle
rearrangement", Mat. Res. Soc. Symp. Proc. 195,477 (1990).
9 J. Liu, W. Y. Shih, R. Kikuchi, and 1. A. Aksay, "On the Clustering of Binary Colloidal

Suspensions," J Colloid & Inter! Sci. 142,369 (1991).

10 W. Y. Shih, W. H. Shih, 1. A. Aksay, "Heteroflocculation in Binary Colloidal Suspensions: Monte
Carlo Simulations," JAm. Ceram. Soc. 79(10) 2587 (1996).
" C.-Y. Yang and W.-H. Shih, "Effect of Acid on the Coating of Boehmite onto SiC Particles in
Aqueous Suspensions," JAm. Ceram. Soc., 82[2]436-40 (1999).
12 c.-y. Yang, W. Y. Shih, and W.-H. Shih, "Gelation, Consolidation, and Rheology of Boehmite-
Coated SiC Suspensions," JAm. Ceram. Soc., 83[8],1879-84 (2000).
13 C. Y. Yang, W. Y. Shih, and W.-H. Shih, "Effects of Boehmite Coating Thickness on the
Consolidation and Rheological Properties of Boehmite-Coated SiC Suspensions," JAm. Ceram. Soc.
84[12]2834-40 (2001)
14 C. Y. Yang, W. Y. Shih, and W.-H. Shih, "Monte Carlo Simulations of the Nucleation and Growth
Process of Colloidal Particles," Phys. Rev. E., 64, 021403, (2001).
15 C. Y. Yang, "Experimental Study and Numerical Simulation on Synthesis, Properties and

Applications of Oxide-Coated Ceramic Particles," PhD Thesis, Drexel University, (2000)

16 L. E. Cross, "Relaxor Ferroelectrics," Ferroelectrics, 151, 305-20 (1994).
394

17 S. L. Swartz and T.R. Shrout, "Fabrication of Perovskite Lead Magnesium Niobate," Mater. Res.
Bull., Vol 17, 1245-1250, (1982).
18 T. R. Shrout and A. Halliyal, "Preparation of Lead-Based Ferroelectric Relaxors for Capacitors,"
Am. Ceram. Soc. Bull., 66[4]704-11 (1987).
19 Huiming Gu, W, Y. Shih, and W.-H. Shih, "A Single-Calcination Synthesis of Pyrochlore-Free
0.9PMN-0.IPT and PMN Ceramics by a Coating Method," in press, JAm. Ceram. Soc.
20 J. W. Vi, W. Y. Shih, and W.-H. Shih, "Effects of Length, Width, and Mode on the Mass Detection
Sensitivity of Piezoelectric Unimorph Cantilevers," J Appl. Phys. 91[3],1680-1686 (2002)
21 J. W. Vi, Wan Y. Shih, R. Mutharasan, and Wei-Heng Shih, "In Situ Cell Detection Using
Piezoelectric PZT-Stainless Steel Cantilevers," submitted to J. Appl. Phys.
Subject Index

ab initio-DFT 357 coatings 41, 51

alcohol electrode 327 cocaine 27
amitriptyline 27 CoCrMo 51
amorphous carbon 175 computer modeling 333
amorphous silicon carbide 249 computerized complex 229
antibacterial 41 conductance 21 9
application 187 conductive particulate composite 219
assisted proton transfer 365 controller 229
bioceramics 1, 83 coordination defects 249
biochemical media 63 copper-carbon 175
biocompatibility 15 core-shell particles 27
bioglass 1 corundum 1
biological materials 1 covalent immobilization 371
biological solution 111 crystallization 169
biomaterials 15, 77 amorphous silica 169
biomedical devices 71 diamond adsorbents III
biomimetic coatings 15 DNA hybridization 297
biosensor 311,327, 371, 377 drug overdose 27
bone tissue 77 electrochemical DNA biosensors 297
borosilicate 211 electrochemical quartz crystal
bupivacaine 27 microbalance (EQCM) 371
C 60 239 electronic structure 249, 257
carbon 51,119, 239 electropolymerisation 371
carbon films 175 endoprosthesis 103
ceramic sensor 279 enzyme electrode 327
charge transfer complexation 27 explosive shock waves 91
coating adhesion strength 103 fabrication 305
coating on particles 377 ferromagnetic I, 239
395
396

flame aerosol process 203 molecular mechanics 357

friction coefficient 175 molecular dynamics simulations 249
FTIR spectroscopy 103 Mulliken population analysis 357
fullerene 239 nanoceramics 229
fullerene coatings 103 nanocrystalline 175
functionalization 187 nanodevices 257
gas sensors 343 nanodiamond 187
Ge microelectrodes 7 I nanolayered coatings 131
geliliquid interfaces 365 nanoparicles 1,27,41,83,203
germanium microwires 71 nanopowders 63
glucose amperometric biosensor 365 nanoscale ZnO/Si planar structure 305
hardness 175 nanostructured 175
high pressure apparatus 229 nanostructured silica 159
high pressure sintering 229 nanotube 257
hip friction pair 103 nanotube production 187
hip joint 51 NEMS 257
humidity 119 non-aggregated 203
hydroxyapatite 1, 15, 77, 83 nucleotides 371
haltitania composites 91 onion-like carbon 187
immobilisation 311 oxygen agglomeration 333
implants 77 photonic devices 291
in vivo toxin removal 27 physical-chemical simulation 63
InN film 291 pi-complexation 27
intraocular lenses 15 piezoelectric ceramics 377
IR 91,103 plasma polymerization 327
iron 181 plasma spraying 149
ITIES 365 polyethylene 103
lithia alumninosilicates 279 polymer 219
localized states 249 polytyramine 371
lyotropic liquid crystalline templates 159 pressure 229, 239
magnetic beads 297 production 187
magnetron sputtering 131, 181, 343 prostheses 51
Matlab 219 quantum capacitance 257
matrixes for biocorrosion studies 181 quantum mechanics 257
mechanical properties 175, 305 quasi-disordered phase 333
MEMS 257 Raman spectroscopy 103, 175,291
mesoporous silica 159 rare-earth metal oxides 169
metal-organic precursors 203 reduction 211
microemulsions 27 rehybridization 357
micro hardness 71, 149 SEM-EDAX 91
microstructure 71, 291, 305 sensitive elements 3 11
microwave plasma 175 serum interaction 91
model 219 Si surface 333
modification 111 silica 203
modulus of elasticity 175 silicon boride 2 I I
molecular dynamics 333 silicon carbide 2 I 1

silicon nitride 63
sol-gel 41, 279
sol-gel processing 377
sorption 77
spherical 203
sputtering 175
structure 175
structural properties 249
structuring 3 11
surface characterization 343
surface chemistry 187
surface energy 15
surface modification 27
surface reconstruction 357
temperature 239
templated particles 27
tensile strength 71
tests on animals 63
theory 257
397
therapeutic overdose 27
thermal and electrochemical
treatment 111
tribo1ogical properties 175
tribological tests 175
thermal barrier coatings 149
thermal evaporation of C60 103
thin films 119, 169, 181
tight-binding model 249
tin dioxide 343, 357
toxic effect 63
toxin-receptor interactions 27
tribology 119
unimorph cantilevers 377
universal force field 357
Van der Waals force 257
vapor deposition 175
zirconia coatings 13 1