1: %\documentclass[aps,twocolumn]{revtex4}
2: \documentclass[aps,twocolumn,showpacs,floatfix]{revtex4}
3:
4: \usepackage{graphicx}
5:
6: \begin{document}
7:
8: \title{Interface relaxation in electrophoretic deposition of polymer
9: chains: Effects of segmental dynamics, molecular weight, and
10: field}
11:
12: \author{Frank W. Bentrem, Jun Xie, and R.~B. Pandey}
13:
14: \affiliation{Department of Physics and Astronomy, The University of Southern Mississippi,
15: Hattiesburg, MS 39406-5046}
16:
17: \date{\today}
18:
19: \begin{abstract}
20:
21: Using different segmental dynamics and relaxation, characteristics of the
22: interface growth is examined in an electrophoretic deposition of polymer
23: chains on a three (2+1) dimensional discrete lattice with a Monte Carlo
24: simulation. Incorporation of faster modes such as crankshaft and
25: reptation movements along with the relatively slow kink-jump dynamics
26: seems crucial in relaxing the interface width. As the continuously
27: released polymer chains are driven (via segmental movements) and
28: deposited, the interface width $W$ grows with the number of time steps
29: $t$, $W \propto t^{\beta},$ ($\beta \sim 0.4$--$0.8)$, which is followed
30: by
31: its saturation to a steady-state value $W_s$. Stopping the release of
32: additional chains after saturation while continuing the segmental
33: movements relaxes the saturated width to an equilibrium value ($W_s \to
34: W_r$). Scaling of the relaxed interface width $W_r$ with the driving field
35: $E$, $W_r \propto E^{-1/2}$ remains similar to that of the steady-state
36: $W_s$ width. In contrast to monotonic increase of the steady-state width
37: $W_s$, the relaxed interface width $W_r$ is found to decay (possibly as a
38: stretched exponential) with the molecular weight.
39:
40: \end{abstract}
41:
42: \pacs{68.35.Ct, 61.41.+e, 81.15.Pq}
43:
44: \maketitle
45:
46: \section{Introduction}
47:
48: The deposition process \cite{barabasi, family} is one of the common
49: methods used for growing and designing composites, polymeric materials,
50: interfaces, and surface coating \cite{wool}.
51: As the polymer chains are driven toward an impenetrable
52: substrate/wall, the polymer density at the substrate grows and the
53: interface
54: develops \cite{bpf, fp1, fp2}. A number of parameters (e.g., temperature,
55: pressure/field, molecular weight of polymer) and processes (e.g., rate
56: of
57: polymer release, segmental dynamics, relaxation, etc.) affect the growth
58: of the
59: polymer density and its interface.
60: Recently we examined the growth of the interface, its scaling,
61: and roughness in an electrophoretic deposition model on a discrete
62: lattice
63: using kink-jump segmental dynamics \cite{bpf} and by including
64: reptation \cite{fp1, fp2}.
65: We have observed many interesting scaling behaviors of the interface
66: width
67: in these continuous deposition processes where the polymer chains
68: are released at a constant rate throughout the computer simulation.
69:
70: As the polymer chains deposit, the polymer density spreads from the
71: substrate
72: toward the source of the polymer chains, and the interface width
73: grows.
74: Even in such a continuous deposition process, the interface width $W$
75: reaches a steady-state (saturated) value ($W_s$) after an appropriate
76: growth period. The steady-state width $W_s$ exhibits interesting scaling
77: with
78: temperature, field, and molecular weight leading to
79: roughening and deroughening \cite{bpf, fp1, fp2}.
80:
81: In this article we examine the characteristics of the density profile
82: and
83: the interface width as we relax the system (which includes the polymer
84: bulk and interface). The study of relaxations at polymer
85: surfaces is relatively new \cite{liu97,wallace01}, and characteristic
86: surface relaxation times $\tau$ have been measured for some polymer materials
87: \cite{riemann96,riemann97,tanaka00}. One approach to achieve such a relaxation
88: in simulations is to stop injecting new polymer chains after an appropriate amount of
89: polymer chains are in the system but continue to allow segmental motion of
90: the chains.
91: Temperature $T$, field $E$, and molecular weight $L_c$ play
92: important
93: roles in relaxing the interface width; here we restrict the analysis to
94: the effects of
95: field
96: and molecular weight.
97: We also probe the relaxation by examining the effects of incorporating
98: the
99: faster modes of movements such as crankshaft and reptation along with
100: the
101: relatively slow kink-jump segmental dynamics \cite{binder}. In the next section
102: we
103: briefly describe the model followed by results and discussion, and
104: finally, a summary/conclusion.
105:
106: \section{Model}
107:
108: We briefly describe the model to point out the differences between the
109: simulation procedure adopted here for segmental dynamics and
110: relaxation
111: and our previous studies \cite{bpf, fp1, fp2}. We consider a discrete
112: lattice of size
113: $L_x \times L \times L$ with a large aspect ratio $L_x/L$. Typically,
114: $L_x = 100$--$200$, $L = 40$, $60$. Polymer chains each of length $L_c$,
115: generated
116: as the trail of a random walk of $L_c$ steps along the lattice with
117: excluded
118: volume constraints, are released from the $x=1$ end of the sample. The
119: lengths used here ($L$, $L_x$, $L_c$) are in units of the lattice
120: constants. An external
121: field $E$ drives the chains from the source near $x=1$ toward the
122: substrate
123: (impenetrable wall) at $x = L_x$.
124: The field couples with the change in energy, $E \Delta x$, where
125: $\Delta x$ is the displacement of the chain node along $x$ direction.
126: In addition to excluded volume, there is
127: a nearest-neighbor polymer-polymer repulsive and polymer-wall
128: attractive
129: interaction. Chains are released at a constant rate and moved with the
130: Metropolis algorithm \cite{metropolis53} using segmental dynamics such as kink-jump,
131: crankshaft \cite{verdier62}, or
132: slithering snake (reptation) or some combination \cite{binder}. Attempt to move each chain
133: node
134: once is defined
135: as one Monte Carlo step (MCS). The simulation is performed for a
136: relatively
137: large number of time steps and for a sufficient number of independent samples
138: to
139: obtain
140: a reliable estimate of averaged physical quantities.
141:
142: In our earlier studies \cite{bpf, fp1, fp2}, polymer chains are
143: released
144: throughout the simulation.
145: In this study, chains are released with a constant rate for a
146: sufficiently
147: long time (typically about three-fourths of the entire simulation time)
148: before
149: stopping the release of new chains into the system.
150: The simulation is then continued to
151: allow chains (already released into the system) to deposit and relax.
152: As discussed below, the relaxation makes a significant difference in
153: characteristics of the interface.
154:
155: \section{Results}
156:
157: \begin{figure}
158: \centerline{\includegraphics*[width=8.5cm]{fig1.eps}}
159: \caption{
160: $W$ (in units of the lattice constant) versus time steps $t$ (MCS) with
161: $K$ segmental
162: dynamics for different fields with 10--20 independent samples.
163: }
164: \label{fig:w_t_k}
165: \end{figure}
166:
167: Figures \ref{fig:w_t_k} and \ref{fig:w_t_kc} show the variation of the
168: interface width with the number of time steps
169: for
170: kink-jump ($K$) (Fig.~\ref{fig:w_t_k}) and kink-jump and crankshaft ($KC$) (Fig.~\ref{fig:w_t_kc})
171: segmental dynamics for different driving fields. Data with both
172: segmental
173: movements are generated by ($i$) depositing chains for about
174: three-fourths
175: of the entire simulation steps as mentioned above and ($ii$) relaxing
176: chains
177: already in the system (polymer bulk and the interface width) for
178: the last
179: one-fourth of the time steps without adding new chains.
180: We see a rapid growth of the interface width $W$ initially (time step
181: $t \le 10^3$) particularly at higher fields before reaching a
182: steady-state
183: value ($W_s$). At the time step, $t = 7.5 \times 10^4$, when the
184: injection of
185: new chains stops, the interface width decays rapidly and attains a
186: constant relaxed value ($W_r$). At low fields ($E = 0.07$), it takes
187: longer
188: for the interface to both grow and decay.
189:
190: \begin{figure}
191: \centerline{\includegraphics*[width=8.5cm]{fig2.eps}}
192: \caption{
193: $W$ (in units of the lattice constant) versus time steps $t$ (MCS) with
194: $KC$ segmental
195: dynamics for different fields with 10--20 independent samples.
196: }
197: \label{fig:w_t_kc}
198: \end{figure}
199:
200: Although the general features of growth and decay of the interface width
201: with $K$ and $KC$ movements appear similar, there are clear differences in
202: the relaxation. Obviously, it is faster for the chains and the interface
203: width to relax with the $KC$ dynamics than with the $K$ move alone. At low
204: field ($E = 0.07$), the interface width is not completely relaxed
205: (approach a constant value) with $K$ dynamics by the end of the simulation
206: run. In fact, a complete relaxation even at higher values of field ($E \ge
207: 0.6$) within the time of simulation with the $K$ dynamics is questionable.
208: Figure~\ref{fig:w_t_relax} shows the decay of the interface width from
209: steady-state to equilibrium with the $K$ dynamics during the relaxation
210: period. One can identify an exponential decay,
211:
212: $$W-W_r \sim e^{-{{t-t_0}\over\tau}},$$
213:
214: \noindent
215: where $t_0$ is the time for stopping the release of new chains, and
216: $\tau$ is the characteristic relaxation time for the interface
217: (determined by the slope).
218:
219: Initial
220: interface
221: growth $W \propto t^{\beta}$, ($\beta \sim 0.3$--1) has been
222: studied in detail \cite{bpf, fp1} and values of the growth exponent
223: $\beta$ are nearly the same with both $K$ and $KC$ dynamics. It is
224: clear,
225: though, that the relaxation of the interface width is much faster with
226: the
227: $KC$ dynamics.
228: However, it is because of the slow relaxation of the interface width with
229: the $K$ dynamics that we are able to comment on the exponential decay of the
230: interface width here.
231:
232: \begin{figure}
233: \centerline{\includegraphics*[width=8.5cm]{fig3.eps}}
234: \caption{
235: Decay of $W$ (in units of the lattice constant) at $E = 0.4$, $T=1$ with
236: $K$ segmental dynamics on a semilog plot.
237: The range of the $Y$ axis ($W-W_r$) lies between 0.4 and 1.8 and has
238: units of the lattice constant.
239: The slope of the fit is provided in the legend. Statistics
240: are the same as for Fig.~\ref{fig:w_t_k}.
241: }
242: \label{fig:w_t_relax}
243: \end{figure}
244:
245: Since the relaxations of chains and the interface are so different between
246: the $K$ and $KC$ dynamics, the major question remains, which
247: dynamics
248: is appropriate? In our opinion, it depends on the situation.
249: If one looks for the well
250: behaved relaxed or steady-state properties such as the interface width,
251: $KC$ dynamics would be preferable over the $K$ movement of chains.
252: Nevertheless, it is important to understand the details of results
253: arising
254: from different dynamics. Lets look at the density profile of polymer
255: with the $K$ motion presented in Fig.~4. We immediately note the
256: difference in density profile at low ($E < 0.6$) and high ($E \ge 0.6$)
257: fields. At low field, the polymer density remains low ($\to 0$) from
258: the source end (the region for releasing new chains) and increases
259: monotonically
260: toward the substrate. On the other hand, at high
261: field there is a large density toward the source end indicating a
262: build-up
263: (accumulation) of polymer chains before they reach the bulk region growing
264: from the substrate. Such a clogging restricts the deposition of polymer
265: chains and the growth becomes independent of the rate of polymer
266: release
267: (i.e., dynamics-limited deposition).
268: Clogging occurs due to slow motion of chains with the $K$ dynamics
269: alone.
270: Thus, care must be exercise in analyzing the interface growth, decay,
271: and its scaling to field at low and high fields where deposition
272: rates
273: differ substantially.
274:
275: \begin{figure}
276: \centerline{\includegraphics*[width=8.5cm]{fig4.eps}}
277: \caption{
278: Polymer density $D$ (fraction of occupied sites) vs.\ $x$ (in units of
279: the lattice constant) with the $K$
280: dynamics for different
281: fields.
282: }
283: \end{figure}
284:
285: With the $KC$ dynamics, there is no clogging at these values of field
286: ($E \le 2.0$) and the interface width relaxes very well within our
287: simulation time. Variation of the relaxed interface width $W_r$ with
288: the field is presented in Fig.\ 5. We see that the interface width
289: ($W_r$) decays with the field with a power law $W_r \propto E^{-1/2}$.
290: Note that the nature of decay of the relaxed interface width $W_r$
291: with
292: the field remains similar to that of the saturated width $W_s$ in
293: steady-state. Such a decaying trend of the interface width with the
294: field
295: is also observed with the $K$ alone in the low field regime ($E <
296: 0.6$).
297: We have also studied the dependence of the interface width ($W_r$) on
298: the molecular weight $L_c$ and find a monotonic decrease as shown in
299: Fig.~6. Note the contrast, while the relaxed width $W_r$ decays
300: with
301: the molecular weight, the saturated steady-state width $W_s$
302: increases.
303: In the steady-state growth, polymer chains are continuously
304: deposited. As a result the incoming chains at the surface are not
305: relaxed.
306: The contribution of unrelaxed chains at the growing surface to the
307: interface
308: width is dominant over the relaxed chains, since the conformation of
309: incoming
310: chains are relatively stretched out along the direction of the field.
311:
312: \begin{figure}
313: \centerline{\includegraphics*[width=8.5cm]{fig5.eps}}
314: \caption{
315: Interface width ($W_s$ and $W_r$ in units of the lattice constant) vs.\
316: field
317: $E$ (in arbitrary units) with
318: $KC$ segmental
319: dynamics. Statistics are the same as for Fig.~\ref{fig:w_t_k}.
320: }
321: \end{figure}
322:
323: \begin{figure}
324: \centerline{\includegraphics*[width=8.5cm]{fig6.eps}}
325: \caption{
326: Interface width ($W_s$ and $W_r$) vs.\
327: $L_c$ with $KC$ and $KCR$ segmental
328: dynamics. $W_s$, $W_r$, and $L_c$ are in units of the lattice
329: constant. Statistics are the same as for Fig.~\ref{fig:w_t_k}.
330: }
331: \end{figure}
332:
333: We have also examined the effect of incorporating the slithering-snake
334: (reptation) motion to kink-jump and crankshaft dynamics ($KCR$). The
335: relaxation of the interface width is relatively faster with the $KCR$ than
336: with $KC$ and $K$ segmental movements. However, the qualitative nature of
337: the variation of the relaxed interface width $W_r$ with the molecular
338: weight $L_c$ is similar to that with $KC$ dynamics (see Fig.\ 6) with
339: somewhat higher magnitude of the widths. It is worth pointing out the
340: difference in the interface widths with $KC$ and $KCR$ segmental dynamics.
341: While the difference in the steady-state width $W_s$ increases with the
342: molecular weight, the difference in relaxed interface width $W_r$
343: decreases. At high molecular weights, it is rather difficult to
344: distinguish the relaxed interface widths. The larger difference in steady
345: state and relaxed interface width ($W_s-W_r$) between $KC$ and $KCR$
346: dynamics suggests that the magnitude of $W_s$ is dominated by the
347: elongation of polymer chains (along the field direction) which is larger
348: with the $KCR$ semental motion. Figure 7 shows the variation of the
349: relaxed interface width $W_r$ with the molecular weight for two different
350: fields with the $KCR$ segmental dynamics. We see that the qualitative
351: nature of the variation remain similar at these fields. Decay of the
352: interface width ($W_r$) with the molecular weight appears to be stretched
353: exponential. With the $KCR$ dynamics, we have also observed a power-law
354: decay of the interface width ($W_r$) with the field similar to Fig.\ 5
355: with the $KC$ dynamics. As expected, the interface width is relatively
356: well relaxed with the $KC$ dynamics, and adding reptation enhances the
357: relaxtion further.
358:
359: \begin{figure}
360: \centerline{\includegraphics*[width=8.5cm]{fig7.eps}}
361: \caption{
362: Interface width ($W_r$) versus $L_c$ with
363: $KCR$ segmental
364: dynamics at $E=0.4$, 0.5, $T=1$. $W_r$ and $L_c$ are in units of
365: the lattice constant. Statistics are the same as
366: for Fig.~\ref{fig:w_t_k}.
367: }
368: \end{figure}
369:
370: \section{Conclusion}
371:
372: A computer simulation study was presented to investigate the effect of
373: segmental dynamics on the growth and decay (relaxation) of the interface width for an
374: electrophoretic deposition model for polymer chains. The results for
375: the
376: dependence of the relaxed interface width ($W_r$) on the molecular
377: weight
378: are quite different from that of the steady-state interface width
379: ($W_s$)
380: reported in previous studies \cite{bpf, fp1} where polymer chains were
381: continuously deposited throughout the simulation.
382: In contrast to an increase of the steady-state width $W_s$, the relaxed
383: width $W_r$ decays with the molecular weight.
384: The power-law decay of the relaxed width with the field
385: ($W_r \propto E^{-1/2}$) remains the same as that of the steady-state
386: width ($W_s$).
387:
388: The relaxation of the width after stopping the addition of more
389: polymer chains to the system depends on the segmental dynamics.
390: For example, with the kink-jump dynamics alone, it is very difficult to
391: reach relaxed interface width within a reasonable simulation time at
392: low
393: fields ($E \le 0.07$). On the other hand, clogging occurs around the
394: entrance
395: area of polymer injection at high fields ($E \ge 0.6$), which reduces
396: the rate
397: of polymer deposition on the substrate. Thus, the scaling of the
398: interface
399: width at high fields should be different from those at low to moderate
400: field values. Inclusion of crankshaft motion leads to faster
401: growth
402: and interface relaxation at all field values we studied. Adding
403: large scale
404: segmental dynamics (reptation) enhances the interface relaxation and
405: reduces
406: the magnitude of the relaxed interface width. The scaling of the
407: interface
408: width with the field and molecular weight is qualitative similar.
409:
410: \acknowledgments
411:
412: We acknowledge partial support from a DOE-EPSCoR grant.
413:
414: \begin{thebibliography}{99}
415: \bibitem{barabasi}
416: A.-L. Barabasi and H.~E. Stanley, {\em Fractal Concepts in
417: Surface Growth} (Cambridge, Cambridge, 1995).
418:
419: \bibitem{family} {\em Dynamics of Fractal Surfaces},
420: edited by F. Family and T. Vicsek (World Scientific, Singapore, 1991).
421:
422: \bibitem{wool} R.~P. Wool, {\em Polymer Interfaces: Structure and
423: Strength} (Hanser, New York, 1995).
424:
425: \bibitem{bpf} F.~W. Bentrem, R.~B. Pandey, and F. Family, Phys. Rev. E,
426: {\bf 62},
427: 914 (2000).
428:
429: \bibitem{fp1} G.~M. Foo and R.~B. Pandey, Phys. Rev. E, {\bf 61}, 1793
430: (2000).
431:
432: \bibitem{fp2} G.~M. Foo and R.~B. Pandey, Biomacromolecules, {\bf 1}, 407
433: (2000).
434:
435: \bibitem{liu97} Y. Liu and T. P. Russell, Macromolecules, {\bf 30},
436: 7768 (1997).
437:
438: \bibitem{wallace01} W. E. Wallace, D. A. Fischer, K. Efimenko, W. Wu,
439: and J. Genzer, Macromolecules, {\bf 34}, 5081 (2001).
440:
441: \bibitem{riemann96} R.-E. Riemann, H.-J. Cantow, C. Friedrich,
442: Poly. Bull., {\bf 36}, 637 (1996).
443:
444: \bibitem{riemann97} R.-E. Riemann, H.-J. Cantow, C. Friedrich,
445: Macromolecules, {\bf 30}, 5476 (1997).
446:
447: \bibitem{tanaka00} K. Tanaka, A. Takahara, and T. Kajiyama,
448: Macromolecules, {\bf 33}, 7588 (2000).
449:
450: \bibitem{binder} {\em Monte Carlo and Molecular Dynamics Simulations in
451: Polymer Science}, edited by K. Binder (Oxford, New York, 1995).
452:
453: \bibitem{metropolis53} M. A. Metropolis, A. W. Rosenbluth,
454: M.~N. Rosenbluth, A.~H. Teller, and E. Teller, J. Chem. Phys., {\bf
455: 21}, 1087 (1953).
456:
457: \bibitem{verdier62} P. H. Verdier and W. T. Stockmayer,
458: J. Chem. Phys., {\bf 36}, 227 (1962).
459:
460: \end{thebibliography}
461:
462: \end{document}
463:
464:
465:
466:
467:
468:
469:
470: