1: %%
2: \tolerance=10000
3: \documentstyle[twocolumn,prb,aps,floats,epsfig]{revtex}
4: \begin{document}
5: \twocolumn[\hsize\textwidth\columnwidth\hsize\csname
6: @twocolumnfalse\endcsname
7: \title{A model of transport nonuniversality in
8: thick-film resistors}
9: \author{C. Grimaldi}
10: \address{Institut de Production et Robotique, LPM,
11: EPFL,
12: CH-1015 Lausanne, Switzerland}
13: \author{T. Maeder}
14: \address{Institut de Production et Robotique, LPM,
15: EPFL,
16: CH-1015 Lausanne, Switzerland}
17: \address{Sensile Technologies SA, PSE, CH-1015 Lausanne, Switzerland}
18: \author{P. Ryser}
19: \address{Institut de Production et Robotique, LPM,
20: EPFL,
21: CH-1015 Lausanne, Switzerland}
22: \author{S. Str\"assler}
23: \address{Institut de Production et Robotique, LPM,
24: EPFL,
25: CH-1015 Lausanne, Switzerland}
26: \address{Sensile Technologies SA, PSE, CH-1015 Lausanne, Switzerland}
27: \maketitle
28:
29: %\centerline \\
30:
31: \begin{abstract}
32: We propose a model of transport in thick-film resistors which naturally
33: explains the observed nonuniversal values of the conductance exponent $t$
34: extracted in the vicinity of the percolation transition.
35: Essential ingredients of the model are the segregated microstructure typical
36: of thick-film resistors and tunneling between the conducting grains.
37: Nonuniversality sets in as consequence of wide distribution of interparticle
38: tunneling distances.
39:
40: PACS numbers: 72.60.+g, 64.60.Fr, 72.80.Tm
41: \end{abstract}
42: \vskip 2pc ]
43:
44:
45: \newpage
46:
47: %\narrowtext
48: \centerline \\
49: Thick-film resistors (TFRs) are glass-conductor
50: composites based on RuO$_2$ (but also Bi$_2$Ru$_2$O$_7$,
51: Pb$_2$Ru$_2$O$_6$, and IrO$_2$) grains mixed and fired with glass powders.\cite{prude1}
52: Besides the widespread use of TFRs in
53: pressure and force sensor applications,\cite{white} their transport properties are
54: of great interest also for basic research.
55: The percolating nature of transport in TFRs
56: has been reported since long time and now
57: it is well documented.\cite{pike,dejeu,carcia1,carcia2,listki,tambo,kusy}
58: As shown in Fig.\ref{fig1} where we reports a selection of previously published data
59: on different TFRs,\cite{carcia1,carcia2,listki,tambo}
60: the conducting phase concentration $x$ dependence of the conductance
61: $G$ of TFRs follows a percolating-like power-law equation of the form:
62: \begin{equation}
63: \label{eq1}
64: G=G_0(x-x_c)^t,
65: \end{equation}
66: where $G_0$ is a prefactor, $x_c$ is the critical concentration below which
67: $G$ vanishes and $t$ is the transport critical exponent.\cite{kirk,stauffer}
68: The values of $G_0$, $x_c$, and $t$ which best fit the experimental data
69: are reported in the inset of Fig.\ref{fig1}.
70:
71: According to the standard theory of transport in isotropic
72: percolating systems,\cite{stauffer}
73: $G_0$ and $x_c$ depend on microscopic details such as the microstructure
74: and the mean value of the junction resistances connecting two
75: neighbouring conducting sites, while, unless the microscopic resistances have a
76: diverging distribution function (see below),
77: the critical exponent $t$ is {\it universal}, {\it i. e.}, it depends only upon
78: the lattice dimensionality $D$. For $D=3$, random resistor network calculations
79: predict $t=t_0\simeq 2.0$,\cite{clerc} in agreement with various granular metal
80: systems,\cite{abeles,lee} or other disordered compounds.\cite{putten}
81:
82: \begin{figure}
83: \protect
84: \centerline{\epsfig{figure=fig1.EPS,width=20pc,clip=}}
85: \caption{Measured conductances on different RuO$_2$
86: (Refs.\protect\onlinecite{carcia2,tambo}) and Bi$_2$Ru$_2$O$_7$ (Refs.\protect\onlinecite{carcia1,listki})
87: TFRs. Solid lines are fits to Eq.(\ref{eq1}) with fitting values reported in the
88: inset. dashed line denotes a power-law with exponent $t=2$.}
89: \label{fig1}
90: \end{figure}
91:
92: As it is clear from Fig.\ref{fig1}, TFRs have values of $t$ ranging from its
93: universal limit $t\simeq 2.0$ (filled squares, Ref.\onlinecite{carcia1})
94: up to very high values like $t\sim 5.0$ (filled diamonds, Ref.\onlinecite{listki})
95: or even higher.\cite{pike}
96: Despite of their clear percolating behavior, TFRs do not fulfil
97: therefore the hypothesis of universality common to many systems and instead belong
98: to a different, quite vast, class of materials which display
99: nonuniversal transport behavior, that is a regime where the transport exponent $t$
100: depends on microscopic details (microstructure etc.).
101: Typical examples of nonuniversal systems are carbon-black--polymer composites,\cite{balb}
102: and materials constituted by insulating regions embedded in a
103: continuous conducting phase.\cite{lee,wu}
104:
105: Despite that TFRs have been historically among
106: the first materials for which transport nonuniversality has been reported,\cite{pike}
107: the microscopic origin of their universality breakdown has not been specifically
108: addressed so far.
109: In this letter we show that the cross-over between universality and
110: nonuniversality reported in Fig.\ref{fig1} can be explained
111: within a single model whose basic features are the peculiar microstructure of TFRs
112: and the tunneling processes between conducting grains.
113:
114:
115: Before describing our model for TFRs, let us first recall the mathematical
116: requisites for universality breakdown in random resistor networks. Consider a regular
117: lattice of sites and assign to each neighbouring couple of sites a bond which
118: has finite conductance $g$ with probability $p$ and zero conductance
119: with probability $1-p$.
120: The resulting conductance distribution function is then:
121: \begin{equation}
122: \label{distri1}
123: \rho(g)=p h(g)+(1-p)\delta (g),
124: \end{equation}
125: where $\delta(g)$ is the Dirac delta-function and $h(g)$ is the distribution
126: function of the finite bond conductances. For well behaved distribution functions
127: $h(g)$, conductivity is universal and follows Eq.(\ref{eq1}) with $t=t_0$.
128: Instead, as first shown by Kogut and Straley,\cite{straley}
129: if $h(g)$ has a power law divergence for small $g$
130: of the form:
131: \begin{equation}
132: \label{distri2}
133: \lim_{g\rightarrow 0} h(g) \propto g^{-\alpha},
134: \end{equation}
135: and $\alpha$ is larger than a critical value $\alpha_c$, then transport is no longer
136: universal. Renormalization group analysis predicts in fact that
137: \begin{equation}
138: \label{nonuni}
139: t=\left\{
140: \begin{array}{ll}
141: t_0 & \mbox{if} \hspace{3mm}\nu+1/(1-\alpha)< t_0 \\
142: \nu+1/(1-\alpha) & \mbox{if}\hspace{3mm} \nu+1/(1-\alpha)> t_0
143: \end{array}
144: \right.,
145: \end{equation}
146: where $\nu\simeq 0.88$ is the
147: correlation-length exponent for a three dimensional lattice.\cite{machta}
148: By using $t_0\simeq 2.0$ we obtain therefore $\alpha_c\simeq 0.107$.
149: Equations (\ref{distri2}) and (\ref{nonuni}) have been shown to arise from
150: a system of insulating spheres embedded in a continuous
151: conducting material (swiss-cheese model),\cite{halperin} and from a
152: tunneling-percolation model
153: with highly fluctuating tunneling distances.\cite{balb}
154: Here we show that Eq.(\ref{distri2}) [and consequently Eq.(\ref{nonuni})]
155: arises naturally from a simple representation of TFRs in terms of
156: their microstructure and elemental transport processes.
157:
158: Let us start by considering the highly non-homogeneous
159: microstructure typical of TFRs. These systems are constituted by a mixture
160: of large glassy particles (typically with size $L$ of order $1$-$3$ $\mu$m) and small
161: conducting grains of size $\Phi$ varying between $\sim 10$ nm up to $\sim 200$ nm.
162: In this situation, the small metallic grains tend to occupy the narrow
163: regions between the much larger insulating zones leading to a filamentary
164: distribution of the conducting phase. A classical model to describe
165: such a segregation effect was proposed already in the 1970's by Pike.\cite{pike}
166: This model treats the
167: glassy particles as cubes of size $L\gg \Phi$ whose edges are occupied by chains
168: of adjacent metallic spheres of diameter $\Phi$. Such chains define channels (bonds)
169: which form a cubic lattice spanning the whole sample. Let us assume for the moment
170: that a bond has probability $p$ of being occupied by a fixed number $n+1$ of spheres
171: and probability $1-p$ of being empty. To each couple of adjacent spheres we assign
172: an inter-sphere conductance $\sigma_i$ ($i=1,\cdots ,n$).
173: A random resistor network can be therefore defined as in Eq.(\ref{distri1})
174: where $h_n(g)$ is the distribution function of the total channel conductance $g$
175: of $n$ conductances $\sigma_i$ in series:
176: \begin{equation}
177: \label{bond1}
178: g^{-1}=\sum_{i=1}^{n}\frac{1}{\sigma_i}.
179: \end{equation}
180: The high values of piezoresistance ({\it i.e.}, the strain sensitivity of transport)
181: typical of TFRs,\cite{prude1} and the low temperature dependence of transport
182: strongly indicate that the main contribution to
183: the overall resistance stems from tunneling processes between neighbouring
184: metallic grains.
185: Hence, if the centers of two neighbouring metallic spheres are separated by a
186: distance $r$, then the intergrain tunneling
187: conductance $\sigma$ is approximatively of the form:
188: \begin{equation}
189: \label{tunnel1}
190: \sigma=\sigma(r)\equiv\sigma_0 e^{-2(r-\Phi)/\xi},
191: \end{equation}
192: where $\sigma_0$ is a constant which we set equal to the unity,
193: $\xi \propto 1/\sqrt{V}$ is the tunneling factor and
194: $V$ is the intergrain barrier potential.
195: Let us make the quite general assumption that the centers of the spheres
196: are set randomly along the channel, so that the distances $r$ change according
197: to the distribution function $P(r)$ for a set of impenetrable spheres arranged randomly
198: in a quasi one dimensional channel. By following Ref.\cite{torquato}, $P(r)$ can be
199: calculated exactly and its explicit expression is:
200: \begin{equation}
201: \label{tran}
202: P(r)=\frac{1}{a_n-\Phi}e^{-(r-\Phi)/(a_n-\Phi)}\Theta(r-\Phi)
203: \end{equation}
204: where $a_n=(1+L/n\Phi)\Phi/2$ is the mean inter-sphere distance and
205: $\Theta$ is the step function. By combining Eq.(\ref{tunnel1}) with Eq.(\ref{tran})
206: the distribution $f(\sigma)$ of the inter-sphere conductances is then:
207: \begin{equation}
208: \label{distri4}
209: f(\sigma)=\int_0^{\infty} \!dr\, P(r)\,\delta[\sigma-\sigma(r)]
210: =(1-\alpha_n)\sigma^{-\alpha_n},
211: \end{equation}
212: where
213: \begin{equation}
214: \label{alfa}
215: \alpha_n=1-\frac{\xi/2}{a_n-\Phi}.
216: \end{equation}
217: To obtain the distribution function $h_n(g)$ of the occupied channels, we first note
218: that Eq.(\ref{bond1}) implyies that $g$ is dominated by the minimum inter-sphere
219: conductance $\sigma_{\rm min}$ among the set of $n$ conductances in series. Hence
220: the small-$g$ limit of $h_n(g)$ is just the distribution function $\tilde{f}$
221: of $\sigma_{\rm min}$:
222: \begin{equation}
223: \label{min}
224: \tilde{f}(\sigma_{\rm min})=
225: n f(\sigma_{\rm min})\left[1-\int_{\sigma_{\rm min}}^1 d\sigma_{\rm min}
226: f(\sigma_{\rm min})\right]^{n-1},
227: \end{equation}
228: which, from Eq.(\ref{distri4}) and by setting $g\simeq \sigma_{\rm min}$, leads to:
229: \begin{equation}
230: \label{distri5}
231: \lim_{g\rightarrow 0} h_n(g)\simeq n(1-\alpha_n)g^{-\alpha_n}.
232: \end{equation}
233:
234:
235:
236:
237: The conducting bond distribution function behaves therefore as Eq.(\ref{distri2})
238: so that for $\alpha_n>\alpha_c\simeq 0.107$ transport universality breaks down and
239: $t>t_0$. For $L=1$ $\mu$m, $\Phi=10$ nm, and $\xi=1$ nm this is achieved already for
240: $n<90$, {\it i. e.}, slightly less then the maximum number $L/\Phi=100$
241: of spheres which can be accommodated inside a channel.
242:
243: Our model of universality breakdown in TFRs can be readily generalized to describe
244: more realistic situations. For example, the number $n$
245: of spheres inside the occupied channels can vary according to a given
246: distribution.
247: It is also straightforward to rewrite Eq.(\ref{tran}) in order
248: to describe cases in which
249: the diameter $\Phi$ of the spheres is not fixed,\cite{lu} or to let the size of the
250: insulating grains to change by assigning a distribution function for $L$.
251: It is then possible to have different scenarios and, more importantly, to
252: obtain a crossover from transport universality ($t=t_0$) to nonuniversality
253: ($t>t_0$) within the same framework. This reminds the experimental situation
254: reported for TFRs and summarized in Fig.\ref{fig1}.
255:
256: Let us now comment on the capability of other existing theories to describe transport
257: universality breakdown in TFRs. At a first glance, the swiss-cheese model
258: of Ref.\onlinecite{halperin} is a natural candidate since the large values
259: of $L/\Phi$ typical of many TFRs may lead to an effective continuous conducting
260: phase filling the voids between the large glassy grains. However, there are examples
261: in which nonuniversality has been reported for TFRs with $L/\Phi$ only of
262: order $\sim 5-10$,\cite{carcia1} a value probably too small to be compatible
263: with the swiss-cheese picture.
264: Even more problematic are the cases for
265: which $t=t_0\simeq 2.0$ has been measured for TFRs with $L/\Phi\sim 100$
266: (see for example Fig.\ref{fig1}), while the swiss-cheese model would have
267: predicted $t>t_0$.
268: Regarding instead the model proposed by Balberg,\cite{balb} for
269: which transport is dominated by
270: random tunneling processes in a percolating network, it is important to point
271: out that it was defined by using a phenomenological distribution function for
272: the nearest-neighbour particle distances very similar to our Eq.(\ref{tran}).
273: Balberg argured that such form of $P(r)$ is a reasonable compromise between
274: the distribution function of spheres
275: randomly placed in three dimensions,\cite{torquato}
276: and the effect of interactions between the conducting and insulating phases.
277: Instead we have shown that Eq.(\ref{tran}),
278: and consequently the power-law divergence of $h(g)$, is a straightforward outcome
279: of the quasi one-dimensional geometry of the conducting channels in
280: the segregation model of TFRs. Despite that our model has been formulated
281: specifically for TRFs, nevertheless it could be applied also to other
282: segregated disordered compounds for which tunneling is the main mechanism
283: of transport and nonuniversality has been reported.\cite{wu}
284:
285:
286: In conclusion, we have proposed a simple tunneling-percolation model capable of
287: describing the observed transport universality breakdown in TFRs.
288: Essential ingredients of the theory are the segregated structure, modelled by
289: quasi-one dimensional channels occupied randomly by the conducting particles, and
290: intergrain tunneling taking place within the channels.
291:
292: This work is part of TOPNANO 21 project n.5557.2.
293:
294:
295: \begin{references}
296:
297: \bibitem{prude1}
298: M. Prudenziati, {\it Handbook of Sensors and Actuators}
299: (Elsevier, Amsterdam, 1994), p.189.
300:
301: \bibitem{white}
302: N. M. White and J. D. Turner,
303: Meas. Sci. Technol. {\bf 8}, 1 (1997).
304:
305: \bibitem{pike}
306: G. E. Pike in {\it Electrical Transport and Optical Properties
307: of Inhomogeneous Media}
308: (J. C. Garland and D. B. Tanner, New York, 1978) p.366.
309:
310: \bibitem{dejeu}
311: W. H. de Jeu, R. W. J. Geuskens, and G. E. Pike,
312: J. Appl. Phys. {\bf 52}, 4128 (1981).
313:
314: \bibitem{carcia1}
315: P. F. Carcia, A. Ferretti, and A. Suna,
316: J. Appl. Phys. {\bf 53}, 5282 (1982).
317:
318: \bibitem{carcia2}
319: P. F. Carcia, A. Suna, and W. D. Childers,
320: J. Appl. Phys. {\bf 54}, 6002 (1983).
321:
322: \bibitem{listki}
323: E. Listkiewicz and A. Kusy,
324: Thin Solid Films {\bf 130}, 1 (1985).
325:
326: \bibitem{tambo}
327: M. Tamborin, S. Piccinini, M. Prudenziati, and B. Morten,
328: Sensors and Actuators A {\bf 58}, 159 (1997).
329:
330: \bibitem{kusy}
331: A. Kusy, Physica B {\bf 240}, 226 (1997).
332:
333: \bibitem{kirk}
334: S. Kirkpatrick, Rev. Mod. Phys. {\bf 45}, 574 (1973).
335:
336: \bibitem{stauffer}
337: D. Stauffer and A. Aharony {\it Introduction to Percolation Theory}
338: (Taylor \& Francis, London, 1992).
339:
340: \bibitem{clerc}
341: G. G. Batrouni, A. Hansen, and B. Larson,
342: Phys. Rev. E {\bf 53}, 2292 (1996);
343: J. P. Clerc, V. A. Podolskiy, and A. K. Sarichev,
344: Eur. Phys. J B {\bf 15}, 507 (2000).
345:
346: \bibitem{abeles}
347: B. Abeles, H. L. Pinch, and J. I. Gittleman,
348: Phys. Rev. Lett. {\bf 35}, 247 (1975).
349:
350: \bibitem{lee}
351: S.-I. Lee, T. W. Noh, X.-D. Chen, and J. R. Gaines,
352: Phys. Rev. B {\bf 34}, 6719 (1986).
353:
354: \bibitem{putten}
355: D. van der Putten {\it et al.}, Phys. Rev. Lett. {\bf 69}, 494 (1992).
356:
357: \bibitem{wu}
358: J. Wu and D. S. McLachlan,
359: Phys. Rev. B {\bf 56}, 1236 (1997).
360:
361: \bibitem{straley}
362: P. M. Kogut and J. Straley, J. Phys. C {\bf 12}, 2151 (1979).
363:
364: \bibitem{machta}
365: J. Machta, R. A. Guyer, and S. M. Moore,
366: Phys. Rev. B {\bf 33}, 4818 (1986).
367:
368: \bibitem{halperin}
369: B. I. Halperin, S. Feng, and P. N. Sen,
370: Phys. Rev. Lett. {\bf 54}, 2391 (1985).
371:
372: \bibitem{balb}
373: I. Balberg, Phys. Rev. Lett. {\bf 59}, 1305 (1987).
374:
375: \bibitem{torquato}
376: J. R. Macdonald, Mol. Phys. {\bf 44}, 1043 (1981);
377: S. Torquato, B. Lu, and J. Rubinstein,
378: Phys. Rev. A {\bf 41}, 2059 (1990).
379:
380: \bibitem{lu}
381: B. Lu and S. Torquato, Phys. Rev. A {\bf 45}, 5530 (1992).
382:
383: \end{references}
384:
385:
386:
387:
388: \end{document}