1: \documentclass{elsart}
2: \usepackage{graphicx}
3: \begin{document}
4: \begin{frontmatter}
5:
6: \title{Statistical modelling of electrochemical deposition of nanostructured
7: hybrid films with ZnO-Eosin Y as a case example}
8: \author[uff]{F. D. A. Aar\~ao Reis\corauthref{cor1}}
9: \ead{reis@if.uff.br}
10: \corauth[cor1]{Fax number: (55) 21-2629-5887 }
11: \author[pmc]{J. P. Badiali}
12: \ead{badiali@ccr.jussieu.fr}
13: \author[enscp]{Th. Pauport\'e}
14: \ead{thierry-pauporte@enscp.fr}
15: \author[enscp]{D. Lincot}
16: \ead{daniel-lincot@enscp.fr}
17: \address[uff]{
18: Instituto de F\'\i sica, Universidade Federal Fluminense,
19: Avenida Litor\^anea s/n, 24210-340 Niter\'oi RJ, Brazil}
20: \address[pmc]
21: {Laboratoire d'Electrochimie et de Chimie Analytique (UMR 7575 - CNRS - ENSCP -
22: Paris 6), Universit\'e P. et M. Curie, 4 Place Jussieu, 75005 Paris, France}
23: \address[enscp]
24: { Laboratoire d'Electrochimie et de Chimie Analytique (UMR 7575 - CNRS - ENSCP -
25: Paris 6), \'Ecole Nationale Sup\'erieure de Chimie de Paris, 11 rue Pierre
26: et Marie Curie, 75231, Paris, France}
27: \date{\today}
28: \maketitle
29:
30: \begin{abstract}
31: We study models
32: % related to the experimental features
33: of electrodeposition of hybrid organic-inorganic films with
34: a special focus on the growth of ZnO with eosin Y.
35: First we propose a rate equation model which assumes that the organic
36: additives form branches with an exposed part above the $ZnO$ deposit, growing
37: with larger rate than the pure film and producing $ZnO$ at the exposed length.
38: This accounts for the generation of ${OH}^-$ ions from reduction of dissolved
39: oxygen near the branches and reactions with ${Zn}^{2+}$ ions to form $ZnO$
40: molecules. The film grows with the same
41: rate of the branches, which qualitatively explains their catalytic effect, and
42: we discuss the role of the additive concentration.
43: Subsequently, we propose a statistical model which represents
44: the diffusion of the hydroxide precursor and of
45: eosin in solution and adopt simple probabilistic rules for the reactions,
46: similarly to diffusion-limited aggregation models.
47: The catalytic effect is represented by the preferencial
48: production of ${OH}^-$ ions near eosin.
49: The model is simulated with relative concentrations in
50: solution near the experimental values. An improvement of the growth rate is
51: possible only with a rather large apparent diffusion coefficient of eosin in
52: solution compared to that of hydroxide precursors. When
53: neighboring eosin clusters competitively grow, the increase in the growth
54: rate and a high eosin loading are observed in the simulated deposits. Those
55: features are in qualitative agreement with experimental results.
56: \end{abstract}
57:
58: \begin{keyword}
59: Zinc oxide \sep Eosin Y \sep cathodic electrodeposition \sep growth model \sep
60: diffusion-limited aggregation
61: \end{keyword}
62:
63: \end{frontmatter}
64:
65: \section{Introduction}
66:
67: The growth of thin films from solutions has emerged as an efficient, low
68: temperature, versatile preparation route which can be used at large scale for
69: the production of high quality materials \cite{revlincot}. One of the interests
70: of these methods
71: is the possibility of adding foreign soluble compounds to the deposition bath
72: and modulating film properties by playing with the interactions which arise
73: between the growing film and these compounds. Among the additives, organic ones
74: are of utmost interest since many of them have been shown to act as templating
75: agents and/or crystal growth directing agents, and in many cases as
76: functionalizing agents for the deposits.
77:
78: These effects are well-documented in the case of zinc oxide. This inorganic
79: compound has attracted much attention due to a broad range of potential high
80: technology applications such as surface acoustic wave filter \cite{emanetoglu},
81: light emitting diodes \cite{konenkamp}, lasers \cite{huang}, varistors
82: \cite{lin}, gas sensors \cite{golego} and solar cells \cite{keis}. $ZnO$ can be
83: prepared as a thin film by chemical deposition methods \cite{vaysieres,tian} or
84: by electrodeposition
85: \cite{izaki,peulon1,peulon2,pauporte1,pauporte2,pauporte3}. The dramatic
86: effects of organic additives on film structures and morphologies have been
87: reported with both deposition routes. For instance, in the presence of citrate,
88: chemically deposited $ZnO$ are formed of stacked nanoplates assembled in a
89: biomimetic manner \cite{tian}. However the most impressive morphological
90: changes have been observed in electrodeposited zinc oxide. In the presence of
91: sodium dodecylsulfate (SDS), the formation of lamellar nanostructures has been
92: shown \cite{choi,michaelis}. The SDS can be removed from the film and highly
93: porous thin films are released \cite{michaelis}. Different dyes, such as eosin
94: Y (EY) \cite{yoshida1,pauporte4,yoshida2,pauporte5}, fluorescein (FL)
95: \cite{okabe}, Tetrasulfonated phthalocyanines (TSPc)
96: \cite{yoshida3,yoshida4,schlettwein,pauporte6} and riboflavin (RI)
97: \cite{karuppuchamy}, have also been shown to incorporate in the film and to
98: give rise to nanostructures. These organic dyes contain negatively charged
99: functions (carboxylate (EY, FL), sulfonate (TSPc) or phosphonate (RI) groups)
100: which allow their direct binding to the oxide crystal surface during the
101: synthesis.
102:
103: While the effects of additives in film shaping are widely illustrated
104: in experimental works, particularly that of the organic ones, the exact role of
105: these compounds in the growth process remains to be clarified in most cases.
106: The aim of the present work is to fill this gap by
107: proposing kinetic and statistical models for zinc oxide electrodeposition in
108: the presence of EY in solution, which focus on a small number of basic
109: features of those processes. The starting point for this modeling is a series of
110: experimental results, some of them also observed in deposition with different
111: organic additives \cite{michaelis}.
112:
113: The first step of electrochemical zinc oxide synthesis is the cathodic
114: reduction of a hydroxide precursor such as molecular oxygen
115: \cite{peulon1,pauporte1}, hydrogen peroxide \cite{pauporte2,pauporte3} or
116: nitrate ions \cite{izaki,yoshida1}:
117: \begin{equation}
118: \frac{1}{2} O_2 + H_2O + 2e^- \rightarrow 2{OH}^- ,
119: \label{reacoxigen}
120: \end{equation}
121: \begin{equation}
122: H_2O_2 + 2e^- \rightarrow 2{OH}^-
123: \label{reacperoxide}
124: \end{equation}
125: or
126: \begin{equation}
127: {NO_3}^- + H_2O + 2e^- \rightarrow {NO_2}^- + 2{OH}^- .
128: \label{reacnitrate}
129: \end{equation}
130: At a temperature slightly above room temperature \cite{goux}, well crystallized
131: zinc oxide with the wurtzite hexagonal structure is precipitated at the
132: electrode surface:
133: \begin{equation}
134: Zn^{2+} + 2 {OH}^- \rightarrow ZnO + H_2O.
135: \label{reaczno}
136: \end{equation}
137: In many cases, it has also been shown that the organic component presents
138: catalytic properties for the deposition process. This catalytic effect is
139: connected to the improvement of $O_2$ reduction (Eq. \ref{reacoxigen}) with EY
140: \cite{yoshida2}, TSPc \cite{pauporte6}, SDS \cite{michaelis}, whereas EY has
141: also been shown to catalyze the $H_2O_2$ reduction (Eq. \ref{reacperoxide})
142: \cite{pauporte4}. Thus, one of the assumptions of our models is a distinguished
143: rate of formation of $ZnO$ near the species that represent aggregated eosin
144: molecules.
145:
146: The zinc oxide films
147: prepared in the presence of EY show large round shaped single crystals of zinc
148: oxide filled with self-assembled dye aggregates, as shown in Fig. 1
149: \cite{pauporte5}. It is observed that the dye can be almost completely
150: removed by a soft chemical
151: treatment \cite{yoshida5,pauporte7}, revealing a network of
152: mesopores which is connected to the surface and can be filled with a
153: solution. This network, formed by aggregated dye molecules, acts as a template
154: for the growth of $ZnO$. In our models, these experimental facts justify the
155: assumption of formation of connected branches of the species that represent the
156: additive molecules in the films.
157:
158: In Sec. II of this paper, we will introduce our first model, which is based on
159: rate equations, and explains some
160: features of the cooperative growth of a $ZnO$ film and additive aggregates. This
161: model is particularly useful for understanding the role of additive
162: concentrations and diffusion coefficients in a qualitative way. Subsequently, in
163: Sec. III we will present a statistical model which
164: represents the main microscopic features of those processes by adopting
165: probabilistic
166: rules of diffusion and aggregation of some chemical species in solution and in
167: the deposit. The hypothesis of a growth mechanism controlled by
168: diffusion-limited aggregation of the eosin molecules was anticipated in Ref.
169: \protect\cite{lincot1}. Simulation results of that statistical model, presented
170: in Sec. IV, provide estimates
171: of growth rates, structures of $ZnO$ and EY deposits and shows a role of
172: diffusion coefficients which are in qualitative agreement with experimental
173: findings. Finally, in Sec. V we
174: summarize our results and present our conclusions.
175:
176: \section{Rate equation model}
177:
178: This model aims at explaining basic kinetic features of the growth of $ZnO$ with
179: organic additives with drastic simplifications of the film structure.
180:
181: First we assume that pure
182: $ZnO$ growth takes place with rate $r_0$ (in nanometers per second), due to
183: reactions of type (\ref{reaczno}) near the film surface. If the average volume
184: occupied by one molecule is $V$, then $r_0/V$ is the growth rate per unit area
185: in that case. Even in the presence of the additives, the roughness of the $ZnO$
186: film surface will be neglected.
187:
188: When the additive is present, we assume that it forms straight structures which
189: grow from the electrode, as shown in Fig. 2. This is consistent with the
190: experimental finding of porous structures when eosin Y was removed from the
191: hybrid films \cite{yoshida5,pauporte7}, but with an initial oversimplification
192: of that structure. The total number of linear
193: branches of the additive is $N_0$, which depends on the nucleation of the
194: additive at the electrode surface in the beginning of the growth process. Each
195: branch is assumed to grow with rate $r_A$ if it is not covered by the $ZnO$
196: deposit ($r_A>r_0$), so that the height of each branch increases as $H=r_A t$.
197: $r_A$ is expected to increase with the concentration $\rho_A$ and the
198: diffusion coefficient $D_A$ of the additive in solution, but the exact
199: dependence on those quantities will not be important at this point.
200:
201: In order to represent the catalytic effect of the additive, we assume that an
202: excess of ${OH}^-$ is produced near the branches, which leads to the
203: production of $ZnO$ at those regions. The rate of production of $ZnO$ is
204: expected to be proportional to the length of those branches which is exposed
205: above the film surface, $\left( H-h\right)$ (see Fig. 2), where $h$ is the
206: $ZnO$ film height. Consequently, the total rate of production of $ZnO$
207: molecules near the additive branches is $N_0k\left( H-h\right)$, where $k$ is a
208: reaction constant. A constant reaction rate in this model accounts for
209: the assumption that the oxygen reduction reaction is not under diffusion
210: control. Also, it is assumed that the $ZnO$ molecules immediately
211: precipitate and aggregate to the flat film surface, leaving the branches
212: uncovered.
213:
214: Under these conditions, if the area of the film is $A$, then the number of $ZnO$
215: molecules deposited per unit time is
216: \begin{equation}
217: \frac{dN}{dt} = \frac{r_0}{V} A + N_0 k { \left( H-h\right)} .
218: \label{dNdt}
219: \end{equation}
220: Here, the first contribution in the right hand side comes from the reactions
221: ocurring near the film surface and the second one comes from the reactions near
222: the branches. The growth rate of the $ZnO$ deposit is
223: $\frac{dh}{dt} = \frac{V}{A} \frac{dN}{dt}$, which gives
224: \begin{equation}
225: \frac{dh}{dt} = r_0 -ah +r_A at \qquad ,\qquad a\equiv \frac{N_0kV}{A} .
226: \label{dhdt}
227: \end{equation}
228:
229: The solution of Eq. (\ref{dhdt}) is
230: \begin{equation}
231: h = r_A t + \frac{\left( r_A-r_0\right)}{a} {\left( e^{-at}-1 \right)} .
232: \label{ht}
233: \end{equation}
234: It gives a constant growth rate $r_A$ for the deposit at long times, with $1/a$
235: being the characteristic time of decay. It means that the film growth rate
236: attains the same growth rate of the branches in the steady state. The rate
237: of production of $ZnO$ near the film surface is $r_0<r_A$, but the exposed
238: height of the branches saturates at $H-h = \frac{r_A-r_0}{a}$, and produces an
239: excess of $ZnO$ molecules in their neighborhood which is enough for branches and
240: film to grow with the same rate.
241:
242: The dependence of $r_A$ on $N_0$ is not explicit in the model, but it exists. As
243: the number of branches per unit area increases, the aggregation of the
244: diffusing additive to them is facilitated and, consequently, their growth rate
245: $r_A$ increases. From the dependence of the saturation value of $H-h$ on $a$,
246: we expect that it decreases as $N_0$ increases: the steady state regime is more
247: rapidly attained ($1/a$ decreases), with a smaller exposed height of the
248: branches. On the other hand, the dependence of $H-h$ on $r_A$ suggests the
249: opposite behavior, i. e. that it increases with $N_0$. The interplay of these
250: effects will determine the overall dependence of $H-h$ on $N_0$. Anyway,
251: this analysis shows the nontrivial consequences of the nucleation processes
252: (which determine $N_0$) on the long time behavior of the system.
253:
254: Now let us consider the effect of the additive concentration in solution,
255: $\rho_A$, assuming that its diffusion coefficient is constant (constant
256: temperature conditions). Since $r_0$ only represents the reactions occurring
257: near the film surface, it is expected to be constant as $\rho_A$ increases. On
258: the other hand, the growth rate of the branches, $r_A$, is some monotonically
259: increasing function of $\rho_A$. For low concentrations,
260: $r_A$ is small, consequently the branches will be covered by $ZnO$.
261: Quantitatively, this is the case where $r_A\left( \rho_A\right) <r_0$, which
262: implies $\rho_A<\rho_c$, where $\rho_c$ is a critical concentration of additive
263: such that
264: $r_A\left( \rho_c\right) =r_0$. For higher concentrations, we have
265: $r_A>r_0$ and the above solution of the rate equation model is valid.
266:
267: The dependence of the growth rate on the additive concentration, with the
268: catalytic effect only for $r_A>r_0$, is illustrated in Fig. 3. There, we assume
269: that $r_A$ linearly increases with
270: $\rho_A$ when the concentration is not too large (first order reaction). This is
271: the regime where our model is expected to work. For large concentrations, Fig.
272: 3 shows a saturation of growth rate, which is a consequence of
273: the limitations to oxigen diffusion towards the surface. Indeed, in this
274: regime, the catalytic effect of eosin to reduce oxygen does not result in
275: an increased production of hydroxide ions anymore. The hydroxide precursor
276: current towards the electrode becomes constant, which also occurs with the
277: growth rate of ZnO, which is directly linked to the availability of hydroxyde
278: ions.
279:
280: The behavior observed in the intermediate EY concentration range is in
281: qualitative agreement with the findings
282: of Ref. \protect\cite{pauporte4} for $ZnO$/EY electrodeposition with
283: hydrogen peroxide as hydroxide precursor: no significant effect in the growth
284: rate is found for EY concentrations below $2\mu M$, but an increase
285: is observed for EY concentrations above that value. On the other hand, no
286: well-defined plateau in the growth rate is observed in electrodeposition with
287: molecular oxygen \cite{pauporte8}, but an increase in the growth rate even for
288: small additive concentrations. This is probably due to nucleation features that
289: are not properly represented in the present simplified model.
290:
291: The main parameters of the rate equation model are the growth rates $r_A$ and
292: $r_0$, which depend on several parameters, such as
293: concentrations of hydroxide precursors and additives in solution, zinc
294: ions concentration and the respective diffusion coefficients. However, the model
295: leads to two important conclusions. First,
296: in order that the catalytic effect is observed, it is necessary that $r_A>r_0$,
297: consequently the concentration and the diffusion coefficient of the additive in
298: solution must be combined in a suitable way to fullfill this condition.
299: Secondly, if the additive branches are able to grow above the film surface,
300: then the film growth rate is the same of those branches in the stationary state
301: ($t\gg 1/a$ in Eq. \ref{ht}), independently of the rate of pure $ZnO$ films
302: growth. These conclusions will guide the simulation work on the microscopic
303: model.
304:
305: \section{Microscopic model}
306:
307: Now we present a more complete model for the electrodeposition of $ZnO$ with
308: organic additives, in which the physical properties of different species in
309: solution and in the deposit are taken into account. This is particularly
310: important to understand the role of diffusion coefficients and to predict the
311: structure of the deposit, as well as to
312: perform quantitative comparisons with experimental data. At the latter point,
313: we will basically refer to $ZnO$/EY films data available, thus the
314: presentation of the model will be inspired by this growth process.
315:
316: We will consider a lattice model with two species executing random walks in
317: solution and two species forming a deposit grown above a flat electrode. All
318: lengths will be given in lattice units. Due to the relative complexity of this
319: model, it will be solved by simulation. A two-dimensional version will be
320: discussed, which we expect to capture the main features of the real process,
321: at the same time of being amenable for computation.
322:
323: The species in the solution, O and E, represent the hydroxide precursor ($O_2$
324: or $H_2O_2$) and eosin Y, respectively.
325: Their different sizes, as shown in Fig. 4a, qualitatively account for the
326: different masses and radius. Their auto-diffusion coefficients are $D_O$
327: and $D_E$, i. e. each O (E) in solution executes $D_O$ ($D_E$) random movements
328: of unit size per unit time. These rates set the time scale of our model.
329:
330: The deposit will be formed by two aggregated species, named ZA and EA,
331: representing respectively the $ZnO$ and the eosin Y molecules in the
332: film, as shown in Fig. 4b. For simplicity, their sizes are the same as those of
333: the species in solution.
334:
335: Although the zinc and hydroxide ions play an essential role in
336: the reactions leading to the formation of $ZnO$ (Eqs. \ref{reacoxigen},
337: \ref{reacperoxide} and \ref{reaczno}), they will not be explicitely represented
338: in our model. Instead, their effects will be described by probabilities of
339: removing O particles from the solution and producing new aggregated particles,
340: thus avoiding the complications of the explicit representation of the above
341: mentioned reactions. Following this reasoning, the growth process is
342: represented by the simplified reactions $O\to ZA$ and $E\to EA$, which can take
343: place with given probabilities and under conditions that are related to the
344: neighborhood of the particles in solution.
345:
346: The growth process begins with a flat electrode (a line of length $L$) and a
347: solution with concentrations $\rho_O$ and $\rho_E$ of O and E particles,
348: respectively.
349:
350: The aggregation of an E particle (reaction $E\to EA$) occurs when that particle
351: has a nearest neighbor of a previously aggregated EA (Fig. 5a) or when it has a
352: nearest neighbor site belonging to the electrode (the initial substrate). The
353: new EA particle permanently sticks at the position where it is
354: formed. The contact of E particles with ZA sites never generates EA particles,
355: which is a key hypothesis for explaining the formation of continuous fibers of
356: eosin Y in the films.
357:
358: The reaction $O\to ZA$ occurs upon contact of a particle O
359: with ZA or EA. When the O particle has a neighbor site ZA or a neighbor site
360: belonging to the
361: electrode (initial substrate), the reaction occurs with probability $p$,
362: otherwise the O particle
363: remains in the solution. When the O particle has a neighboring EA, that reaction
364: occurs with probability $1$. These rules are illustrated in Fig. 5b.
365:
366: Contrary to the EA particles, ZA particles are allowed to precipitate and
367: diffuse on the film surface after their formation. In order to represent the
368: main features expected for this
369: process and save computational time, a simple diffusion mechanism will be
370: mimicked immediately after the ZA particle is created: it precipitates to the
371: topmost empty site of its current column, searches for the
372: point with the highest number of nearest neighbors ZA whithin a certain radius
373: $R$ (defined below), and permanently sticks at that point. These processes are
374: illustrated in Fig. 5c.
375:
376: The above stochastic rules are motivated by some experimental features of the
377: $ZnO$/EY films growth. The rules for formation of EA particles represent
378: the initial nucleation of EY at the electrode and the subsequent formation
379: of EY clusters upon reduction of the molecules that reach the film
380: surface. Since EA is not formed upon contact of E with ZA, there
381: will be no isolated EA particle in the ZA matrix, which accounts for the fact
382: that almost all EY can be removed from the hybrid films, leaving them with a
383: porous structure \cite{yoshida5,pauporte7}. As regards the formation of ZA
384: particles, the higher
385: probability of occurrence after contact of O with EA follows from the
386: assumption that oxygen precursors reduction is enhanced near the eosin
387: clusters, which facilitates the formation of $ZnO$ in those regions. This
388: accounts for the catalytic effect of eosin Y, also observed experimentally.
389:
390: For the diffusion of a ZA particle on the ZnO film surface after its formation,
391: we consider a surface diffusion
392: coefficient $D_A$, which gives a radius for searching the final
393: aggregation position as $R=\sqrt{D_A\tau}$, where $\tau$ is the
394: current average time for one layer deposition. This time is updated during the
395: simulation, while $D_A$ is kept fixed. The condition to choose the final
396: aggregation point follows the same ideas of the Wolf-Villain model for thin film
397: deposition \cite{wv}, and are
398: reasonable for formation of a crystalline structure due to the trend of
399: increasing the binding energy.
400:
401: The height of the solution above the film surface is kept constant during the
402: simulation of the growth process. New particles (O or E) are left at random
403: positions of the top layer of the solution immediately after an aggregation
404: event (rections $E\to EA$, $O\to ZA$), in order to maintain constant
405: concentrations. This height was $300$ lattice units in our simulations. The
406: diffusion layer near the film surface, in which concentrations of O and E were
407: reduced, was of order of $100$ lattice
408: units or less. These conditions are reasonable to represent experiments in which
409: a rotating disk electrode was used, thus avoiding the growth
410: of the diffusion layer and the subsequent decrease of the growth rate.
411:
412: We recall that the rules for aggregation of E particles are the same of the
413: original diffusion-limited aggregation (DLA) model of Witten and Sander
414: \cite{witten}, in which growth took place from a single seed. Extensions to the
415: case of an initial flat surface were also considered by several authors
416: \cite{meakin,racz,jullien,burlatsky}. To be more precise, due to the collective
417: diffusion mechanism of E particles in solution, the growth kinetics of E is
418: equivalent to that of the multiparticle biased DLA (MBDLA) in the condition of
419: zero bias \cite{sanchez,castro1,castro2,schwarzacher}. Indeed, MBDLA was
420: proposed as a model
421: for electrodeposition of $CoP$ alloys \cite{castro1,riveiro}. On the other
422: hand, the aggregation rules of ZA particles, which include surface relaxation,
423: suggest Edwards-Wilkinson growth in that case \cite{ew,barabasi}. However, in
424: the cooperative growth
425: model, a nontrivial interplay of both processes is observed, as shown below.
426:
427: \section{Results for the microscopic model}
428:
429: In all simulations presented here, we considered $D_O=1$, and in most cases the
430: lattice lenghts are $L=512$ and the diffusion coefficient of aggregated ZA is
431: $D_A=0.005$. Some data for $L=2048$ were also
432: collected in order to search for possible finite-size effects, mainly when
433: the thickest deposits were grown.
434: In experiments, the maximum ratio of concentrations of eosin and hydroxide
435: precursor in
436: solution is in the range ${10}^{-3}-{10}^{-2}$, thus in this model we worked
437: with ratios between $0$ (no eosin) and ${10}^{-2}$.
438:
439: The results of the rate equation model of Sec. II suggest that the growth rate
440: $r_A$ of the additive (eosin) is proportional to $\rho_E D_E$, while
441: the growth rate of the pure $ZnO$ deposit, $r_0$, is proportional to $p \rho_O
442: D_O$. In the following, the model parameters will be tuned
443: in the light of these relations and the conclusions of the model
444: of Sec. II.
445:
446: First the model was simulated considering $D_E=1/8$ (thus $D_E/D_O\sim 0.1$), in
447: order to account for the
448: much larger mass of eosin when compared to the hydroxide precursors.
449: $D_A=0.05$ was considered in all
450: simulations, some of them with $\rho_O=2\times {10}^{-2}$,
451: $\rho_E=2\times {10}^{-4}$, and others with $\rho_O=5\times {10}^{-3}$,
452: $\rho_E=5\times {10}^{-5}$. Several different values of $p$ were considered,
453: with $p\geq {10}^{-5}$.
454:
455: In Fig. 6 we show a region of a deposit obtained with $p= {10}^{-4}$,
456: $\rho_Z=5\times {10}^{-3}$ and $\rho_E=5\times {10}^{-5}$, at
457: $t=3\times {10}^5$. Some EA
458: particles are present at the initial electrode surface, but they are rapidly
459: covered by the ZA layer. Subsequently, there is no possibility for uptaking
460: other eosin molecules in the film, then the
461: final growth rate is determined by ZA aggregation. From the results obtained
462: with such small diffusion coefficient of E in solution, we conclude that this
463: is not the suitable condition to observe the catalytic effect of the additive.
464:
465: Subsequently, we assumed that E particles had a significantly large diffusion
466: coefficient, $D_E=1$, which is equal to the coefficient of Z.
467: In Figs. 7a, 7b and 7c, we show the time evolution of a region of a deposit
468: grown with $\rho_O=2\times {10}^{-2}$,
469: $\rho_E=2\times {10}^{-4}$ and $p={10}^{-3}$. Although many EA particles attach
470: to the electrode in the beginning of the process, they frequently
471: become covered by ZA particles after some time, even with a small
472: probability of O-ZA aggregation. The growth process is improved only at isolated
473: regions where clusters of EA particles succeed to grow, as shown in
474: Figs. 7a-7c.
475:
476: When compared to the case without E particles, the average height
477: of the deposit increases more than twice near the clusters of Figs. 7a-c, for
478: the same time of growth. However, the large distance between the surviving
479: branches leads to a very large surface roughness at long lengthscales.
480: Moreover, this roughness rapidly increases in time because the mounds growing
481: with large rates are separated by long valleys which slowly grow. These
482: features are not observed in experiments, for which we refer to Fig. 1 as a
483: typical example. We also tested smaller concentrations of $\rho_O$, but keeping
484: the relative concentration between E and O fixed (${10}^{-2}$), obtaining
485: similar results (growth of EA branches may be even more difficult for the same
486: $p$). We conclude that larger values of $D_E$ must be tested in order to find
487: the suitable conditions to represent the experimental results.
488:
489: Following this reasoning, we also considered the case $D_E=2$, and obtained
490: results that qualitatively agree with experimental findings. In this case,
491: small
492: concentrations of E particles are able to improve the growth process, with the
493: creation of a large number of nucleation centers in the eletrode, the
494: formation of a large number of branches and a competition between them as the
495: deposit grows, as well as a much smaller surface roughness.
496: The formation of an eosin interconnected network inside the ZnO matrix is clear,
497: which also agrees with the main features of the experimental nanostructures.
498:
499: Fig. 8 illustrates the beginning of the growth of a region of a deposit with
500: $\rho_O=5\times {10}^{-3}$,
501: $\rho_E=5\times {10}^{-5}$, $D_A=0.05$ and $p=0.25$.
502: The average height for $t=3\times {10}^6$ (Fig. 8c) is nearly $3$ times larger
503: than the height of the film grown without E particles during the same time, and
504: the eosin loadings in the deposits (ratio between number of EA particles and
505: ZA particles) of Figs. 8a-8c increase from $0.64\%$ to
506: $0.85\%$. Simulation at longer times show sthat these quantities attain
507: different steady state values, so that the features of Fig. 8 may be
508: interpreted as typical of the initial growth process.
509:
510: One interesting point revealed in Fig. 8 is the competition between the growing
511: EA clusters. The central cluster of Figs. 8a and 8b grew slower than the
512: neighboring ones and, consequently, was covered by particles ZA in Fig. 8c,
513: while the larger surviving clusters created more branches.
514:
515: In Fig. 9 we show a region of a deposit grown until $t=2\times {10}^7$
516: with the same parameters of those in Fig. 8, except that $D_A=0.01$. The
517: surface roughness is larger in this case due to the inhomogeneous growth of EA
518: branches and smaller diffusion coefficient of aggregated particles. However,
519: other quantities are not much different from those obtained with $D_A=0.05$,
520: such as the growth rate at steady state conditions, which is nearly $3.3$ times
521: larger than the rate of film grown without E particles. It clearly shows the
522: catalytic effect of the eosin clusters. Another important quantity is the
523: EA loading in the film, which is nearly $2\%$. This is a high value
524: compared to the relative concentration of $1\%$ in solution.
525:
526: The above results show that a realistic description of the growth of
527: $ZnO$/EY films must take into account a remarkably large diffusion
528: coefficient for eosin in solution. A quantitative comparison with the ratio
529: $D_O/D_E$ of the model is not reliable because the model contains a small
530: number of adjustable parameters, while
531: the efficiency of other processes such as $ZnO$ precipitation are not taken into
532: account. However, it is noticeable that an unexpectedly large diffusion
533: coefficient of EY is also obtained experimentally:
534: $3.4\times {10}^{-5} cm^2 s^{-1}$ and $1.4\times {10}^{-5} cm^2 s^{-1}$ for
535: $O_2$ and EY, respectively, in the classical deposition condition of the films,
536: that is at $70{}^oC$ and in chloride medium \cite{pauporte4}. Consequently, the
537: experimental ratio between diffusion coefficients is near $2.4$, a much smaller
538: value than the ratio between the masses of EY and $O_2$, which is $20$.
539:
540: The order of magnitude of the growth rate increase in the films of Figs. 8 and
541: 9, when compared to the films
542: without EA particles, is also consistent with experiments: a factor $3$ was
543: obtained in Ref. \protect\cite{pauporte4}, with hydrogen peroxide as the oxygen
544: precursor, when $\rho_E/\rho_O = {10}^{-3}$, while a factor
545: $5.5$ was obtained in Ref. \protect\cite{pauporte8}, with molecular oxygen
546: precursor, when $\rho_E/\rho_O = {10}^{-2}$. Moreover, the high eosin loading
547: of $2.1\%$ was observed in the latter experiments \cite{pauporte8}, which also
548: agrees with the predictions of our model.
549:
550: We also recall that the rules of the model were proposed so that all EA
551: branches are connected to the electrode - see the images in Figs. 7, 8 and 9.
552: At long times, it is also
553: observed that most EA particles belong to some of the branches which are
554: exposed above the film surface - see Figs. 8 and 9. Accordingly, a small
555: fraction of EA particles is hidden in the deposit. This also agrees with the
556: experimental finding that eosin can be almost completely removed by a soft
557: chemical treatment of the hybrid films \cite{yoshida5,pauporte7}.
558:
559: Finally, it is interesting to point out that the configurations of EA
560: branches are similar to those encountered in a model of diffusion-controlled
561: deposition by Burlatsky et al \cite{burlatsky}, whose growth mechanisms are the
562: same adopted for the E particles in our model. However, while the screening
563: process of the larger clusters was the only reason for small clusters to stop
564: growing in their model, here this effect is enhanced by the deposition
565: of ZA particles, which may eventually cover an EA cluster and suppress its
566: growth (see e. g. Fig. 8c). It is also important to notice the similarity
567: between the EA clusters and the electrodeposited copper aggregates of Refs.
568: \protect\cite{leger1,leger2}, which clearly show such screening effects.
569:
570: From the point of view of competitive growth models, the results of the present
571: paper also show nontrivial features. It is usually observed that one of the
572: competing dynamics is dominant at long times and large lengthscales, such as in
573: the widely studied EW-KPZ crossover \cite{GGG,NT,AF,chamereis} or in random to
574: correlated growth \cite{rdcor}. However, here it was shown that a less trivial
575: association of these different dynamics to different chemical species leads to
576: a cooperative behavior where features of both dynamics are to some extent
577: preserved. In the present model, the competing dynamics are those of MBDLA
578: without bias for the EA particles, and of the EW model for the ZA particles, as
579: discussed in Sec. IV.
580:
581: \section{Conclusion}
582:
583: We proposed a rate equation model and a microscopic statistical model to
584: represent some features of $ZnO$ electrodeposition with organic additives,
585: mainly focusing the application to electrodeposition with eosin Y.
586:
587: The rate equation model assumes the formation of branches of the additive which
588: are extended above the film surface, and is useful to understand the basic
589: features of the growth process. Under conditions that the branches grow faster
590: than the pure $ZnO$ deposit, it was shown that both structures grow with the
591: growth rate of the former.
592:
593: With the statistical model, we were able to reproduce several qualitative
594: features of the
595: electrodeposition of $ZnO$/EY films by assuming that it may be viewed as a
596: diffusion-limited aggregation process with an interplay between different
597: chemical species in solution and a catalytic effect of the eosin clusters for
598: the formation of $ZnO$. Among the predictions of the model which are in
599: qualitative agreement with experimental findings, it is
600: important to mention: 1) the requirement of particularly large diffusion
601: coefficients of EY in solution, so that it is able to improve the growth
602: process with small concentrations; 2) the formation of
603: branches of EY connected to the surface
604: and filling the film structure without significant large scale inhomogeneities;
605: 3) the increase in relative growth rates and the high EY loadings.
606: Consequently, we believe that this model incorporated most of the basic features
607: of the $ZnO$ electrodeposition with EY, and may be extended in order to
608: provide quantitatively good results or to represent related growth processes.
609:
610: \vskip 1cm
611:
612: {\bf Acknowledgements}
613:
614: The authors thank Dr. Aur\'elie Goux for the preparation of the film shown in
615: Fig.1. FDAAR thanks Laboratoire de Electrochimie et Chimie Analytique where part
616: of this work was done, for the
617: hospitality, and acknowledges support by CNPq (Brazil) and CNRS (France).
618: DL and TP acknowledge the fruitful collaboration with Prof. Tsukasa Yoshida
619: devoted to the electrodeposition of ZnO-Eosin films.
620:
621:
622: %~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
623: %~~~~~~~~~~~~~~~~~~~ REFERENCES ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
624: %~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
625:
626: \begin{thebibliography}{99}
627:
628: \bibitem{revlincot}
629: D. Lincot, Thin Solid Films {\bf 487} (2005) 40.
630:
631: \bibitem{emanetoglu}
632: N. W. Emanetoglu, C. Gorla, Y. Liu, S. Liang, and Y. Lu, Mater. Sci. Semicond.
633: Process {\bf 2} (1999) 247.
634:
635: \bibitem{konenkamp}
636: R. Konenkamp, R.C. Word, and M. Godinez, Nanoletters {\bf 5} (2005) 2005.
637:
638: \bibitem{huang}
639: M.H. Huang, S. Mao, H. Feick, H. Yan, Y. Wu, H. Kind, E. Weber, R. Russo, and P.
640: Yang, Science {\bf 292} (2001) 1897.
641:
642: \bibitem{lin}
643: Y. Lin, Z. Zhang, Z. Tang, F. Yuan, and J. Li, Adv. Mater. Opt. Electron. {\bf
644: 9}
645: (1999) 205.
646:
647: \bibitem{golego}
648: N. Golego, S.A. Studenikin, and M. Cocivera, J. Electrochem. Soc. {\bf 147}
649: (2000)
650: 1592.
651:
652: \bibitem{keis}
653: K. Keis, C. Bauer, G. Boschloo, A. Hagfelt, K. Westermark, H. Rensmo, and H.
654: Siegbahn, J. Photochem. Photobiology A {\bf 148} (2002) 57.
655:
656: \bibitem{vaysieres}
657: L. Vaysieres, Adv. Mater. {\bf 15} (2003) 464.
658:
659: \bibitem{tian}
660: Z.R. Tian, J.A. Voigt, J. Liu, B. Mckenzie, M. J. Macdermott, M.A. Rodriguez,
661: H. Konishi, and H. Xu, Nature Mater. {\bf 2} (2003) 821.
662:
663: \bibitem{izaki}
664: M. Izaki and T. Omi, Appl. Phys. Lett. {\bf 68} (1996) 2439.
665:
666: \bibitem{peulon1}
667: S. Peulon and D. Lincot, Adv. Mater. {\bf 8} (1996) 166
668:
669: \bibitem{peulon2}
670: S. Peulon and D. Lincot, J. Electrochem. Soc. {\bf 145} (1998) 864.
671:
672: \bibitem{pauporte1}
673: T. Pauport\'e and D. Lincot, Appl. Phys. Lett. {\bf 75} (1999) 3817.
674:
675: \bibitem{pauporte2}
676: T. Pauport\'e and D. Lincot, J. Electrochem. Soc. {\bf 148} (2001) C310.
677:
678: \bibitem{pauporte3}
679: T. Pauport\'e and D. Lincot, J. Electroanal. Chem. {\bf 517} (2001) 54.
680:
681: \bibitem{choi}
682: K.S. Choi, H.C. Lichtenegger, and G.D. Stucky, J. Am. Chem. Soc. {\bf 124}
683: (2002)
684: 12402.
685:
686: \bibitem{michaelis}
687: E. Michaelis, D. W\"ohrle, J. Rathousky, and M. Wark, Thin Solid Films {\bf 497}
688: (2006) 163.
689:
690: \bibitem{yoshida1}
691: T. Yoshida, K. Terada, D. Schlettwein, T. Oekermann, T. Sugiura, and H. Minoura,
692: Adv. Mater. {\bf 12} (2000) 1214.
693:
694: \bibitem{pauporte4}
695: T. Pauport\'e, T. Yoshida, A. Goux, and D. Lincot, J. Electroanal. Chem. {\bf
696: 534} (2002) 55.
697:
698: \bibitem{yoshida2}
699: T. Yoshida, T. Pauport\'e, D. Lincot, T. Oekermann, and H. Minoura, J.
700: Electrochem.
701: Soc. {\bf 150} (2003) C608.
702:
703: \bibitem{pauporte5}
704: T. Pauport\'e, T. Yoshida, R. Cort\`es, M. Froment , and D. Lincot, J. Phys.
705: Chem.
706: B, {\bf 107} (2003) 10077.
707:
708: \bibitem{okabe}
709: K. Okabe, T. Yoshida, T. Sugiura, and H. Minoura, Trans. Mater. Research. Soc.
710: Jap.,
711: {\bf 26} (2001) 523.
712:
713: \bibitem{yoshida3}
714: T. Yoshida, K. Miyamoto, N. Hibi, T. Sugiura, H. Minoura, D. Schlettwein, T.
715: Oekermann, G. Schneider, and D. W\"ohrle, Chem. Lett. {\bf 27} (1998) 599.
716:
717: \bibitem{yoshida4}
718: T. Yoshida, M. Tochimoto, D. Schlettwein, D. W\"ohrle, T. Sugiura, and H.
719: Minoura,
720: Chem. Mater. {\bf 11} (1999) 2657.
721:
722: \bibitem{schlettwein}
723: D. Schlettwein, T. Oekermann, T. Yoshida, M. Tochimoto, and H. Minoura, J.
724: Electroanal. Chem. {\bf 481} (2000) 42.
725:
726: \bibitem{pauporte6}
727: T. Pauport\'e, F. Bedioui, and D. Lincot, J. Mat. Chem. {\bf 15} (2005) 1552
728:
729: \bibitem{karuppuchamy}
730: S. Karuppuchamy, T. Yoshida, T. Sugiura, and H. Minoura, Thin Solid Films {\bf
731: 397}
732: (2001) 63.
733:
734: \bibitem{goux}
735: A. Goux, T. Pauport\'e, J. Chivot, and D. Lincot, Electrochim. Acta {\bf 50}
736: (2005)
737: 2239.
738:
739: \bibitem{yoshida5}
740: T. Yoshida, M. Iwaya, H. Ando, T. Oekermann, K. Nonomura, D. Schlettwein, D.
741: W\"ohrle, and H. Minoura, Chem. Comm. (2004) 400.
742:
743: \bibitem{pauporte7}
744: T. Pauport\'e, T. Yoshida, D. Komatsu, and H. Minoura, Electrochem. Solid State
745: Lett. {\bf 9} (2006) H16.
746:
747: \bibitem{lincot1}
748: D. Lincot, T. Pauport\'e, A. Goux, V. Lair, and T. Yoshida,
749: Meet. Abstr. - Electrochem. Soc. {\bf 501}, 460 (2006).
750:
751: \bibitem{pauporte8}
752: T. Pauport\'e, unpublished results.
753:
754: \bibitem{wv}
755: D. Wolf and J. Villain, Europhys. Lett. {\bf 13} (1990) 389.
756:
757: \bibitem{witten}
758: T. A. Witten and L. M. Sander, Phys. Rev. Lett. {\bf 47} (1981) 1400.
759:
760: \bibitem{meakin}
761: P. Meakin, Phys. Rev. A {\bf 27}, 2616 (1983); {\bf 30} (1984) 4207.
762:
763: \bibitem{racz}
764: Z. Racz and T. Vicsek, Phys. Rev. Lett. {\bf 51} (1983) 2383.
765:
766: \bibitem{jullien}
767: R. Jullien, M. Kolb, and R. Botet, J. Physique (Paris) {\bf 45} (1984) 395.
768:
769: \bibitem{burlatsky}
770: S. F. Burlatsky, G. S. Oshanin, and M. M. Elyashevich, Phys. Lett. A {\bf 151},
771: (1990) 538.
772:
773: \bibitem{sanchez}
774: A. S\'anchez, M. J. Bernal, and J. M. Riveiro, Phys. Rev. E {\bf 50} (1994)
775: R2427 .
776:
777: \bibitem{castro1}
778: M. Castro, R. Cuerno, A. S\'anchez and F. Dom\'{\i}nguez-Adame, Phys. Rev. E
779: {\bf 57} (1998) R2491.
780:
781: \bibitem{castro2}
782: M. Castro, R. Cuerno, A. S\'anchez and F. Dom\'{\i}nguez-Adame, Phys. Rev. E
783: {\bf 62} (2002) 161.
784:
785: \bibitem{schwarzacher}
786: W. Schwarzacher, J. Phys.: Condens. Matter {\bf 16} (2004) R859.
787:
788: \bibitem{riveiro}
789: J. M. Riveiro and M. J. Bernal, J. Non-Cryst. Solids {\bf 160} (1993) 18.
790:
791: \bibitem{ew}
792: S. F. Edwards and D. R. Wilkinson, Proc. R. Soc. London {\bf 381} (1982) 17.
793:
794: \bibitem{barabasi}
795: A. L. Barab\'asi and H. E. Stanley, {\it Fractal concepts in surface growth}
796: (Cambridge University Press, Cambribge, England, 1995).
797:
798: \bibitem{leger1}
799: C. L\'eger, J. Elezgaray, and F. Argoul, Phys. Rev. Lett. {\bf 78} (1997) 5010.
800:
801: \bibitem{leger2}
802: C. L\'eger, J. Elezgaray, and F. Argoul, Phys. Rev. Lett. {\bf 58} (1998) 7700.
803:
804: \bibitem{GGG}
805: B. Grossmann, H. Guo, and M. Grant, Phys. Rev. A {\bf 43} (1991) 1727.
806:
807: \bibitem{NT}
808: T. Nattermann and L.-H. Tang, Phys. Rev. A {\bf 45} (1992) 7156.
809:
810: \bibitem{AF}
811: J. G. Amar and F. Family, Phys. Rev. A {\bf 45} (1992) R3373.
812:
813: \bibitem{chamereis}
814: A. Chame and F. D. A. Aar\~ao Reis, Phys. Rev. E {\bf 66} (2002)
815: 051104.
816:
817: \bibitem{rdcor}
818: F. D. A. Aar\~ao Reis, Phys. Rev. E {\bf 73} (2006) 021605.
819:
820: \end{thebibliography}
821:
822:
823: \vfill\eject
824:
825: \begin{figure}[!h]
826: \centering
827: \includegraphics[clip,
828: width=\textwidth,
829: height=0.85\textheight,
830: angle=0]{ez1f1.ps}
831: \caption{\label{fig1} FESEM cross-sectional view of a $ZnO$/EY thin film
832: prepared by electrodeposition.}
833: \end{figure}
834:
835:
836: \begin{figure}[!h]
837: \includegraphics[clip,width=0.80\textwidth,
838: height=0.25\textheight,angle=0]{ez1f2.eps}
839: \caption{\label{fig2} Scheme for the rate equation model, with the $ZnO$ film of
840: height $h$
841: growing from the electrode below it, and straight branches of an additive of
842: height $H$.}
843: \end{figure}
844:
845: \vskip 5cm
846:
847: \begin{figure}
848: \includegraphics[clip,width=0.70\textwidth,
849: height=0.35\textheight,angle=0]{ez1f3.eps}
850: \caption{\label{fig3} The solid curve shows the $ZnO$ film growth rate as a
851: function of the additive concentration in the rate equation model. The dashed
852: curve is an extension of the solid curve for $\rho_A <\rho_c$ [curve
853: $r_A\left( \rho_A \right)$]. For higher concentrations of the additive, the
854: diffusion of the oxygen precursor becomes limiting, which corresponds to the
855: plateau in the growth rate.}
856: \end{figure}
857:
858: \begin{figure}[!h]
859: \includegraphics[clip,width=0.80\textwidth,
860: height=0.42\textheight,angle=0]{ez1f4.eps}
861: \caption{\label{fig4} (a) Species in the solution of the statistical model are
862: represented by squares
863: of sizes equal to $1$ lattice unit (O) and $3$ lattice units (E). (b)
864: Aggregated species (ZA and EA) are represented by squares of the same sizes of
865: the solution.}
866: \end{figure}
867:
868: \begin{figure}[!h]
869: \includegraphics[clip,width=1.0\textwidth,
870: height=0.45\textheight,angle=0]{ez1f5.eps}
871: \caption{\label{fig5} (a) Rule for aggregation of E particles ($E\to EA$), which
872: is possible only upon contact with an EA particle.
873: (b) Probabilistic rules for aggregation of O particles ($O\to ZA$). The
874: aggregation occur with probability $1$ in contact with EA.
875: (c) Precipitation and diffusion of a ZA particle (indicated by an arrow)
876: immediately after its creation.
877: $R$ is the radius to search for the final aggregation point. $\times$ indicates
878: the point to be chosen in this case, which presents the largest number of
879: neighbors ZA.
880: }
881: \end{figure}
882:
883: \begin{figure}[!h]
884: \includegraphics[clip,width=\textwidth,
885: height=0.68\textheight,angle=0]{ez1f6.eps}
886: \caption{\label{fig6} Section of lateral size $128$ of a deposit grown with
887: $D_O=1$, $D_E=1/8$, $\rho_O=0.005$, $\rho_E=0.00005$, $D_A=0.05$ and
888: $p={10}^{-4}$ at $t=3\times {10}^5$, with the solution above it. The horizontal
889: line at the bottom represents the electrode.}
890: \end{figure}
891:
892: \begin{figure}[!h]
893: \includegraphics[clip,width=\textwidth,
894: height=0.68\textheight,angle=0]{ez1f7.eps}
895: \caption{\label{fig7} Initial evolution of a section of lateral size $256$ of a
896: deposit grown with $D_O=1$, $D_E=1$, $\rho_O=0.02$, $\rho_E=0.0002$, $D_A=0.05$
897: and $p=0.001$. Growth times are:
898: (a) $t=5\times {10}^5$, (b) $t={10}^6$ and (c) $t=1.5\times {10}^6$. The
899: horizontal
900: line at the bottom represents the electrode.}
901: \end{figure}
902:
903: \begin{figure}[!h]
904: \includegraphics[clip,width=\textwidth,
905: height=0.68\textheight,angle=0]{ez1f8.eps}
906: \caption{\label{fig8} Initial evolution of a section of lateral size $256$ of a
907: deposit grown with $D_O=1$, $D_E=2$, $\rho_O=0.005$, $\rho_E=0.00005$,
908: $D_A=0.05$ and $p=0.25$. Growth times are:
909: (a) $t={10}^6$, (b) $t=2\times {10}^6$ and (c) $t=3\times {10}^6$. The
910: horizontal
911: line at the bottom represents the electrode.}
912: \end{figure}
913:
914: \begin{figure}[!h]
915: \includegraphics[clip,width=\textwidth,
916: height=0.68\textheight,angle=0]{ez1f9.eps}
917: \caption{\label{fig9} Section of lateral size $1024$ of a
918: deposit grown with $D_O=1$, $D_E=2$, $\rho_O=0.005$, $\rho_E=0.00005$,
919: $D_A=0.01$ and $p=0.25$, at $t=2\times {10}^7$. The
920: horizontal
921: line at the bottom represents the electrode.}
922: \end{figure}
923:
924: \end{document}
925:
926:
927: