Professional Documents
Culture Documents
978 94 010 0157 1
978 94 010 0157 1
The Series is published by lOS Press, Amsterdam, and Kluwer Academic Publishers in conjunction
with the NATO Scientific Affairs Division
Sub-Series
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
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.
Preface ......................................................................................................................... ix
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 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.
~
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.
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
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].
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 .
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.
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
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.
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
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.
2.1 BIOMATERIALS
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
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
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
120,1------------------------------------,
60 .~~~---L~~L-~--~~--~~~---L-.~~IJ'
o 0.001 0.01 0.1 0.5 5 10
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
~ 100%
~
Cl..
~
<>
d 50%
0
'00
<lJ
..<::
"0
-<
0"10
10 5 0.1 om 0.001
Initial PC concentration, g C l
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
SS Ti Zr D FblZ PI SS Ti Zr PI D
Non-reduced Reduced
1200 + - - - -
800 -j-------~-----------__I
400 -------.-----.--------~-_i
ss Zr Ti D
Materials
~----------------------- ---------~
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
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
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
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
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
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
Bupivacaine
Kill by inhibiting heart function
(Local Anesthetic)
~
~,/CH3 N
JH Amitriptyline
3 (Antidepressant)
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.
Figure 3. Example procedure for the synthesis of silica using a 2,6-dimethylaniline derivative template
OIL-IN-WATER MICROEMULSIONS
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
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.
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
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.
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
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.
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
Figure 10. Bupivacaine removal from solutions by silica nanoparticles modified with 1t receptor
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.
Biodegradable Shell
(stability & ~.. t~,,,.f".'IhI, Lipid Phase (microemulsion)
(rapid concentration of toxin
from blood)
Molecular Pores
(toxin passage into nanopartide)
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
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.
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
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.
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.
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.
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 Figure 8. Relative number of Escherichia Coli
survived after antibacterial drop-test on the as prepared survived after the antibacterial drop-test on the
l3Ag87SiO, sol-gel thin coatings and after thermal xCuI(IOO-x)SiO, coatings thermally treated under
treatment at 500°C under oxidative and reductive oxidative conditions contained (a) 2.Smol%, (b)
conditions. 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.
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.
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].
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].
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.
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
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).
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].
.,----------_._----
(a)
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.
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.
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
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
1800
Hplast (kg/mm')
1600
+ 40 N (22 ·C, 32 %RH)
1400
1200
1000
800
o ~ ~ ~ D 1~1~1~1~
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
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
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.
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.
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 2. Results of filtrate analysis (mass.%) after interaction ofShN4 powders with
biochemical media
(2) reaction.
(3)
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
Table 4. Change of biochemical indexes in blood serum and rats' lung tissue
resulting from animals' priming with AIN powders
CONCLUSION
REFERENCES
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.
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)
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.
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.
a b
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.
I • Mi¢_1=tt_
,.lruli_ .....
2-Ej><>xW" _ "
4""O¢'m~'jnt
$.M""~
6 - O:"".."',,*km
4. Conclusion
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
Abstract
1. Introduction
3. Results
The results of studying the biomaterial ability to adsorb the synthetic normal
saline solutions and antibacterial drugs solutions are submitted in the table 2.
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.
4. Conclusion
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
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.
1. Introduction
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.
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
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
Abstract
1- Introduction:
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.
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.
I
,.
.;;;
c
~
1
70 60 50 .0 30
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
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].
(e) (d)
fb)
(e)
(a)
(h)
(g) (k)
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
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)
After
3000
SWA
T "I.
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).
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
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
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:
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
1. Introduction
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.
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
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
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
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
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
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
1. ABSTRACT
2. Keywords
3. INTRODUCTION
4. EXPERIMENTAL
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.
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
3,0
2,4
1,8 3
..
1.2
0.6
~,."
o lOO 200 300 400 500 600
Temperature, °c
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.
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
MJ. McNALLAN
Department of Civil and Materials Engineering
University ofIllinois at Chicago
Chicago. IL 60607
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
2. Experimental
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
2.4
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
=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.
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 ~----~----~----,------.-----.----,
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.
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
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.
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].
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
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.
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
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).
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
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
Abstract
1. Introduction
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
3.1. MICROSTRUCTURE
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.
15 1.5 15 1.5
10 10 10 10
,...
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.
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.
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
0001 001 01 10
Pore nldiu.. [pm]
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
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
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.
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
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.
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.
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
, 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.
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
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
1. Abstract
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.
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
(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.
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.
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
Abstract
1. Introduction
2. Experimental procedure
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.
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].
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).
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
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
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.
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
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.
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
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
1. Introduction
2. Nanodiamonds
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.
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.
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
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.
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 closing of the ~
Q"P","~
~-
The diamond cores ONION-LIKE CARBON (OLC)
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
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
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
1. Introduction
2. Experimental Procedure
~
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
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
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).
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
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].
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
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
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 •
1. Introduction
3. Experimental Results
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 •
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.
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
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
~\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
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
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
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
(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
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)
·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:
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
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. %
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
4. Conclusions
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
1. Introduction
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.
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.
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
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
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
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
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
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
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
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.
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 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:
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
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
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
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
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
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
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
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
C(X)<p2
2 - W(x; E, 0:) (1)
(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.
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:
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.
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).
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).
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)
(6)
5 Repulsion is due to the Pauli principle and will not be addressed here.
266
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)
(8)
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:
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).
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.
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,
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).
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).
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).
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.
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'):
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.
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.
This analytical form of solution for the device electrostatics is very useful
for calculating electrostatic forces in various NEMS devices.
5. Summary
Acknowledgements
References
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
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.
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.
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
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
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
Peak !l-l5.06 ·C
~. rtB3.911 ·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
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..
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
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 ~
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.
(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
§ "-
~
~
~ ~
~
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.
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
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
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
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
ABSTRACT
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
!
. , 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.
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
Abstract
1. Introduction
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
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
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
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
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.
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.
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
(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
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
3000
2500
2000
1500
1000
500
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.
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.
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
1. Introduction
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
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).
Monomer Tank
Electrode Needle Valve
Plasma Reactor
Vacuum Pump
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.
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 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)
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).
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
1. Introduction
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) 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
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.
%
40
30
20
10
~ ____ ~ ____ ~ ____ ~ N
2 3 4
Si-O-Sb
9(1J)
20
10
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
4. Conclusion
Acknowledgments
References
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
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
3. Results
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.
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.
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)
~~~~c
~
'"
u
c:
.;: 100 ---
11l
E ~~------
r/)
c: !
I 100' C
~ I
I
I
I
i
0 i
500 1000 1500 2000
f...(nm)
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.
element on the layer structure, in order explain the change in the electrical
behaviour.
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
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. 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.
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.
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.
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.
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.
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].
-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
Abstract
1. Introduction
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
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'
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.
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.
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
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
1. Introduction
tyramine polytyraminc
2. Experimental
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, ,. =!! '#)
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 '
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.
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·'.
4. Conclusions
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. 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
~
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
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
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
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
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 /
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.
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 ~-
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
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
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
(6)
For Lim «Me, the resonance frequency shift, Llf", due to the added mass, Lim, is thus
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.
• •
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
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
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