1: \documentclass[twocolumn,showpacs,amsmath,amssymb,aps,prb,floatfix,nofootinbib,superscriptaddress]{revtex4}

2: \usepackage{graphicx,graphics,times}% Include figure files

3: \usepackage{bm}% bold math

4: \begin{document}

5: \title

6: {Etched Glass Surfaces, Atomic Force Microscopy and Stochastic

7: Analysis}

8: \author

9: { G. R. Jafari $^{a,b}$, M. Reza Rahimi Tabar $^{c,d}$, A. Iraji zad

10: $^{c}$, G. Kavei $^f$}

11: \address

12: {\it $^a$ Department of Physics, Shahid Beheshti University, Evin,

13: Tehran 19839, Iran\\

14: $^b$ Department of Nano-Science, IPM,

15: P. O. Box 19395-5531, Tehran, Iran\\

16: $^c$  Department of Physics, Sharif University of

17: Technology, P. O. Box 11365-9161, Tehran, Iran \\

18: $^d$ CNRS UMR 6529, Observatoire de la C$\hat o$te d'Azur, BP

19: 4229, 06304 Nice Cedex 4, France\\

20:  $^e$ Material and Energy,

21: Research Center, P.O. Box 14155-4777, Tehran, Iran}

22: 

23: 

24: 

25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

26: %ABSTRACT

27: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

28: \begin{abstract}

29: 

30: The effect of etching time scale of glass surface on its statistical

31: properties has been studied using atomic force microscopy technique.

32: We have characterized the complexity of the height fluctuation of a

33: etched surface by the stochastic parameters such as intermittency

34: exponents, roughness, roughness exponents, drift and diffusion

35: coefficients and find their variations in terms of the etching time.

36: 

37: %PACS: {}

38: 

39: \end{abstract}

40: 

41: \maketitle

42: 

43: \section{Introduction}

44: 

45: The complexity of rough surfaces is subject of a large variety of

46: investigations in different fields of science

47: \cite{Barabasi,Davies}. Surface roughness has an enormous influence

48: on many important physical phenomena such as contact mechanics,

49: sealing, adhesion, friction and self-cleaning paints and glass

50: windows, \cite{Bo,Zhao}. A surface roughness of just a few

51: nanometers is enough to remove the adhesion between clean and

52: (elastically) hard solid surfaces \cite{Bo}. The physical and

53: chemical properties of surfaces and interfaces are to a significant

54: degree determined by their topographic structure. The technology of

55: micro fabrication of glass is getting more and more important

56: because glass substrates are currently being used to fabricate micro

57: electro mechanical system (MEMS) devices \cite{Won}. Glass has many

58: advantages as a material for MEMS applications, such as good

59: mechanical and optical properties. It is a high electrical

60: insulator, and it can be easily bonded to silicon substrates at

61: temperatures lower than the temperature needed for fusion bonding

62: \cite{Melvin}. Also micro and nano-structuring of glass surfaces is

63: important for the production of many components and systems such as

64: gratings, diffractive optical elements, planar wave guide devices,

65: micro-fluidic channels and substrates for (bio) chemical

66: applications \cite{Cheng}. Wet etching is also well developed for

67: some of these applications

68: \cite{Knotter,Spierings,Schuitema,Glebov,Jafari1,Silikas,Irajizad}.

69: 

70:  One of the main problems in the rough surface is the scaling

71: behavior of the moments of height $h$ and  evolution of the

72: probability density function (PDF) of $h$, i.e. $P(h, x)$ in terms

73: of the length scale $x$. Recently some authors have been able to

74: obtain a Fokker-Planck equation describing the evolution of the

75: probability distribution function in terms of the length scale, by

76: analyzing some stochastic phenomena, such as rough surfaces

77: \cite{Jafari2,Waechter,Sangpour}, turbulent system \cite{Renner},

78: financial data \cite{Renner2},  cosmic background radiation

79: \cite{Ghasemi} and heart interbeats \cite{pei04} etc. They noticed

80: that the conditional probability density of field increment

81: satisfies the Chapman-Kolmogorov equation. Mathematically, this is

82: a necessary condition for the fluctuating data to be a Markovian

83: process in the length (time) scales \cite{Risken}. \vskip 1cm

84: 

85:  In this work,

86: we investigate the etching process as a stochastic process. We

87: measure the intermittency exponents of height structure function,

88: roughness, roughness exponents and Kramers-Moyal`s (KM)

89: coefficients. Indeed we consider the etching time $t$, as an

90: external parameter, to control the statistical properties of a

91: rough surface and find their variations with $t$. It is shown that

92: the first and second KM`s coefficients have well-defined values,

93: while the third and fourth order coefficients tend to zero. The

94: first and second KM`s coefficients for the fluctuations of $h(x)$,

95: enables us to explain the height fluctuation of the etched glass

96: surface.

97: 

98: \section{Experimental}

99: 

100:  We started with glass microscope slides as a sample. Only one

101: side of samples was etched by HF solution for different

102:  etching time (less than 20 minutes). HF concentration was $\%40$ for

103: all the experiments. The surface topography of the etched glass

104: samples in the scale ($< 5 \mu m$) was obtained using an AFM (Park

105: Scientific Instruments). The images in this scale were collected in

106: a constant force mode and digitized into $256$$\times$$256$ pixels.

107: A commercial standard pyramidal $Si_3N_4$ tip was used. A variety of

108: scans, each with size $L$, were recorded at random locations on the

109: surface. Figure 1 shows typical AFM image with resolutions of about

110: $20 nm$.

111: 

112: \section{Statistical quantities}

113: 

114: \subsection{Multifractal Analysis and the Intermittency Exponent }

115: 

116:  Assuming statistical translational invariance, the

117: structure functions $S^{q}(l)=<|h(x+l)-h(x)|^{q}>$, (moments of the

118: increment of the rough surface height fluctuation $h(x)$) will

119: depend only on the space deference of heights $l$, and has a power

120: law behavior if the process has the scaling property:

121: \begin{equation}

122: S^{q}(l)=<|h(x+l)-h(x)|^{q}> \propto

123: S^{q}(L_{0})(\frac{l}{L_{0}})^{\xi(q)}

124: \end{equation}

125: where $L_{0}$ is the fixed largest length scale of the system, $<

126: \cdot \cdot \cdot>$ denotes statistical average (for non-overlapping

127: increments of length $l$), $q$ is the order of the moment (we take

128: here $q > 0$), and $\xi(q)$ is the exponents of structure function.

129: The second moment is linked to the slope $\beta$ of the Fourier

130: power spectrum: $\beta=1+\xi_{2}$. The main property of a

131: multifractal processes is that it is characterized by a non-linear

132: $\xi_{q}$ function verses $q$. Monofractals are the generic result

133: of this linear behavior. For instance, for Brownian motion (Bm)

134: $\xi_{q} = q/2$, and for fractional Brownian motion (fBm) $\xi_{q}

135: \propto q$.

136: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

137: \begin{figure}

138: \includegraphics[width=8truecm]{fig1.eps}

139: %\epsfxsize=8truecm\epsfbox{fig1.eps}

140:  \narrowtext \caption{AFM surface image of etched glass

141: film with size  $5\times 5 \mu m^2 $  after 12 minutes.}

142: \end{figure}

143: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

144: 

145: \subsection {Roughness and Roughness Exponents}

146: 

147: It is also known that to derive the quantitative information of the

148: surface morphology one may consider a sample of size $L$ and define

149: the mean height of growing film $\overline{h}$ and its {\it

150: variance}, $\sigma$ by:

151: \begin{equation}\label{w}

152: \sigma(L,t) =(\langle (h-\overline{h})^2\rangle)^{1/2}

153: \end{equation}

154: where $t$ is etching time and $\langle\cdots\rangle$ denotes an

155: averaging over different samples, respectively. Moreover, etching

156: time is a factor which can apply to control the surface roughness of

157: thin films.

158: 

159:  Let us now calculate also the roughness exponent of the

160: etched glass. Starting from a flat interface (one of the possible

161: initial conditions), it is conjectured that a scaling of space by

162: factor $b$ and of time by factor $b^z$ ($z$ is the dynamical scaling

163: exponent), rescales the variance, $\sigma$ by factor $b^{\chi}$ as

164: follows \cite{Barabasi}:

165: \begin{equation}\label{scaling}

166: \sigma(bL,b^zt)=b^{\alpha}\sigma(L,t)

167: \end{equation}

168: which implies that

169: \begin{equation}

170: \sigma(L,t)=L^{\alpha}f(t/L^z).

171: \end{equation}

172: If for large $t$ and fixed $L$ $(x=t / L^z \rightarrow \infty)$

173: $\sigma$ saturate.  However, for fixed large $L$ and $t\ll L^z$, one

174: expects that correlations of the height fluctuations are set up only

175: within a distance $t^{1/z}$ and thus must be independent of $L$.

176: This implies that for $x \ll 1$, $f(x)\sim x^{\beta}$ with

177: $\beta=\alpha / z$. Thus dynamic scaling postulates that

178: \begin{eqnarray}

179: \sigma(L,t)\propto

180: \left\{%

181: \begin{array}{ll}

182:    t^{\beta}, & \hbox{t$\ll L^z$;}\\

183:    L^{\alpha}, & \hbox{t$\gg L^{z}$}. \\

184: \end{array}%

185: \right.

186: \end{eqnarray}

187: The roughness exponent $\alpha$ and the dynamic exponent $\beta$

188: characterize the self-affine geometry of the surface and its

189: dynamics, respectively.

190: 

191: The common procedure to measure the roughness exponent

192:  of a rough surface is use of the surface structure function depending

193: on the length scale $l$ which is defined as:

194: \begin{eqnarray}\label{Structure}

195: S^{2}(l)=\langle|h(x+l)-h(x)|^2\rangle.

196: \end{eqnarray}

197: It is equivalent to the statistics of height-height correlation

198: function $C(l)$ for stationary surfaces, i.e.

199: $S^{2}(l)=2\sigma^2(1-C(l))$. The second order structure function

200: $S(l)$, scales with $l$ as $ l^{2\alpha}$ \cite{Barabasi}.

201: 

202: 

203: \subsection{The Markov Nature of Height Fluctuations: Drift and

204: Diffusion Coefficients}

205: 

206: We check whether the data of height fluctuations follow a Markov

207: chain and, if so, measure the Markov length scale $l_M$. As is

208: well-known, a given process with a degree of randomness or

209: stochasticity may have a finite or an infinite Markov length scale

210: \cite{Markov}. The Markov length scale is the minimum length

211: interval over which the data can be considered as a Markov process.

212: To determine the Markov length scale $l_M$, we note that a complete

213: characterization of the statistical properties of random

214: fluctuations of a quantity $h$ in terms of a parameter $x$ requires

215: evaluation of the joint PDF, i.e. $P_N(h_1,x_1;....;h_N,x_N)$, for

216: any arbitrary $N$. If the process is a Markov process (a process

217: without memory), an important simplification arises. For this type

218: of process, $P_N$ can be generated by a product of the conditional

219: probabilities $P(h_{i+1},x_{i+1}|h_i,x_i)$, for $i=1,...,N-1$. As a

220: necessary condition for being a Markov process, the

221: Chapman-Kolmogorov equation,

222: \begin{eqnarray}

223:   &&P(h_2,x_2|h_1,x_1)= \cr \nonumber\\

224:   &&\int \hbox{d} (h_i)\,

225:   P(h_2,x_2|h_i,x_i)\, P(h_i,

226:   x_i| h_1,x_1)

227: \end{eqnarray}

228: should hold for any value of $x_i$, in the interval $ x_2<x_i<x_1$

229: \cite{Risken}.

230: 

231: The simplest way to determine $l_M$ for homogeneous surface is the

232: numerical calculation of the quantity,

233: $S=|P(h_2,x_2|h_1,x_1)-\int\hbox{d}

234: h_3P(h_2,x_2|h_3,x_3)\,P(h_3,x_3|h_1,x_1)|$, for given $h_1$ and

235: $h_2$, in terms of, for example, $x_3-x_1$ and considering the

236: possible errors in estimating $S$. Then, $l_M=x_3-x_1$ for that

237: value of $x_3-x_1$ such that, $S=0$ \cite{Markov}.

238: 

239:  It is well-known, the Chapman-Kolmogorov

240: equation yields an evolution equation for the change of the

241: distribution function $P(h,x)$ across the scales $x$. The

242: Chapman-Kolmogorov equation formulated in differential form yields a

243: master equation, which can take the form of a Fokker-Planck equation

244: \cite{Risken,Markov}:

245: \begin{eqnarray}\label{Fokker}

246:   \frac {\partial}{\partial x} P(h,x)=

247:  [-\frac{\partial }{\partial h}

248:   D^{(1)}(h,x)

249:   +\frac{\partial^2 }{\partial h^2} D^{(2)}(h,x)]

250:   P(h,x).

251: \end{eqnarray}

252: The drift and diffusion coefficients $D^{(1)}(h, r)$, $D^{(2)}(h,

253: r)$ can be estimated directly from the data and the moments

254: $M^{(k)}$ of the conditional probability distributions:

255: \begin{eqnarray}\label{D(k)}

256:   && D^{(k)}(h,x) = \frac{1}{k!}

257:    {\rm lim}_{r \rightarrow 0}  M^{(k)} \cr \nonumber \\

258:   && M^{(k)} = \frac{1}{r}  \int dh'

259:   (h'-h)^k P(h',x+r|

260:   h,x).

261: \end{eqnarray}

262: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

263: \begin{figure}

264: %\begin{center}

265: \includegraphics[width=8truecm]{fig2.eps}

266: %\epsfxsize=8truecm\epsfbox{fig2.eps}

267:  \narrowtext \caption{

268: Scaling of the structure functions in log-log plot for moments less

269: than 8. (from bottom to top).}

270: %\end{center}

271: \end{figure}

272: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

273: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

274: \begin{figure}

275: %\begin{center}

276: \includegraphics[width=8truecm]{fig3.eps}

277: %\epsfxsize=8truecm\epsfbox{fig3.eps}

278:  \narrowtext \caption{

279: The results of scaling exponent $\xi_{q}$ which is clearly linear

280: vs. q.}

281: %\end{center}

282: \end{figure}

283: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

284: The coefficients $D^{(k)}(h,x)$`s are known as Kramers-Moyal

285: coefficients.  According to Pawula`s theorem \cite{Risken}, the

286: Kramers-Moyal expansion stops after the second term, provided that

287: the fourth order coefficient $D^{(4)} (h,x)$ vanishes \cite{Risken}.

288:  The forth order coefficients $D^{(4)}$

289:  in our analysis was found to be about $ {D^{(4)}} \simeq 10^{-4} {D^

290: {(2)}}$. In this approximation, we can ignore the coefficients

291: $D^{(n)}$ for $n \geq 3$. We note that this Fokker-Planck equation

292: is equivalent to the following Langevin equation (using the Ito

293: interpretation) \cite{Risken}:

294: \begin{equation}\label{Langevin}

295:   \frac{\partial}{\partial x}  h(x)=D^{(1)}(h,x) +

296:   \sqrt{D^{(2)}(h,x)}f(x)

297: \end{equation}

298: where $f(x)$ is a random force, zero mean with gaussian statistics,

299: $\delta$-correlated in $x$, i.e. $\langle

300: f(x)f(x')\rangle=2\delta(x-x')$. Furthermore, with this last

301: expression, it becomes clear that we are able to separate the

302: deterministic and the noisy components of the surface height

303: fluctuations in terms of the coefficients $D^{(1)}$ and $D^{(2)}$.

304: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

305: \begin{figure}

306: %\begin{center}

307: \includegraphics[width=8truecm]{fig4.eps}

308: %\epsfxsize=8truecm\epsfbox{fig4.eps}

309: \narrowtext \caption{ Log-Log

310: plot of selection structure function of the etched glass surfaces.}

311: %\end{center}

312: \end{figure}

313: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

314: 

315: \section{Results and Discussion}

316: 

317: Now, using the introduced statistical parameters in the previous

318: sections, it is possible to obtain some quantitative information

319: about the effect of etching time on surface topography of the glass

320: surface. To study the effect of the etching time on the surface

321: statistical characteristics, we have utilized AFM  imaging technique

322: in order to obtain microstructural data of the etched glass surfaces

323: at the different etching time in the HF. Figure 1 shows the AFM

324: image of etched glass after $12$ minuets etched. To investigate the

325: scaling behavior of the moments of $\delta h_{l}=h(x+l)-h(x)$, we

326: consider the samples that they reached to the stationary state. This

327: means that their statistical properties do not change with time. In

328: our case the samples with etching time more than $20$ minutes are

329: almost stationary. Figure 2 shows the log-log plot of the structure

330: functions verses length scale $l$ for different orders of moments.

331: The straight lines show that the moments of order $q$ have the

332: scaling behavior. We have checked the scaling relation up to moment

333: $q=10$.  The resulting intermittency exponent $\xi_{q}$ is shown in

334: figure 3. It is evident that $\xi_{q}$ has a linear behavior. This

335: means that the height fluctuations are mono-fractal behavior. We

336: also directly estimated the scaling exponent of the linear term

337: $l^{qH}/<(h(x+l)-h(x))^{q}>$ and obtain the following values for the

338: samples with 20 minuets etching time, $\xi_{1} = 0.70 \pm 0.04$ and

339: $\xi_{2}= 1.40 \pm 0.04$. This means etching memorize fractal

340: feature during etching. Therefore using the scaling exponent

341: $\xi_{2}$ we obtain the roughness exponent $\alpha$ as $ \xi_{2}/2=

342: 0.70 \pm 0.04$.

343: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

344: \begin{figure}

345: \begin{center}

346: \includegraphics[width=8truecm]{fig5.eps}

347: %\epsfxsize=8truecm\epsfbox{fig5.eps}

348:  \narrowtext \caption{ Drift

349: coefficients of the surfaces at different etching time less than 20

350: minutes.}

351: \end{center}

352: \end{figure}

353: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

354: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

355: \begin{figure}

356: \begin{center}

357: \includegraphics[width=8truecm]{fig6.eps}

358: %\epsfxsize=8truecm\epsfbox{fig6.eps}

359: \narrowtext \caption{ Diffused

360: coefficients of the surface at different etching time less than 20

361: minutes.}

362: \end{center}

363:  \end{figure}

364: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

365:  Figure 4 presents the structure function $S(l)$ of

366: the surface at the different etching time, using equation

367: (\ref{Structure}). It is also possible to evaluate the grain size

368: dependence to the etching time, using the correlation length

369: achieved by the structure function represented in figure 4. The

370: correlation lengths increase with etching time. Its value has a

371: exponential behavior $448(1-\exp(-0.15t))nm$. Also we find that the

372: dynamical exponent is given by $\beta = 0.6\pm 0.1$. Also we

373: measured the variation of the Markov length with etching time $t$

374: (min), and obtain $l_{M}=40+3t$ (nm) for time scales $t<20 ~min$.

375: 

376: Finally to obtain the stochastic equation of the height fluctuations

377: behavior of the surface, we need to measure the Keramer- Moyal

378: Coefficients. In our analysis the forth order coefficients $D^{(4)}$

379: is less than Second order coefficients, $D^{(2)}$, about $ {D^{(4)}}

380: \simeq 10^{-4} {D^ {(2)}}$. In this approximation, we ignore the

381: coefficients $D^{(n)}$ for $n \geq 3$. So, to discuss the surfaces

382: it just needs to measure the drift coefficient

383: $D^{(1)}(\frac{h}{\sigma})$ and diffusion coefficient

384: $D^{(2)}(\frac{h}{\sigma})$ using Eq. (\ref{D(k)}). Figures 5 and 6

385: show the drift coefficient $D^{(1)}(\frac{h}{\sigma})$ and diffusion

386: coefficients $D^{(2)}(\frac{h}{\sigma})$ for the surfaces at the

387: different etching time, respectively. It can be shown that the drift

388: and diffusion coefficients have the following behavior,

389: \begin{eqnarray}\label{D1}

390: D^{(1)} (\frac{h}{\sigma},t)=-f^{(1)}(t)\frac{h}{\sigma}

391: \end{eqnarray}

392: \begin{eqnarray}\label{D2}

393: D^{(2)} (\frac{h}{\sigma},t)=f^{(2)}(t)(\frac{h}{\sigma})^2

394: \end{eqnarray}

395: The two coefficients $f^{(1)}(t)$ and $f^{(2)}(t)$ increase with the

396: $\frac{h}{\sigma}$ then is saturated. Using the data analysis we

397: obtain that they are linear verses time (min): $f^{(1)}(t)=0.005 t$

398: and $f^{(2)}(t)=0.0003 t$ for time scales $t<20 ~min$. To better

399: comparing the parameter of samples we divided the heights to their

400: variances.  In this case, maximum and minimum of heights are about

401: plus 1 and mines 1, respectively. Comparing samples with etching

402: times 2 and 6 minutes, shows $f^{(1)}$ increases 300 percent after 4

403: minutes (from 2 min to 6 min) from $f^{(1)}(t=2 \times 60)=0.6$ to

404: $f^{(1)}(t=6 \times 60)=1.8$. Also, $f^{(2)}$ is $0.006$ and $0.018$

405: after 2 and 6  minutes, respectively.

406: 

407: \section{Conclusions}

408: 

409: We have investigated the role of etching time, as an external

410: parameter, to control the statistical properties of a rough

411: surface.  We have shown that in the saturate state the structure

412: of topography has fractal feature with fractal dimension

413: $D_{f}=1.30$. In addition, Langevin characterization of the etched

414: surfaces enable us to regenerate the rough surfaces grown at the

415: different etching time, with the same statistical properties in

416: the considered scales \cite{Jafari2}.

417: 

418: \section{Acknowledgment}

419: 

420: We would like to thank S. M. Mahdavi for his useful comments and

421: discussions and Also P. Kaghazchi and M. Shirazi for samples

422: preparation.

423: %!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

424: %!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

425: 

426: \begin{thebibliography}{99}%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

427: \bibitem{Barabasi} A.L. Barabasi and H.E. Stanley, {\it Fractal

428: Concepts in Surface Growth} (Cambridge  University Press, New York,

429: 1995).

430: \bibitem{Davies} S. Davies, P. Hall , J. Roy. Stat. Soc. B 61 (1999) 3.

431: \bibitem{Bo} A. G. Peressadko, N. Hosoda, and B. N. J. Persson, PRL 95, 124301

432: (2005), B N J Persson, O Albohr, U Tartaglino, A I Volokitin and E

433: Tosatti, J. Phys.: Condens. Matter 17 (2005) R1–R62.

434: \bibitem{Zhao} Zhao Y-P,Wang L S and Yu T X, J. Adhes. Sci. Technol. 17 519

435: (2003)

436: \bibitem{Won} Won Ick Jang, Chang Auck Choi, Myung Lae Lee, Chi Hoon

437: Jun and Youn Tae Kim, J. Micromech. Microeng. 12 (2002) 297–306.

438: \bibitem{Melvin} M. Bu , T. Melvin a, G.  J. Ensell, J. S. Wilkinson,

439: A. G.R. Evans, Sensors and Actuators A 115:pp. 476-482 (2004).

440: \bibitem{Cheng}Yu-Cheng Lin, Hsiao-Ching Ho, Chien-Kai Tseng and Shao-Qin Hou 2001

441: J. Micromech. Microeng. 11 189-194

442: \bibitem{Knotter} D.M. Knotter, J. Am. Chem. Soc. 122 (2000) 4345.

443: \bibitem{Spierings} G.A.C.M. Spierings, J. Mater. Sci. 28 (1993) 6261.

444: \bibitem{Schuitema} R. Schuitema, et al, Light scattering at rough

445: interfaces of thin film solar cells to

446:  improve the efficiency and stability, IEEE/ProRISC99, pp 399-404

447: (1999).

448: 

449: % 11-20

450: 

451: \bibitem{Glebov} L. B. Glebov, et al, Photo induced chemical etching of

452: silicate and borosilicate glasses, Glasstech.

453:  Ber. Glass Sci. Technol. 75 C2 pp 298 - 301 (2002).

454: \bibitem{Jafari1} G. R. Jafari, S. M. Mahdavi, A. Iraji zad, and P.

455: Kaghazchi, Surface And Interface Analysis; 37: 641 –645 (2005).

456: \bibitem{Silikas} N. Silikas, k.E.R. England, D.C Wattes, K.D Jandt, J.

457: Dentistry 27 (1999) 137.

458: \bibitem{Irajizad}  A. Irajizad, G. Kavei, M. Reza Rahimi Tabar, and

459: S.M. Vaez Allaei, J. Phys.: Condens. Matter {\bf 15}, 1889 (2003).

460: \bibitem{Jafari2} G.R. Jafari, S.M. Fazeli, F. Ghasemi, S.M. Vaez

461: Allaei, M. Reza Rahimi Tabar, A. Irajizad, and G. Kavei, Phys. Rev.

462: Lett. {\bf 91}, 226101 (2003).

463: \bibitem{Waechter} M. Waechter, F. Riess, Th. Schimmel, U. Wendt and J.

464: Peinke, Eur. Phys. J. B 41, 259-277 (2004).

465: \bibitem{Sangpour} P. Sangpour, G. R. Jafari, O. Akhavan, A.Z.

466: Moshfegh, and M. Reza Rahimi Tabar, Phys. Rev. B 71, 155423 (2005).

467: \bibitem{Renner} Christoph Renner, Joachim Peinke, and Rudolf

468: Friedrich, Journal of Fluid Mechanics, 433:383–409, 2001.

469: \bibitem{Renner2} Ch. Renner, J. Peinke, R. Friedrich, Physica A 298, 499 (2001).

470: \bibitem{Ghasemi} F. Ghasemi, A. Bahraminasab, S. Rahvar, and M. Reza

471: Rahimi Tabar, Preprint arxiv:astro-phy/0312227, 2003.

472: \bibitem{pei04}

473: F. Ghasemi, J. Peinke, M. Sahimi and M. Reza Rahimi Tabar, Eur.

474: Phys. J. B 47, 411(2005)

475: \bibitem{Risken} H. Risken, {\it The Fokker-Planck equation} (Springer,

476: Berlin, 1984).

477: \bibitem{Markov} R. Friedrich, J. Zeller, and J. Peinke, Europhysics Letters 41, 153

478: (1998).

479: 

480: \end{thebibliography}

481: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

482: %%%%%%%%%%%%%%%%%%%%%%%%%%

483: \end{document}

484: