Professional Documents
Culture Documents
Full Download Book Scattering Methods and Their Application in Colloid and Interface Science PDF
Full Download Book Scattering Methods and Their Application in Colloid and Interface Science PDF
Otto Glatter
Department of Inorganic Chemistry,
Graz University of Technology,
Graz, Austria
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
Copyright r 2018 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopying, recording, or any information storage and retrieval system, without
permission in writing from the publisher. Details on how to seek permission, further information about the
Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance
Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other
than as may be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
understanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using
any information, methods, compounds, or experiments described herein. In using such information or methods
they should be mindful of their own safety and the safety of others, including parties for whom they have a
professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability
for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or
from any use or operation of any methods, products, instructions, or ideas contained in the material herein.
British Library Cataloguing-in-Publication Data
A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data
A catalog record for this book is available from the Library of Congress
ISBN: 978-0-12-813580-8
Scattering methods like small-angle X-ray and neutron scattering (SAXS and SANS), as
well as static and dynamic light scattering (SLS and DLS), have become leading techniques
for the characterization of structure and dynamics in soft matter. Scattering methods are
integral techniques; this means that the investigated area, the scattering volume, is several
orders of magnitude larger than the structures of interest. In most cases, the scattering vol-
ume has dimensions in the regime of millimeters, while the structures of interest can range
from nanometers to several micrometers. This means that typically one measures the con-
tribution of a larger number of scattering centers or particles, even in the case of dilute
systems. This method has the advantage of a relatively strong signal, quick acquisition of
statistically relevant data, and the experiments give a representative overview of the exist-
ing structure. The disadvantage is that no information about individual particles is avail-
able, and that the systems are in most cases not oriented. Complementary techniques with
a high spatial resolution like electron microscopy or atomic force microscopy are, there-
fore, of great importance to get a complete picture of the systems under investigation, but
will not be discussed here.
Soft matter, also called soft condensed matter, includes a wide range of systems and
materials. It is maybe easiest to define what is not included in soft matter: solids, gases,
simple fluids, and crystals. However, liquid crystalline systems are also soft matter, as
well as many complex fluids, colloidal dispersions, and gels. Soft matter science includes
colloids, polymers, and surfactant systems. Surfactants and other amphiphilic systems are
of increasing importance, due to their ability to form self-assembled systems, a key point
in the formation of nanostructured materials and nanotechnology. Molecular self-assembly
leads automatically to nanostructured systems. However, the length scales of interest
range from nanometers up to several micrometers, and to macroscopic dimensions in hier-
archically organized systems. This also leads to a wide range of interesting time-scales of
dynamic behavior, ranging from nanoseconds to several hours when fluid systems show
arrested dynamics. Soft matter science has a strong overlap with biosciences, physics, and
chemistry, especially with polymer research and more generally, with materials sciences.
One or the other aspect of these fields is taught in most universities, but scattering meth-
ods are hardly part of the curriculum.
It is impossible to cover all aspects of scattering methods and their application to soft
matter in one book. Here, I want to give a somewhat detailed introduction into SAXS,
SANS, SLS, and DLS based on my personal experience in these fields over more than 40
years, including typical examples manifesting the possibilities and limitations of these
methods. This contribution is not meant as a review of the existing literature, and is cer-
tainly biased by my personal view. Therefore, some related research fields are completely
missing, like grazing-incidence small-angle scattering (GISAXS and GISANS), inelastic
ix
x PREFACE
Chapter 10, Light Scattering From Large Particles: Lorenz 2 Mie Theory, discusses SLS
from large particles. Large means their size is in the order of the wavelength of light.
Interaction of light is much more complicated in this case, as there exist possibilities of
surface resonances. The corresponding theory—known as Lorenz Mie theory—is dis-
cussed in some detail. Such large particles have a very high scattering power, and there-
fore multiple scattering is an important issue here. Finally, design principles of SLS
instruments are discussed.
The last scattering technique, DLS, is discussed extensively in Chapter 11, Dynamic
Light Scattering (DLS), including basic principles, experimental details, and evaluation
techniques. DLS is mostly used as a fast technique for particle sizing in dilute systems,
using translational diffusion measured in a homodyne experiment. However, it is also pos-
sible to learn about particle interactions from concentration series, and finally one can
study solidifying systems with a so-called ergodic-to-nonergodic transition (also some-
times called glass transition, jamming, or gelation). Rotational diffusion can be measured
for nonisotropic scatterers like rod-like particles by depolarized DLS (DDLS), measuring
only the depolarized component of the scattered light. The electrophoretic mobility is
accessible from electrophoretic DLS (EDLS) measurements performed in a so-called het-
erodyne mode. Moving or flowing samples, like soot particles in a flame, can be measured
in specialized DLS setups.
Even though some examples have been given throughout the previous chapters, the
subsequent two chapters will give more examples to demonstrate the applicability of the
scattering techniques. Chapter 12, Dilute Systems: Practical Aspects—Applications, is
restricted to dilute systems, while in Chapter 13, Concentrated, Interacting Systems:
Practical Aspects, Applications, examples of applications to interacting dense systems are
given.
Glasses, liquid crystals, and gels are typical examples of dense and highly viscous sys-
tems. Some applications in this direction are discussed in Chapter 14, Glasses, Liquid
Crystals, and Gels.
In the Appendix the basic principles of statistical thermodynamics and their relation-
ship to particle correlation functions are summarized to assist the theoretical description
of the structure factor in Chapter 4, Concentration Effects, Interactions.
As already mentioned, this presentation is far from complete, but it should give a good
entry into and overview of the scientific field of scattering techniques, and their applica-
tion to soft matter. Finally, the reference list should help the reader to find the relevant
original literature for further in-depth studies.
C H A P T E R
1
Interference,
RayleighDebyeGans Theory
When an electromagnetic wave is sent through a thin slab of material, the electric field
will introduce a polarization of the atoms, leading to the formation of little dipoles. The elec-
tric field oscillates with a frequency that is defined by the wavelength of the radiation sent
into the material, so the dipoles will oscillate in a forced motion with the same frequency.
Around the turn of the last century, H.A. Lorentz1 developed a classical theory of the
optical properties of matter. He treated electrons and ions as simple harmonic oscillators
(springs), which are moved by the driving force of the electromagnetic field. The results of
this procedure are qualitatively identical to those resulting from a quantum mechanical
treatment of these problems, even though quantities are interpreted differently in the clas-
sical and in quantum mechanical theory.2 The principles of the Lorentz model are essential
to understand the connections between spectroscopy and scattering, but we want to focus
here on the scattering process only. The interested reader will find an excellent overview
of the Lorentz model in Marshall (1978).2
It can be shown that a wave is not only partially absorbed by matter when passing
through a thin slab of material (exponential decrease), but also gives rise to a scattering
field, as every accelerated electric charge emits an electric field with an amplitude propor-
tional to the acceleration.3 In spectroscopic experiments we are interested in absorption
phenomena, i.e., we check the frequency-dependent absorption of radiation to find the
resonances of the systems. This frequency range close to the resonance is called the
Lorentz limit. In scattering methods we try to stay far away from these resonance frequen-
cies, ω0. For low frequencies, i.e., ω{ω0, we are dealing with the Rayleigh limit, which
holds for visible light, and for high frequencies, i.e., ωcω0, we are dealing with the
Thomson limit, which holds for X-rays. In both regimes we are interested in the scattering
wavelets that are emitted from these oscillating dipoles.
At the moment, it is only important to notice that the energy of X-rays is high enough
to be above the resonance regime ω0 for most soft matter materials and to polarize all
electrons, also those in the innermost shells of the atoms, and therefore every electron in
the illuminated volume will give rise to a scattered wave. In the case of visible light the
situation is very different. The energy is lower than that of X-rays by a factor of about
4.000 (defined by the ratio of the wavelengths). This lower energy corresponds to lower
frequencies (ω , ω0), and can only polarize the electrons in the outer shell, typically the
valence electrons. Fortunately we have easy access to this polarizability as it is propor-
tional to the refractive index of the material. This fact will be discussed in more detail in
Chapter 9, Static Light Scattering From Small Particles. Here, we focus on X-ray scattering
and we ignore all prefactors relating the amplitude of the scattered wave to the amplitude
of the incoming plane wave. We set the scattered field amplitude as equal to one, because
we are not interested in the amplitude of the single scattering waves, but we want to dis-
cuss how they sum up (interfere) to the total scattering field. We discuss a virtual particle
in a vacuum with a volume V and a distribution of electrons that results in an electron den-
sity ρ(r) (number of electrons per volume element dV) at position r.
We now let a plane wave hit this particle. All scattering waves are coherent, i.e., there is
a fixed phase relation between the incident radiation and the scattered waves, and both
fields have the same frequency. Incoherent (Compton) scattering can be neglected for our
scattering process with X-rays, but is important in the case of neutrons.
The different waves will differ only in their relative phases ϕ, which will depend on the
position of the scattering centers relative to each other. The phase difference relative to an
arbitrary reference beam is given by the difference of the path lengths multiplied by the
wave number k 5 2π/λ, where λ is the wavelength of the radiation in the material. For a
traveling wave, the wave number gives the change in phase per length. This phase change
is, of course, equal to 2π if the wave has traveled one wavelength.
We choose an arbitrary reference point O in the body, and consider a beam through
this point as a reference beam. The direction of the incident beam is defined by the unit
- -
vector s0 s0 ; the direction of the scattered beam is given by s s (see Fig. 1.1). The
path-length difference of the reference beam and a beam through the point P is:
a 2 b 5 rs0 2 rs 5 2 r ðs 2 s0 Þ; (1.1)
so we get the phase difference ϕ by multiplication with the wave number:
ϕ 5 2 ð2π=λÞ r ðs 2 s0 Þ (1.2)
We now introduce the scattering vector q with:
q 5 ð2π=λÞ ðs 2 s0 Þ: (1.3)
and so we get:
ϕ 5 2qr (1.4)
The difference vector (s 2 s0) lies symmetrically with the incident and the scattered
beam, or orthogonal to the “mirror plane” (dashed line in Fig. 1.2) and its magnitude is
2sin(θ/2), where θ is the scattering angle. Contrary to what older textbooks on small-angle
scattering tell us, we do not use 2θ as the scattering angle; such a definition would contra-
dict the usual definition in light scattering!
The scattering vector q has the same direction as (ss0), and its magnitude is given by
the magnitude of (s 2 s0), 2sin(θ/2) (see Fig. 1.2) multiplied by the wave number 2π/λ:
4π θ
q q 5 sin (1.5)
λ 2
For small angles sinθ/2 may be replaced by θ/2, i.e., q, the magnitude of q is propor-
tional to the scattering angle θ (to be given in radians). The dimension is [length]21,
depending on the choice of λ. We use nanometers for the wavelength, so q has the dimen-
sion nm21. It should be noted that many other symbols are used in literature for the scat-
tering vector, like h, k, s, Q, etc., sometimes with a slightly different definition (missing
factor 2π). For all theoretical considerations, we shall use the scattering vector instead of
the scattering angle. This makes all our results independent of the wavelength used. It
should be kept in mind, however, that a certain q value will lead to different scattering
angles for different wavelengths! To summarize: the scattering vector has a magnitude
proportional to the scattering angle θ, has the dimension nm21, and its direction is not the
direction of the detector but points into the bisecting line between the incoming beam and
the detector direction.
Now let us come back to the phase difference ϕ. The scalar vector product qr in
Eq. (1.4) means that only the component of r that is in the direction of q is relevant for the
phase. This implies that all points in a plane perpendicular to q will have the same phase.
Scattering might, therefore, be regarded as a “reflection” by a set of planes (dashed line in
Fig. 1.2). This concept, though of great importance in crystallography, will not be used in
the following.
The product ks 5 k is often used as wave vector and the wave vector change q 5 k 2 k0 is
also called momentum transfer. According to de Broglie the relationship between the wave
properties and the particulate properties can be expressed by the wavelength λ that is
associated with a particle of mass m and velocity v:
h
λ5 (1.6)
mv
where h is Plank’s constant. So we see that the momentum, mv, can be written as:
mv 5
h
5
λ 2π
h
2πλ 5 h̄ k (1.7)
where h̄ 5 h/2π. This equation holds for neutrons as well as for photons and for any elec-
tromagnetic radiation. We see that the momentum of the field or of the neutron is a vector
pointing in the direction of propagation and its magnitude is h̄ω=c 5 h̄k. The elastically
scattered wave has the same energy or wavelength as before and so the same wave num-
ber k, but the direction of the wave vector has changed from s0 to s. The corresponding
momentum transfer is h̄q (see Fig. 1.3). We may even think of a completely classical anal-
ogy of a billiard ball moving in s0 direction with a momentum of h̄k0 5 h̄ks0 (whatever h̄k
might be). In order to change its direction into s with the same speed, i.e., momentum, as
h̄k 5 h̄ks we have to knock it with another ball with momentum h̄q, i.e., we have to trans-
fer the momentum h̄q to our ball.
We can now calculate the total scattered field from our particle by summing up all sec-
ondary waves considering their relative phases. This can easily be done by using complex
notation. A complex number with magnitude 1 and phase ϕ is represented by eiϕ or
e2iqr. The number of electrons (and the number of wavelets) per unit volume is given
by the electron density ρ(r). A small volume element dr (or dV) at position r will contain
ρ(r)dr electrons. So the summation can be replaced by integration over the whole vol-
ume V of the particle:
ð
Es ðqÞ 5 ρðrÞe2iqr dr (1.8)
V
In mathematical terms, the scattered field is the Fourier transformation of the electron-
density distribution of the object.
We call the space containing all position vectors r the real space, and the space contain-
ing the vectors q the reciprocal space. This is a direct consequence of the similarity theorem
of Fourier transformation.4 The reciprocity between real and reciprocal space can also be
explained by using Eq. (1.8). The two spaces are connected by the phase qr only; the result
would be the same if r were enlarged and q diminished by the same factor. So large
RAYLEIGHDEBYEGANS THEORY
The so-called RayleighDebyeGans (RDG) theory of scattering is nothing other than the
rigorous application of the idea of coherent interference to scattering, assuming that the
electromagnetic wave propagates into and through the particle regularly and homo-
geneously, i.e., the magnitude of the electric field is the same in all parts of the particle.
The RDG theory describes well SAXS and SANS, as well as static light scattering of small
particles (size much smaller than the wavelength) and low contrast. The name small angle
comes from the fact that the wavelength of X-rays and neutrons is usually much smaller
than the size of the particles under investigation.
In the following we are no longer interested in the scattering of a particle in vacuum,
as we are dealing with soft matter, but in scattering from scattering centers embedded
in a homogeneous medium such as surfactants or macromolecules and their aggregates
in solution, and also defects in glasses, and pores in noncrystalline amorphous solids,
i.e., soft condensed matter where the scattering centers exist in a surrounding medium
of (nearly) constant electron density ρ. The incident field is a plane wave in the direc-
tion s0:
E1 ðr; tÞ 5 E0 eiðks0 r2ωtÞ (1.9)
The scattered waves have a spherical wave form:
eiðkR2ωtÞ
Es ðR; tÞ 5 Es0
R
sinϑ Φ q; t (1.10)
where R is the vector from the scattering center O to the point-like detector; and sinϑ is
the polarization term for polarized light, where ϑ is the angle between the direction of
polarization and the direction to the detector. This polarization factor is mostly relevant
for light-scattering experiments such as those that are usually performed with a polar-
ized laser source. In order to detect the scattered light intensity in a horizontal plane
the laser must be vertically polarized, i.e., ϑ 5 90 degrees, and therefore sinϑ 5 1.0. For
unpolarized light this factor has to be replaced by (1 1 cos2θ)/2, with θ being the scat-
tering angle. Most X-ray sources emit unpolarized light. However, in SAXS experiments
the scattering angle is small enough so this factor can be neglected, but it must be
considered for SWAXS (combined small- and wide-angle X-ray scattering) experiments.
where we have omitted the polarization term since (1 1 cos2θ)/2 is close enough to 1 for
the usual range of scattering angles in SAXS. In order to find the total scattered field we
have to integrate over the whole illuminated volume V:
ð
e2 eiðkR2ωtÞ
Es ðR; tÞ 5 E0 2
mc
R V
ρðr; tÞe2iqr dr: (1.16)
We can now express the density ρ(r, t) by its mean ρ and its fluctuations Δρ(r, t):
ρðr; tÞ 5 ρ 1 Δρðr; tÞ (1.17)
The contribution of the first term, caused by the mean ρ integrated over the whole scat-
tering volume V, peaks strongly at q 5 0, as the typical size of the scattering volume is in
the order of millimeters; this is huge compared to the wavelength λ. Therefore, this term
is hidden in the primary beam which has to be blocked in the experiment by a beam stop.
So we are left with the second term only, and for the scattering at q 6¼ 0 we get only the
contributions of the fluctuations Δρ(r, t) around the mean ρ:
ð
e2 eiðkR2ωtÞ
Es ðR; tÞ 5 E0 2 Δρðr; tÞe2iqr dr (1.19)
mc R V
So we see that the scattered field Es is the Fourier transformation of the density fluctua-
tions Δρ. This result leads to the fact that the contrast can also be negative (whenever the
density of a particle is lower than that of the solvent).
Inhomogeneities of the solvent are usually very small, and their contributions are taken
into account experimentally by subtracting the scattering of the pure solvent from the scat-
tering of the solution (see Chapter 8: Numerical Methods).
In addition, we should keep in mind that ρ is the electron density of X-rays, but repre-
sents the scattering-length density of neutrons (see Chapter 5: Absolute Intensity, SAXS,
and SANS) and the refractive index for light scattering (see Chapter 9: Static Light
Scattering From Small Particles) (Fig. 1.4).
Now we omit for simplicity all prefactors (including the fixed value of R) from
Eq. (1.19), replace the vector R pointing to the detector by the corresponding scattering
vactor q, and get:
N ð
X
Es q; t ~ Δρðrk ; tÞe2iq½Rk ðtÞ1rk drk
k51 Vk
XN ð
5 Δρðrk ; tÞe2iqrk drk e2iqRk ðtÞ (1.21)
k51 Vk
XN
5 Fk q; t e2iqRk ðtÞ
k51
where we have introduced the field Fk scattered by the particle k with volume Vk as:
ð
Fk ðq; tÞ 5 Δρðrk ; tÞe2iqrk drk (1.22)
Vk
At this step one might wonder what happened to the time-dependent phase factor
ei(kRωt). The time average over any periodic function with arbitrary frequency is
/2 (hcos2 (ωt 1 ϕ)it 5 1/2), so the intensity of the incident field I0 is simply equal to E02/2;
1
DILUTE SYSTEMS
We start with the simplest possible case; a very dilute system of scatterers. This
assumption not only guarantees that the condition of single scattering is fulfilled automati-
cally, but also leads to a situation in which all scatterers are uncorrelated, i.e., every particle
If the system is very dilute and the scatterers are uncorrelated, the phase factors in the
double sum go through many multiples of 2π, and so these contributions cancel each other
and we get:
XN
Is q 5 jFk q j2 ; (1.25)
k51
i.e., the scattered intensity is simply the sum over the average intensities scattered by all N
individual particles, and contains information about structure and shape of these particles
averaged over orientations and particle sizes and shapes.
Now we introduce the convolution square Δ~ρ2 ðrÞ of the density fluctuations:
ð
γðrÞ Δρ~2 ðrÞ 5 Δρ ðr1 Þ Δρ ðr1 2 rÞ dr1 (1.31)
V
So we see that the scattered intensity is essentially the Fourier transformation of the
spatial ACF of the density fluctuations ΔρðrÞ of the particle.
Now we have to perform the spatial averaging process. This means that the autocorrela-
tion function depends only on the magnitude of the distance r 5 |r|, and the scattering
intensity is a function of the magnitude q of the scattering vector q only. It is obvious that
information about the system is lost by this averaging process; a point that is important
for the evaluation and interpretation of scattering experiments. Also, the three-
dimensional Fourier transformation (Eq. 1.32) reduces to the one-dimensional, centro-
symmetric form. The phase factor must be replaced by its average taken over all directions
of r and so we get:
ð
γðrÞ 5 ρ~2 ðrÞ 2 VðρÞ2 5 Δρ~2 ðrÞ 5 Δρ ðr1 Þ Δρ ðr1 2 rÞ dr1 (1.33)
V
The convolution square Δρ~ ðrÞ is calculated as the overlap between a particle fixed in
2
space and its identical ghost shifted by r. If Δρ(r) is constant inside the particle, Δρ~2 ðrÞ is
given by the square of this contrast, (Δρ)2 times the overlap volume between the particle
and its shifted ghost, see Porod (1948).9
The spatially averaged convolution square γ ðrÞ 5 Δρ~2 ðrÞ results from the same process;
the ghost is shifted by a distance r 5 |r|, but we have to average over all possible direc-
tions in space. This correlation function γ(r) was introduced in the field of small-angle
scattering by Debye and Bueche (1949),10 and in its normalized form as the characteristic
function γ 0(r) by Porod (1951)11 (Fig. 1.5).
FIGURE 1.6 Spatially averaged convolution square γðrÞ 5 Δρ~2 ðrÞ: here we
shift the ghost again over the distance r 5 |r|, but take the average over all
directions in space.
Let us take a vector r with magnitude r that takes all orientations with equal probability. We
define the angle between q and r with ϕ. This angle varies between 0 and π radians. Any rota-
tion of r around q with fixed ϕ is described by the angle ψ. For any possible direction r we shall
find a direction r so that hsinqri 5 0, and we can deal with hcosqri instead of he2iqri (remember
Euler’s identity: e2iqr 5 cosqr 2 i sinqr). The product qr can therefore be described as qr cos ϕ,
and the probability of finding the endpoint of r between ϕ and ϕ 1 dϕ is sinϕdϕ.
So we have:
ð ð
1 2π π
hcosqri 5 cos qr cos ϕ sin ϕ dϕ dψ
4π Ψ50 ϕ50
ð
1 π
5 cos qr cos ϕ sin ϕ dϕ
2 ϕ50
and by introducing the pair 2 distance distribution function (PDDF) p(r) as:
ð
2
2 2
pðrÞ 5 r γ ðrÞ 5 r Δρ~ ðrÞ 5 r 2
Δρ ðr1 Þ Δρ ðr1 2 rÞ dr1
V
(1.38)
we obtain:
ðN
sin ðqrÞ
IðqÞ 5 4π pðrÞ dr: (1.39)
0 qr
The correlation function and the PDDF p(r) can, in principle, be calculated from I(q) by
the inverse transformation:
ð
1 N sinqr
γ ðrÞ 5 IðqÞq2 dq (1.40)
2π 0 qr