\documentclass[12pt]{article}
\usepackage{wrapfig}
\usepackage{amsmath}
\usepackage{graphicx}
\message{
*************************************************************************
Copyright, 19952017, all rights reserved, Mark Jarrell (Dept.of
Physics and Atronomy, Louisiana State University, LA 70803). This
material may not be reproduced for profit, modified or published in any
form (this includes electronic redistribution) without the prior written
permission of the author listed above.
*************************************************************************
}
\title{Chapter 4: Crystal Lattice Dynamics}
\author{Debye}
\input{../defs}
\def\baselinestretch{1.4}
\bdo
\maketitle
\large
\tableofcontents
%\Large
\pagebreak
A crystal lattice is special due to its long range order. As you
explored in the homework, this yields a sharp diffraction
pattern, especially in 3d. However, lattice vibrations are
important. Among other things, they contribute to
\begin{itemize}
\item the thermal conductivity of insulators is due to dispersive
lattice vibrations, and it can be quite large (in fact, diamond has
a thermal conductivity which is about 6 times that of metallic copper).
\item in scattering they reduce of the spot intensities, and
also allow for inelastic scattering where the energy of the scatterer
(i.e.\ a neutron) changes due to the absorption or creation of a phonon
in the target.
\item electronphonon interactions renormalize the properties of
electrons (make them heavier).
\item superconductivity (conventional) comes from multiple electronphonon
scattering between timereversed electrons.
\end{itemize}
\section{An Adiabatic Theory of Lattice Vibrations}
At first glance, a theory of lattice vibrations would appear impossibly
daunting. We have $N\approx 10^{23}$ ions interacting strongly (with
energies of about ($e^2/\A$)) with $N$ electrons. However, there is a
natural expansion parameter for this problem, which is the ratio of the
electronic to the ionic mass:
\beq
\frac{m}{M} \ll 1
\eeq
which allows us to derive an accurate theory.
Due to Newton's third law, the forces on the ions and electrons are comparable $F\sim e^2/a^2$, where $a$ is the lattice constant. If we
imagine that, at least for small excursions, the forces binding the electrons
and the ions to the lattice may be modeled as harmonic oscillators,
then
\beq
F \sim e^2/a^2 \sim m\omega_{electron}^2a \sim M\omega_{ion}^2a
\eeq
This means that
\beq
\frac{\omega_{ion}}{\omega_{electron}}\sim\lep\frac{m}{M}\rip^{1/2}
\sim 10^{3} \;\;\mbox{to}\;\;10^{2}
\eeq
Which means that the ion is essentially stationary during the
period of the electronic motion.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.49\textwidth,clip=true]{setup.pdf}}
\caption[]{\em{Nomenclature for the lattice vibration problem. $\s_{n,\alpha}$
is the displacement of the atom $\alpha$ within the nth unit cell from
its equilibrium position, given by $\r_{n,\alpha}=\r_{n}+\r_{\alpha}$,
where as usual, $\r_n=n_1\aone+n_2\atwo+n_3\athree$. }}
\label{fig:setup}
\end{figure}
For this reason we may make an adiabatic approximation:
\begin{itemize}
\item we treat the ions as stationary at locations $\R_1,\cdots\R_N$
and determine the electronic ground state energy, $E(\R_1,\cdots\R_N)$.
This may be done using standard abinitio band structure techniques
(DFT, GGA, etc.).
\item we then use this as a potential for the ions; i.e.. we
recalculate $E$ as a function of the ionic locations, always
assuming that the electrons remain in their ground state.
\end{itemize}
Thus the potential energy for the ions
\beq
\phi(\R_1,\cdots\R_N)= E(\R_1,\cdots\R_N)+\mbox{the ionion interaction}
\eeq
We will define the zero potential such that when all $\R_n$ are at
their equilibrium positions, $\phi=0$. Then
\beq
H=\sum_n\frac{P_n^2}{2M} +\phi(\R_1,\cdots\R_N)
\eeq
Typical lattice vibrations involve small atomic excursions of the order
$0.1\A$ or smaller, thus we may expand about the equilibrium position of
the ions.
\beq
\phi(\{r_{n\alpha i}+ s_{n\alpha i}\})=
\phi(\{r_{n\alpha i}\}) +
\pde{\phi}{r_{n\alpha i}}s_{n\alpha i} +
\frac{1}{2}\spdm{\phi}{r_{n\alpha i}}{r_{m\beta j}}s_{n\alpha i}s_{m\beta j}
\eeq
The first two terms in the sum are zero; the first by definition, and
the second is zero since it is the first derivative of a potential
being evaluated at the equilibrium position. We will define the matrix
\beq
\Phi_{n\alpha i}^{m\beta j}=\spdm{\phi}{r_{n\alpha i}}{r_{m\beta j}}
\eeq
From the different conservation laws (related to symmetries) of the system
one may derive some simple relationships for $\Phi$. We will discuss these
in detail later.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.85\textwidth,clip=true]{periodicpot.pdf}}
\caption[]{\em{Since the coefficients of potential between the atoms linked by
the blue lines (or the red lines) must be identical,
$\Phi_{n\alpha i}^{m\beta j}=\Phi_{0\alpha i}^{(mn)\beta j}$.}}
\label{fig:periodicpot}
\end{figure}
However, one must be introduced now, that
is, due to translational invariance.
\beq
\Phi_{n\alpha i}^{m\beta j}=\Phi_{0\alpha i}^{(mn)\beta j}
=\spdm{\phi}{r_{0\alpha i}}{r_{(nm)\beta j}}
\eeq
ie, it can only depend upon the distance. This is important for
the next subsection.
\subsection{The Equation of Motion}
From the derivative of the potential, we can calculate the
force on each site
\beq
F_{n\alpha i}= \pde{\phi(\{r_{m\beta j}+s_{m\beta j}\})}{s_{n\alpha i}}
\eeq
so that the equation of motion is
\beq
\Phi_{n\alpha i}^{m\beta j} s_{m\beta j} = M_\alpha \ddot{s}_{n\alpha i}
\eeq
If there are $N$ unit cells, each with $r$ atoms, then this gives
$3Nr$ equations of motion. We will take advantage of the periodicity of
the lattice by using Fourier transforms to achieve a significant
decoupling of these equations. Imagine that the coordinate $s$ of each
site is decomposed into its Fourier components. Since the equations
are linear, we may just consider one of these components to derive our
equations of motion in Fourier space
\beq
s_{n\alpha i}=\frac{1}{\sqrt{M_\alpha}} u_{\alpha i}(\q)
e^{i(\q\cdot\r_n\omega t)}
\label{eq:displacement}
\eeq
where the first two terms on the rhs serve as the polarization vector
for the oscillation, $u_{\alpha i}(\q)$ is independent of $n$ due
to the translational invariance of the system.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.7\textwidth,clip=true]{propagate.pdf}}
\caption[]{\em{$u_{\alpha i}(\q)$ is independent of $n$ so that a lattice
vibration can propagate and respect the translational invariance of the
lattice.}}
\label{fig:propagate}
\end{figure}
In a real system the
real $s$ would be composed of a sum over all $\q$ and polarizations.
With this substitution, the equations of motion become
\beq
\omega^2 u_{\alpha i}(\q)
=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{n\alpha i}^{m\beta j}
e^{i\q\cdot(\r_m\r_n)}
u_{\beta j}(\q)
\;\;\;\mbox{sum repeated indices}\,.
\eeq
Recall that $\Phi_{n\alpha i}^{m\beta j}=\Phi_{0\alpha i}^{(mn)\beta j}$
so that if we identify
\beq
D_{\alpha i}^{\beta j}
=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{n\alpha i}^{m\beta j}
e^{i\q\cdot(\r_m\r_n)}
=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0\alpha i}^{p\beta j}
e^{i\q\cdot(\r_p)}
\eeq
where $\r_p=\r_m\r_n$, then the equation of motion becomes
\beq
\omega^2 u_{\alpha i}(\q)
=
D_{\alpha i}^{\beta j}
u_{\beta j}(\q)
\eeq
or
\beq
\lep D_{\alpha i}^{\beta j} \omega^2\delta_{\alpha i}^{\beta j}\rip
u_{\beta j}(\q)
=
0
\label{eq:secular}
\eeq
which only has nontrivial ($u\neq 0$) solutions if
$ \det \lep \D(\q)\omega^2\I \rip =0$. For each $\q$ there are
$3r$ different solutions (branches) with eigenvalues $\omega^{(n)}(\q)$
(or rather $\omega^{(n)}(\q)$ are the root of the eigenvalues).
The dependence of these eigenvalues $\omega^{(n)}(\q)$ on $\q$ is known
as the dispersion relation.
\subsection{Example, a Linear Chain}
\begin{figure}[htb]
\centerline{\includegraphics[width=0.8\textwidth,clip=true]{M1M2.pdf}}
\caption[]{\em{A linear chain of oscillators composed of a twoelement
basis with different masses, $M_1$ and $M_2$ and equal strength springs
with spring constant $f$.}}
\label{fig:M1M2}
\end{figure}
Consider a linear chain of oscillators composed of a twoelement
basis with different masses, $M_1$ and $M_2$ and equal strength springs
with spring constant $f$. It has the potential energy
\beq
\phi=\frac12 f\sum_n
\lep s_{n,1}s_{n,2}\rip^2 + \lep s_{n,2}s_{n+1,1}\rip^2\,.
\eeq
We may suppress the indices $i$ and $j$, and search for a solution
\beq
s_{n\alpha}=\frac{1}{\sqrt{M_\alpha}} u_{\alpha}(\q)
e^{i(\q\cdot\r_n\omega t)}
\eeq
to the equation of motion
\beq
\omega^2 u_{\alpha}(\q) = D_{\alpha}^{\beta} u_{\beta}(\q)
\;\;
\mbox{where}
\;\;
D_{\alpha}^{\beta}
=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0\alpha}^{p,\beta}
e^{i\q\cdot(\r_p)}
\eeq
and,
\beq
\Phi_{n,\alpha}^{m,\beta}
=\spdm{\phi}{r_{0,\alpha}}{r_{(nm),\beta}}
\eeq
where nontrivial solutions are found by solving $ \det \lep \D(\q)\omega^2\I \rip =0$.
The potential matrix has the form
\beq
\Phi_{n,1}^{n,1}=\Phi_{n,2}^{n,2}=2f
\eeq
\beq
\Phi_{n,1}^{n,2}=\Phi_{n,2}^{n,1}=\Phi_{n,1}^{n1,2}=\Phi_{n,2}^{n+1,1}=f\,.
\eeq
This may be Fourier transformed on the space index $n$ by inspection, so that
\beqa
D_{\alpha}^{\beta}
&=&
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0\alpha}^{p\beta}
e^{i\q\cdot(\r_p)}\nonumber \\
&=&
\lep\begin{array}{cc}
\frac{2f}{M_1} & \frac{f}{\sqrt{M_1M_2}}\lep 1+e^{iqa}\rip \\
\frac{f}{\sqrt{M_1M_2}}\lep 1+e^{+iqa}\rip & \frac{2f}{M_2}
\end{array}\rip
\eeqa
Note that the matrix $\D$ is hermitian, as it must be to yield real,
physical, eigenvalues $\omega^2$ (however, $\omega$ can still be imaginary
if $\omega^2$ is negative, indicating an unstable mode). The secular
equation $ \det \lep \D(\q)\omega^2\I \rip =0$ becomes
\beq
\omega^4 \omega^2 2f\lep\frac{1}{M_1}+\frac{1}{M_2}\rip
+\frac{4f^2}{M_1M_2}\sin^2(qa/2)=0\,,
\eeq
with solutions
\beq
\omega^2= f\lep\frac{1}{M_1}+\frac{1}{M_2}\rip
\pm
f\sqrt{
\lep\frac{1}{M_1}+\frac{1}{M_2}\rip^2 \frac{4}{M_1M_2}\sin^2(qa/2)
}
\eeq
\begin{figure}[htb]
\centerline{\includegraphics[width=0.7\textwidth,clip=true]{dispersion.pdf}}
\caption[]{\em{Dispersion of the linear chain of oscillators shown in
Fig.~\ref{fig:M1M2} when $M_1=1$, $M_2=2$ and $f=1$. The upper branch $\omega_+$
is called the optical and the lower branch is the acoustic mode.}}
\label{fig:dispersion}
\end{figure}
This equation simplifies significantly in the $q\to 0$ and $q/a\to\pi$ limits.
In units where $a=1$, and where the reduced mass
$1/\mu=\lep\frac{1}{M_1}+\frac{1}{M_2}\rip$,
\beq
\lim_{q\to 0} \omega_{}(q)=qa\sqrt{\frac{f\mu}{2M_1M_2}}
\;\;\;\;\;\;
\lim_{q\to 0} \omega_{+}(q)=\sqrt{\frac{2f}{\mu}}
\eeq
and
\beq
\omega_(q=\pi/a)= \sqrt{2f/M_2}\,.
\;\;\;\;\;\;
\omega_+(q=\pi/a)= \sqrt{2f/M_1}
\eeq
As a result, the $+$ mode is quite flat; whereas the $$ mode varies from
zero at the Brillouin zone center $q=0$ to a flat value at the edge
of the zone. This behavior is plotted in Fig.~\ref{fig:dispersion}.
It is also instructive to look at the eigenvectors, since they
will tell us how the atoms vibrate. Let's look at the optical mode
at $q=0$, $\omega_+(0)=\sqrt{2f/\mu}$. Here,
\beq
\D=\left(\begin{array}{cc}
2f/M_1 & 2f/\sqrt{M_1M_2}\\
2f/\sqrt{M_1M_2} & 2f/M_2
\end{array}\right)\,.
\eeq
Eigenvectors are nontrivial solutions to $(\omega^2\I\D)\u=0$,
or
\beq
0=\lep\begin{array}{cc}
2f/\mu2f/M_1 & 2f/\sqrt{M_1M_2}\\
2f/\sqrt{M_1M_2} & 2f/\mu2f/M_2
\end{array}\rip
\lep\begin{array}{c}
u_1\\u_2\end{array}\rip
\,.
\eeq
with the solution $u_1=\sqrt{M_2/M_1}u_2$. In terms of the actual
displacements Eqs.\ref{eq:displacement}
\beq
\frac{s_{n1}}{s_{n2}}=\sqrt{\frac{M_2}{M_1}}\frac{u_1}{u_2}
\eeq
or $s_{n1}/s_{n2}=M_2/M_1$
so that the two atoms in the basis are moving out of phase with
amplitudes of motion inversely proportional to their masses.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.75\textwidth,clip=true]{M1M2_opt.pdf}}
\caption[]{\em{Optical Mode (bottom) of the linear chain (top).}}
\label{fig:optical}
\end{figure}
These modes are described as optical modes since these atoms, if oppositely
charged, would form an oscillating dipole which would couple to optical
fields with $\lambda\sim a$. Not all optical modes are optically
active.
\subsection{The Constraints of Symmetry}
We know a great deal about the dispersion of the lattice
vibrations without solving explicitly for them. For example, we know that
for each $q$, there will be $dr$ modes (where $d$ is the lattice dimension,
and $r$ is the number of atoms in the basis). We also expect (and implicitly
assumed above) that the allowed frequencies are real and positive. However,
from simple mathematical identities, the pointgroup and translational
symmetries of the lattice, and its timereversal invariance, we can learn
more about the dispersion without solving any particular problem.
The basic symmetries that we will employ are
\begin{itemize}
\item The translational invariance of the lattice and reciprocal
lattice.
\item The point group symmetries of the lattice and reciprocal
lattice.
\item Timereversal invariance.
\end{itemize}
\subsubsection{Symmetry of the Dispersion}
\paragraph{Complex Properties of the dispersion and Eigenmodes}
First, from the symmetry of the second derivative, one may show that
$\omega^2$ is real. Recall that the dispersion is determined by the
secular equation $ \det \lep \D(\q)\omega^2\I \rip =0$, so if
$\D$ is hermitian, then its eigenvalues, $\omega^2$, must be real.
\beqa
D_{\alpha i}^{*\beta j}
&=&
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0\alpha i}^{p\beta j}
e^{i\q\cdot(\r_p)}\\
&=&
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0,\alpha,i}^{p,\beta, j}
e^{i\q\cdot(\r_p)}
\eeqa
Then, due to the symmetric properties of the second derivative
\beq
D_{\alpha i}^{*\beta j}
=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{p,\beta, j}^{0,\alpha,i}
e^{i\q\cdot(\r_p)}=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0,\beta, j}^{p,\alpha,i}
e^{i\q\cdot(\r_p)}=
D_{\beta j}^{\alpha i}
\eeq
Thus, $\D^{T*}=\D^{\dag}=\D$ so $\D$ is hermitian and its eigenvalues
$\omega^2$ are real. This
means that either $\omega$ are real or they are pure imaginary. We will
assume the former. The latter yields pure exponential growth of our
Fourier solution, indicating an instability of the lattice to a secondorder
structural phase transition.
Timereversal invariance allows us to show related results.
We assume a solution of the form
\beq
s_{n\alpha i}=\frac{1}{\sqrt{M_\alpha}} u_{\alpha i}(\q)
e^{i(\q\cdot\r_n\omega t)}
\eeq
which is a plane wave. Suppose that the plane wave is moving to the
right so that $\q=\xh q_x$, then the plane of stationary
phase travels to the right with
\beq
x=\frac{\omega}{q_x}t\,.
\eeq
Clearly then changing the sign of $q_x$ is equivalent to taking
$t\to t$. If the system is to display proper timereversal invariance,
so that the plane wave retraces its path under timereversal, it must have
the same frequency when time, and hence $\q$, is reversed, so
\beq
\omega(\q)=\omega(\q)\,.
\eeq
Note that this is fully equivalent to the statement that
$D_{\beta j}^{\alpha i}(\q) = D_{\beta j}^{*\alpha i}(\q)$
which is clear from the definition of $\D$.
Now, return to the secular equation, Eq.~\ref{eq:secular}.
\beq
\lep D_{\alpha i}^{\beta j}(\q) \omega^2(\q)\delta_{\alpha i}^{\beta j}\rip
\epsilon_{\beta j}(\q)
=
0
\eeq
Lets call the (normalized) eigenvectors of this equation $\epsilon$. They
are the elements of a unitary matrix which diagonalizes $\D$. As a
result, they have orthogonality and completeness relations
\beq
\sum_{\alpha,i}
\epsilon_{\alpha,i}^{*(n)}(\q)\epsilon_{\alpha,i}^{(m)}(\q)=\delta_{m,n}
\;\;\;\;\mbox{orthogonality}
\eeq
\beq
\sum_{n}
\epsilon_{\alpha,i}^{*(n)}(\q)\epsilon_{\beta,j}^{*(n)}(\q)=
\delta_{\alpha,\beta}\delta_{i,j}
\eeq
If we now take the complex conjugate of the secular equation
\beq
\lep D_{\alpha i}^{\beta j}(\q) \omega^2(\q)\delta_{\alpha i}^{\beta j}\rip
\epsilon_{\beta j}^*(\q)
=
0
\eeq
Then it must be that
\beq
\epsilon_{\beta j}^*(\q) \propto \epsilon_{\beta j}(\q)\,.
\eeq
Since the $\{\epsilon\}$ are normalized the constant of proportionality may
be chosen as one
\beq
\epsilon_{\beta j}^*(\q) = \epsilon_{\beta j}(\q)\,.
\eeq
\paragraph{PointGroup Symmetry and the Dispersion}
A point group operation takes a crystal back to an identical
configuration. Both the original and final lattice must have the
same dispersion. Thus, since the reciprocal lattice has the same point group
as the real lattice, the dispersion relations have the same point group
symmetry as the lattice.
For example, the dispersion must share the periodicity of the Brillouin zone.
From the definition of $\D$
\beq
D_{\alpha i}^{\beta j}(\q)
=
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0\alpha i}^{p\beta j}
e^{i\q\cdot(\r_p)}
\eeq
it is easy to see that $D_{\alpha i}^{\beta j}(\q+\G)=
D_{\alpha i}^{\beta j}(\q)$ (since $\G\cdot\r_p=2\pi n$, where $n$ is
an integer). I.e., $\D$ is periodic in kspace, and so its eigenvalues
(and eigenvectors) must also be periodic.
\beq
\omega^{(n)}(\k+\G)=\omega^{(n)}(\k)
\eeq
\beq
\epsilon_{\beta j}(\k+\G) = \epsilon_{\beta j}(\k)\,.
\eeq
\subsubsection{Symmetry and the Need for Acoustic modes}
Applying basic symmetries, we can show that an elemental lattice
(that with $r=1$) must have an acoustic model.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.49\textwidth,clip=true]{shift.pdf}}
\caption[]{\em{If each ion is shifted by $s_{1,1, i}$, then the lattice
energy is unchanged.}}
\label{fig:shift}
\end{figure}
First, look at the translational invariance of $\Phi$ . Suppose we make
an overall shift of the lattice by an arbitrary displacement
$s_{n,\alpha, i}$ for all sites $n$ and elements of the basis $\alpha$
(i.e.\ $s_{n,\alpha, i}=s_{1,1, i}$).
Then, since the interaction is
only {\em{between}} ions, the energy of the system should remain unchanged.
\beqa
\delta E&=& \frac12\sum_{m,n,\alpha,\beta,i,j} \Phi_{0,\alpha,i}^{mn,\beta,j}
s_{n,\alpha, i}s_{m,\beta, j} =0 \\
&=& \frac12 \sum_{mn\alpha,\beta,i,j}\Phi_{0,\alpha,i}^{mn,\beta,j}
s_{1,1, i}s_{1,1, j}\\
&=& \frac12 \sum_{i,j}s_{1,1, i}s_{1,1, j}\sum_{mn\alpha,\beta}
\Phi_{0,\alpha,i}^{mn,\beta,j}
\eeqa
Since we know that $s_{1,1, i}$ is finite, it must be that
\beq
\sum_{m,n,\alpha,\beta} \Phi_{0,\alpha,i}^{mn,\beta,j}=
\sum_{p,\alpha,\beta}\Phi_{0,\alpha,i}^{p,\beta,j}=0
\eeq
Now consider a strain on the system $V_{m,\beta,j}$, described
by the strain matrix $m_{\beta,j}^{\alpha,i}$
\beq
V_{m,\beta,j}=\sum_{\alpha,i}m_{\beta,j}^{\alpha,i}s_{m,\alpha, i}
\eeq
\begin{figure}[htb]
\centerline{\includegraphics[width=0.75\textwidth,clip=true]{squish.pdf}}
\caption[]{\em{After a stress is applied to a lattice, the movement
of each ion (strain) is not only in the direction of the applied
stress. The response of the lattice to an applied stress is
described by the strain matrix.}}
\label{fig:squish}
\end{figure}
After the stress has been applied, the atoms in the bulk of the sample are
again in equilibrium (those on the surface are maintained in equilibrium
by the stress), and so the net force must be zero. Looking at the
central ($n=0$) atom this means that
\beq
0=F_{0,\alpha,i}=\sum_{m,\beta,j,\gamma,k}\Phi_{0,\alpha,i}^{m,\beta,j}m_{\beta,j}^{\gamma,k}s_{m,\gamma, k}
\eeq
Since this applies for an arbitrary strain matrix $m_{\beta,j}^{\gamma,k}$,
the coefficients for each $m_{\beta,j}^{\gamma,k}$ must be zero
\beq
\sum_{m}\Phi_{0,\alpha,i}^{m,\beta,j}s_{m,\gamma, k}=0
\eeq
An alternative way (cf.\ Callaway) to show this is to recall that the
reflection symmetry of the lattice requires that $\Phi_{0,\alpha,i}^{m,\beta,j}$
be even in $m$; whereas, $s_{m,\gamma, k}$ is odd in $m$. Thus the
sum over all $m$ yields zero.
Now let's apply these constraints to $\D$ for an elemental lattice where
$r=1$, and we may suppress the basis indices $\alpha$.
\beq
D_{i}^{j}(\q)
=
\frac{1}{M}\sum_p
\Phi_{0, i}^{p, j}
e^{i\q\cdot(\r_p)}
\eeq
For small $q$ we may expand $D$
\beq
D_{i}^{j}(\q)
=
\frac{1}{M}\sum_p
\Phi_{0, i}^{p, j}
\lep 1+ i\q\cdot(\r_p) \frac12 \lep \q\cdot(\r_p)\rip^2 +\cdots \rip
\eeq
We have shown above that the first two terms in this series are zero.
Thus,
\beq
D_{i}^{j}(\q)
\approx
\frac{1}{2M}\sum_p
\Phi_{0, i}^{p, j}
\lep i\q\cdot(\r_p)\rip^2
\eeq
Thus, the leading order (small $\q$) eigenvalues $\omega^2(\q)\sim q^2$.
I.e. they are acoustic modes. We have shown that all elemental lattices
must have acoustic modes for small $\q$.
In fact, one may show that all harmonic lattices in which the
energy is invariant under a rigid translation of the entire lattice must
have at least one acoustic mode. We will not prove this, but rather
make a simple argument. The rigid translation of the lattice corresponds
to a $\q=0$ translational mode, since no energy is gained by this translation,
it must be that $\om_s(\q=0)=0$ for the branch $s$ which contains this
mode. The acoustic mode may be obtained by perturbing (in $\q$) around
this point. Physically this mode corresponds to all of the elements of
the basis moving together so as to emulate the motion in the elemental
basis.
\section{The Counting of Modes}
In the sections to follow, we need to perform sums (integrals) of functions
of the dispersion over the crystal momentum states $\k$ within the reciprocal
lattice. However, the translational and point group symmetries of the crysal,
often greatly reduce the set of points we must sum. In addition, we often
approximate very large systems with hypertoroidal models with periodic
boundary conditions. This latter approximation becomes valid as the
system size diverges so that the surface becomes of zero measure.
\subsection{Periodicity and the Quantization of States}
A consequence of approximating our system as a finitesized periodic system
is that we now have a discrete sum rather than an integral over $\k$.
Consider a onedimensional finite system with $N$ atoms and periodic boundary
conditions. We seek solutions to the phonon problem of the type
\beq
s_n=\epsilon(q)e^{i(qr_n\omega t)}\;\;\;\mbox{where }r_n=na
\eeq
and we require that
\beq
s_{n+N}=s_n
\eeq
or
\beq
q(n+N)a=qna+2\pi m\;\;\;\mbox{where $m$ is an integer}
\eeq
Then, the allowed values of $q=2\pi m/Na$. This will allow us to convert
the integrals over the Brillouin zone to discrete sums, at least for cubic
systems; however, the method is easily generalized for other Bravais
lattices.
\subsection{Translational Invariance: First Brillouin Zone}
We can use the translational invariance of the crystal to reduce
the complexity of sums or integrals of functions of the dispersion over
the crystal momentum states. As shown above, translationally invariant
systems have states which are not independent. It is useful then to define
a region of kspace which contains only independent states. Sums over k
may then be confined to this region. This region is defined as the smallest
polyhedron centered at the origin of the reciprocal lattice and enclosed by
perpendicular bisectors of the G's is called the Brillouin zone
(cf.~Fig.~\ref{fig:firstbz}).
\begin{figure}
\centerline{\includegraphics[width=0.7\textwidth,clip=true]{firstbz.pdf}}
\caption[]{\em{
The First Brillouin Zone. The end points of all vector pairs that
satisfy the Bragg condition $\k\k_0=\G_{hkl}$ lie on the perpendicular
bisector of $\G_{hkl}$. The smallest polyhedron centered at the origin
of the reciprocal lattice and enclosed by perpendicular bisectors of
the G's is called the first Brillouin zone.
}}
\label{fig:firstbz}
\end{figure}
Typically, we choose to include only half of the bounding surface
within the first Brillouin zone, so that it can also be defined as the
set of points which contains only independent states.
From the discussion in chapter 3 and in this chapter, it is also clear that
the reciprocal lattice vectors have some interpretation as momentum. For
example, the Laue condition requires that the change in momentum of the
scatterer be equal to a reciprocal lattice translation vector. The end
points of all vector pairs that satisfy the Bragg condition $\k\k_0=\G_{hkl}$
lie on the perpendicular bisector of $\G_{hkl}$. Thus, the FBZ is also the
set of points which cannot satisfy the Bragg condition.
\subsection{Point Group Symmetry and Density of States}
Two other tricks to reduce the complexity of these sums are
worth mentioning here although they are discussed in detail elswhere.
The first is the use of the point group symmetry of the system.
It is clear from their definition in chapter 3, the reciprocal lattice
vectors have the same point group symmetry as the lattice. As we discussed
in chapter 2, the knowledge of the group elements and corresponding
degeneracies may be used to reduce the sums over $\k$ to the irreducible
wedge within the the First Brillioun zone. For example, for a cubic
system, this wedge is only $1/2^33!$ or $1/48$th of the the FBZ!
The second is to introduce a phonon density of states to reduce the
multidimensional sum over $k$ to a onedimensional integral over energy.
This will be discussed in chapter 5.
\section{Normal Modes and Quantization}
In this section we will derive the equations of motion for the lattice,
determine the canonically conjugate variables (the the sense of Lagrangian
mechanics), and use this information to both first and second quantize
the system.
Any lattice displacement may be expressed as a sum over the eigenvectors
of the dynamical matrix $\D$.
\beq
s_{n,\alpha,i} =\frac{1}{\sqrt{M_\alpha N}}\sum_{\q,s}
Q_s(\q,t) \ep_{\alpha,i}^s(\q)e^{i\q\cdot\r_n}
\eeq
Recall that $\ep_{\alpha,i}^s(\q)$ are distinguished from
$\u_{\alpha,i}^s(\q)$ only in that they are normalized. Also since
$\q+\G$ is equivalent to $\q$, we need sum only over the first Brillouin
zone. Finally we will assume that $Q_s(\q,t)$ contains the harmonic time
dependence and since $s_{n,\alpha,i}$ is real $Q_s^*(\q)=Q_s(\q)$.
We may rewrite both the kinetic and potential energy of the system as
sums over $Q$. For example, the kinetic energy of the lattice
\beqa
T & = & \frac12 \sum_{n,\alpha,i}M_\alpha\lep\dot{s}_{n\alpha,i}\rip^2\\
& = & \frac{1}{2N}\sum_{n,\alpha,i}\sum_{\q,\k,r,s}
\dot{Q}_r(\q)\ep_{\alpha,i}^r(\q)e^{i\q\cdot\r_n}
\dot{Q}_s(\k)\ep_{\alpha,i}^s(\k)e^{i\k\cdot\r_n}
\eeqa
Then as
\beq
\frac{1}{N}\sum_n e^{i(\k+\q)\cdot\r_n}=\delta_{\k,\q}
\;\;\;\mbox{and}\;\;\;
\sum_{\alpha,i} \ep_{\alpha,i}^r\ep_{\alpha,i}^{*s}=\delta_{rs}
\eeq
the kinetic energy may be reduced to
\beq
T=\frac12 \sum_{\q,r}\left\dot{Q}_r(\q)\right^2
\eeq
The potential energy may be rewritten in a similar fashion
\beqa
V&=&\frac12 \sum_{n,m,\alpha,\beta,i,j}
\Phi_{n,\alpha,i}^{m,\beta,j} s_{n,\alpha,i}s_{m,\beta,j} \nonumber\\
&=& \frac12 \sum_{n,m,\alpha,\beta,i,j}
\frac{\Phi_{0,\alpha,i}^{mn,\beta,j}}{N\sqrt{M_\alpha M_\beta}}\nonumber\\
&&\sum_{\q,\k,s,r}
Q_s(\q,t) \ep_{\alpha,i}^s(\q)e^{i\q\cdot\r_n}
Q_r(\k,t) \ep_{\beta,j}^r(\k)e^{i\k\cdot\r_m}
\eeqa
Let $\r_l=\r_m\r_n$
\beqa
V&=&\frac12 \sum_{n,l,\alpha,\beta,i,j}
\frac{\Phi_{0,\alpha,i}^{l,\beta,j}}{N\sqrt{M_\alpha M_\beta}}\nonumber\\
&&\sum_{\q,\k,s,r}
Q_s(\q,t) \ep_{\alpha,i}^s(\q)e^{i\q\cdot\r_n}
Q_r(\k,t) \ep_{\beta,j}^r(\k)e^{i\k\cdot(\r_l+\r_n)}
\eeqa
and sum over $n$ to obtain the delta function $\delta_{\k,\q}$
so that
\beq
V=\frac12 \sum_{l,\alpha,\beta,i,j,s,r}
Q_s(\k)\ep_{\alpha,i}^s(\k)Q_r(\k)\ep_{\beta,j}^r(\k)
\frac{1}{\sqrt{M_\alpha M_\beta}}
\Phi_{0,\alpha,i}^{l,\beta,j}e^{i\k\cdot\r_l}\,.
\eeq
Note that the sum over $l$ on the last three terms yields $\D$,
so that
\beq
V=\frac12 \sum_{l,\alpha,\beta,i,j,s,r} D_{\alpha,i}^{\beta,j}(\k)
Q_s(\k)\ep_{\alpha,i}^s(\k)Q_r(\k)\ep_{\beta,j}^r(\k)\,.
\eeq
Then, since $\sum_{\beta,j}D_{\alpha i}^{\beta j}(\k)\ep_{\beta j}^r(\k)=
\omega^2_r(\k)\ep_{\alpha,i}^r(\k)$ and $\ep_{\alpha,i}^s(\k)=
\ep_{\alpha,i}^{*s}(\k)$,
\beq
V=\frac12 \sum_{\alpha,i,\k,r,s}\ep_{\alpha,i}^r(\k)\ep_{\alpha,i}^{*s}(\k)
\omega_r^2(\k)Q_s^*(\k)Q_r(\k)
\eeq
Finally, since
$\sum_{\alpha,i}\ep_{\alpha,i}^r(\k)\ep_{\alpha,i}^{*s}(\k)=\delta_{r,s} $
\beq
V=\frac12 \sum_{\k,s} \omega_s^2(\k) \leftQ_s(\k)\right^2
\eeq
Thus we may write the Lagrangian of the ionic system as
\beq
L=TV=\frac12 \sum_{\k,s}
\lep
\left\dot{Q}_s(\k)\right^2

\omega_s^2(\k) \leftQ_s(\k)\right^2
\rip\,,
\eeq
where the $Q_s(\k)$ may be regarded as canonical coordinates, and
\beq
P_r^*(\k)=\pde{L}{Q_r(\k)}=\dot{Q}_s^*(\k)
\eeq
(no factor of 1/2 since $Q_s^*(\k)=Q_s(\k)$) are the canonically
conjugate momenta.
The equations of motion are
\beq
\frac{d}{dt}\lep\pde{L}{\dot{Q}^*_s(\k)}\rip\pde{L}{{Q}^*_s(\k)}
\;\;\;\mbox{or}\;\;\;
\ddot{Q}_s(\k)+\om_s^2(\k)Q_s(\k)=0
\eeq
for each $\k,s$. These are the equations of motion for $3rN$ independent
harmonic oscillators. Since going to the Qcoordinates accomplishes the
decoupling of these equations, the $\{Q_s(\k)\}$ are referred to as normal coordinates.
\subsection{Quantization and Second Quantization}
P.A.M. Dirac laid down the rules of quantization, from Classical
HamiltonJacobi classical mechanics to Hamiltonianbased quantum mechanics
following the path (Dirac p.8489):
\begin{enumerate}
\item First, identify the classical canonically conjugate set of variables
$\{q_i,p_i\}$
\item These have Poisson Brackets
\beq
\{\{u,v\}\}=\sum_i\lep\pde{u}{q_i}\pde{v}{p_i}\pde{u}{p_i}\pde{v}{q_i}\rip
\eeq
\beq
\{\{q_i,p_j\}\}=\delta_{i,j}\;\;\; \{\{p_i,p_j\}\}=\{\{q_i,q_j\}\}=0
\eeq
\item Then define the quantum Poisson Bracket (the commutator)
\beq
[u,v]=uvvu=i\hbar\{\{u,v\}\}
\eeq
\item In particular, $[q_i,p_j]=i\hbar\delta_{i,j}$, and
$[q_i,q_j]=[p_i,p_j]=0$.
\end{enumerate}
Thus, following Dirac, we may now quantize the normal coordinates
\beq
[Q_r^*(\k),P_s(\q)]=i\hbar\delta_{\k,\q}\delta_{r,s}
\;\;\;\mbox{where the other commutators vanish}\,.
\label{eq:qcom}
\eeq
Furthermore, since we have a system of $3rN$ uncoupled harmonic
oscillators we may immediately second quantize by introducing
\beqa
a_s(\k)&=&\frac{1}{\sqrt{2\hbar}}\lep\sqrt{\om_s(\k)}Q_s(\k)+
\frac{i}{\sqrt{\om_s(\k)}}P_s(\k)\rip\\
a_s^{\dag}(\k)&=&\frac{1}{\sqrt{2\hbar}}\lep\sqrt{\om_s(\k)}Q_s^*(\k)
\frac{i}{\sqrt{\om_s(\k)}}P_s^*(\k)\rip\,,
\eeqa
or
\beqa
Q_s(\k)&=&\sqrt{\frac{\hbar}{2\om_s(\k)}}\lep a_s(\k)+a_s^{\dag}(\k)\rip\\
P_s(\k)&=&i\sqrt{\frac{\hbar\om_s(\k)}{2}}\lep a_s(\k)a_s^{\dag}(\k)\rip
\eeqa
Where
\beq
[a_s(\k),a_r^{\dag}(\q)]=\delta_{r,s}\delta_{\q,\k}
\;\;\;[a_s(\k),a_r(\q)]=[a_s^{\dag}(\k),a_r^{\dag}(\q)]=0
\label{eq:acom}
\eeq
This transformation $\{Q,P\}\to\{a,a^{\dag}\}$ is canonical, since is
preserves the commutator algebra Eq.~\ref{eq:qcom}, and the Hamiltonian
becomes
\beq
H=\sum_{k,s}\hbar\om_s(\k)\lep a_s^{\dag}(\k)a_s(\k) + \frac12 \rip
\eeq
which is a sum over $3rN$ independent quantum oscillators, each one
referred to as a phonon mode!
The number of phonons in state $(\k,s)$ is given by the operator
\beq
n_s(\k)=a_s^{\dag}(\k)a_s(\k)
\eeq
and $a_s^{\dag}(\k)$ and $a_s(\k)$ create and destroy phonons respectively,
in the state $(\k,s)$
\beqa
a_s^{\dag}(\k)
\left n_s(\k)\right> & = & \sqrt{n_s(\k)+1}\leftn_s(\k)+1\right>\\
a_s(\k)\leftn_s(\k)\right> & = & \sqrt{n_s(\k)}\leftn_s(\k)1\right>
\eeqa
If $\left0\right>$ is the normalized state with no phonons present,
then the state with $\{n_s(\k)\}$ phonons in each state $(\k,s)$ is
\beq
\left\{n_s(\k)\}\right>
=
\leb
\lep\prod_{\k,s}\frac{1}{n_s(\k)!}\rip^{\frac12}
\rib
\prod_{\k,s}\lep a_{s}^{\dag}(\k)\rip^{{n_s(\k)}}\left0\right>
\eeq
Finally the lattice point displacement
\beq
s_{n,\alpha,i}=\frac{1}{\sqrt{M_\alpha N}}\sum_{\q,s}\sqrt{\frac{\hbar}{2\om_s(\q)}}
\lep a_s(\q)+ a_s^{\dag}(\q)\rip \ep^s_{\alpha,i}(\q)e^{i\q\cdot\r_n}
\eeq
will be important in the next section, especially with respect to
zeropoint motion (i.e.\ $\left< s^2 \right>_{T=0}\neq 0$).
\section{Theory of Neutron Scattering}
\begin{figure}[htb]
\centerline{\includegraphics[width=0.75\textwidth,clip=true]{neutron.pdf}}
\caption[]{\em{Neutron Scattering. The De Broglie wavelength of the
neutrons must be roughly the same size as the lattice constants in
order to learn about the lattice structure and its vibrational modes
from the experiment. This dictates the use of thermal neutrons.}}
\label{fig:neutron}
\end{figure}
To ``see'' the lattice with neutrons, we want their De Broglie
wavelength $\lambda=h/p$
\beq
\lambda_{\rm{neutron}} =\frac{0.29\A}{\sqrt{E}}\;\;\;\;
\mbox{E measured in eV}
\eeq
to be of the same length as the intersite distance on the lattice.
This means that their kinetic energy $E\approx \frac12 Mv^2\approx 0.1$eV,
or $E/k_b\approx 1000K$; i.e.\ thermal neutrons.
Since the neutron is chargeless, it only interacts with the
atomic nucleus through a shortranged nuclear interaction
(ignoring any spinspin interaction). The range of this interaction
is ~1 Fermi ($10^{13}$cm.) or about the radius of the atomic nucleus.
Thus
\beq
\lambda\sim\A\gg \mbox{range of the interaction}\sim 10^{13}{\rm{cm.}}
\eeq
Thus the neutron cannot "see" the detailed structure of the
nucleus, and so we may approximate the neutronion interaction
potential as a contact interaction
\beq
V(\r)=\sum_{\r_n} V_n \delta(\r\r_n)
\eeq
i.e., we may ignore the angular dependence of the scattering
factor f.
\subsection{Classical Theory of Neutron Scattering}
Due to the importance of lattice vibrations, which are inherently
quantum in nature, there is a limit to what we can learn from a
classical theory of diffraction. Nevertheless it is useful to
compare the classical result to what we will develop for the
quantum problems.
For the classical problem we will assume that the lattice is elemental
($r=1$) and start with a generalization of the formalism developed in the last
chapter
\beq
I\propto \left\rho(\K,\Omega)\right^2
\eeq
where $\K=\k_0\k$ and $\Omega=\om_0\om$. Furthermore, we take
\beq
\rho(\r(t))\propto\sum_n\delta(\r\r_n(t))
\eeq
where
\beq
\r_n(t)=\r_n+\s_n(t)\;\;\mbox{and}
\;\;\s_n(t)=\frac{1}{\sqrt{M}}\u(\q)e^{i(\q\cdot\r_n\om(\q)t)}
\eeq
describes the harmonic motion of the smode with wavevector q.
\beq
\rho(\K,\Omega)\propto\sum_n\int dt
e^{i\leb\K\cdot(\r_n+\s_n(t))\Omega t\rib}\,.
\eeq
For $\K\sim 2\pi/\A$ and $\s_n(t)\ll \A$ we may expand
\beq
\rho(\K,\Omega)\propto\sum_n\int dt
e^{i\leb\K\cdot(\r_n)\Omega t\rib}\lep 1+i\K\cdot\s_n(t) +\cdots\rip
\eeq
The first term yields a finite contribution only when
\beq
\K=\k_0\k=\G
\;\;\mbox{and}\;\;
\Omega=\omega_0\omega=0
\eeq
which are the familiar Bragg conditions for elastic scattering.
The second term, however, yields something new. It only
yields a finite result when
\beq
\K\pm\q=\k_0\k\pm\q=\G
\;\;\mbox{and}\;\;
\Omega\pm\om_s(\q)=\om_0\om\pm\om_s(\q)=0
\eeq
When multiplied by $\hbar$ , these can be interpreted as conditions for
the conservation of (crystal) momentum and energy when the
scattering event involves the creation (destruction) of a lattice
excitation (phonon). These processes are called Stokes and antistokes
processes, respectively, and are illustrated in Fig.~\ref{fig:stokes}.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.75\textwidth,clip=true]{stokes.pdf}}
\caption[]{\em{Stokes and antistokes processes in inelastic neutron
scattering involving the creation or absorption of a lattice phonon.}}
\label{fig:stokes}
\end{figure}
Clearly, the antiStokes process can only happen at finite temperatures
where real (as opposed to virtual) phonons are excited. Thus, our classical
formalism does not correctly describe the temperature dependence of the
scattering. Several other things are missing in the classical theory,
including:
\begin{itemize}
\item Security in the validity of the result.
\item The effects of zeropoint motion.
\item Correct temperature dependence.
\end{itemize}
\subsection{Quantum Theory of Neutron Scattering}
To address these concerns, we will do a fully quantum calculation.
Several useful references for this calculation include
\begin{itemize}
\item Ashcroft and Mermin, Appendix N, p.\ 790)
\item Callaway, p. 36.
\item Hook and Hall (for experiment) Ch.~12 p.342
\end{itemize}
We will imagine that the scattering shown in Fig.~\ref{fig:quantum_neutron}
occurs in a box of volume $V$.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.85\textwidth,clip=true]{quantum_neutron.pdf}}
\caption[]{\em{The initial (left) and final (right) states of the
neutron and lattice during a scattering event. The initial system state
is given by $\Psi_0=\phi_0\psi_0$, with energy $\ep_0=E_0+\hbar\om_0$
where $\om_0=k_0^2/2M$. The final system state is given by
$\Psi_f=\phi_f\psi_f$, with energy $\ep_f=E_f+\hbar\om_f$
where $\om_f=k_f^2/2M$. }}
\label{fig:quantum_neutron}
\end{figure}
The momentum transfer, from the neutron to the lattice is $\K=\k_0\k_f$ and
the energy transfer which is finite for inelastic scattering is
$\hbar\Omega=\hbar\lep\om_0\om_f\rip$. Again we will take the neutronlattice
interaction to be local
\beq
V(\r)=\sum_{r_n} V(\r\r_n)=\frac{1}{N}\sum_{\q,n}V(\q)e^{i\q\cdot(\r\r_n)}
=\int\frac{d^3q}{V} V_0\sum_n e^{i\q\cdot(\r\r_n)}
\eeq
where the locality of the interaction ($V(\r\r_n)\propto\delta(\r\r_n)$)
indicates that $V(\q)=V(0)=V_0$. Consistent with Aschcroft and Mermin, we
will take
\beq
V(\r) =\frac{2\pi\hbar^2 a}{M} \frac{1}{V}\sum_{\r_n}
\int d^3qe^{i\q\cdot(\r\r_n)}
\label{eq:ionneutron}
\eeq
where $a$ is the scattering length, and $V_0=\frac{2\pi\hbar^2 a}{M}$ is
chosen such that the total cross section $\sigma =4\pi a^2$.
To formulate our quantum theory, we will use Fermi's golden rule
for time dependent perturbation theory. (This is fully
equivalent to the lowestorder Born approximation). The
probability per unit time for a neutron to scatter from state $\k_0$ to
$\k_f$ is given by
\beqa
P&=&\frac{2\pi}{\hbar}\sum_f\delta(\ep_0\ep_f)\left\left<\Psi_0\leftV\right
\Psi_f\right>\right^2\\
&=& \frac{2\pi}{\hbar}\sum_f\delta(E_0+\hbar\om_0E_f\hbar\om_f)\nonumber\\
&&\left\frac{1}{V}\int d^3r e^{i(\k\k_0)\cdot\r}
\left<\phi_0\leftV(\r)\right\phi_f\right>\right^2
\eeqa
If we now substitute in the ionneutron potential Eq.~\ref{eq:ionneutron},
then the integral over $\r$ will yield a delta function
$V\delta(\q+\k\k_0)$ which allows the $\q$ integral to be evaluated
\beq
P=a^2\frac{(2\pi\hbar)^3}{(MV)^2}\sum_f\delta(E_0E_f+\hbar\Omega)
\left\sum_{\r_n}
\left<\phi_0\lefte^{i\K\cdot\r_n}\right\phi_f\right>\right^2
\eeq
Now, before proceeding to a calculation of the differential cross
section $\frac{d\sigma}{d\Omega d\E}$ we must be able
to convert this probability (rate) for eigenstates into a flux of neutrons
of energy $E$ and momentum $\p$.
A differential volume element of momentum space $d^3p$ contains
$Vd^3p/(2\pi\hbar)^3$ neutron states. While this is a natural consequence
of the uncertainty principle, it is useful to show this in a more
quantitative sense: Imagine a cubic volume $V = L^3$ with periodic
boundary conditions so that for any state $\Psi$ in $V$,
\beq
\Psi(x+L,y,z)=\Psi(x,y,z)
\eeq
If we write $\Psi(\r)=\frac{1}{N}\sum_{\q}e^{i\q\cdot\r}\Psi(\q)$,
then it must be
that
\beq
q_x L =2\pi m\;\;\;\mbox{where $m$ is an integer}
\eeq
with similar relations for the $y$ and $z$ components.
So for each volume element of qspace $\lep\frac{2\pi}{L}\rip^3$ there is
one such state.
In terms of states $\p=\hbar\q$, the volume of a state is
$(2\pi\hbar/L)^3$. Thus $d^3p$
contains $Vd^3p/(2\pi\hbar)^3$ states.
The incident neutron flux of states (velocity times density) is
\beq
\j=\frac{\hbar\k_0}{M}\left\Psi_0\right^2=
\frac{\hbar\k_0}{M}\left\frac{1}{\sqrt{V}}e^{i\k_0\cdot\r}\right^2
=\frac{\hbar\k_0}{MV}
\eeq
Then since the number of neutrons is conserved
\beq
j\frac{d\sigma}{dE d\Omega}dE d\Omega =
\frac{\hbar k_0}{MV}\frac{d\sigma}{dE d\Omega}dE d\Omega=
PV\frac{d^3p}{(2\pi\hbar)^3} = PV\frac{p^2 dp d\Omega}{(2\pi\hbar)^3}
\eeq
And for thermal (nonrelativistic) neutrons $E=p^2/2M$, so $dE=pdp/M$, and
\beq
\frac{\hbar k_0}{MV}\frac{d\sigma}{dE d\Omega}dE d\Omega=
PV\frac{\hbar kM dE d\Omega}{(2\pi\hbar)^3}
\eeq
or
\beq
\frac{d\sigma}{dE d\Omega}=P\frac{k}{k_0}\frac{(MV)^2}{(2\pi\hbar)^3}\,.
\eeq
Substituting in the previous result for $P$
\beq
\frac{d\sigma}{dE d\Omega}=\frac{k}{k_0}\frac{(MV)^2}{(2\pi\hbar)^3}
a^2\frac{(2\pi\hbar)^3}{(MV)^2}\sum_f\delta(E_0E_f+\hbar\Omega)
\left\sum_{\r_n}
\left<\phi_0\lefte^{i\K\cdot\r_n}\right\phi_f\right>\right^2
\eeq
or
\beq
\frac{d\sigma}{dE d\Omega}=\frac{k}{k_0}\frac{Na^2}{\hbar} S(\K,\Omega)
\eeq
where
\beq
S(\K,\Omega)=\frac{1}{N}\sum_f\delta(E_0E_f+\hbar\Omega)
\left\sum_{\r_n}
\left<\phi_0\lefte^{i\K\cdot\r_n}\right\phi_f\right>\right^2\,.
\eeq
We may deal with the Dirac delta function by substituting
\beq
\delta(\Omega)=\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\Omega t}\,.
\eeq
so that
\beqa
S(\K,\Omega)&=&\frac{1}{N}\sum_f
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep (E_0E_f)/\hbar+\Omega\rip t}
\nonumber\\
&&\sum_{\r_n,\r_m}
\left\left<\phi_0\lefte^{i\K\cdot\r_n}\right\phi_f\right>\right
\left\left<\phi_f\lefte^{i\K\cdot\r_m}\right\phi_0\right>\right\,.
\eeqa
then as
\beq
e^{iHt/\hbar}\left\phi_l\right>=e^{iE_lt/\hbar}\left\phi_l\right>
\eeq
where $H$ is the lattice Hamiltonian, we can write this as
\beqa
S(\K,\Omega)&=&\frac{1}{N}\sum_f
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\Omega t}
\sum_{\r_n,\r_m}
\left<\phi_0\lefte^{iHt/\hbar}e^{i\K\cdot\r_n}e^{iHt/\hbar}\right\phi_f
\right>\nonumber\\
&&\left<\phi_f\lefte^{i\K\cdot\r_m}\right\phi_0\right>\,,
\eeqa
and the argument in the first expectation value is the timedependent
operator $e^{i\K\cdot\r_n}$ in the Heisenberg representation
\beq
e^{i\K\cdot\r_n(t)}=e^{iHt/\hbar}e^{i\K\cdot\r_n}e^{iHt/\hbar}\,.
\eeq
Thus,
\beqa
S(\K,\Omega)&=&\frac{1}{N}\sum_f
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\Omega t}
\sum_{\r_n,\r_m}
\left<\phi_0\lefte^{i\K\cdot\r_n(t)}\right\phi_f\right>
\left<\phi_f\lefte^{i\K\cdot\r_m}\right\phi_0\right>\nonumber\\
&=&\frac{1}{N}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\Omega t}
\sum_{\r_n,\r_m}
\left<\phi_0\lefte^{i\K\cdot\r_n(t)}e^{i\K\cdot\r_m}\right\phi_0\right>\,.
\eeqa
Now since $\r_n(t)=\r_n+\s_n(t)$ (with $\r_n$ time independent),
\beq
S(\K,\Omega)=\frac{1}{N} \sum_{n,m}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep \K\cdot(\r_n\r_m)+\Omega t\rip}
\left<\phi_0\lefte^{i\K\cdot\s_n(t)}e^{i\K\cdot\s_m}\right\phi_0\right>\,.
\eeq
This formula is correct at zero temperature. In order to describe finite
$T$ effects (ie., antistokes processes involving phonon absorption) we
must introduce a thermal average over all states
\beq
\left<\phi_0\leftA\right\phi_0\right> \to
\left =
\sum_le^{\beta E_l}
\left<\phi_l\leftA\right\phi_l\right>/\sum_le^{\beta E_l}\,.
\eeq
With this substitution,
\beq
S(\K,\Omega)=\frac{1}{N} \sum_{n,m}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep \K\cdot(\r_n\r_m)+\Omega t\rip}
\left< e^{i\K\cdot\s_n(t)}e^{i\K\cdot\s_m}\right>\,.
\eeq
and $S(\K,\Omega)$ is called the dynamical structure factor.
\subsubsection{The DebyeWaller Factor}
To simplify this relation further, recall that the
exponentiated operators within the brackets are linear functions
of the creation and annihilation operators $a^{\dag}$ and $a$.
\beq
\s_{n,\alpha}(t)=\frac{1}{\sqrt{M_\alpha N}}
\sum_{\q,s}\sqrt{\frac{\hbar}{2\om_s(\q)}} {\bf{\ep}}^s_{\alpha}(\q)
\lep a_s(\q)(t)+ a_s^{\dag}(\q)(t)\rip e^{i\q\cdot\r_n}
\eeq
So that, in particular
$\left<\s_{n,\alpha,i}(t)\right>= \left=0$.
Then let $A=i\K\cdot \s_{n,\alpha,i}(t)$ and $B=i\K\cdot \s_{m,\alpha,i}(0)$
and suppose that the expectation values
of $A$ and $B$ are small. Then
\beqa
\left< e^A e^B \right >
&=& \left<(1+ A +\frac12 A^2+\cdots)
(1+ B +\frac12 B^2+\cdots)\right>\nonumber\\
&\approx& \left<1+ A + B + AB +\frac12 A^2+\frac12 B^2+\cdots \right >\nonumber\\
&\approx& 1+ \frac12 \left<2AB + A^2+ B^2 \right >+\cdots\nonumber\\
&\approx& e^{\frac12 \left<2AB + A^2+ B^2 \right >}
\eeqa
This relation is in fact true to all orders, as long as $A$ and $B$
are linear functions of $a^{\dag}$ and $a$ . (c.f.\ \AM{792}, Callaway pp.\
4148).
Thus
\beq
\left< e^{i\K\cdot\s_n(t)}e^{i\K\cdot\s_m}\right>=
e^{\frac12 \left< \lep\K\cdot\s_n(t)\rip^2\right>}
e^{\frac12 \left< \lep\K\cdot\s_m\rip^2\right>}
e^{\left< \K\cdot\s_n(t)\K\cdot\s_m\right>}\,.
\eeq
Since the Hamiltonian has no time dependence, and the lattice is
invariant under translations $r_n$
\beq
\left< e^{i\K\cdot\s_n(t)}e^{i\K\cdot\s_m}\right>=
e^{\left< \lep\K\cdot\s_n\rip^2\right>}
e^{\left< \K\cdot\s_{nm}(t)\K\cdot\s_0\right>}\,,
\eeq
where the first term is called the DebyeWaller factor $e^{2W}$.
\beq
e^{2W}=e^{\left<\lep \K\cdot\s_n\rip^2\right>}\,.
\eeq
Thus letting $l=nm$
\beq
S(\K,\Omega)= e^{2W}\sum_{l}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep \K\cdot\r_l+\Omega t\rip}
e^{\left< \K\cdot\s_{l}(t)\K\cdot\s_0\right>}\,.
\eeq
Here the DebyeWaller factor contains much of the crucial quantum
physics. It is finite, even at $T=0$ due to zeropoint
fluctuations, and since $\left< \K\cdot\s_n\right>^2$ will increase with
temperature, the total strength
of the Bragg peaks will diminish with increasing $T$. However, as
long as a crystal has longranged order, it will remain finite.
\subsubsection{Zerophonon Elastic Scattering}
One may disentangle the elastic and inelastic processes by expanding the
exponential in the equation above.
\beq
e^{\left< \K\cdot\s_{l}(t)\K\cdot\s_0\right>}
=\sum_m\frac{1}{m!} \lep \left< \K\cdot\s_{l}(t)\K\cdot\s_0\right>\rip^m
\eeq
If we approximate the exponential by 1, ie.\ take only the first, $m=0$
term, then
\beq
S_0(\K,\Omega)= e^{2W}\sum_{l}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep \K\cdot\r_l\Omega t\rip}
\,.
\eeq
And we recover the lowest order classical result (modified by the
DebyeWaller factor) which gives us the Bragg conditions that
$S_0(\K,\Omega)$ is only finite when $\K=\G$ and $\Omega=\om_0\om_f=0$.
\beq
S_0(\K,\Omega)= e^{2W} \delta(\Omega)N\sum_\G \delta_{\K,\G}
\,,
\eeq
\beq
\frac{d\sigma_0}{dE d\Omega}=\frac{k}{k_0}\frac{Na^2}{\hbar}
e^{2W} \delta(\Omega)N\sum_\G \delta_{\K,\G}
\eeq
However, now the scattering intensity is reduced by the DebyeWaller
factor $e^{2W}$, which accounts for zeropoint motion and thermal
fluctuations.
\subsubsection{OnePhonon Inelastic Scattering}
When $m=1$, then the scattering involves either the absorption
or creation of a phonon. To evaluate
\beq
S_1(\K,\Omega)= e^{2W}\sum_{l}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep \K\cdot\r_l+\Omega t\rip}
\left< \K\cdot\s_{l}(t)\K\cdot\s_0(0)\right>\,.
\eeq
we need
\beq
\s_{n,\alpha}(t)=\frac{1}{\sqrt{M_\alpha N}}
\sum_{\q,s}\sqrt{\frac{\hbar}{2\om_s(\q)}} {\bf{\ep}}^s_{\alpha}(\q)
\lep a_s(\q,t)+ a_s^{\dag}(\q,t)\rip e^{i\q\cdot\r_n}
\eeq
in the Heisenberg representation, and therefore we need,
\beqa
a(\q,t) &=& e^{iHt/\hbar} a(\q) e^{iHt/\hbar} \nonumber \\
&=& e^{i(\om(\q) t a^{\dag}(\q) a(\q))} a(\q)
e^{i(\om(\q) t a^{\dag}(\q) a(\q)} \nonumber\\
&=& a(\q) e^{i\lep\om(\q) t (a^{\dag}(\q) a(\q)1)\rip}
e^{i(\om(\q) t a^{\dag}(\q) a(\q)} \nonumber\\
&=& a(\q) e^{i\om(\q) t}
\eeqa
where we have used the fact that $(a^{\dag}a)^na=(a^{\dag}a)^{n1}a^{\dag}aa
=(a^{\dag}a)^{n1}(aa^{\dag}1)a=(a^{\dag}a)^{n1}a(a^{\dag}a1)=
a(a^{\dag}a1)^n$.
Similarly $a^{\dag}(\q,t)=a^{\dag}(\q) e^{i\om(\q) t}$. Thus,
\beq
\s_{n,\alpha}(t)=\frac{1}{\sqrt{M_\alpha N}}\sum_{\q,s}
\frac{\sqrt{\hbar}e^{i\q\cdot\r_n}}{\sqrt{2\om_s(\q)}} {\bf{\ep}}^s_{\alpha}(\q)
\lep a_s(\q)e^{i\om_s(\q)t}+ a_s^{\dag}(\q)e^{i\om_s(\q)t}\rip
\eeq
and
\beq
\s_{0,\alpha}(0)=\frac{1}{\sqrt{M_\alpha N}}
\sum_{\p,r}\sqrt{\frac{\hbar}{2\om_r(\p)}} {\bf{\ep}}^r_{\alpha}(\p)
\lep a_r(\p)+ a_r^{\dag}(\p)\rip
\eeq
Recall, we want to evaluate
\beq
S_1(\K,\Omega)= e^{2W}\sum_{l}
\int_{\infty}^\infty \frac{dt}{2\pi} e^{i\lep \K\cdot\r_l+\Omega t\rip}
\left< \K\cdot\s_{l}(t)\K\cdot\s_0(0)\right>\,.
\eeq
Clearly, the only terms which survive in
$\left< \K\cdot\s_{l}(t)\K\cdot\s_0(0)\right>$ are those with
$r=s$ and $\p=\q$. Furthermore, the sum over $l$ yields a delta
function $N\sum_{\G}\delta_{\K+\q,\G}$. Then as $\ep(\G\k)=\ep(\k)
=\ep^*(\k)$, and $\om(\G\k)=\om(\k)$,
\beqa
S_1(\K,\Omega)&=& e^{2W}\int_{\infty}^\infty \frac{dt}{2\pi}
e^{i \Omega t}\sum_{\q,\G,s}
\frac{\hbar\left\K\cdot\ep(\K)\right^2}{2\om_s(\q)M}
\delta_{\K+\q,\G}\\
&&\leb
e^{i\om_s(\q)t}\left< a_s(\K)a_s^{\dag}(\K)\right> +
e^{i\om_s(\q)t}\left< a_s^{\dag}(\K)a_s(\K)\right>
\rib\nonumber
\eeqa
The occupancy of each mode $n(\q)$ is given by the Bose factor
\beq
\left< n(\q) \right> =\frac{1}{e^{\beta\om(\q)}1}
\eeq
So, finally
\beqa
S_1(\K,\Omega)&=&
e^{2W} \sum_s \frac{\hbar}{2M\om_s(\K)}
\left\K\cdot{\bf{\ep}}^s(\K)\right^2\\
&&\leb(1+n_s(\K)) \delta(\Omega+\om_s(\K)) +
n_s(\K)\delta(\Omega+\om_s(\K))\rib
\,.\nonumber
\eeqa
For the first term, we get a contribution only when
$\Om\om_s(\K)=\om_0\om_f\om_s(\K)=0$; ie., the final energy of the
neutron is smaller than the initial energy. The energy is lost in the
creation of a phonon. Note that this can happen at any temperature,
since $(1+n_s(\K))\neq 0$ at any $T$. The second term is only finite when
$\Om+\om_s(\K)=\om_0\om_f+\om_s(\K)=0$; ie., the final energy of the
neutron is larger than the initial energy. The additional energy comes
from the absorption of a phonon. Thus phonon absorption is only allowed
at finite temperatures, and in fact, the factor $n_s(\K)= 0$ at zero
temperature. These terms correspond to the Stokes and antiStokes
processes, respectively, illustrated in Fig.~\ref{fig:stokes2}.
\begin{figure}[htb]
\centerline{\includegraphics[width=0.8\textwidth,clip=true]{stokes2.pdf}}
\caption[]{\em{Stokes and antistokes processes in inelastic neutron
scattering involving the creation or absorption of a lattice phonon.
The antistokes process can only occur at finiteT, when $n_s(\K)\neq 0$.}}
\label{fig:stokes2}
\end{figure}
If we were to continue our expansion of the exponential to
larger values of $m$, we would find multiplephonon scattering processes.
However, these terms are usually of minimal contribution to the total
cross section, due to the fact that the average ionic excursion $s$
is small, and are usually neglected.
\edo