1: %% \documentstyle[aps,multicol,epsf,citesort]{revtex}
2: %% \documentstyle[aps,twocolumn,epsf]{revtex}
3: %% \documentstyle[aps,preprint,epsf,tighten,citesort]{revtex}
4: %% \draft
5: %% \renewcommand{\baselinestretch}{2}
6: %% \documentstyle[preprint,aps]{revtex}
7: %% \draft
8: %% \tightenlines
9: %%
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12:
13:
14: \documentstyle[aps,preprint,epsf,citesort]{revtex}
15: %\documentstyle[aps,multicol,epsf,citesort]{revtex}
16:
17:
18: \begin{document}
19:
20: \title{Topological Defects in Size-Dispersed Solids}
21:
22: \author{M. Reza Sadr-Lahijany,$^1$ Purusattam Ray,$^{2,1}$
23: and H. E. Stanley$^1$}
24:
25: \address{$^1$ Center for Polymer Studies and Department of Physics\\
26: Boston University, Boston, Massachusetts 02215\\
27: $^2$ The Institute of Mathematical Sciences, CIT Campus, Chennai 600
28: 113, India}
29:
30: \date{submitted to Phys. Rev. Letters 7 Jan 2000}
31:
32: \maketitle
33:
34: \begin{abstract}
35:
36: We study the behavior of the topological defects in the inherent
37: structures of a two-dimensional binary Lennard-Jones system as the size
38: dispersity varies. We find that topological defects arising from the
39: particle size dispersity are responsible for destabilizing the solid as
40: follows: (i) for particle density $\rho \leq 0.9$, the solid melts
41: through intermediate states of decreasing hexatic order arising from the
42: proliferation of unbounded dislocations, (ii) for $\rho > 0.9$, the
43: dislocations form grain boundaries, dividing the system into
44: micro-crystallites and destroying the translational and orientational
45: order.
46:
47: \end{abstract}
48:
49: \bigskip\bigskip
50:
51: %\begin{multicols}{2}
52:
53: Topological defects play a crucial role in melting of a solid,
54: especially in two dimensions. In two dimensions (2D) these defects,
55: present in bound pairs at low temperature solid phase, are believed to
56: unbind and destroy the crystalline order as temperature is raised
57: causing melting~\cite{nelson2}. Like temperature, size-dispersity
58: (inhomogeneity in particle size) {\it disfavors\/} crystalline order,
59: and can even convert a solid to a liquid~\cite{Barrat-Hansen,srs}. Here
60: we study the defect morphology in the inherent structures of a 2D
61: Lennard-Jones system with a bimodal distribution of particle sizes.
62:
63: We simulate $N=10^4$ particles interacting with a truncated
64: ``shifted-force Lennard-Jones'' pair-potential in
65: 2D. We choose half the particles to be smaller than the rest, and
66: define the size dispersity $\Delta$ to be the ratio of the
67: difference in their sizes to the mean size. We start by placing the
68: particles randomly on the sites of a triangular lattice, embedded in a
69: rectangular box of edges $L_x$ and $L_y$ with aspect ratio
70: $L_x/L_y=\sqrt 3/2$ (to accommodate close-pack hexagonal structure
71: without distortion). We apply periodic boundary condition and use the
72: velocity Verlet method to integrate Newton's equation of motion. Units
73: are set by choosing the mass of the particles and the Lennard-Jones (LJ)
74: energy and length scales to be unity. The density $\rho$ is
75: the ratio of the area occupied by the particles to the box area.
76:
77: We equilibrate a state, defined by $(\rho,\Delta)$, at a constant
78: temperature $T=1$ using Berendsen's thermostat; at this temperature the
79: 2D solid can form for low enough dispersity. We run our simulation at
80: constant energy until the temperature $T$, the pressure $P$, and the
81: energy $E$ stabilize with less than 1\% fluctuation (typically for
82: $2\times 10^5$ time steps, where the time step is 0.01 in LJ units). We
83: also check that the average particle displacement is at least a few
84: times the average particle size. To obtain higher density states, we
85: increase $\rho$ in steps of 0.01 from 0.85 (liquid state) to 1.05 by
86: gradually compressing the box, keeping the aspect ratio fixed, and
87: equilibrating the system at each of these densities.
88:
89: We perform defect analysis on 100 equilibrated configurations for each
90: of the state points ($\rho,\Delta$). Clear identification of geometrical
91: defects is difficult in simulations due to the presence of many
92: ``virtual defects'' which arise because of vibrational excitations. To
93: overcome this difficulty, we analyze the inherent structure of each
94: configuration \cite{stillinger}, obtained by removing the vibrational
95: excitations---or equivalently by minimizing (locally) the potential
96: energy using the conjugate gradient method.
97:
98: In order to find the defects, we construct a Voronoi cell around each
99: particle, thereby uniquely defining its nearest neighbors. The ordered
100: close-packed structure of the 2D solid is hexagonal, so each particle
101: $i$ at position $\mbox{\bf r}_i$ has $n_i=6$ neighbors. A defect arises
102: when a particle has $n_i\ne 6$, which generates a ``topological charge''
103: $q_i\equiv n_i-6$ \cite{clark}. Defects which are nearest neighbors are
104: grouped as defect clusters with total charge $Q\equiv\sum_i q_i$ and
105: total dipole moment $\mbox{\bf P}\equiv\sum_i \mbox{\bf r}_iq_i$, where
106: the sum is over all defects $i$ in the cluster. Figure~\ref{defsfig}
107: shows an example of defects in an otherwise ideal triangular lattice.
108:
109: We find that the defects fall into three categories: (i) {\it
110: Monopoles}. Clusters with $Q\neq 0$. The simplest case is a
111: disclination---a size-one cluster (Fig.~\ref{defsfig}a). (ii) {\it
112: Dipoles}. Clusters with $\mbox{\bf P}\neq 0$. The simplest case is a
113: dislocation, a size-two cluster (Figs.~\ref{defsfig}a
114: and~\ref{defsfig}b), composed of a ``bound pair'' of neighboring defects
115: of opposite charge. (iii) {\it Blobs}. Clusters with $Q=\mbox{\bf
116: P}=0$. The most common case is a quadrupole (Fig.~\ref{defsfig}a) made
117: of a ``bound pair'' of neighboring dislocations with oppositely-oriented
118: dipole moments (see, e.g., \cite{nelson3,Rubinstein}).
119:
120: Figure~\ref{r9defs} shows typical snapshots of the inherent structure at
121: $\rho=0.9$ for three values of $\Delta$. The corresponding phase points
122: ($\rho,\Delta$) represent solid, hexatic and liquid phases \cite{srs}.
123: In the low-dispersity solid phase [Figs.~\ref{r9defs}a and
124: \ref{r9defs}b], defects occur mostly in the form of blobs (quadrupoles).
125: Such defects, being without charge or dipole moment, cause little
126: distortion in the nearby order and so are energetically inexpensive. We
127: find that the defects in the inherent structure of the solid phase
128: aggregate to form domains separated by nearly defect-free regions,
129: suggesting that there is an effective attraction between defects. This
130: effective attraction may arise from packing constraints since in dense
131: packing, formation of a local large-amplitude defect is improbable. As
132: $\Delta$ increases, more defects in the form of dipoles are created.
133:
134: At large defect density, free dislocations appear. Figure~\ref{r9defs}c
135: shows their presence, in the hexatic phase. A dislocation destroys the
136: long-range translational order as it introduces an extra half row
137: (Fig.~\ref{defsfig}b) that can only terminate in another dislocation
138: with equal but opposite dipole moment. Translational order is destroyed
139: over the range of separation of the dislocation pair \cite{nelson2}.
140: Dislocations, however, retain orientational order. We find in
141: configurations such as Fig.~\ref{r9defs}c that 2D translational order is
142: lost (due to the abundance of dipoles). However, orientational order
143: shows an algebraic decay---the characteristic features of a hexatic
144: phase \cite{nelson2}. The system breaks up into crystalline patches of
145: finite length $\xi_t$ (the translational correlation length) which are
146: shifted but not rotated with respect to one another, so the range of
147: orientational correlation $\xi_6$ far exceeds $\xi_t$. On further
148: increasing $\Delta$, more defects are created: many monopoles appear
149: which destroy the orientational order and the system melts
150: [Fig.~\ref{r9defs}d] (see \cite{nelson3,Rubinstein}).
151:
152: In the hexatic phase, we find a steep increase in the number of dipoles
153: (Fig.~\ref{xi6ndef}a) between $\Delta_{\mbox{\scriptsize SH}}$ (the
154: solid-hexatic transition value for $\Delta$) and
155: $\Delta_{\mbox{\scriptsize HL}}$ (the hexatic-liquid transition value
156: for $\Delta$). Also, we detect a gentler increase in the number of
157: monopoles as the hexatic-liquid transition at
158: $\Delta=\Delta_{\mbox{\scriptsize HL}}$ is approached. KTHNY
159: theory~\cite{nelson2} predicts that in the liquid phase, the
160: orientational correlation function $C_6(r)$ decays exponentially,
161: $C_6(r) \sim e^{-r/\xi_6}$, with an orientational correlation length
162: $\xi_6$ that diverges as the liquid-hexatic transition is approached
163: \cite{nelson2}. Figure~\ref{xi6ndef}b reveals a rapid increase of
164: $\xi_6$ as the liquid-hexatic phase boundary is approached from the
165: liquid side.
166:
167: Figures~\ref{r1defs}b--d show typical snapshots of the inherent
168: structure at $\rho = 1.0$, for $\Delta = 0.04$, 0.1 and 0.12. A first
169: order solid-liquid transition is found at the value
170: $\Delta_{\mbox{\scriptsize SL}}\approx 0.1$ \cite{srs}, the solid-liquid
171: transition value for $\Delta$. Our defect analysis shows that for
172: $\Delta<\Delta_{\mbox{\scriptsize SL}}$, there are free
173: dislocations, and near the transition these defects line up in
174: ``strings'' to form long chains of large-angle grain boundaries
175: (Fig.~\ref{r1defs}c). As $\Delta\to\Delta_{\mbox{\scriptsize SL}}$,
176: these chains percolate, fragmenting the system into micro-crystallites
177: rotated with respect to each other. Thus the grain boundaries
178: simultaneously destroy both the translational and the rotational
179: order. In the theory of grain-boundary-induced melting
180: \cite{Halperin-Nelson,chui}, the Landau free-energy expansion yields a
181: first-order solid-liquid transition with the absence of any hexatic
182: phase. We find that the formation of defects and the proliferation of
183: grain boundaries occurs abruptly at $\Delta_{\mbox{\scriptsize
184: SL}}\approx 0.1$. Within the resolution of our simulations, we do not
185: see any hexatic phase \cite{srs}.
186:
187: In summary, we have seen that size-dispersity induces topological
188: defects which in turn destroy crystalline order, and that the mechanism
189: of dispersity-induced melting displays surprising parallels with the
190: mechanism proposed for the case of temperature-induced melting.
191: Depending on the value of $\rho$, dispersity-induced melting can be
192: either a first-order transition or a continuous transition (with an
193: intervening hexatic phase), and the defect morphologies display
194: completely different behavior in the two cases.
195:
196: \begin{references}
197:
198: \bibitem{nelson2}
199: D. R. Nelson, in {\it Phase Transitions and Critical Phenomena, Vol.~7},
200: edited by C. Domb \& J. L. Lebowitz (Academic, London, 1983), pp.~1.
201:
202: \bibitem{Barrat-Hansen}
203: J. L. Barrat and J. P. Hansen, J. Physique (France) {\bf 47}, 1547
204: (1986).
205:
206: \bibitem{srs}
207: M. R. Sadr-Lahijany, P. Ray and H. E. Stanley, Phys.~Rev.~Lett. {\bf
208: 79}, 3206 (1997).
209:
210: \bibitem{stillinger}
211: F. H. Stillinger and T. A. Weber, Science {\bf 225}, 983 (1984).
212:
213: \bibitem{clark}
214: M. A. Glaser and N. A. Clark, {\it Advances in Chemical Physics, Volume
215: LXXXIII\/} (John Wiley and Sons, New York, 1993), pp.~543.
216:
217: \bibitem{nelson3}
218: D. R. Nelson, M. Rubinstein and F. Spaepen, Philos. Mag. A {\bf 46}, 105
219: (1982).
220:
221: \bibitem{Rubinstein}
222: M. Rubinstein and D. R. Nelson, Phys. Rev. B {\bf 26}, 6254 (1982).
223:
224: \bibitem{Halperin-Nelson}
225: B. I. Halperin and D. R. Nelson,
226: Phys. Rev. Lett. {\bf 41}, 121 (1978)
227:
228: \bibitem{chui}
229: S. T. Chui, Phys. Rev. B {\bf 48}, 178 (1983).
230:
231: %\bibitem{seul}
232: %M. Seul and C. A. Murray, Science {\bf 262}, 558 (1993)
233:
234: %% \bibitem{allen-tildesley}
235: %% M. P. Allen and D. J. Tildesley, {\it Computer Simulation of Liquids\/}
236: %% (Oxford University Press, New York, 1989).
237:
238: %% \bibitem{PressXX}
239: %% W.~H.~Press, S.~A.~Teukolsky, W.~T.~Vetterling and B.~P.~Flannery, {\it
240: %% Numerical Recipes in C} (Cambridge University Press, Cambridge, 1992).
241:
242: %% \bibitem{nabarro}
243: %% F. R. N. Nabarro, {\it Theory of Crystal Dislocations\/} (Clarendon
244: %% Press, Oxford, 1967).
245:
246: %% \bibitem{Barker}
247: %% G. C. Barker and M. J. Grimson, J. Phys. Cond. Mat. {\bf 1}, 2779
248: %% (1989).
249:
250: %% \bibitem{Kosterlitz-Thoules}
251: %% J. M. Kosterlitz and D. J. Thoules, in
252: %% {\it Progress in Low-Temperature Physics, Vol. 7B}, edited by D. F.
253: %% Brewer (North-Holland, Amsterdam, 1979), pp.~371.
254:
255: %% \bibitem{Young}
256: %% A. P. Young,
257: %% Phys. Rev. B {\bf 19}, 1855 (1979)
258:
259: %% \bibitem{zahn}
260: %% K. Zahn, R. Lenke and G. Maret,
261: %% Phys. Rev. Lett. {\bf 82},
262: %% 2721 (1999).
263:
264: %% \bibitem{fisher}
265: %% D. S. Fisher, B. I. Halperin and R. Morf, Phys. Rev. B {\bf 20}, 4692
266: %% (1979).
267:
268: \end{references}
269:
270: %% Acknowledgment: We thank A.~Scala for seminal contributions in the
271: %% initial stages of this work, D.~R.~Nelson for extremely helpful
272: %% suggestions, and NSF for financial support.
273:
274: \begin{figure}[htb]
275: \narrowtext
276: \caption{
277: Identification of defects taken from our simulated system. Neutral
278: ($n_i=6$) particles are shown as dots. Solid squares and triangles
279: denote defects, with $n_i=7$ ($q=+1$) and $n_i=5$ ($q=-1$)
280: respectively. Neighboring defects form different types of defect
281: clusters as explained below. (a) The large open square denotes a
282: positive monopole. It represents a cluster of size one, which is the
283: positive defect it contains, thus it is a disclination. The large open
284: triangle denotes a negative monopole. It represents a cluster of size
285: three, which consists of two negative and one positive defects. The
286: arrows denote dipoles; all but the largest arrow (pointing ``East'')
287: represent clusters of size two, and hence all but one dipole are
288: dislocations. The largest arrow denotes a dipole, the vector sum
289: of two adjacent disclinations forming a cluster of size four with a
290: non-zero net dipole moment. Finally, the parallelogram contains a
291: cluster of two $n_i=5$ ($q=-1$) and two $n_i=7$ ($q=+1$) defects with
292: zero net charge and zero net dipole moment (hence a quadrupole). (b) A
293: dislocation dipole (solid arrow). Also shown is the corresponding
294: Burger's vector (dashed arrow), as the vector needed to close an
295: equilateral (in this case $3\times3$) contour around the
296: dislocation. The thin solid line marks the extra row which ends at the
297: dipole. The length of the dipole arrow represents the magnitude of the
298: dipole $\mbox {\bf P}$, proportional to the length of the
299: Burger's vector.}
300: \label{defsfig}
301: \end{figure}
302:
303: \begin{figure}[htb]
304: \caption{Defects of four different $\rho=0.9$ configurations.
305: The symbols used for the defects are same as in
306: Figure~\protect\ref{defsfig}. (a) A $\Delta=0.04$ configuration; this
307: configuration is in the solid phase. The magnified portion includes two
308: dislocations and a cluster of size $10$ (encircled by the dashed line)
309: with an overall dipole moment. (b) The inherent structure for the same
310: $\Delta=0.04$ configuration. Most defects in (a) and (b) form
311: quadrupoles, and some form dipoles. Notice that upon forming the
312: inherent structure of a configuration, the particles with six neighbors
313: move locally to form an ordered triangular lattice structure. Also many
314: of the defect clusters, mostly low energy quadrupoles and some dipoles
315: and monopoles, disappear. There remain only a few defect clusters,
316: topological defects that cannot be annihilated by local translations
317: performed in generating the inherent structure; e.g.
318: in the magnified region, most of the defects have disappeared; there
319: remains only one dislocation, the vector sum of the three
320: previous dipoles. (c) Inherent structure for $\Delta=0.06$
321: configuration (in the hexatic phase). The arrows
322: denote dipoles. Note that free dislocation dipoles appear. (d)
323: Inherent structure for $\Delta=0.09$ configuration
324: (in the liquid phase). Many disclination
325: monopole defects, shown as open triangles and squares, appear. For clarity,
326: individual defect particles and blobs (including
327: therefore quadrupoles) are not shown in c and d. Dipole clusters
328: are represented by the dipole vector for each cluster.}
329: \label{r9defs}
330: \end{figure}
331:
332: \begin{figure}[htb]
333: \caption{(a) The number of different defect clusters as a
334: function of $\Delta$ along the $\rho=0.9$ isochore. There is first a
335: ten-fold increase in the number of {\it dipoles\/} associated with the
336: solid-hexatic transition $\Delta_{\mbox{\scriptsize SH}}$. Then a large
337: increase in the number of {\it monopoles\/} occurs near the
338: hexatic-liquid transition $\Delta_{\mbox{\scriptsize HL}}$. (b)
339: Orientational correlation length $\xi_6$ as a function of $\Delta$ along
340: the same $\rho=0.9$ isochore. Upon approaching the hexatic phase from
341: the liquid phase (i.e., as $\Delta\to 0.06^+$), $\xi_6$ diverges.}
342: \label{xi6ndef}
343: \end{figure}
344:
345: \begin{figure}[htb]
346: \caption{Defects of the $\rho=1.0$ system. (a) A $\Delta=0.04$
347: configuration, which is in the solid phase. The defects are magnified in
348: the insets. This configuration contains only two dipoles and one low
349: energy quadrupole. Note that the leftmost and rightmost defects (two
350: dipoles) cancel, a feature characteristic of the solid phase. (b) The
351: quadrupole disappears upon generating the inherent structure.
352: (c) Inherent structure for a $\Delta=0.095$ configuration which is in
353: the solid phase near the melting point. (d) Inherent structure
354: for a $\Delta=0.115$ configuration, which is in the liquid phase. Note
355: that the defects form lines (grain boundaries) that partition the system
356: into sub-regions of solid order (micro-crystallites). Unlike the liquid
357: in Fig.~\protect\ref{r9defs}d, there are few disclination
358: monopoles. What destroys the solid order is the existence of the grain
359: boundaries, while in Fig.~\protect\ref{r9defs}d what destroys the solid
360: order is the existence of both free dislocations and free
361: disclinations.}
362: \label{r1defs}
363: \end{figure}
364:
365: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
366:
367: %\end{document}
368:
369: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
370:
371:
372: \begin{figure}[htb]
373: \centerline{
374: \vbox {
375: \epsfxsize=7.0cm
376: \epsfbox{mondipquad.black.eps}
377: }
378: }
379: \centerline{
380: \vbox {
381: \epsfxsize=7.0cm
382: \epsfbox{burgers.black.eps}
383: }
384: }
385: \end{figure}
386:
387: %\newpage
388:
389:
390: \begin{figure}[htb]
391: \narrowtext
392: \centerline{
393: \vbox {
394: \vspace*{0.1cm} \epsfxsize=7.0cm
395: \epsfbox{r9d4.black.eps} } }
396: \end{figure}
397:
398: %\newpage
399: \begin{figure}[htb]
400: \narrowtext
401: \centerline{
402: \vbox {
403: \vspace*{0.1cm} \epsfxsize=7.0cm
404: \epsfbox{r9d4is.black.eps} } }
405: \end{figure}
406:
407: %\newpage
408: \begin{figure}[htb]
409: \narrowtext
410: \centerline{
411: \vbox {
412: \vspace*{0.1cm} \epsfxsize=7.0cm
413: \epsfbox{r9d6is.black.eps} } }
414: \end{figure}
415:
416: %\newpage
417: \begin{figure}[htb]
418: \narrowtext
419: \centerline{
420: \vbox {
421: \vspace*{1cm} \epsfxsize=7.0cm
422: \epsfbox{r9d9is.black.eps} } }
423: \end{figure}
424: %\newpage
425:
426: \begin{figure}[htb]
427: \narrowtext
428: \centerline{
429: \hbox {
430: \epsfxsize=7.0cm \epsfbox{ndefr9.eps} }
431: }
432: \end{figure}
433: %\newpage
434: \begin{figure}[htb]
435: \narrowtext
436: \centerline{
437: \hbox {
438: \epsfxsize=7.0cm \epsfbox{xi6r9.eps} }
439: }
440: \end{figure}
441: %\newpage
442:
443: \begin{figure}[htb]
444: \narrowtext
445: \centerline{
446: \hbox {
447: \vspace*{1cm} \epsfxsize=7.0cm
448: \epsfbox{r1d4.black.eps} } }
449: \end{figure}
450: %\newpage
451: \begin{figure}[htb]
452: \narrowtext
453: \centerline{
454: \hbox {
455: \vspace*{1cm} \epsfxsize=7.0cm
456: \epsfbox{r1d4is.black.eps} } }
457: \end{figure}
458: %\newpage
459: \begin{figure}[htb]
460: \narrowtext
461: \centerline{
462: \hbox {
463: \vspace*{1cm} \epsfxsize=7.0cm
464: \epsfbox{r1d95is.black.eps} } }
465: \end{figure}
466: %\newpage
467: \begin{figure}[htb]
468: \narrowtext
469: \centerline{
470: \hbox {
471: \vspace*{1cm} \epsfxsize=7.0cm
472: \epsfbox{r1d115is.black.eps} } }
473: \end{figure}
474:
475: %\end{multicols}{2}
476:
477: \end{document}
478:
479: