ViewVC Help
View File | Revision Log | Show Annotations | View Changeset | Root Listing
root/group/trunk/multipole/dielectric.tex
Revision: 4353
Committed: Tue Sep 15 15:56:07 2015 UTC (9 years, 10 months ago) by mlamichh
Content type: application/x-tex
File size: 40716 byte(s)
Log Message:
Added new tex file for dielectric.

File Contents

# User Rev Content
1 mlamichh 4353 \documentclass[%
2     aip,
3     jcp,
4     amsmath,amssymb,
5     preprint,%
6     % reprint,%
7     %author-year,%
8     %author-numerical,%
9     ]{revtex4-1}
10    
11     \usepackage{graphicx}% Include figure files
12     \usepackage{dcolumn}% Align table columns on decimal point
13     \usepackage{multirow}
14     \usepackage{bm}% bold math
15     \usepackage{natbib}
16     \usepackage{times}
17     \usepackage{mathptmx}
18     \usepackage[version=3]{mhchem} % this is a great package for formatting chemical reactions
19     \usepackage{url}
20     \usepackage{braket}
21    
22     \begin{document}
23    
24     \title{Real space electrostatics for multipoles. III. Dielectric Properties}
25    
26     \author{Madan Lamichhane}
27     \affiliation{Department of Physics, University
28     of Notre Dame, Notre Dame, IN 46556}
29     \author{Thomas Parsons}
30     \affiliation{Department of Chemistry and Biochemistry, University
31     of Notre Dame, Notre Dame, IN 46556}
32     \author{Kathie E. Newman}
33     \affiliation{Department of Physics, University
34     of Notre Dame, Notre Dame, IN 46556}
35     \author{J. Daniel Gezelter}
36     \email{gezelter@nd.edu.}
37     \affiliation{Department of Chemistry and Biochemistry, University
38     of Notre Dame, Notre Dame, IN 46556}
39    
40     \date{\today}% It is always \today, today,
41     % but any date may be explicitly specified
42    
43     \begin{abstract}
44     Note: This manuscript is a work in progress.
45    
46     We report on the dielectric properties of the shifted potential
47     (SP), gradient shifted force (GSF), and Taylor shifted force (TSF)
48     real-space methods for multipole interactions that were developed in
49     the first two papers in this series. We find that some subtlety is
50     required for computing dielectric properties with the real-space
51     methods, particularly when using the common fluctuation formulae.
52     Three distinct methods for computing the dielectric constant are
53     investigated, including the standard fluctuation formulae,
54     potentials of mean force between solvated ions, and direct
55     measurement of linear solvent polarization in response to applied
56     fields and field gradients.
57     \end{abstract}
58    
59     \maketitle
60    
61     \section{Introduction}
62    
63     Over the past several years, there has been increasing interest in
64     pairwise methods for correcting electrostatic interactions in computer
65     simulations of condensed molecular
66     systems.\cite{Wolf99,Zahn02,Kast03,Beckd.A.C._Bi0486381,Ma05,Fennell06}
67     These techniques were developed from the observations and efforts of
68     Wolf {\it et al.} and their work towards an $\mathcal{O}(N)$
69     Coulombic sum.\cite{Wolf99} Wolf's method of cutoff neutralization is
70     able to obtain excellent agreement with Madelung energies in ionic
71     crystals.\cite{Wolf99} One of the most difficult tests of any new
72     electrostatic method is the fidelity with which that method can
73     reproduce the bulk-phase polarizability or equivalently, the
74     dielectric properties of a fluid. Of particular interest is the
75     static dielectric constant, $\epsilon$. Using the Ewald sum under
76     tin-foil boundary conditions, $\epsilon$ can be calculated for systems
77     of non-polarizable substances via
78     \begin{equation}
79     \epsilon = 1 + \frac{\langle M^2\rangle}{3\epsilon_0k_\textrm{B}TV},
80     \label{eq:staticDielectric}
81     \end{equation}
82     where $\epsilon_0$ is the permittivity of free space and $\langle
83     M^2\rangle$ is the fluctuation of the system dipole
84     moment.\cite{Allen:1989fp} The numerator in the fractional term in equation
85     (\ref{eq:staticDielectric}) is the fluctuation of the simulation-box
86     dipole moment, identical to the quantity calculated in the
87     finite-system Kirkwood $g$ factor ($G_k$):
88     \begin{equation}
89     G_k = \frac{\langle M^2\rangle}{N\mu^2},
90     \label{eq:KirkwoodFactor}
91     \end{equation}
92     where $\mu$ is the dipole moment of a single molecule of the
93     homogeneous system.\cite{Neumann:1983mz,Neumann:1983yq,Neumann84,Neumann85} The
94     fluctuation term in both equation (\ref{eq:staticDielectric}) and
95     (\ref{eq:KirkwoodFactor}) is calculated as follows,
96     \begin{equation}
97     \begin{split}
98     \langle M^2\rangle &= \langle\bm{M}\cdot\bm{M}\rangle
99     - \langle\bm{M}\rangle\cdot\langle\bm{M}\rangle \\
100     &= \langle M_x^2+M_y^2+M_z^2\rangle
101     - (\langle M_x\rangle^2 + \langle M_y\rangle^2
102     + \langle M_z\rangle^2).
103     \end{split}
104     \label{eq:fluctBoxDipole}
105     \end{equation}
106     This fluctuation term can be accumulated during a simulation; however,
107     it converges rather slowly, thus requiring multi-nanosecond simulation
108     times.\cite{Horn04} In the case of tin-foil boundary conditions, the
109     dielectric/surface term of the Ewald summation is equal to zero.
110    
111     \section{Dielectric constants for real-space electrostatics}
112    
113     The new real-space methods require some care when computing dielectric
114     properties because the interaction between molecular dipoles depends
115     on the level of approximation being utilized. Consider the cases of
116     Stockmayer (dipolar) soft spheres that are represented either by two
117     closely-spaced point charges or by a single point dipole (see
118     Fig. \ref{fig:stockmayer}).
119     \begin{figure}
120     \includegraphics[width=3in]{DielectricFigure}
121     \caption{With the real-space electrostatic methods, the effective
122     dipole tensor, $\mathbf{T}$, governing interactions between
123     molecular dipoles is not the same for charge-charge interactions as
124     for point dipoles.}
125     \label{fig:stockmayer}
126     \end{figure}
127     In the case where point charges are interacting via an electrostatic
128     kernel, $v(r)$, the effective {\it molecular} dipole tensor,
129     $\mathbf{T}$ is obtained from two successive applications of the
130     gradient operator to the electrostatic kernel,
131     \begin{equation}
132     \mathbf{T}_{\alpha \beta}(r) = \nabla_\alpha \nabla_\beta \left(v(r)\right) = \delta_{\alpha \beta}
133     \left(\frac{1}{r} v^\prime(r) \right) + \frac{r_{\alpha}
134     r_{\beta}}{r^2} \left( v^{\prime \prime}(r) - \frac{1}{r}
135     v^{\prime}(r) \right)
136     \end{equation}
137     where $v(r)$ may be either the bare kernel ($1/r$) or one of the
138     modified (Wolf or DSF) kernels. This tensor describes the effective
139     interaction between molecular dipoles ($\mathbf{D}$) in Gaussian
140     units as $-\mathbf{D} \cdot \mathbf{T} \cdot \mathbf{D}$.
141    
142     When utilizing the new real-space methods for point dipoles, the
143     tensor is explicitly constructed,
144     \begin{equation}
145     \mathbf{T}_{\alpha \beta}(r) = \delta_{\alpha \beta} v_{21}(r) +
146     \frac{r_{\alpha} r_{\beta}}{r^2} v_{22}(r)
147     \end{equation}
148     where the functions $v_{21}(r)$ and $v_{22}(r)$ depend on the level of
149     the approximation. Although the Taylor-shifted (TSF) and
150     gradient-shifted (GSF) models produce to the same $v(r)$ function for
151     point charges, they have distinct forms for the dipole-dipole
152     interactions.
153    
154     Neumann and Steinhauser showed~\cite{Neumann:1983mz,Neumann:1983yq}
155     that the relative dielectric permittivity $\epsilon$ is given by the
156     general fluctuation formula,
157     \begin{equation}
158     \frac{\epsilon - 1}{\epsilon +2} = \frac{4 \pi}{3} \frac{\left<M^2\right>}{3 V k_B
159     T} \left( 1 - \frac{3}{4 \pi} \frac{\epsilon -1}{\epsilon + 2}
160     \tilde{T}(0) \right)
161     \end{equation}
162     where $\left<M^2\right>$ is the mean fluctuation of the square of the
163     net dipole moment of the simulation cell, $V$ is the volume of the
164     cell, and $\tilde{T}(0) = \tilde{T̃}_{xx}(0) = \tilde{T̃}_{yy}(0) =
165     \tilde{T̃}_{zz}(0)$ is the $\mathbf{k} = 0$ limit of Fourier transform
166     of the diagonal term of the effective molecular dipolar tensor,
167     \begin{equation}
168     \tilde{T}(0) = \int_V \mathbf{T}(\mathbf{r}) d\mathbf{r}
169     \end{equation}
170     where the integral is carried out over the relevant geometry for the
171     interaction. For the real-space methods, the integration can be done
172     easily up to the imposed cutoff radius, while for the Ewald sum, the
173     results also depend on details of the $k$-space
174     calculation.\cite{Neumann:1983yq}
175    
176     For molecules composed of point charges interacting via the bare
177     $(1/r)$ kernel, the simple cutoff (SC) and minimum image (MI)
178     approaches give $\tilde{T}(0) = 0$ which means that the
179     Clausius-Mosotti equation,
180     \begin{equation}
181     \frac{4\pi}{3}\frac{\left< M^2 \right>}{3 V k_B T} = \frac{\epsilon
182     -1}{\epsilon+2}
183     \end{equation}
184     is the relevant fluctuation formula for the relative permittivity.
185    
186     Zahn {\it et al.}\cite{Zahn02} showed that for the damped shifted
187     force charge-charge kernel, $\tilde{T}(0) = 4 \pi / 3$. This was later
188     generalized by Izvekov {\it et al.} for all point-charge kernels which
189     have forces that go to zero at a cutoff radius and which maintain a
190     pole of first order at $r=0$.\cite{Izvekov:2008wo} When $\tilde{T}(0)
191     = 4 \pi/ 3$, the expression for the dielectric constant reduces to the
192     widely-used {\it conducting boundary} formula,
193     \begin{equation}
194     \frac{4\pi}{3}\frac{\left< M^2 \right>}{3 V k_B T} = \frac{\epsilon -
195     1}{3}
196     \end{equation}
197     It is convenient to define a quantity $A = \frac{3}{4 \pi}
198     \tilde{T}(0)$, and to combine all of the fluctuation formulae
199     together:
200     \begin{equation}
201     \frac{4 \pi}{3} \frac{\left< M^2 \right>}{3 V k_B T} =
202     \left\{\begin{array}{ll}
203     \frac{\epsilon-1}{\epsilon+2} \left[1- \frac{\epsilon-1}{\epsilon+2}
204     A\right]^{-1} & \mathrm{General~Case} \\
205     \frac{\epsilon-1}{\epsilon+2} & A \rightarrow 0 \mathrm{~limit} \\
206     \frac{\epsilon-1}{3} & A \rightarrow 1 \mathrm{~limit}
207     \end{array}\right.
208     \end{equation}
209    
210     The Clausius-Mossotti $(A\rightarrow 0)$ approach is subject to noise
211     and error magnification,\cite{Allen:1989fp} and the conducting
212     boundary approach $(A \rightarrow 1)$ has become widely used as a
213     result. If the electrostatic method being used has $A < 1$, the
214     relative dielectric permittivity $\epsilon$ can be estimated once the
215     conducting boundary fluctuation result has been found,
216     \begin{equation}
217     \epsilon = \frac{(A+2) (\epsilon_\text{CB}-1)+3} {(A-1) (\epsilon_\text{CB}-1)+3}
218     \end{equation}
219     Note that this expression becomes quite numerically sensitive when the
220     value of $A$ deviates significantly from 1.
221    
222     We have derived a set of expressions for the value of $A$ for the new
223     real space methods that are shown in Table \ref{tab:A}.
224    
225     \begin{table}
226     \caption{Expressions for the dielectric correction factor ($A$) for the
227     real-space electrostatic methods in terms of the damping parameter
228     ($\alpha$) and the cutoff radius ($r_c$). The Ewald-Kornfeld result
229     derived in Refs. \onlinecite{Adams:1980rt,Adams:1981fr,Neumann:1983yq} is shown for comparison using the Ewald
230     convergence parameter ($\kappa$) and the real-space cutoff value ($r_c$). }
231     \label{tab:A}
232     \resizebox{\columnwidth}{!}{%
233     \begin{tabular}{l|c|c|c|c|c|}
234     & \multicolumn{3}{c|}{damped} & \multicolumn{2}{c|}{undamped} \\ \cline{2-4} \cline{5-6}
235     Method & $A_\mathrm{charges}$ & $A_\mathrm{dipoles}$ & $A_\mathrm{quad} (length^{-2})$ & $A_\mathrm{charges}$ & $A_\mathrm{dipoles}$ \\
236     \hline
237     Spherical Cutoff (SC) & $\mathrm{erf}(r_c \alpha) - \frac{2 \alpha r_c}{\sqrt{\pi}} e^{-\alpha^2 r_c^2}$ & 0 &$-\frac{8{\alpha}^5{r_c}^3 }{5\sqrt{\pi}}e^{-{\alpha}^2{r_c}^2 }$ &0 &0\\
238     Shifted Potental (SP) & $ \mathrm{erf}(r_c \alpha) - \frac{2 \alpha r_c}{\sqrt{\pi}} e^{-\alpha^2 r_c^2}$ & $\mathrm{erf}(\alpha r_c)-\frac{2 \alpha r_c }{\sqrt{\pi }} \left(1 + \frac{2 \alpha^2 r_c^2}{3} \right) e^{-\alpha ^2 r_c^2} $&$-\frac{16{\alpha}^7{r_c}^5 }{15\sqrt{\pi}}e^{-{\alpha}^2{r_c}^2 }$ &0 &0\\
239     Gradient-shifted (GSF) & 1 & $\mathrm{erf}(\alpha r_c)-\frac{2 \alpha r_c}{\sqrt{\pi}} \left(1 + \frac{2 \alpha^2 r_c^2}{3} + \frac{\alpha^4 r_c^4}{3}\right)e^{-\alpha^2 r_c^2} $ &$-\frac{4{\alpha}^7{r_c}^5 }{15\sqrt{\pi}}(-1+2\alpha ^2 r_c^2)e^{-\alpha^2 r_c^2}$ &1 &0\\
240     Taylor-shifted (TSF) & 1 & 1 &$\frac{6\;\mathrm{erf}(r_c \alpha)}{{r_c}^2} + \frac{4{\alpha}{r_c}}{15\sqrt{\pi}{r_c}^2}\left(45 + 30\alpha ^2 {r_c}^2 + 12\alpha^4 {r_c}^4 + 3\alpha^6 {r_c}^6 + 2 \alpha^8 {r_c}^8\right)e^{-\alpha^2 r_c^2}$ &1 & 1\\
241     Ewald-Kornfeld (EK) & $\mathrm{erf}(r_c \kappa) - \frac{2 \kappa r_c}{\sqrt{\pi}} e^{-\kappa^2 r_c^2}$ & $\mathrm{erf}(r_c \kappa) - \frac{2 \kappa r_c}{\sqrt{\pi}} e^{-\kappa^2 r_c^2}$ &- & - & - \\\hline
242     \end{tabular}%
243     }
244     \end{table}
245    
246     \section{Boltzman average of the dipole polarization density}
247     Consider a dipolar fluid with polarization density $\mathbf{P}(r)$ at
248     equlibrium at temperature $T$. The average of polarization density is
249     given by,
250     \begin{equation}
251     \left<\mathbf{P}(r)\right> = \frac{\int\mathbf{P}(r) e^{-H/k_B T} d
252     \Gamma^N}{\int e^{-H/k_B T} d \Gamma^N}
253     \end{equation}
254     where $d \Gamma^N = dr_1...dr_N dp_1...dp_N$ is the volume element in
255     phase space, and $H$ is the Hamiltonian of the $N$-particle system in
256     the absence of applied external field. The Hamiltonian of the system
257     is given by,
258     \begin{equation}
259     H = \frac{1}{2}\sum_{i=1}^N m_i {\dot{\mathbf{r}}_i}^2 + \sum_{i<j} U(\mathbf{r}_i - \mathbf{r}_j)
260     \end{equation}
261     where $m_i$ is mass of site $i$, and $U(\mathbf{r}_i - \mathbf{r}_j)$
262     is the interaction potential between sites $i$ and $j$.
263    
264     Let $E'$ be the applied external electric field. The Hamiltonian for
265     the polarized system then becomes
266     \begin{equation}
267     H' = H - {M}_\alpha {E'}_\alpha
268     \end{equation}
269     where $\mathbf{M}$ is the total dipole moment in zero field. The
270     polarization density changes in the presence of the external field
271     $\mathbf{E'}$. The modified polarization is given by
272     \begin{equation}
273     \left< P'_\alpha(r) \right>_{E'} = \frac{\int P_\alpha(r) e^{-H'/k_B
274     T} d \Gamma^N}{\int e^{-H'/k_B T} d \Gamma^N} \equiv \frac{F}{G}
275     \end{equation}
276     For small values of the applied electric field we can expand the
277     polarization density as,
278     \begin{equation}
279     \left< P'_\alpha(r) \right>_{E'} = \left< P'_\alpha(r) \right> + \frac{\partial
280     \left< P'_\alpha(r) \right>}{\partial E'_\beta }\bigg|_{E'=0} E'_\beta
281     \end{equation}
282     and
283     \begin{equation}
284     \frac{\partial{\left<P'_\alpha(r)\right>}}{\partial{E'_\beta}}\bigg|_{E'=0} = \frac{F'G-FG'}{G^2}
285     \end{equation}
286     where
287     \begin{equation}
288     \begin{split}
289     F' & = \frac{1}{k_BT} \int {P}_\alpha M_\beta\; e^{-H/k_B T} d\Gamma^N\\
290     \textit{and} \\
291     G' & = \frac{1}{k_BT} \int M_\beta\; e^{-H/k_B T} d\Gamma^N
292     \end{split}
293     \end{equation}
294     Therefore
295     \begin{equation}
296     \braket{{P'_\alpha(r)}}_{E'}=\braket{{P'_\alpha(r)}}+\frac{1}{k_B T}(\braket{P_{\alpha}M_{\beta}}-\braket{P_{\alpha}}\braket{M_{\beta}}){E'}_\beta,
297     \end{equation}
298     where $\braket{A} = \frac{\int A\; e^{-H/k_B T}d\Gamma^N}{\int e^{-H/k_B T}d\Gamma^N}$ is the ensemble average of a physical quantity $A$ at temperature $T$. Now the polarization density due to applied external field is given by,
299     \begin{equation}
300     {\Delta P}_\alpha =\frac{1}{k_B T}\left(\braket{P_{\alpha}M_{\beta}}-\braket{P_{\alpha}}\braket{M_{\beta}}\right){E'}_\beta.
301     \end{equation}
302     Therefore the total polarization for the system can be obtained by integrating polarization density over the volume,
303     \begin{equation}
304     {\Delta M}_\alpha =\frac{1}{k_B T} \left(\braket{M_{\alpha}M_{\beta}}-\braket{M_{\alpha}}\braket{M_{\beta}}\right){E'}_\beta.
305     \end{equation}
306     If we consider both applied field and polarization in the z-direction,
307     \begin{equation}
308     {\Delta M}_z =\frac{1}{k_B T}\left(\braket{{M_{z}}^2}-\braket{M_{z}}^2\right){E'}_z.
309     \end{equation}
310     In general applied field $ \braket{{M_{z}}^2} = \frac{1}{3}\braket{{\mathbf{M}}^2}$ and $ \braket{{M_{z}}} = \frac{1}{3}\braket{{\mathbf{M}}}$ therefore,
311     \begin{equation}
312     {\Delta P}_z = \frac{{\Delta M}_z}{V} = \epsilon_o \frac{\braket{{\mathbf{M}}^2}-\braket{{\mathbf{M}}}^2}{3 \epsilon_o \; k_B T V}{E'}_z.
313     \label{eq:fluc}
314     \end{equation}
315     Comparing equation (\ref{eq:fluc}) with $ \Delta P_z = \epsilon_o \chi \ {E'}_z$ the fluctuation formula for susceptibility ($\chi$) can be written as,
316     \begin{equation}
317     \chi = \frac{\braket{{\mathbf{M}}^2}-\braket{{\mathbf{M}}}^2}{3 \epsilon_o \; k_B T V}
318     \end{equation}
319     \section{Boltzmann average of a quadrupole moment}
320     The average quadrupole moment at temperature T in the presence of uniform applied
321     field gradient is given by,
322     \begin{equation}
323     \braket{q_{\alpha\beta}} \;=\; \frac{\int d\Omega\; e^{\frac{-U(r)}{k_B T}}q_{\alpha\beta}}{\int d\Omega\; e^{\frac{-U(r)}{k_B T}}} \;=\; \frac{\int d\Omega\; e^{\frac{q_{\mu\nu}\;\partial_\nu E_\mu}{k_B T}}q_{\alpha\beta}}{\int d\Omega\; e^{\frac{q_{\mu\nu}\;\partial_\nu E_\mu}{k_B T}}},
324     \label{eq:1}
325     \end{equation}
326     where $\int \Omega = \int_0^{2\pi} \int_0^\pi \int_0^{2\pi}
327     sin\theta\; d\theta\ d\phi\ d\psi$ is the integration over Euler
328     angles and $ U(r) = -q_{\mu\nu}\;\partial_\nu E_\mu $ is the energy of
329     a quadrupole in the gradient of the
330     applied field. The energy and quadrupole moment can be transformed
331     into body frame using following relation,
332     \begin{equation}
333     \begin{split}
334     q_{\alpha\beta} &= \eta_{\alpha\alpha'}{q}^* _{\alpha'\beta'}\eta_{\beta'\beta} \\
335     U(r) &= - q:\vec{\nabla}\vec{E} = - q_{\mu\nu}\;\partial_\nu E_\mu = -\eta_{\mu\mu'}{q}^*_{\mu'\nu'}\eta_{\nu'\nu}\;\partial_\nu E_\mu.
336     \end{split}
337     \label{eq:2}
338     \end{equation}
339     Here the starred tensors are the components in the body fixed
340     frame. Substituting equation (\ref{eq:2}) in the equation (\ref{eq:1})
341     and taking linear terms in the expansion we get,
342     \begin{equation}
343     \braket{q_{\alpha\beta}} = \frac{ \int d\Omega \left(1 + \frac{\eta_{\mu\mu'}{q}^*_{\mu'\nu'}\eta_{\nu'\nu}\;\partial_\nu E_\mu }{k_B T}\right)q_{\alpha\beta}}{ \int d\Omega \left(1 + \frac{\eta_{\mu\mu'}{q}^*_{\mu'\nu'}\eta_{\nu'\nu}\;\partial_\nu E_\mu }{k_B T}\right)},
344     \label{eq:3}
345     \end{equation}
346     where $\eta_{\alpha\alpha'}$ is the rotation matrix that transforms
347     the body fixed to the space fixed frame,
348     \[\eta_{\alpha\alpha'}
349     = \left(\begin{array}{ccc}
350     cos\phi\; cos\psi - cos\theta\; sin\phi\; sin\psi & cos\theta\; cos\psi\; sin\phi + cos\phi\; sin\psi & sin\theta\; sin\phi \\
351     -cos\psi\; sin\phi - cos\theta\; cos\phi \; sin\psi & cos\theta\; cos\phi\; cos\psi - sin\phi\; sin\psi & cos\phi\; cos\theta \\
352     sin\theta\; sin\psi & -cos\psi\; sin\theta & cos\theta
353     \end{array} \right).\]
354     Integration of 1st and 2nd terms in the denominator gives $8 \pi^2$
355     and $8 \pi^2 /3\;\vec{\nabla}.\vec{E}\; Tr(q^*) $ respectively. The
356     second term vanishes for charge free space
357     (i.e. $\vec{\nabla}.\vec{E} \; = \; 0)$. Similarly integration of the
358     1st term in the numerator produces
359     $8 \pi^2 /3\; Tr(q^*)\delta_{\alpha\beta}$ and the 2nd term produces
360     $8 \pi^2 /15k_B T (3{q}^*_{\alpha'\beta'}{q}^*_{\beta'\alpha'} -
361     {q}^*_{\alpha'\alpha'}{q}^*_{\beta'\beta'})\partial_\alpha E_\beta$,
362     if $\vec{\nabla}.\vec{E} \; = \; 0$,
363     $ \partial_\alpha E_\beta = \partial_\beta E_\alpha$ and
364     ${q}^*_{\alpha'\beta'}= {q}^*_{\beta'\alpha'}$. Therefore the
365     Boltzmann average of a quadrupole moment can be written as,
366    
367     \begin{equation}
368     \braket{q_{\alpha\beta}}\; = \; \frac{1}{3} Tr(q^*)\;\delta_{\alpha\beta} + \frac{{\bar{q_o}}^2}{15k_BT}\;\partial_\alpha E_\beta,
369     \label{eq:4}
370     \end{equation}
371     here ${\bar{q_o}}^2=
372     3{q}^*_{\alpha'\beta'}{q}^*_{\beta'\alpha'}-{q}^*_{\alpha'\alpha'}{q}^*_{\beta'\beta'}$
373     which is a net quadrupole moment of a molecule.
374    
375     To find the average quadrupole moment for the entire system, we can
376     multiply above equation with number of molecules $N$.
377     \begin{equation}
378     \braket{Q_{\alpha\beta}} \;=\; N\braket{q_{\alpha\beta}}\; = \; \frac{N}{3} Tr(q^*)\;\delta_{\alpha\beta} + \frac{{N\;\bar{q_o}}^2}{15k_BT}\;\partial_\alpha E_\beta
379     \label{eq:5}
380     \end{equation}
381     The analogous relation in a system of the $N$ quadrupoles is,
382     \begin{equation}
383     \braket{Q_{\alpha\beta}} \; = \; \frac{1}{3}\;Tr(Q)\;\delta_{\alpha\beta} + \frac{{\braket{Q^2}}_{en} - {\braket{Q}}_{en}^2}{15k_BT}\;\partial_\alpha E_\beta
384     \label{eq:6}
385     \end{equation}
386     where $\braket{{Q^2}}_{en}$ and $\braket{Q}_{en}^2$ are the ensemble
387     average of the square of the magnitude of the box quadrupole moment
388     and the square of the ensemble average of the box quadrupole moment
389     respectively. The ensemble averages in the above equation can be
390     obtained using relations,
391     ${\braket{Q^2}}_{en} =
392     {\braket{3Q_{\alpha'\beta'}Q_{\beta'\alpha'}-Q_{\alpha'\alpha'}Q_{\beta'\beta'}}}_{en}$
393     and
394     ${\braket{Q}}_{en}^2 = 3{\braket{Q_{\alpha'\beta'}}}_{en}\;
395     \braket{Q_{\beta'\alpha'}}_{en} -{\braket{Q_{\alpha'\alpha'}}}_{en}\;
396     {\braket{Q_{\beta'\beta'}}}_{en}$
397     respectively. The non-zero value of the $\braket{Q}_{en}^2$ is due to
398     the applied gradient, therefore its contribution is subtracted to find
399     the actual fluctuation in the box quadrupole moment. From the
400     equations (\ref{eq:5}) and (\ref{eq:6}) we get,
401     \begin{equation}
402     \begin{split}
403     \braket{{Q^2}}_{en}-{\braket{Q}_{en}}^2 &= N\; {\bar{q_o}}^2\\
404     & and \\
405     \braket{Q_{\alpha\beta}}\; &= \; N\braket{q_{\alpha\beta}}.
406     \end{split}
407     \label{eq:7}
408     \end{equation}
409    
410    
411    
412     \section{Relation between quadrupolar polarization and applied field gradient}
413     In the presence of the field gradient $\partial_\alpha {E}_\beta $, a
414     non-vanishing quadrupolar polarization (quadrupole moment per unit
415     volume) $\bar{Q}_{\alpha\beta}$ will be induced in the entire volume
416     of a sample. The total field at any point $\vec{r}$ in the sample is
417     given by,
418     \begin{equation}
419     \partial_\alpha E_\beta(\vec{r}) = \partial_\alpha {E^o}_\beta(\vec{r}) + \frac{1}{4\pi \epsilon_o}\int T_{\alpha\beta\gamma\delta}(\vec{r}-\vec{r'})\;\bar{Q}_{\gamma\delta}(\vec{r'})\; d^3r'
420     \label{eq:20}
421     \end{equation}
422     where $\partial_\alpha {E^o}_\beta$ is the applied field gradient and $ T_{\alpha\beta\gamma\delta}$ is the quadrupole-quadrupole interaction tensor which is obtained from applications of four successive gradient operator to the electrostatic kernal $v(r)$.
423     \begin{equation}
424     \begin{split}
425     T_{\alpha\beta\gamma\delta}(r) &=\nabla_\alpha \nabla_\beta \nabla_\gamma \nabla_\delta\;v(r) \\ &= \left(\delta_{\alpha\beta}\delta_{\gamma\delta} + \delta_{\alpha\gamma}\delta_{\beta\delta}+ \delta_{\alpha\delta}\delta_{\beta\gamma}\right)v_{41}(r) + \left(\delta_{\alpha\beta} r_\gamma r_\delta + 5 \; permutations \right) \frac{v_{42}(r)}{r^2} \\ &+ r_\alpha r_\beta r_\gamma r_\delta\; \frac{v_{43}(r)}{r^4},
426     \end{split}
427     \label{quadRadial}
428     \end{equation}
429     where $v_{41}(r),\; v_{42}(r), \; and v_{43}(r)$ are defined in Paper I of the series (Refer Paper I) which have different functional forms for different cutoff methods. The integral in equation
430     (\ref{eq:20}) can be divided into two parts, $|\vec{r}-\vec{r'}|< R $
431     and $|\vec{r}-\vec{r'}|> R $ where $R \rightarrow 0 $. Since the total
432     field gradient vanishes at the singularity (see Appendix \ref{singular}),
433     equation (\ref{eq:20}) can be written as,
434     \begin{equation}
435     \partial_\alpha E_\beta(\vec{r}) = \partial_\alpha {E^o}_\beta(\vec{r}) +
436     \frac{1}{4\pi \epsilon_o}\lim_{R \to
437     0}\int\limits_{|\vec{r}-\vec{r'}|> R }
438     T_{\alpha\beta\gamma\delta}(|\vec{r}-\vec{r'}|)\;\bar{Q}_{\gamma\delta}(\vec{r'})\;
439     d^3r'.
440     \label{eq:23}
441     \end{equation}
442     If $r = r'$ is excluded from the integration, the gradient of the electric can be expressed in terms of traceless quadrupole moment as below (Refer eq 2.7 of Logan I) ,
443     \begin{equation}
444     \partial_\alpha E_\beta(\vec{r}) = \partial_\alpha {E^o}_\beta(\vec{r}) + \frac{1}{12\pi \epsilon_o}\int\limits_{|\vec{r}-\vec{r'}|> R } T_{\alpha\beta\gamma\delta}(|\vec{r}-\vec{r'}|)\;\bar{\Theta}_{\gamma\delta}(\vec{r'})\; d^3r',
445     \label{eq:22}
446     \end{equation}
447     where $\Theta_{\alpha\beta} = 3Q_{\alpha\beta} - \delta_{\alpha\beta}Tr(Q)$
448     is the traceless quadrupole moment. The total quadrupolar polarization is written as,
449     \begin{equation}
450     \bar{Q}_{\alpha\beta}(\vec{r}) = \frac{1}{3}\delta_{\alpha\beta}\;Tr(\bar{Q})+\epsilon_o {\chi}_Q\;\partial_\alpha E_\beta(\vec{r}),
451     \label{eq:21}
452     \end{equation}
453     where ${\chi}_Q = \frac{\braket{{Q^2}} - \braket{Q}^2}{15\epsilon_o Vk_BT}$ is the susceptibility of the quadrupolar fluid. In terms of traceless quadrupole moment, equation (\ref{eq:21}) can be written as,
454     \begin{equation}
455     \frac{1}{3}\bar{\Theta}_{\alpha\beta}(\vec{r}) = \epsilon_o {\chi}_Q \; \partial_\alpha E_\beta (\vec{r})= \epsilon_o {\chi}_Q \left(\partial_\alpha {E^o}_\beta(\vec{r}) + \frac{1}{12\pi \epsilon_o}\int\limits_{|\vec{r}-\vec{r'}|> R } T_{\alpha\beta\gamma\delta}(|\vec{r}-\vec{r'}|)\;\bar{\Theta}_{\gamma\delta}(\vec{r'})\; d^3r'\right)
456     \label{eq:22}
457     \end{equation}
458     For toroidal boundary conditions, both $\partial_\alpha E_\beta$ and
459     $\bar{\Theta}_{\alpha\beta}$ are uniform over the entire space. After
460     performing a Fourier transform (see the Appendix in the Neumann Paper),
461     we get,
462     \begin{equation}
463     \frac{1}{3}\widetilde{\bar{\Theta}}_{\alpha\beta}(\vec{k})=
464     \epsilon_o {\chi}_Q \;\left[{\partial_\alpha
465     {E^o}_\beta}(\vec{k})+ \frac{1}{12\pi
466     \epsilon_o}\;\widetilde{T}_{\alpha\beta\gamma\delta}(\vec{k})\;
467     {\bar{\Theta}}_{\gamma\delta}(\vec{k})\right]
468     \label{eq:24}
469     \end{equation}
470     Since the quadrupolar polarization is in the direction of the applied
471     field
472     $\widetilde{\bar{\Theta}}_{\gamma\delta}(\vec{k}) =
473     \widetilde{\bar{\Theta}}_{\alpha\beta}(\vec{k})$
474     and
475     $\widetilde{T}_{\alpha\beta\gamma\delta}(\vec{k}) =
476     \widetilde{T}_{\alpha\beta\alpha\beta}(\vec{k})$.
477     (Note: This is true for orientational polarization but might not be
478     true for polarization due to changes in the atomic distribution. For
479     the orientational case, we need only to work with individual
480     components of the matrix.)
481     \begin{equation}
482     \begin{split}
483     \frac{1}{3}\widetilde{\bar{\Theta}}_{\alpha\beta}(\vec{k}) &= \epsilon_o \bar{\chi}_Q \left[{\partial_\alpha E^o_\beta}(\vec{k})+ \frac{1}{12\pi \epsilon_o}\widetilde{T}_{\alpha\beta\alpha\beta}(\vec{k})\;\widetilde{\bar{\Theta}}_{\alpha\beta}(\vec{k})\right] \\
484     &= \epsilon_o {\chi}_Q\;\left(1-\frac{1}{4\pi} {\chi}_Q
485     \widetilde{T}_{\alpha\beta\alpha\beta}(\vec{k})\right)^{-1}
486     {\partial_\alpha E^o_\beta}(\vec{k})
487     \end{split}
488     \label{eq:25}
489     \end{equation}
490     Matrix contraction of quadrupole-quadrupole tensor and quadrupole
491     moment gives the identity matrix. Therefore each component can be
492     treated as scalar. If the field gradient is homogeneous over the
493     entire volume, ${\partial_ \alpha E_\beta}(\vec{k}) = 0 $ except at
494     $ \vec{k} = 0$, hence it is sufficient to know
495     $ \widetilde{T}_{\alpha\beta\alpha\beta}(\vec{k})$ at $ \vec{k} =
496     0$. Therefore equation (\ref{eq:25}) can be written as,
497     \begin{equation}
498     \begin{split}
499     \frac{1}{3}\widetilde{\bar{\Theta}}_{\alpha\beta}({0}) &= \epsilon_o {\chi}_Q\left(1-\frac{1}{4\pi} {\chi}_Q\;\widetilde{T}_{\alpha\beta\alpha\beta}({0})\right)^{-1} \partial_\alpha E^o_\beta({0})
500     \end{split}
501     \label{eq:26}
502     \end{equation}
503     where $ \widetilde{T}_{\alpha\beta\alpha\beta}(\vec{0})$ can be evaluated as,
504     \begin{equation}
505     \widetilde{T}_{\alpha\beta\alpha\beta}({0}) = \int \widetilde{T}_{\alpha\beta\alpha\beta}\;(\vec{r})d^3r
506     \label{eq:27}
507     \end{equation}
508     The susceptibility is modified as,
509     \begin{equation}
510     \Xi_Q = \bar{\chi}_Q\left(1-\frac{1}{4\pi} \bar{\chi}_Q\;\widetilde{T}_{\alpha\beta\alpha\beta}({0})\right)^{-1}.
511     \label{eq:28}
512     \end{equation}
513     In terms of traced quadrupole moment equation (\ref{eq:25}) can be written as,
514     \begin{equation}
515     {\bar{Q}}_{\alpha\beta} = \frac{1}{3}\delta_{\alpha\beta}\;Tr(\bar{Q}) + \epsilon_o\; \Xi_Q\; \partial_\alpha E^o_\beta
516     \label{eq:29}
517     \end{equation}
518     From equations (\ref{eq:6}), (\ref{eq:28}), and (\ref{eq:29}) we get,
519     \begin{equation}
520     \begin{split}
521     &\frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\; =\; {\chi}_Q\left(1-\frac{1}{4\pi} {\chi}_Q\;\widetilde{T}_{\alpha\beta\alpha\beta}({0})\right)^{-1}, \\
522     &{\chi}_Q \;=\; \frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\left(1 + \frac{1}{4\pi} \frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\;\widetilde{T}_{\alpha\beta\alpha\beta}({0})\right)^{-1}
523     \end{split}
524     \end{equation}
525     Finally in terms of quadrupolar correction factor ($A_{quad}$) the quadrupolar susceptibility is written as,
526     \begin{equation}
527     {\chi}_Q \;=\; \frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\left(1 + \frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\; A_{quad}\right)^{-1}
528     \end{equation}
529     The $A_{quad}$ has dimension of the $length^{-2}$ is different for different cutoff methods which is listed in Table I.
530     \section{The effective dielectric constant of a quadrupolar fluid}
531     The dielectric constant of a material is a measure of the screening of
532     the electrostatic energy when a material is placed in the
533     electrostatic field. If $U_{field}(r)$ and $U_{int}(r)$ are the total
534     electrostatic energies of a system due to the applied field and
535     interaction of the quadrupole with the gradient of the field
536     respectively. The dielectric constant can be written as,
537     \begin{equation}
538     \epsilon = \frac{U_{field} + U_{int}}{U_{field}}
539     \label{eq:31}
540     \end{equation}
541     The energy due to field can be written as,
542     \begin{equation}
543     U_{field}(r) = \int_V u(r)dv = \int_V \frac{1}{2}\epsilon_o {|E^o|}^2 dv,
544     \label{eq:32}
545     \end{equation}
546     where $u(r)$ is the energy density for applied electric field and $V$ is the total volume
547     of the system. The interaction energy can be assumed to be created by assembling quadrupoles from infinitely far away against the existing field gradient(Analogy Jackson's equations 4.83 - 4.94). Therefore
548     \begin{equation}
549     U_{int}(r) = -\frac{1}{2}\int_V \bar{Q}_{\alpha\beta}\;\partial_\beta E^o_{\alpha}\; dv,
550     \label{eq:33}
551     \end{equation}
552     where $\bar{Q}_{\alpha\beta}$ is the quadrupolar polarization. The quadrupoloarization can be expressed as linear function of the gradient of the applied field as below,
553     \begin{equation}
554     \bar{Q}_{\alpha\beta} = \epsilon_{o}\; {\chi_Q}\; \partial_\alpha E_\beta \approx \epsilon_{o}\; {\chi_Q}\; \partial_\alpha E^o_\beta ,
555     \label{eq:34}
556     \end{equation}
557     where ${\chi_Q} $ is a quadrupolar susceptibility. The local field gradient can be ignored in the region of low polarizability. In terms of the box quadrupole
558     moment it can be expressed as,
559     \begin{equation}
560     \chi_Q \;=\; \frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\left(1 + \frac{\braket{{Q^2}}_{en} - \braket{Q}_{en}^2}{15 \epsilon_o Vk_BT}\;A_{quad}\right
561     )^{-1}.
562     \label{eq:36}
563     \end{equation}
564     Using above relations (equation (\ref{eq:32}\;-\;\ref{eq:34}) in the
565     equation (\ref{eq:31}) produce,
566     \begin{equation}
567     \epsilon = 1 + {\chi}_Q \frac{\int_V |\partial_\alpha E^o_\beta|^2 dv}{\int_V{|E^o|}^2 dv}.
568     \end{equation}
569    
570     \newpage
571    
572     \appendix
573     \section{Point-multipolar interactions with a spatially-varying electric field}
574    
575     We can treat objects $a$, $b$, and $c$ containing embedded collections
576     of charges. When we define the primitive moments, we sum over that
577     collections of charges using a local coordinate system within each
578     object. The point charge, dipole, and quadrupole for object $a$ are
579     given by $C_a$, $\mathbf{D}_a$, and $\mathsf{Q}_a$, respectively.
580     These are the primitive multipoles which can be expressed as a
581     distribution of charges,
582     \begin{align}
583     C_a =&\sum_{k \, \text{in }a} q_k , \label{eq:charge} \\
584     D_{a\alpha} =&\sum_{k \, \text{in }a} q_k r_{k\alpha}, \label{eq:dipole}\\
585     Q_{a\alpha\beta} =& \frac{1}{2} \sum_{k \, \text{in } a} q_k
586     r_{k\alpha} r_{k\beta} . \label{eq:quadrupole}
587     \end{align}
588     Note that the definition of the primitive quadrupole here differs from
589     the standard traceless form, and contains an additional Taylor-series
590     based factor of $1/2$. In Paper 1, we derived the forces and torques
591     each object exerts on the others.
592    
593     Here we must also consider an external electric field that varies in
594     space: $\mathbf E(\mathbf r)$. Each of the local charges $q_k$ in
595     object $a$ will then experience a slightly different field. This
596     electric field can be expanded in a Taylor series around the local
597     origin of each object. A different Taylor series expansion is carried
598     out for each object.
599    
600     For a particular charge $q_k$, the electric field at that site's
601     position is given by:
602     \begin{equation}
603     E_\gamma + \nabla_\delta E_\gamma r_{k \delta}
604     + \frac {1}{2} \nabla_\delta \nabla_\varepsilon E_\gamma r_{k \delta}
605     r_{k \varepsilon} + ...
606     \end{equation}
607     Note that the electric field is always evaluated at the origin of the
608     objects, and treating each object using point multipoles simplifies
609     this greatly.
610    
611     To find the force exerted on object $a$ by the electric field, one
612     takes the electric field expression, and multiplies it by $q_k$, and
613     then sum over all charges in $a$:
614    
615     \begin{align}
616     F_\gamma &= \sum_{k \textrm{~in~} a} q_k \lbrace E_\gamma + \nabla_\delta E_\gamma r_{k \delta}
617     + \frac {1}{2} \nabla_\delta \nabla_\varepsilon E_\gamma r_{k \delta}
618     r_{k \varepsilon} + ... \rbrace \\
619     &= C_a E_\gamma + D_{a \delta} \nabla_\delta E_\gamma
620     + Q_{a \delta \varepsilon} \nabla_\delta \nabla_\varepsilon E_\gamma +
621     ...
622     \end{align}
623    
624     Similarly, the torque exerted by the field on $a$ can be expressed as
625     \begin{align}
626     \tau_\alpha &= \sum_{k \textrm{~in~} a} (\mathbf r_k \times q_k \mathbf E)_\alpha \\
627     & = \sum_{k \textrm{~in~} a} \epsilon_{\alpha \beta \gamma} q_k
628     r_{k\beta} E_\gamma(\mathbf r_k) \\
629     & = \epsilon_{\alpha \beta \gamma} D_\beta E_\gamma
630     + 2 \epsilon_{\alpha \beta \gamma} Q_{\beta \delta} \nabla_\delta
631     E_\gamma + ...
632     \end{align}
633    
634     The last term is essentially identical with form derived by Torres del
635     Castillo and M\'{e}ndez Garrido,\cite{Torres-del-Castillo:2006uo} although their derivation
636     utilized a traceless form of the quadrupole that is different than the
637     primitive definition in use here. We note that the Levi-Civita symbol
638     can be eliminated by utilizing the matrix cross product in an
639     identical form as in Ref. \onlinecite{Smith98}:
640     \begin{equation}
641     \left[\mathsf{A} \times \mathsf{B}\right]_\alpha = \sum_\beta
642     \left[\mathsf{A}_{\alpha+1,\beta} \mathsf{B}_{\alpha+2,\beta}
643     -\mathsf{A}_{\alpha+2,\beta} \mathsf{B}_{\alpha+1,\beta}
644     \right]
645     \label{eq:matrixCross}
646     \end{equation}
647     where $\alpha+1$ and $\alpha+2$ are regarded as cyclic permuations of
648     the matrix indices. In table \ref{tab:UFT} we give compact
649     expressions for how the multipole sites interact with an external
650     field that has exhibits spatial variations.
651    
652     \begin{table}
653     \caption{Potential energy $(U)$, force $(\mathbf{F})$, and torque
654     $(\mathbf{\tau})$ expressions for a multipolar site embedded in an
655     electric field with spatial variations, $\mathbf{E}(\mathbf{r})$.
656     \label{tab:UFT}}
657     \begin{tabular}{r|ccc}
658     & Charge & Dipole & Quadrupole \\ \hline
659     $U$ & $C \phi(\mathbf{r})$ & $-\mathbf{D} \cdot \mathbf{E}(\mathbf{r})$ & $- \mathsf{Q}:\nabla \mathbf{E}(\mathbf{r})$ \\
660     $\mathbf{F}$ & $C \mathbf{E}(\mathbf{r})$ & $+\mathbf{D} \cdot \nabla \mathbf{E}(\mathbf{r})$ & $+\mathsf{Q} : \nabla\nabla\mathbf{E}(\mathbf{r})$ \\
661     $\mathbf{\tau}$ & & $\mathbf{D} \times \mathbf{E}(\mathbf{r})$ & $+2 \mathsf{Q} \times \nabla \mathbf{E}(\mathbf{r})$
662     \end{tabular}
663     \end{table}
664     \section{Gradient of the field due to quadrupolar polarization}
665     \label{singular}
666     In this section, we will discuss the gradient of the field produced by
667     quadrupolar polarization. For this purpose, we consider a distribution
668     of charge ${\rho}(r)$ which gives rise to an electric field
669     $\vec{E}(r)$ and gradient of the field $\vec{\nabla} \vec{E}(r)$
670     throughout space. The total gradient of the electric field over volume
671     due to the all charges within the sphere of radius $R$ is given by
672     (cf. Jackson equation 4.14):
673     \begin{equation}
674     \int_{r<R} \vec{\nabla}\vec{E}\;d^3r = -\int_{r=R} R^2 \vec{E}\;\hat{n}\; d\Omega
675     \label{eq:8}
676     \end{equation}
677     where $d\Omega$ is the solid angle and $\hat{n}$ is the normal vector
678     of the surface of the sphere which is equal to
679     $sin[\theta]cos[\phi]\hat{x} + sin[\theta]sin[\phi]\hat{y} +
680     cos[\theta]\hat{z}$
681     in spherical coordinates. For the charge density ${\rho}(r')$, the
682     total gradient of the electric field can be written as (cf. Jackson
683     equation 4.16),
684     \begin{equation}
685     \int_{r<R} \vec{\nabla}\vec{E}\; d^3r=-\int_{r=R} R^2\; \vec{\nabla}\Phi\; \hat{n}\; d\Omega =-\frac{1}{4\pi\;\epsilon_o}\int_{r=R} R^2\; \vec{\nabla}\;\left(\int \frac{\rho(r')}{|\vec{r}-\vec{r'}|}\;d^3r'\right) \hat{n}\; d\Omega
686     \label{eq:9}
687     \end{equation}
688     The radial function in the equation (\ref{eq:9}) can be expressed in
689     terms of spherical harmonics as (cf. Jackson equation 3.70),
690     \begin{equation}
691     \frac{1}{|\vec{r} - \vec{r'}|} = 4\pi \sum_{l=0}^{\infty}\sum_{m=-l}^{m=l}\frac{1}{2l+1}\;\frac{{r^l_<}}{{r^{l+1}_>}}\;{Y^*}_{lm}(\theta', \phi')\;Y_{lm}(\theta, \phi)
692     \label{eq:10}
693     \end{equation}
694     If the sphere completely encloses the charge density then $ r_< = r'$ and $r_> = R$. Substituting equation (\ref{eq:10}) into (\ref{eq:9}) we get,
695     \begin{equation}
696     \begin{split}
697     \int_{r<R} \vec{\nabla}\vec{E}\;d^3r &=-\frac{R^2}{\epsilon_o}\int_{r=R} \; \vec{\nabla}\;\left(\int \rho(r')\sum_{l=0}^{\infty}\sum_{m=-l}^{m=l}\frac{1}{2l+1}\;\frac{{r'^l}}{{R^{l+1}}}\;{Y^*}_{lm}(\theta', \phi')\;Y_{lm}(\theta, \phi)\;d^3r'\right) \hat{n}\; d\Omega \\
698     &= -\frac{R^2}{\epsilon_o}\sum_{l=0}^{\infty}\sum_{m=-l}^{m=l}\frac{1}{2l+1}\;\int \rho(r')\;{r'^l}\;{Y^*}_{lm}(\theta', \phi')\left(\int_{r=R}\vec{\nabla}\left({R^{-(l+1)}}\;Y_{lm}(\theta, \phi)\right)\hat{n}\; d\Omega \right)d^3r
699     '
700     \end{split}
701     \label{eq:11}
702     \end{equation}
703     The gradient of the product of radial function and spherical harmonics
704     is given by (cf. Arfken, p.811 eq. 16.94):
705     \begin{equation}
706     \begin{split}
707     \vec{\nabla}\left[ f(r)\;Y_{lm}(\theta, \phi)\right] = &-\left(\frac{l+1}{2l+1}\right)^{1/2}\; \left[\frac{\partial}{\partial r}-\frac{l}{r} \right]f(r)\; Y_{l, l+1, m}(\theta, \phi)\\ &+ \left(\frac{l}{2l+1}\right)^{1/2}\left[\frac
708     {\partial}{\partial r}+\frac{l}{r} \right]f(r)\; Y_{l, l-1, m}(\theta, \phi).
709     \end{split}
710     \label{eq:12}
711     \end{equation}
712     Using equation (\ref{eq:12}) we get,
713     \begin{equation}
714     \vec{\nabla}\left({R^{-(l+1)}}\;Y_{lm}(\theta, \phi)\right) = [(l+1)(2l+1)]^{1/2}\; Y_{l,l+1,m}(\theta, \phi) \; \frac{1}{R^{l+2}},
715     \label{eq:13}
716     \end{equation}
717     where $ Y_{l,l+1,m}(\theta, \phi)$ is the vector spherical harmonics
718     which can be expressed in terms of spherical harmonics as shown in
719     below (cf. Arfkan p.811),
720     \begin{equation}
721     Y_{l,l+1,m}(\theta, \phi) = \sum_{m_1, m_2} C(l+1,1,l|m_1,m_2,m)\; {Y_{l+1}}^{m_1}(\theta,\phi)\; \hat{e}_{m_2},
722     \label{eq:14}
723     \end{equation}
724     where $C(l+1,1,l|m_1,m_2,m)$ is a Clebsch-Gordan coefficient and
725     $\hat{e}_{m_2}$ is a spherical tensor of rank 1 which can be expressed
726     in terms of Cartesian coordinates,
727     \begin{equation}
728     {\hat{e}}_{+1} = - \frac{\hat{x}+i\hat{y}}{\sqrt{2}},\quad {\hat{e}}_{0} = \hat{z},\quad and \quad {\hat{e}}_{-1} = \frac{\hat{x}-i\hat{y}}{\sqrt{2}}
729     \label{eq:15}
730     \end{equation}
731     The normal vector $\hat{n} $ can be expressed in terms of spherical tensor of rank 1 as shown in below,
732     \begin{equation}
733     \hat{n} = \sqrt{\frac{4\pi}{3}}\left(-{Y_1}^{-1}{\hat{e}}_1 -{Y_1}^{1}{\hat{e}}_{-1} + {Y_1}^{0}{\hat{e}}_0 \right)
734     \label{eq:16}
735     \end{equation}
736     The surface integral of the product of $\hat{n}$ and
737     ${Y_{l+1}}^{m_1}(\theta, \phi)$ gives,
738     \begin{equation}
739     \begin{split}
740     \int \hat{n}\;{Y_{l+1}}^{m_1}\;d\Omega &= \int \sqrt{\frac{4\pi}{3}}\left(-{Y_1}^{-1}{\hat{e}}_1 -{Y_1}^{1}{\hat{e}}_{-1} + {Y_1}^{0}{\hat{e}}_0 \right)\;{Y_{l+1}}^{m_1}\; d\Omega \\
741     &= \int \sqrt{\frac{4\pi}{3}}\left({{Y_1}^{1}}^* {\hat{e}}_1 +{{Y_1}^{-1}}^* {\hat{e}}_{-1} + {{Y_1}^{0}}^* {\hat{e}}_0 \right)\;{Y_{l+1}}^{m_1}\; d\Omega \\
742     &= \sqrt{\frac{4\pi}{3}}\left({\delta}_{l+1, 1}\;{\delta}_{1, m_1}\;{\hat{e}}_1 + {\delta}_{l+1, 1}\;{\delta}_{-1, m_1}\;{\hat{e}}_{-1}+ {\delta}_{l+1, 1}\;{\delta}_{0, m_1} \;{\hat{e}}_0\right),
743     \end{split}
744     \label{eq:17}
745     \end{equation}
746     where ${Y_{l}}^{-m} = (-1)^m\;{{Y_{l}}^{m}}^* $ and
747     $ \int {{Y_{l}}^{m}}^*\;{Y_{l'}}^{m'}\;d\Omega =
748     \delta_{ll'}\delta_{mm'} $.
749     Non-vanishing values of equation \ref{eq:17} require $l = 0$,
750     therefore the value of $ m = 0 $. Since the values of $ m_1$ are -1,
751     1, and 0 then $m_2$ takes the values 1, -1, and 0, respectively
752     provided that $m = m_1 + m_2$. Equation \ref{eq:11} can therefore be
753     modified,
754     \begin{equation}
755     \begin{split}
756     \int_{r<R} \vec{\nabla}\vec{E}\;d^3r = &- \sqrt{\frac{4\pi}{{3}}}\;\frac{1}{\epsilon_o}\int \rho(r')\;{Y^*}_{00}(\theta', \phi')[ C(1, 1, 0|-1,1,0)\;{\hat{e}_{-1}}{\hat{e}_{1}}\\ &+ C(1, 1, 0|-1,1,0)\;{\hat{e}_{1}}{\hat{e}_{-1}}+C(
757     1, 1, 0|0,0,0)\;{\hat{e}_{0}}{\hat{e}_{0}} ]\; d^3r'.
758     \end{split}
759     \label{eq:18}
760     \end{equation}
761     After substituting ${Y^*}_{00} = \frac{1}{\sqrt{4\pi}} $ and using the
762     values of the Clebsch-Gorden coefficients: $ C(1, 1, 0|-1,1,0) =
763     \frac{1}{\sqrt{3}}, \; C(1, 1, 0|-1,1,0)= \frac{1}{\sqrt{3}}$ and $
764     C(1, 1, 0|0,0,0) = -\frac{1}{\sqrt{3}}$ in equation \ref{eq:18} we
765     obtain,
766     \begin{equation}
767     \begin{split}
768     \int_{r<R} \vec{\nabla}\vec{E}\;d^3r &= -\sqrt{\frac{4\pi}{{3}}}\;\frac{1}{\epsilon_o}\int \rho(r')\;d^3r'\left({\hat{e}_{-1}}{\hat{e}_{1}}+{\hat{e}_{1}}{\hat{e}_{-1}}-{\hat{e}_{0}}{\hat{e}_{0}}\right)\\
769     &= - \sqrt{\frac{4\pi}{{3}}}\;\frac{1}{\epsilon_o}\;C_{total}\;\left({\hat{e}_{-1}}{\hat{e}_{1}}+{\hat{e}_{1}}{\hat{e}_{-1}}-{\hat{e}_{0}}{\hat{e}_{0}}\right).
770     \end{split}
771     \label{eq:19}
772     \end{equation}
773     Equation (\ref{eq:19}) gives the total gradient of the field over a
774     sphere due to the distribution of the charges. For quadrupolar fluids
775     the total charge within a sphere is zero, therefore
776     $ \int_{r<R} \vec{\nabla}\vec{E}\;d^3r = 0 $. Hence the quadrupolar
777     polarization produces zero net gradient of the field inside the
778     sphere.
779    
780     \newpage
781    
782     \bibliography{multipole}
783    
784     \end{document}