1: \documentstyle[12pt,psfig,subfig]{article}
2: \setlength{\textheight}{23cm}
3: \setlength{\textwidth}{15cm}
4: \pagestyle{plain}
5: \setlength{\topmargin}{-2cm}
6: \raggedbottom
7: \setlength{\parskip}{0.13cm}
8: \begin{document}
9: \bibliographystyle{unsrt}
10: \begin{center}
11: {\Large{\bf Vibrational dynamics of solid poly(ethylene oxide)}} \\
12: \vspace{2cm}
13: {\large{\bf M. Krishnan\footnote{email:mkrishna@jncasr.ac.in}, and S. Balasubramanian\footnote{Corresponding Author. email:bala@jncasr.ac.in}}} \\
14: \vspace{0.5cm}
15: Chemistry and Physics of Materials Unit, \\
16: Jawaharlal Nehru Centre for Advanced Scientific Research,\\
17: Jakkur, Bangalore 560064, India.\\
18: \end{center}
19: \begin{center}
20: {\large Abstract}
21: \end{center}
22: Molecular dynamics (MD) simulations of crystalline poly(ethylene oxide) (PEO) have been carried out in
23: order to study its vibrational properties.
24: The vibrational density of states has been calculated using a normal mode
25: analysis (NMA) and also through the velocity autocorrelation function of the atoms.
26: Results agree well with experimental spectroscopic data.
27: System size effects in the crystalline state, studied through a comparison between
28: results for 16 unit cells and that for one unit cell has shown important differences
29: in the features below 100~cm$^{-1}$.
30: Effects of interchain interactions are examined by a comparison of the
31: spectra in the condensed state to that obtained for an isolated oligomer
32: of ethylene oxide.
33: Calculations of the local character of the modes indicate the presence of collective
34: excitations for frequencies lower than 100~cm$^{-1}$, in which around 8 to
35: 12 successive atoms of the polymer backbone participate. The
36: backbone twisting of helical chains about their long axes is dominant
37: in these low frequency modes.
38:
39: \vspace*{0.5cm}
40: \newpage
41:
42: \par
43: \section{Introduction}
44: \par
45: Solid polymer electrolytes (SPE), composed of inorganic salts solvated in solid, high
46: molecular weight polymer matrices, have been the focus of intense theoretical and
47: experimental research because of their
48: applications as solid state batteries~\cite{halley_power,wakihara}.
49: In spite of their wide technological applications, the precise mechanism of
50: conduction in these materials is still unclear and remains a pursuit of
51: interest. In non-polymeric crystalline and glassy electrolytes,
52: the ionic species hop from one site to another within a rigid host frame.
53: However, the conduction mechanism in solid polymer electrolytes is
54: believed to be different~\cite{brucebook,bruce1,angell_jpcm}. NMR studies of
55: line shape and relaxation rates in
56: poly(ethylene oxide)-lithium salt complexes have demonstrated a relationship between
57: the motion of the Li$^+$ ions
58: and the segmental motion of the PEO chains~\cite{donoso}.
59: This notion gains support from other experimental
60: studies such as Field-Gradient NMR spin echo technique~\cite{boden},
61: and Quasi-elastic Neutron Scattering (QENS)~\cite{qens}.
62: SPE, in
63: general, contain both amorphous and crystalline regions, with conduction
64: being facile in the amorphous regions. This is probably
65: due to differences in the chain dynamics~\cite{angell}.
66: The work of Armand and coworkers has shown that ion mobility in solid
67: electrolytes is of a continuous, diffusive type,
68: in the amorphous regions~\cite{berthier}.
69: NMR, differential scanning calorimetry and electrical conductivity
70: studies have demonstrated that both cations and anions are mobile in the
71: amorphous phase~\cite{gorecki}.
72: Recently, Bruce and coworkers have shown
73: that a SPE with a ratio of ether oxygen to lithium of
74: 6:1, exhibits higher conductivity in its crystalline phase as
75: compared to its amorphous phase and that the ion transport is dominated by
76: the cations~\cite{bruce2}.
77: In such complexes,
78: cross-linking between a pair of polymer chains
79: results in a channel like structure. The ion moves in this channel,
80: aided by the
81: segmental motion of the polymer chain pair.
82:
83: A large number of simulations have been carried out on PEO and PEO-salt
84: complexes, both in their crystalline and amorphous phases over the last decade.
85: Molecular dynamics studies by de Leeuw et al, using an united atom model (UAM) for
86: PEO chains in their molten state, have examined the motion of methylene groups
87: which they term as segmental motion~\cite{leeuw}.
88: Neyertz et al have performed extensive molecular dynamics simulations of bulk
89: crystalline PEO~\cite{sylvie-94,thesis}, PEO melts~\cite{neyertzmelt}, crystalline NaI-PEO~\cite{sylvie-acta95}
90: and amorphous NaI-PEO~\cite{sylvie-cps95} systems. The simulations of Muller-Plathe and coworkers
91: on PEO-LiI complexes~\cite{florian-95} have demonstrated that the ether oxygens belonging to consecutive
92: monomers of a PEO chain coordinate to the same $Li^{+}$ ion and that this segmental coordination is
93: argued to be the driving force for salt dissolution.
94: Laasonen and Klein have performed MD simulations of both crystalline and amorphous PEO-NaI
95: complexes~\cite{laasonen}. Their study showed that ion pairing is most
96: probable in the crystalline rather than in the amorphous phase.
97: Halley and coworkers
98: have studied the structure of amorphous PEO using the Parrinello-Rahman
99: method recently~\cite{halley2001}.
100: MD simulations have also been employed by
101: Smith and coworkers to
102: elucidate the structure and dynamics of poly(propylene oxide)
103: melts~\cite{smith_peo}, PEO-salt complexes~\cite{smith} and aqueous
104: PEO solutions~\cite{smith_h2o}, using a potential model derived from
105: {\em ab initio} calculations~\cite{smith_spectrochim}.
106:
107: It is our endeavor here to characterize the vibrational modes of the polymer backbone
108: in the pristine polymer, with an emphasis on the low frequency modes.
109: This could enhance our understanding on the exact relationship between the
110: dynamics of the polymer and that of the cation in the PEO-salt complexes.
111: We carry out this study with the
112: thought that understanding the evolution of the vibrational modes from the crystalline
113: phase of pure PEO to its nature in the glassy state without and with the salt is
114: crucial in mapping the precise mechanism of ion transport in these systems~\cite{carini,mustarelli}.
115: Normal
116: mode analyses to characterize vibrational spectra have been employed successfully
117: to describe the dynamics of glassy systems~\cite{elliott_papers}, and are likely
118: to be employed in the study of jammed granular
119: materials~\cite{nagel}. These calculations have also helped our understanding of vibrational
120: modes of polyethylene~\cite{hess_avg}, of proteins in solution~\cite{karplus}, and in
121: determining the flexibility of proteins, and in the thermodynamics of hydration water
122: in protein solutions~\cite{verma_jpcb}.
123: Unlike polyethylene, PEO adopts a distorted
124: helical conformation in its crystalline state, with the distortion arising from strong
125: intermolecular interactions. It is thus important that such normal mode calculations
126: be performed for the crystalline state, rather than for an isolated oligomer.
127: In this paper, we limit
128: ourselves to the study of the vibrational spectrum of PEO in its crystalline state.
129: Analyses of these modes under other state and phase
130: conditions will be examined in future. In anticipation of our results,
131: we have identified the existence of segments in the polymer backbone consisting
132: of around 8 to 12 atoms, to exhibit significant atomic displacements, for vibrational modes
133: with frequencies up to around 100~cm$^{-1}$, and that the modes with frequencies
134: below 60~cm$^{-1}$ involve twisting of the skeletal backbone.
135: The paper is divided as follows. Details of the simulation and the methods of
136: analyses are presented in the next section. Later, we discuss the
137: results obtained from our study. Details on obtaining the elements of the dynamical matrix
138: are provided in the Appendix.
139:
140: \par
141: \section{Details of Simulation}
142: \par
143:
144: The unit cell
145: of crystalline PEO is monoclinic with the PEO chains in a (7/2)
146: distorted helical conformation with TTG sequence~\cite{takahashi}. An unitcell consists of four PEO
147: chains with 21 backbone atoms in each chain. The MD runs reported here
148: were initiated from configurations generated from this crystal structure.
149: The two ends of a polymer chain were assumed to be bonded across the simulation cell
150: boundary~\cite{sylvie-94,laasonen} to eliminate end effects.
151: MD simulations were performed primarily for a system that we
152: describe as $\langle422\rangle$, which contained 4 unitcells along the a-axis, 2 unitcells along
153: the b-axis and 2 unitcells along the c-axis, containing a total of 3136 atoms.
154: To study system size effects and effects of inter-chain interactions, we have
155: performed additional simulations of only one unitcell, and also
156: of one isolated finite length polymer chain. Thus the simulations for the $\langle422\rangle$ cells and
157: that for the one unitcell contained chains that had no ends, while that for the isolated
158: molecule, contained a chain with two ends.
159: An all atom model (AAM), with explicit
160: consideration of hydrogen atoms was used, to properly account for the
161: steric interactions expected to be dominant in the crystalline state. The simulations
162: were carried out in the canonical ensemble at 5K for the NMA and in the
163: constant pressure ensemble at 300K and 1~atmosphere to test
164: the stability of the simulated crystal.
165: Temperature control was achieved
166: by the use of Nose-Hoover chain thermostats~\cite{martyna}, using the PINY-MD
167: program~\cite{piny-md}. Long range interactions were treated using the Ewald
168: method with an $\alpha$ value of 0.3~\AA$^{-1}$, and 2399 reciprocal space
169: points were included in the Ewald sum~\cite{hansen}.
170: The methylene groups were treated as rigid entities, enabling us to employ
171: a timestep of 1~fs.
172: To obtain the vibrational density of states,
173: we performed these calculations at a temperature of 5K, where the atoms could be
174: expected to be near their equilibrium positions.
175: The equilibration period for the single molecule, the unitcell, and the $\langle422\rangle$ system were
176: 65~ps, 100~ps, and 750~ps, respectively. These were followed by an analysis run of duration 30~ps
177: during which the coordinates and velocities
178: of each particle were stored at regular intervals.
179: Velocities were dumped
180: every time step to obtain the power spectrum of their time correlation
181: functions and the coordinates were dumped every 10 fs.
182:
183: The simulations were performed with
184: a force field
185: obtained from the work of Neyertz~\cite{sylvie-94} with the
186: torsional parameters of the CHARMm model~\cite{charmm}.
187: The potential parameters are given in Table 1. The non-bonded interactions were
188: truncated at 12\AA~for the $\langle422\rangle$ crystal, at 3.25\AA~for one unitcell, and at 12\AA~for
189: the single molecule runs.
190: The lower cutoff value for the unitcell run is necessitated by the fact that the
191: interaction cutoff should be
192: less than half of the minimum distance between any two opposite faces
193: of the simulation cell.
194:
195: The potential energy of the
196: system can be expanded in terms of atomic displacements from an equilibrium configuration as,
197: \begin{equation}
198: U\left(q_{1},q_{2},...,q_{n}\right) = U\left(q_{01},q_{02},...,q_{0n}\right) +
199: \left(\frac{\partial U}{\partial q_{i}}\right)_{0} \eta_{i} +
200: \frac{1}{2}\left(\frac{\partial^{2} U}{\partial q_{i} \partial q_{j}}\right)_{0}
201: \eta_{i} \eta_{j} + ...
202: \end{equation}
203: where the subscript 0 represents the equilibrium configuration and $\eta_{i}$
204: are the deviations of the coordinates from equilibrium, represented as,
205: \begin{equation}
206: q_{i} = q_{0i} + \eta_{i}
207: \end{equation}
208: The elements of the Hessian matrix are given by
209: \begin{equation}
210: H^{\alpha\beta}_{ij} = \frac{1}{\sqrt{m_{i}m_{j}}} \left(\frac{\partial^{2}U}{\partial \beta_{j}
211: \partial \alpha_{i}}\right)
212: \end{equation}
213: where i,j represent particle indices and $\alpha$, $\beta$ represent the spatial
214: coordinates x,y,z. $m_{i}$ is the mass of particle i.
215: A simple scheme to obtain some of these Hessian elements efficiently is
216: provided in the Appendix. {\em All
217: Hessian elements obtained from such analytical expressions were checked against
218: numerical second derivatives~\cite{abramovitz} within our normal mode analysis code, and were
219: found to match.}
220: The eigenvalues and eigenvectors of the Hessian matrix were examined to
221: understand the vibrational dynamics of PEO.
222: The frequency, $\nu_{s}$ ,of a particular mode of vibration, s,
223: is related to its eigen value, $\lambda_{s}$ , by
224: \begin{equation}
225: \lambda_{s} = \left(2\pi\nu_{s}\right)^{2}
226: \end{equation}
227:
228: With these set of interactions, the initial pressure of the $\langle422\rangle$ system was
229: found to be around 4000 atmospheres. Hence, we
230: performed a MD run in the NPT ensemble for 200 ps under ambient conditions. The change
231: in volume was found to be 1.3\% from that of the experimental crystal.
232: Time correlation functions of atomic velocities were calculated,
233: and were Fourier transformed to obtain the power spectra.
234: These were compared with the spectrum obtained from the NMA.
235: The spectra were convoluted with a Gaussian function of width 4cm$^{-1}$ so as to
236: provide a width to the spectral features.
237: The helical axis of a PEO chain possesses a 7-fold rotational
238: symmetry. In addition, 7 C$_2$ axes lie perpendicular to it, making the molecular symmetry of
239: PEO to be D$_7$~\cite{takahashi}. We have characterized the symmetry of the normal
240: modes by studying the transformation of atomic displacements,
241: on application of the symmetry operations for this group. Specifically,
242: the operation by the C$_2$ axes enables a distinction between the three irreducible
243: representations, A$_1$, A$_2$, and E.
244:
245: We have also characterized the spatial extent of the modes of vibration using a quantity called the
246: local character, defined as~\cite{karplus,elliott}
247: \begin{equation}
248: Local character\left(j\right) = \sum_{i=1}^{3N}u_{ij}^{4}
249: \end{equation}
250: where $u_{ij}$ is the i$^{th}$ component of the j$^{th}$ eigenvector.
251: The local character value can range from 0 to 1 and it determines
252: the extent of localization of a particular mode.
253:
254: To obtain quantitative information on the length of the segment involved in the
255: low frequency modes, we have defined a
256: quantity called the Continuous Segment Size (CSS). We calculate the displacement of
257: the $k^{th}$ atom due to a $i^{th}$ mode using the expression
258: \begin{equation}
259: <\delta r^{2}_{ik}> = k_{\rm B}T \frac{\bracevert \vec{u}_{i}^{k}\bracevert^{2}}{m_{k}\omega_{i}^{2}}
260: \end{equation}
261: where $m_{k}$ is mass of $k^{th}$ atom, $\vec{u}_{i}^{k}$ is
262: the vector formed by the components of the i$^{th}$
263: eigen vector contributed by the k$^{th}$ atom, $T$ is temperature, and $\omega_{i}$ is the
264: frequency of the i$^{th}$ normal mode.
265: We then check if the displacement is greater than a specified cutoff.
266: If $n$ successive backbone atoms of a chain each have displacements greater
267: than the displacement cutoff, they are defined to constitute a segment with CSS = $n$. Similarly,
268: CSS was calculated for all possible modes with
269: different vibrational frequencies from which the average segment size for a given
270: frequency was calculated.
271:
272: \section{Results and Discussion}
273: For crystalline systems, a close
274: match between the experimentally determined cell parameters and that obtained from simulations
275: is a crucial first step in the veracity of the parameters used.
276: We show in Figure~1, the time
277: evolution of the cell parameters at 300K and 1~atmospheres, generated by a MD run
278: in the constant pressure ensemble using the Parrinello-Rahman method~\cite{parr-rahman}.
279: The $b$ and the $\beta$ parameters of the unit cell exhibit a relaxation from the
280: zero time experimental value, due possibly to relatively minor deficiencies in the
281: interaction model used to represent this system.
282: The constancy of the cell parameters for over 100~ps, shows that the simulated crystal is in
283: a stable state, and that the potential parameters are indeed able to reproduce
284: the high demands of the crystal symmetry.
285: Table 2 compares the cell parameters obtained from our simulations to the experimental data.
286:
287: An interesting feature associated with polymer dynamics is the variation of the
288: various vibrational modes of a chain encountered during its transformation from
289: the isolated state to the crystalline state. The vibrational spectra of a single
290: molecule, one unitcell and that of the $\langle422\rangle$ crystal are compared in Figure~2a.
291: When the chains assemble to form a crystal, each chain might prefer a new conformational
292: state relative to its structure in the isolated state. A comparison of the vibrational
293: spectra between that of one isolated molecule of finite length and of the $\langle422\rangle$ system
294: provides information on the effect of intermolecular interactions.
295: As expected, the vibrational states of the isolated molecule showed
296: marked differences from the crystalline state, particularly in the low frequency regions,
297: where large amplitude, collective motions are predominant (see later).
298: However, there are no significant
299: changes in the high frequency regions of the spectrum. The spectrum for the one unitcell
300: compares well with that of the $\langle422\rangle$ crystal. However, the features in the VDOS of the
301: $\langle422\rangle$ crystal is much better resolved, particularly at low frequencies. For instance,
302: the features at around 40~cm$^{-1}$ and
303: 75~cm$^{-1}$ are clearly evident in the larger system than in the spectrum for the unitcell,
304: pointing to effects of long range interactions. The spectrum for the $\langle422\rangle$ system is shown
305: in an expanded scale in Figure~2b. The split in the feature below 100~cm$^{-1}$ is evident
306: and compares well with experimental infrared (IR) spectra~\cite{rabolt} that shows
307: features at 37~cm$^{-1}$, 52~cm$^{-1}$, 81~cm$^{-1}$ and 107~cm$^{-1}$. We do observe a prominent shoulder
308: at 108~cm$^{-1}$ which has been attributed to modes involving C-O internal rotation earlier~\cite{rabolt}.
309: A comparison of the simulated vibrational density of states with the experimental IR and Raman
310: spectra, exhibited in Figure~2c, shows that the potential model captures well nearly all the
311: vibrational modes~\cite{rabolt,dacosta,branca}. Note that the simulated spectrum is the raw
312: density of states and does not contain any other terms that are needed to calculate the
313: experimental spectra. Thus only the peak positions need to be compared, and not their intensities.
314: Almost all the features found in the IR and Raman spectra seem to be present in the simulated
315: VDOS.
316: Another noteworthy feature of the spectrum is the
317: absence of modes with imaginary frequencies which would have corresponded to either the
318: presence of atoms away
319: from equilibrium locations or to a mismatch between the empirical potential function and the
320: crystal structure.
321:
322: We have also calculated the vibrational spectrum through the
323: Fourier transformation of the velocity auto correlation function of the atoms. This
324: is compared with the VDOS obtained by the NMA in Figure~3.
325: The agreement between the spectra is excellent except for features above
326: 1200~cm$^{-1}$. Also, the peak at around 1470~cm$^{-1}$ is
327: missing in the spectrum obtained through the velocity autocorrelation function (VACF).
328: Visualization of the atomic displacements associated with this mode revealed
329: HCH bending in methylene groups.
330: Since all the CH$_{2}$ groups were
331: constrained to be rigid during the MD runs for the VACF
332: analysis, the absence of a peak in the VDOS obtained
333: from VACF, in this region can be rationalized. The comparison of the VDOS obtained from
334: the two methods is good, despite the fact that the NMA method is only a harmonic
335: approximation to the potential. Close examination of the two shows a marginal (approximately
336: 4-5~cm$^{-1}$) shift to higher frequencies in the spectrum obtained by the NMA method
337: relative to that from the VACF.
338:
339: We visualized the eigenvectors corresponding to different vibrational modes to assign
340: the nature of atomic displacements
341: that are responsible for the spectral features. In general, our assignments are consistent with earlier
342: calculations~\cite{yoshihara,lam}.
343: In Figure~4a and Figure~4b, we
344: display a few of the modes.
345: The zero frequency mode characterizes rigid body translation. The features at around 40~cm$^{-1}$,
346: have earlier been attributed to chain deformations~\cite{rabolt}. Based
347: on visualization of atomic displacements shown in Figure~4a, we assign these modes specifically to
348: the twisting of the polymer backbone. The mode at 88~cm$^{-1}$ arises from torsional
349: motion around the C-O bond. The A$_1$ mode at 216~cm$^{-1}$ can be assigned to torsions around the
350: C-C bond,
351: while the 510~cm$^{-1}$ feature involves bending of CCO triplets. COC bending is
352: observed at 952~cm$^{-1}$ while the
353: wagging motion of the methylene groups is found to be present at 1244~cm$^{-1}$, in good
354: agreement with IR and Raman measurements~\cite{yoshihara}. The symmetry of the modes are also
355: shown in Figures~4 agree well with experimental assignments~\cite{yoshihara,rabolt,lam}.
356:
357: For
358: characterizing the normal modes further, we have calculated the local
359: character for each mode~\cite{karplus,elliott} which is shown in Figure~5. In the
360: range, 0 to 160~cm$^{-1}$, the local character value is very small, implying
361: the participation of a large number of atoms.
362: However, the local character does not provide any information
363: on the proximity of the atoms that exhibit significant displacement in a mode. Thus
364: this quantity has to be augmented by a further analysis of the concomitancy or
365: proximity of atoms excited in a mode.
366:
367: We have developed such an index that takes into account the connectivity of the
368: atoms involved in a mode.
369: We determine the number of successive atoms, $n$,
370: that participate in a mode of a given frequency, by calculating the
371: distribution of segment sizes, f(n), for that vibrational mode. As a
372: representative example, the distribution of the CSS for one such vibrational
373: mode of frequency 44~cm$^{-1}$ is shown in Figure~6 for four selected
374: displacement cutoffs. It is our contention that a large number of atoms in
375: proximity to each other participate in these modes. This is evident from the
376: non-zero value of f(n) for n values in the range of 5 to 10, for some of
377: the cutoff values.
378:
379: The average segment
380: size was calculated by evaluating the following summation,
381: \begin{equation}
382: <CSS> = \frac{\sum_{n=5}^{42} n f\left(n\right)}{\sum_{n=5}^{42} f\left(n\right)}
383: \end{equation}
384: In our study, we define a segment as a chain of connected atoms whose displacements
385: are larger than a specified cutoff, with n $\geq $
386: 5. Values with n $<$ 5, which correspond to intramolecular excitations that
387: arise out of bonded interactions like the
388: stretch, bend and torsion, have been omitted in $<CSS>$ calculations, as their
389: origin is trivially known. It should also be noted that the exact values of the average segment
390: size will depend on the chosen displacement cutoff. Hence, we have performed
391: these calculations for a variety of cutoffs and these are exhibited in Figure~7
392: as a function of the frequency of the modes.
393: Notice that the segment size for the zero frequency modes, {\em i.e.,}
394: translations, is 42, which is the number of backbone atoms in a given chain.
395: It is also evident from the figure that the average segment size decays
396: with increase in frequency.
397: Figure~7 also shows
398: the variation of $<$CSS$>$ as a function of frequency, for four different
399: displacement cutoffs. The displacement cutoff which spans the frequency range
400: relevant to collective motion, where the the local character value is almost
401: zero is the one of significance.
402: It can be seen that in the frequency range of 10~cm$^{-1}$ and 100~cm$^{-1}$, which
403: is the range of collective motion, around 8 to 12 successive atoms of the backbone
404: are involved in the excitations.
405:
406: \section{Conclusions}
407: We have studied the vibrational dynamics of crystalline poly(ethylene oxide)
408: using molecular dynamics and normal mode analysis. The vibrational density of
409: states obtained from NMA matches well with that obtained from the Fourier
410: transformation of velocity autocorrelation function and also with experimental
411: IR and Raman data~\cite{rabolt,dacosta,branca}. The VDOS of an isolated
412: PEO chain of finite length showed marked differences from that of the crystal
413: in the low frequency region where collective modes are
414: predominant. We have also explored system size effects by comparing the VDOS
415: obtained from a simulation containing 16 unitcells and that of one unitcell. The spectrum
416: obtained from the former is much better resolved, particularly at low frequencies,
417: where three clear features, at 44~cm$^{-1}$, 70~cm$^{-1}$, and 88~cm$^{-1}$ are
418: observed.
419: The results of our calculations agree well with assignments of mode symmetry
420: by earlier workers, that used standard methods for a single chain of PEO, or for an
421: unit cell~\cite{yoshihara,rabolt,lam}. Such calculations have the added advantage,
422: over the MD route presented here, of being able to obtain dispersion of the vibrational modes.
423: However, the method described here can be used to characterize these vibrations in
424: the amorphous and liquid phases of PEO.
425:
426: The normal modes
427: obtained from the present analysis were characterized by the local character
428: indicator and by a new quantity that determines the number of concomitant
429: atoms excited
430: by a mode, called the CSS. In the range 0 to 160~cm$^{-1}$, the value of
431: the local
432: character indicator was very small, indicating the participation of a large number of
433: atoms in the vibrational modes in this range. A distribution of segment sizes was
434: calculated for each mode from which the average continuous segment size was
435: calculated as a function of the vibrational frequency. The frequency dependence of
436: $<$CSS$>$ clearly shows collective modes to be present for frequencies less than
437: 100~cm$^{-1}$, in which around 8 to 12 successive
438: atoms of the backbone participate. This quantitative analysis is also corroborated
439: by visualization of the atomic displacements for the low frequency modes.
440:
441: The phase behavior of PEO and the evolution of these vibrational modes as
442: a function of temperature and their properties in the amorphous phase will form
443: the objectives of our study in future.
444:
445: \section{Acknowledgements}
446: We thank Prof. S. Vasudevan for enlightening discussions, and Dr. Preston Moore for
447: the visualization program used in Fig. 4.
448: \newpage
449: \par
450:
451: \section{Appendix 1}
452: \par
453: The form of the potential used in our simulations is generic to macromolecular interactions,
454: and can be found in Ref.~\cite{sylvie-94}. Here, we provide only the torsional and
455: Coulomb terms.
456: \begin{equation}
457: U_{coulomb} = \frac{q_{i}q_{j}}{r_{ij}}
458: \end{equation}
459: and
460: \begin{equation}
461: U_{torsion} = f\left(cos\phi\right) = \sum^{6}_{i=0} a_{i}cos^{i}\phi
462: \end{equation}
463: We provide here a procedure
464: to obtain the Hessian elements, which is easy to program. To our knowledge, such a scheme has
465: not been outlined so far~\cite{case}, and hence we provide it here for pedantic reasons.
466: We limit these details to contributions from
467: the torsional interactions, the reciprocal space part and the real space
468: part of the Ewald sum.
469: Contributions from other terms in the potential energy are much simpler to
470: derive and are not given here.
471:
472: \subsection{Hessian from the torsional interaction}
473: The torsional angle $\phi$ between the planes formed by the bond vectors
474: $\vec{r}_{12}$, $\vec{r}_{23}$, and $\vec{r}_{34}$ is,
475: \begin{equation}
476: cos\phi = \hat{A} \cdot \hat{B} = A_{\gamma}B_{\gamma}
477: \label{eq: cosphi}
478: \end{equation}
479: where the Einstein summation convention has been used and
480: \begin{equation}
481: \hat{A}= \frac{\vec{r}_{12} \times \vec{r}_{23}}{\bracevert \vec{r}_{12} \times \vec{r}_{23} \bracevert },~~~and~~~
482: \hat{B}= \frac{\vec{r}_{23} \times \vec{r}_{34}}{\bracevert \vec{r}_{23} \times \vec{r}_{34} \bracevert }
483: \end{equation}
484: The second derivative
485: of $U_{torsion}$ with respect to a spatial coordinate $\alpha_{n}$, where
486: $\alpha$ can be either x, y, or z and n=1, 2, ..., N, can be written as,
487: \begin{equation}
488: \frac{\partial ^{2}U_{torsion}} {\partial \beta_{m}\partial \alpha_{n}}=\left(\frac{\partial f} {\partial cos\phi}\right)\left(\frac{\partial ^{2}cos\phi} {\partial \beta_{m}\partial \alpha_{n}}\right) + \left(\frac{\partial cos\phi}{\partial \alpha_{n}}\right)\left(\frac{\partial cos\phi}{\partial \beta_{m}}\right)\left(\frac{\partial ^{2}f} {\partial cos\phi ^{2}}\right)
489: \label{eq: 2U}
490: \end{equation}
491: The derivates of $cos\phi$ can be calculated as follows.
492: \begin{equation}
493: \frac{\partial ^2 cos\phi}{\partial \beta_{m} \partial \alpha_{n}} = A_{\gamma}\frac{\partial ^2 B_{\gamma}}{\partial \beta_{m} \partial \alpha_{n}} + \frac{\partial B_{\gamma}}{\partial \alpha_{n}}\frac{\partial A_{\gamma}}{\partial \beta_{m}} + B_{\gamma}\frac{\partial ^2 A_{\gamma}}{\partial \beta_{m} \partial \alpha_{n}} + \frac{\partial A_{\gamma}}{\partial \alpha_{n}}\frac{\partial B_{\gamma}}{\partial \beta_{m}}
494: \label{eq: 2cosphi}
495: \end{equation}
496:
497: We can write $A_{\gamma}$ as
498: \begin{equation}
499: A_{\gamma} = \frac{N_{\gamma}^{A}}{D_{A}}
500: \end{equation}
501: with
502: \begin{equation}
503: N_{\gamma}^{A} = \epsilon_{\gamma\nu\xi}\left(\vec{r}_{12}\right)_{\nu}\left(\vec{r}_{23}\right)_{\xi}
504: \end{equation}
505: and
506: \begin{equation}
507: D_{A} = \left[ \sum_{l}\left(N_{l}^{A}\right)^{2}\right]^{\frac{1}{2}}
508: \end{equation}
509: where $\gamma$, $\nu$, and $\xi$ stand for any of the three indices x, y, z and $\epsilon_{\gamma\nu\xi}$
510: is the antisymmetric Levi-Civita tensor of rank 3. A similar expression can be written
511: for $B_{\gamma}$.
512:
513: The first derivative of $A_{\gamma}$ are,
514: \begin{equation}
515: \frac{\partial A_{\gamma}}{\partial \alpha_{n}} =
516: \frac{1}{D_{A}}\frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}} -
517: \frac{N_{\gamma}^{A}}{D_{A}^{3}}\sum_{l} N_{l} \frac{\partial N_{l}^{A}}{\partial \alpha_{n}}
518: \label{eq: 1A-gamma}
519: \end{equation}
520: where,
521: \begin{equation}
522: \frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}} = \sum_{\nu = x,y,z}\epsilon_{\gamma\nu\alpha}\left[\left(\vec{r}_{12}\right)_{\nu}\left(\delta_{n3}-\delta_{n2}\right) - \left(\vec{r}_{23}\right)_{\nu}\left(\delta_{n2}-\delta_{n1}\right)\right]
523: \end{equation}
524: The second derivative of $A_{\gamma}$ is calculated as follows.
525: \begin{equation}
526: \frac{\partial}{\partial \beta_{m}}\frac{\partial A_{\gamma}}{\partial \alpha_{n}} =
527: \frac{\partial \left[\frac{1}{D_{A}}\frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}}\right]}{\partial \beta_m}
528: -
529: \frac{\partial \left[\frac{N_{\gamma}^{A}}{D_{A}^{3}}\sum_{l} N_{l} \frac{\partial N_{l}^{A}}{\partial \alpha_{n}}\right]}
530: {\partial \beta_m}
531: \label{eq: 2A-gamma}
532: \end{equation}
533: with
534: \begin{equation}
535: \frac{\partial \left[\frac{1}{D_{A}}\frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}}\right]}{\partial \beta_m} =
536: = \frac{1}{D_{A}}\frac{\partial}{\partial \beta_{m}}\frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}} - \frac{1}{D_{A}^{2}}\frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}}\frac{\partial D_{A}}{\partial \beta_{m}}
537: \end{equation}
538: \begin{eqnarray}
539: \frac{\partial \left[\frac{N_{\gamma}^{A}}{D_{A}^{3}}\sum_{l} N_{l} \frac{\partial N_{l}^{A}}{\partial \alpha_{n}}\right]}
540: {\partial \beta_m}
541: & = & \frac{N_{\gamma}^{A}}{D_{A}^{3}}
542: \left[\sum_{l}\left[ N_{l}\frac{\partial ^{2}N_{l}}{\partial \beta_{m}
543: \partial \alpha_{n}} + \left(\frac{\partial N_{l}}{\partial \alpha_{n}}\right)
544: \left(\frac{\partial N_{l}}{\partial \beta_{m}}\right)\right]\right] \nonumber \\
545: & + & \left(\sum_{l}N_{l}\frac{\partial N_{l}}{\partial \alpha_{n}}\right)
546: \left(\frac{1}{D_{A}^{3}}\frac{\partial N_{\gamma}}{\partial \beta_{m}} -
547: \frac{3}{D_{A}^{4}}N_{\gamma}\frac{\partial D_{A}}{\partial \beta_{m}}\right)
548: \end{eqnarray}
549: where,
550: \begin{equation}
551: \frac{\partial}{\partial \beta_{m}}\frac{\partial N_{\gamma}^{A}}{\partial \alpha_{n}} = \epsilon_{\gamma\beta\alpha} \left[\left(\delta_{m2}-\delta_{m1}\right)\left(\delta_{n3}-\delta_{n2}\right)- \left(\delta_{m3}-\delta_{m2}\right)\left(\delta_{n2}-\delta_{n1}\right)\right]
552: \end{equation}
553: Using equations (\ref{eq: 2cosphi}), (\ref{eq: 1A-gamma}), and (\ref{eq: 2A-gamma}) we can calculate $\frac{\partial^{2}cos\phi}{\partial \beta_{m} \partial \alpha_{n}}$ which
554: can be substituted in equation (\ref{eq: 2U}) to get the torsional contribution to the Hessian.
555:
556: \subsection{Hessian from the Coulomb interaction}
557: The reciprocal space energy in the Ewald sum method is,
558: \begin{equation}
559: U_{reci} = \frac{1}{V\epsilon_{0}} \sum_{\vec{k} \neq 0}
560: \frac{e^{\frac{-k^{2}}{4 \zeta^2}}}{k^{2}}
561: \left(a_k^2 + b_k^2\right)
562: \end{equation}
563: where,
564: \begin{equation}
565: a_k = \sum_{j=1}^{N} q_{j} cos\left(\vec{k} \cdot \vec{r}_{j}\right), ~~~~and
566: ~~~~~~ b_k = \sum_{j=1}^{N} q_{j} sin\left(\vec{k} \cdot \vec{r}_{j}\right)
567: \end{equation}
568:
569: Here, $V$ denotes the volume of the simulation cell, $\zeta$ determines the width of the
570: Gaussian charge distribution centered on the point charges in the Ewald sum method, $\vec{k}$ denote
571: the reciprocal lattice vectors, and $\epsilon_0$, the permittivity of free space.
572:
573: It can be shown that
574:
575: \begin{eqnarray}
576: \frac{\partial ^{2}U_{reci}}{\partial \beta_{m} \partial \alpha_{n}}
577: & = & \frac{1}{V\epsilon_{0}} \sum_{\vec{k} \neq 0}
578: \frac{e^{\frac{-k^{2}}{4 \zeta^2}}}{k^{2}}
579: \left\{ 2q_{n}k_{\alpha}\left[-k_{\beta}\delta_{mn}\left(a_k cos\left(\vec{k}
580: \cdot \vec{r}_{n}\right) + b_k sin\left(\vec{k} \cdot \vec{r}_{n}\right)\right)\right.\right. \nonumber \\
581: & + & \left.\left. q_{m}k_{\beta} cos\left(\vec{k} \cdot \left(\vec{r}_{n} -
582: \vec{r}_{m}\right)\right)\right] \right\}
583: \end{eqnarray}
584:
585: The expression for the real space energy in the Ewald summation is,
586: \begin{equation}
587: U_{real} = \frac{1}{4\pi\epsilon_{0}}\sum_{i=1}^{N}\sum_{j>i}^{N}q_{i}q_{j}\rho
588: \end{equation}
589: where,
590: \begin{equation}
591: \rho = \frac{erfc\left(\zeta r_{ij}\right)}{r_{ij}}
592: \end{equation}
593: The second derivatives of $U_{real}$ are
594: \begin{equation}
595: \frac{\partial}{\partial \beta_{m}}\left(\frac{\partial U_{real}}{\partial \alpha_{n}}\right) = \frac{1}{4\pi\epsilon_{0}}\sum_{i=1}^{N}\sum_{j>i}^{N}q_{i}q_{j}\frac{\partial}{\partial \beta_{m}}\left(\frac{\partial \rho}{\partial \alpha_{n}}\right)
596: \label{eq: 2U-real}
597: \end{equation}
598: with
599: \begin{equation}
600: \frac{\partial}{\partial \beta_{m}}\left(\frac{\partial \rho}{\partial \alpha_{n}}\right) = \left(\frac{\partial \rho}{\partial r_{ij}}\right) \frac{\partial}{\partial \beta_{m}}\left(\frac{\partial r_{ij}}{\partial \alpha_{n}}\right) + \left(\frac{\partial r_{ij}}{\partial \alpha_{n}}\right) \frac{\partial}{\partial \beta_{m}}\left(\frac{\partial \rho}{\partial r_{ij}}\right)
601: \end{equation}
602:
603: \begin{equation}
604: \frac{\partial}{\partial \beta_{m}}\left(\frac{\partial r_{ij}}{\partial \alpha_{n}}\right) = \frac{\left(\delta_{jn}-\delta_{in}\right)\left(\delta_{jm}-\delta_{im}\right)}{r_{ij}}\left[\delta_{\alpha \beta}-\frac{\left( \alpha_{j}-\alpha_{i}\right)\left( \beta_{j}-\beta_{i}\right)}{r_{ij}^{2}}\right]
605: \end{equation}
606:
607: It can be shown that
608: \begin{equation}
609: \frac{\partial \rho}{\partial r_{ij}} = -\left(s_{1} + s_{2}\right)
610: \end{equation}
611: and
612: \begin{equation}
613: \frac{\partial^{2} \rho}{\partial r_{ij}^{2}} = 2\zeta^{2}s_{1}r_{ij} + \frac{2\left(s_{1} + s_{2}\right)}{r_{ij}}
614: \label{eq: 2rho-rij}
615: \end{equation}
616: with,
617: \begin{equation}
618: s_{1} = \frac{2\zeta}{\sqrt \pi} \frac{e^{-\zeta^{2}r_{ij}^{2}}}{r_{ij}}
619: ~~~~~~and ~~~~~~~ s_{2} = \frac{erfc\left(\zeta r_{ij}\right)}{r_{ij}^{2}}
620: \end{equation}
621:
622: These can be used in Eq.~\ref{eq: 2U-real} to obtain the Hessian elements.
623:
624: \def\cpl{Chem. Phys. Lett.}
625: \newpage
626: \begin{thebibliography}{99}
627: \bibitem{halley_power} J.W. Halley, and Y. Duan, J. Power Sources {\bf 89}, 139 (2000).
628: \bibitem{wakihara} M. Wakihara, Mat. Sci. Engg. {\bf R33}, 109 (2001).
629: \bibitem{brucebook} P.G. Bruce, {\em Solid State Electrochemistry}, Cambridge University Press, Cambridge, 1995.
630: \bibitem{bruce1} P.G. Bruce, and C.A. Vincent, J. Chem. Soc. Faraday Trans. {\bf 89}, 3187 (1993).
631: \bibitem{angell_jpcm} X.G. Sun, W. Xu, S.S. Zhang, and C.A. Angell, J. Phys. Cond. Matt. {\bf 13},
632: 8235 (2001).
633: \bibitem{donoso} J.P. Donoso, T.J. Bonagamba, H.C. Panepucci, L.N. Oliveira, W. Gorecki, C. Berthier and M. Armand, J. Chem. Phys. {\bf 98}, 10026 (1993).
634: \bibitem{boden} I.M. Ward, N. Boden, J. Cruickshank, and S.A. Leng,
635: Electrochim. Acta. {\bf 40}, 2071 (1995).
636: \bibitem{qens} G. Mao, R.F. Perea, W.S. Howells, D.L. Price, and M.-L. Saboungi, Nature
637: (London) {\bf 405}, 163 (2000);
638: M.-L. Saboungi, D.L. Price, G. Mao, R.F. Perea, O. Borodin, and G.D. Smith,
639: M. Armand, and W.S. Howells, Solid State Ionics {\bf 147}, 225 (2002);
640: \bibitem{angell} C.A. Angell, Solid State Ionics {\bf 9-10}, 3 (1983).
641: \bibitem{berthier} C. Berthier, W. Gorecki, M. Minier, M.B. Armand, J.M. Chanbagno,
642: and P. Rigaud, Solid State Ionics {\bf 11}, 91 (1983).
643: \bibitem{gorecki} W. Gorecki, P. Donoso, C. Berthier, M. Mali, J. Roos,
644: D. Brinkmann, M.B. Armand, Solid State Ionics {\bf 28-30},1018 (1988).
645: \bibitem{bruce2} Z. Gadjourova, Y.G. Andreev, D.P. Tunstall, and P.G. Bruce, Nature {\bf 412}, 520 (2001)
646: \bibitem{leeuw} B. Mos, P. Verkerk, A. van Zon, and S.W. de Leeuw, Physica B {\bf 276}, 351 (2000);
647: S.W. de Leeuw, A. Van Zon, and G.J. Bel, Electrochim. Acta {\bf 46}, 1419 (2001);
648: J.J. de Jonge, A. van Zon, and S.W. de Leeuw, Solid State Ionics {\bf 147}, 349 (2002).
649: \bibitem{sylvie-94} S. Neyertz, D. Brown, and J.O. Thomas, J. Chem. Phys. {\bf 101}, 10064 (1994).
650: \bibitem{thesis} S. Neyertz, Ph.D. thesis, Uppsala University, Uppsala, 1995.
651: \bibitem{neyertzmelt} S. Neyertz, and D. Brown, J. Chem. Phys. {\bf 102}, 9725 (1995).
652: \bibitem{sylvie-acta95} S. Neyertz, D. Brown, and J.O. Thomas, Electrochim. Acta {\bf 40}, 2063 (1995).
653: \bibitem{sylvie-cps95} S. Neyertz, D. Brown, and J.O. Thomas, Comput. Polym. Sci. {\bf 5}, 107 (1995).
654: \bibitem{florian-95} F. Muller-Plathe, and W.F. van Gunsteren, J. Chem. Phys. {\bf 103}, 4745 (1995).
655: \bibitem{laasonen} K. Laasonen, and M.L. Klein, J. Chem. Soc. Faraday Trans. {\bf 91}, 2633 (1995).
656: \bibitem{halley2001} J.W. Halley, Y. Duan, B. Nielsen, P.C. Redfern, and L.A. Curtiss,
657: J. Chem. Phys. {\bf 115}, 3957 (2001); B. Lin, P.T. Boinske, and J.W. Halley, J. Chem. Phys.
658: {\bf 105}, 1668 (1996).
659: \bibitem{smith_peo} P. Ahlstr$\ddot{o}$m, G. Wahnstr$\ddot{o}$m, P. Carlsson, O. Borodin,
660: and G.D. Smith, J. Chem. Phys. {\bf 112}, 10669 (2000).
661: \bibitem{smith} O. Borodin, and G.D. Smith, Macromolecules {\bf 31}, 8396 (1998); Macromolecules {\bf 33}, 2273 (2000);
662: \bibitem{smith_h2o} O. Borodin, F. Trouw, D. Bedrov, and G.D. Smith, J. Phys. Chem. B {\bf 106}, 5184 (2002);
663: O. Borodin, D. Bedrov, and G.D. Smith, J. Phys. Chem. B {\bf 106}, 5194 (2002);
664: G.D. Smith, D. Bedrov, and O. Borodin, Phys. Rev. Lett. {\bf 85}, 5583 (2000).
665: \bibitem{smith_spectrochim} G.D. Smith, O. Borodin, M. Pekny, B. Annis, D. Londono, and R.L. Jaffe,
666: Spectrochim. Acta A {\bf 53}, 1273 (1997).
667: \bibitem{carini} G. Carini, G. D'Angelo, G. Tripodo, A. Bartolotta, and G. Di Marco
668: Phys. Rev. B {\bf 54}, 15056-15063 (1996).
669: \bibitem{mustarelli} P. Mustarelli, C. Capiglia, E. Quartarone, C. Tomasi,
670: P. Ferloni, and L. Linati, Phys. Rev. B {\bf 60}, 7228 (1999).
671: \bibitem{elliott_papers} S.I. Simdyankin, M. Dzugutov, S.N. Taraskin, and S.R. Elliott,
672: Phys. Rev. B {\bf 63}, 184301 (2001); S.N. Taraskin, and S.R. Elliott, Phys. Rev. B
673: {\bf 61}, 12017 (2000); S.N. Taraskin, and S.R. Elliott, Phys. Rev. B {\bf 59}, 8572 (1999).
674: \bibitem{nagel} C.S. O'Hern, S.A. Langer, A.J. Liu, and S.R. Nagel, Phys. Rev. Lett. {\bf 86},
675: 111 (2001).
676: \bibitem{hess_avg} K. Fukui, B.G. Sumpter, D.W. Noid, C. Yang, and R.E. Tuzun,
677: J. Phys. Chem. B, {\bf 104}, 526 (2000); D.W. Noid, K. Fukui, B.G. Sumpter, C. Yang, and R.E. Tuzun, \cpl {\bf 316}, 285 (2000).
678: \bibitem{karplus} H.W.T. van Vlijmen and M. Karplus, J. Phys. Chem. B {\bf 103}, 3009 (1999).
679: \bibitem{verma_jpcb} N.P. Barton, C.S. Verma, L.S.D. Caves, J. Phys. Chem. B {\bf 107}, 2170 (2003);
680: J. Phys. Chem. B {\bf 106}, 11036 (2002).
681: \bibitem{takahashi} Y. Takahashi and H. Tadokoro, Macromolecules {\bf 6}, 672 (1973).
682: \bibitem{martyna} G.J. Martyna, M.L. Klein, and M. Tuckerman, J. Chem. Phys. {\bf 97}, 2635 (1992).
683: \bibitem{piny-md} M.E. Tuckerman, D.A. Yarne, S.O. Samuelson, A.L. Hughes,
684: and G. Martyna, Comput. Phys. Commun. {\bf 128}, 333 (2000).
685: \bibitem{hansen} J.-P. Hansen, in {\em Molecular Dynamics Simulations of Statistical Mechanics Systems},
686: Ed. G. Ciccotti, and W.G. Hoover, (North-Holland, Amsterdam, 1986).
687: \bibitem{charmm} A.D. MacKerell Jr. {\em et al}, J. Phys. Chem. B
688: {\bf 102}, 3586 (1998).
689: \bibitem{abramovitz} M. Abramowitz, and I.A. Stegun, {\em Handbook of
690: Mathematical Functions}, (Dover, New York, 1970).
691: \bibitem{elliott} S.N. Taraskin, and S.R. Elliott, Phys. Rev. B. {\bf 56}, 8605 (1997).
692: \bibitem{parr-rahman} M. Parrinello, and A. Rahman, Phys. Rev. Lett. {\bf 45}, 1196 (1980).
693: \bibitem{rabolt} J.F. Rabolt, K.W. Johnson, and R.N. Zitter, J. Chem. Phys. {\bf 61}, 504, (1974).
694: \bibitem{dacosta} V.M. Da Costa, T.G. Fiske, and L.B. Coleman, J. Chem. Phys. {\bf 101}, 2746, (1994).
695: \bibitem{branca} C. Branca, A. Faraone, S. Magazu, G. Maisano, P. Migliardo, and V. Villari,
696: J. Mol. Liquids {\bf 87}, 21 (2000).
697: \bibitem{yoshihara} T. Yoshihara, H. Tadokoro, S. Murahashi, J. Chem. Phys. {\bf 41}, 2902 (1964).
698: \bibitem{lam} K. Song, and S. Krimm, J. Polym. Sci. Polym. Phys. Ed. {\bf 28}, 35, (1990).
699: \bibitem{case} D.A. Case, Curr. Opin. Struct. Biol. {\bf 4}, 285 (1994).
700: \end{thebibliography}
701:
702: \newpage
703:
704:
705: \begin{table}
706: \caption{Parameters of the interaction potential~\cite{sylvie-94,charmm}.}
707: \begin{center}
708: \begin{tabular}{|c|c|c|c|c|c|c|c|}
709: \hline
710: Stretch & \multicolumn{3}{c|}{$r_{0}$ [\AA]} & \multicolumn{4}{c|}{$k_{\rm r}$ [K \AA$^{-2}$]} \\ \hline
711: C-C & \multicolumn{3}{c|}{1.53} & \multicolumn{4}{c|}{237017.0} \\
712: C-O & \multicolumn{3}{c|}{1.43} & \multicolumn{4}{c|}{171094.8} \\
713: C-H & \multicolumn{3}{c|}{1.09} & \multicolumn{4}{c|}{Constrained} \\
714: H-H & \multicolumn{3}{c|}{1.78} & \multicolumn{4}{c|}{Constrained} \\
715: \hline
716: \multicolumn{8}{|c|}{} \\
717: \hline
718: Bend & \multicolumn{3}{c|}{$\theta_{0}$} & \multicolumn{4}{c|}{$k_{\theta}$ [K rad$^{-2}$]} \\
719: \hline
720: C-O-C &\multicolumn{3}{c|}{112$^{o}$} &\multicolumn{4}{c|}{110255.50}\\
721: O-C-C &\multicolumn{3}{c|}{110$^{o}$} &\multicolumn{4}{c|}{76942.34}\\
722: O-C-H &\multicolumn{3}{c|}{109.5$^{o}$} &\multicolumn{4}{c|}{30193.20}\\
723: H-C-C &\multicolumn{3}{c|}{110$^{o}$} &\multicolumn{4}{c|}{45199.50}\\
724: \hline
725: \multicolumn{8}{|c|}{} \\
726: \hline
727: Torsions & $a_0$ [K] & $a_1$ [K] & $a_2$ [K]& $a_3$ [K] & $a_4$ [K]& $a_5$ [K]&
728: $a_6$ [K] \\
729: \hline
730: C-C-O-C & -50.322 & 150.966 & 0.0 & -201.288 & 0.0 & 0.0 & 0.0 \\
731: C-O-C-H & -50.322 & 150.966 & 0.0 & -201.288 & 0.0 & 0.0 & 0.0 \\
732: H-C-C-H & 83.03 &-249.090 & 0.0 & 332.120 & 0.0 & 0.0 & 0.0 \\
733: O-C-C-O & 265.70 &-1826.19 & 2144.72 & 3901.46 & -1667.17 & 142.91 & 1480.98\\
734: H-C-C-O & 98.128 &-294.384 & 0.0 & 392.512 & 0.0 & 0.0 & 0.0 \\
735: \hline
736: \multicolumn{8}{|c|}{} \\
737: \hline
738: Non-bonded & \multicolumn{3}{c|}{A [K]} & B [$\AA^{-1}$] &\multicolumn{3}{c|}{C [K$\AA^{6}$]} \\
739: \hline
740: C...C & \multicolumn{3}{c|}{15909350.62} & 3.3058 & \multicolumn{3}{c|}{325985.916}\\
741: C...O & \multicolumn{3}{c|}{21604039.75} & 3.6298 & \multicolumn{3}{c|}{177536.016}\\
742: C...H & \multicolumn{3}{c|}{7571800.37} & 3.6832 & \multicolumn{3}{c|}{91334.430} \\
743: O...O & \multicolumn{3}{c|}{29337172.46} & 4.0241 & \multicolumn{3}{c|}{96668.562} \\
744: O...H & \multicolumn{3}{c|}{10282092.97} & 4.0900 & \multicolumn{3}{c|}{49718.136} \\
745: H...H & \multicolumn{3}{c|}{3603659.064} & 4.1580 & \multicolumn{3}{c|}{25563.576} \\
746: \hline
747: \multicolumn{8}{|c|}{} \\
748: \hline
749: \multicolumn{8}{|c|}{Atomic charges [e]} \\
750: \hline
751: \multicolumn{4}{|c|}{q$_{\rm C}$} & \multicolumn{4}{c|}{0.103} \\
752: \multicolumn{4}{|c|}{q$_{\rm O}$} & \multicolumn{4}{c|}{-0.348} \\
753: \multicolumn{4}{|c|}{q$_{\rm H}$} & \multicolumn{4}{c|}{0.0355} \\
754: \hline
755: \end{tabular}
756: \end{center}
757: \end{table}
758:
759: \begin{table}
760: \caption{Lattice parameters obtained from simulation compared to experiment~\cite{takahashi}.}
761: \begin{center}
762: \begin{tabular}{|c|c|c|}\hline
763: Lattice parameter & Simulation & Experiment \\ \hline
764: a $[\AA]$ & 8.08 & 8.05 \\
765: & & \\
766: b $[\AA]$ & 13.17 & 13.04 \\
767: & & \\
768: c $[\AA]$ & 18.45 & 19.48 \\
769: & & \\
770: $\alpha [^{o}]$ & 89.98 & 90.0 \\
771: & & \\
772: $\beta [^{o}]$ & 123.01 & 125.40 \\
773: & & \\
774: $\gamma [^{o}]$ & 89.99 & 90.0 \\ \hline
775: \end{tabular}
776: \end{center}
777: \end{table}
778: \clearpage
779:
780: \newpage
781: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
782: \begin{subfigures}
783: \begin{figure}
784: \centerline{\psfig{figure=fig1a.eps,height=5.5in,angle=270}}
785: \vspace*{1.0cm}
786: \caption{}
787: \end{figure}
788:
789: \newpage
790:
791: \begin{figure}
792: \centerline{\psfig{figure=fig1b.eps,height=5.5in,angle=270}}
793: \vspace*{1.0cm}
794: \caption{}
795: \vspace*{1.0cm}
796: {\bf Figure 1:}~~~Plot of instantaneous cell parameters of PEO crystal obtained from simulations :
797: (a) cell lengths, (b) cell angles. The zero time configuration corresponds to
798: the experimental crystal structure.
799: \end{figure}
800: \end{subfigures}
801: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
802:
803: \newpage
804: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
805: \begin{subfigures}
806: \begin{figure}
807: \centerline{\psfig{figure=fig2a.eps,height=5.2in,angle=270}}
808: \vspace{1.0cm}
809: \caption{}
810: \end{figure}
811:
812: \newpage
813:
814: \begin{figure}
815: \centerline{\psfig{figure=fig2b.eps,height=5.0in,angle=270}}
816: \vspace{1.0cm}
817: \caption{}
818: \end{figure}
819:
820: \newpage
821: \begin{figure}
822: \centerline{\psfig{figure=fig2c.eps,height=5.0in,angle=270}}
823: \caption{}
824:
825: \vspace*{0.5cm}
826: {\bf Figure 2:}~~~(a) The vibrational density of states of a single molecule (bottom), for one unitcell (middle) and for the
827: $\langle422\rangle$ crystal (top) of PEO, each obtained from normal mode analysis.
828: The two sections of each
829: spectrum are normalized independently to enable better comparison among the three system sizes.
830: The spectra in Figures~2a, 2b, and 2c are convoluted with a Gaussian function of width 4cm$^{-1}$.
831: (b) The vibrational density of states of the $\langle422\rangle$ crystal, obtained by NMA shown in expanded scale.
832: (c) Vibrational density of states obtained from the NMA method (bottom panels)
833: are compared with experimental Raman (Dashed lines) and Infra-red (Continuous lines)
834: absorption intensities.
835: Experimental data presented in the top left and top right panels
836: were obtained from Ref.~\cite{rabolt} and Ref.~\cite{yoshihara} respectively.
837: Note that the simulated spectrum is the raw
838: density of states and does not contain any other terms that are needed to calculate the
839: experimental absorption spectra. Thus only the peak positions are comparable, and not their intensities.
840: \end{figure}
841: \end{subfigures}
842: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
843:
844: \newpage
845: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
846: \begin{figure}
847: \centerline{\psfig{figure=fig3.eps,height=4.5in,angle=270}}
848: \vspace*{2.0cm}
849: \caption{
850: The vibrational density of states of $\langle422\rangle$ PEO crystal obtained from normal mode
851: analysis (top panels) is compared with that obtained as the power spectrum of the
852: velocity autocorrelation function of all atoms (bottom panels).
853: The two sections of each spectrum are normalized independently to enable better
854: comparison among the two methods.
855: The spectra are convoluted with a Gaussian function of width 4cm$^{-1}$.}
856: \end{figure}
857: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
858:
859: \newpage
860: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
861: \begin{subfigures}
862: \begin{figure}
863: \centerline{\psfig{figure=fig4a.eps,height=5.0in,angle=270}}
864: \vspace{0.5cm}
865: \caption{}
866: \end{figure}
867:
868: \newpage
869: \begin{figure}
870: \centerline{\psfig{figure=fig4b.eps,height=5.0in,angle=270}}
871: \vspace*{0.5cm}
872: \caption{}
873: \vspace*{0.5cm}
874:
875: {\bf Figure 4:}~~~~(a),(b) : Atomic displacements of representative normal modes of the PEO crystal.
876: For each mode, one of the chains which exhibit
877: significant atomic displacement is shown, for clarity. Red (or Black) spheres denote carbon atoms and
878: white spheres denote oxygens. Hydrogen atoms are not shown for clarity in all the modes, except the
879: one with frequency 1244~cm$^{-1}$. Arrows denote the directions of atomic displacements and their
880: lengths are scaled up for the purpose of visualization. The mode
881: at 0~cm$^{-1}$ is the rigid body translation of the chain, and that at
882: 1244~cm$^{-1}$ arises from wagging of CH$_2$ groups. Frequencies in cm$^{-1}$ are as
883: provided in the figures. The representation of these modes in the D$_7$ group, are as
884: follows, with frequencies in cm$^{-1}$ given in paranthesis: A$_2$ (38), A$_2$ (44),
885: E (88), A$_1$ (216), E (510), E (952), and E (1244).
886: \end{figure}
887: \end{subfigures}
888: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
889:
890: \newpage
891: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
892: \begin{figure}
893: \centerline{\psfig{figure=fig5.eps,height=4.5in,angle=270}}
894: \vspace*{2.0cm}
895: \caption{
896: Local character indicator for the normal modes of PEO crystal. The inset
897: shows the frequency range where
898: collective motion is most probable with very small local character indicator values.}
899: \end{figure}
900: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
901:
902: \newpage
903: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
904: \begin{figure}
905: \centerline{\psfig{figure=fig6.eps,height=4.5in,angle=270}}
906: \vspace*{2.0cm}
907: \caption{
908: The segment size distribution, f(n), corresponding to the mode at 44~cm$^{-1}$ of the PEO crystal.
909: The distribution is shown for four different displacement cutoffs.
910: Such distributions are used to calculate the average continuous segment size ($<$CSS$>$) which
911: is defined in equation (7).}
912: \end{figure}
913: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
914:
915: \newpage
916: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
917: \begin{figure}
918: \centerline{\psfig{figure=fig7.eps,height=4.5in,angle=270}}
919: \vspace*{2.0cm}
920: \caption{
921: Variation of the average continuous segment size ($<$CSS$>$) with respect to the
922: vibrational frequencies for different displacement cutoffs.}
923: \end{figure}
924: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
925:
926: \end{document}
927: