1:
2: %\input{tcilatex}
3:
4:
5: \documentclass[aps,prl,twocolumn,10 pt,showpacs]{revtex4}
6: \usepackage{amssymb}
7:
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9: \usepackage{graphicx}
10:
11: %TCIDATA{OutputFilter=Latex.dll}
12: %TCIDATA{LastRevised=Wed Jul 07 16:34:49 2004}
13: %TCIDATA{<META NAME="GraphicsSave" CONTENT="32">}
14:
15: \language0
16:
17:
18: \begin{document}
19:
20: \title{Out of equilibrium electronic transport properties of a misfit cobaltite thin film}
21: \author{A.Pautrat, H.W. Eng, W.Prellier}
22: \affiliation{CRISMAT, UMR 6508 du CNRS et de l'ENSI-Caen,6 Bd
23: Mar\'echal Juin, 14050 Caen, France.}
24:
25: \begin{abstract}
26: We report on transport measurements in a thin film of the 2D
27: misfit Cobaltite $Ca_{3}Co_{4}O_{9}$. Dc magnetoresistance
28: measurements obey the modified variable range hopping law expected
29: for a soft Coulomb gap. When the sample is cooled down, we observe
30: large telegraphic-like fluctuations. At low temperature, these
31: slow fluctuations have non Gaussian statistics, and are stable
32: under a large magnetic field. These results suggest that the low
33: temperature state is a glassy electronic state. Resistance
34: relaxation and memory effects of pure magnetic origin are also
35: observed, but without aging phenomena. This indicates that these
36: magnetic effects are not glassy-like and are not directly coupled
37: to the electronic part.
38:
39: \end{abstract}
40:
41:
42:
43: \pacs{71.27.+a,72.70.+m,72.20.My}
44:
45: \newpage
46: \maketitle
47: Layered cobaltites have recently
48: received a growing
49: interest due to their interesting properties. Their rather low resistivity
50: and large Seebeck coefficient at high temperature lead to interesting thermoelectric
51: properties \cite{sylvie1}. Superconductivity has been also observed in $Na_{x}CoO_{2}$, $yH_{2}O$ \cite{supra}. From a more fundamental point of view, different interesting points have been noted. For example, specific heat evidences a large effective mass in the $(Na,Ca)Co_{2}O_{4}$ family \cite{ando},
52: a good indication of strongly correlated electronic properties. Strong electronic correlations were also recently reported in $Ca_{3}Co_{4}O_{9}$ using electronic transport measurements under pressure \cite{limelette}. $Ca_{3}Co_{4}O_{9}$ is a misfit incommensurate layered system ,
53: with small modulated atomic positions in each layer ([$CoO_2$] and
54: [$Ca_{2}CoO_{3}$]) \cite{riri}. Furthermore, there is a
55: possibility of frustrated magnetic interactions in a kagome
56: lattice in the $CoO_2$ layers \cite{koshi}. Ferrimagnetic or very
57: weak ferromagnetic-like properties are also observed at low
58: temperature \cite{sylvie}. In addition, $\mu^+$SR experiments have
59: been interpreted with the presence of a spin density wave at low
60: temperature \cite{Sujiyama}, giving a clue for a gap opening near
61: the Fermi level leading to the observed "Fermi liquid"-insulator
62: transition. Nevertheless, the low temperature transport behavior
63: is not really understood, specially because the ground state of
64: the system is quite complex. The eventual link between all these
65: observations and the measured transport properties has to be
66: clarified. The above mentioned results suggest that low
67: dimensionality, frustration and disorder, together with strong
68: electronic correlations, can be important features for the
69: physical properties of the $Ca_{3}Co_{4}O_{9}$ system. Since they
70: are known to reinforce fluctuations and metastability, one can
71: expect a strong influence on the transport properties. To our
72: knowledge, this has not been verified by the usual macroscopic
73: transport measurements. Because of the statistical averaging, the
74: discrete nature of the underlying "mesoscopic" processes can be
75: masked in a bulk sample but revealed in a small-area system. In
76: this case, the transport measurements can be used as an efficient
77: probe to bring some light on the ground states of the system
78: \cite{mike}.
79:
80: In this paper, we report a study of DC magnetotransport properties and of
81: resistance fluctuations in a microbridge of a $Ca_{3}Co_{4}O_{9}$
82: thin film at low temperature. We particularly focus on the origin
83: (magnetic or electronic) of these resistance fluctuations and
84: relaxation.
85:
86: A $2000 \AA$ epitaxial thick film of $Ca_{3}Co_{4}O_{9}$ was used
87: for the measurements. The film was deposited on (0001)
88: $Al_{2}O_{3}$ using the pulsed laser deposition technique. The
89: details of the optimization, growth conditions and structural
90: characterizations have been described previously \cite{eng}. Prior
91: to the transport measurements, a silver layer was firstly
92: deposited via thermal evaporation onto the film, and secondly a
93: gold layer via RF sputtering. Microbridges ($L=200$ $\mu m$
94: $\times$ $W=50$ $\mu m$) were then patterned with UV
95: photolithography and argon ion etching. Thin aluminum contact
96: wires were finally used to connect the areas to the electrodes
97: with a wire bonding system. Measurements were made with the four
98: probes method. For the following results, a maximum of 100 $nA$
99: low noise current was used and any significant current dependence
100: for low current was observed, i.e., quasi-linear response
101: conditions were utilized. The magnetic field was applied along the
102: c-axis of the film, perpendicular to the substrate plane.
103:
104: As it is observed in bulk samples, the $Ca_{3}Co_{4}O_{9}$ film
105: exhibits both a strong increase of the resistance and a large
106: negative magnetoresistance at low temperature \cite{masset} (we
107: measure $\Delta R/R \gtrsim$ -0.6 at 14 $T$ and 2 $K$). The
108: analysis of the transport properties in this low temperature part
109: is the main goal of this experiment. One of the most observable
110: characterization of a localized electronic state is the functional
111: form of the dc resistance. A simple activated law does not
112: describe the data, but a variable range hopping (VRH) expression
113: of the form $R=R_0. exp(T^{*}/T)^{\mu}$ ($\mu <$ 1) gives a very
114: good fit (Fig. 1) for a large temperature range (2 $K$ to 90 $K$).
115: Interestingly, one finds a much better agreement using $\mu =$ 1/2
116: than using the other exponents (1/3, 1/4) which can be expected
117: for the Mott VRH regime \cite{mott}. This value has been confirmed
118: by analyzing the so-called Zabrodski plot \cite{zabro}. This
119: latter consists in plotting $log(\partial R /
120: \partial (1/T))$ as function of $log(T)$, the slope of which gives
121: $\mu=$ 0.47 $\pm$ 0.02 with a robust precision and mainly without
122: any presupposition on the form of $R(T)$. The exponent $\mu =$ 1/2
123: is the most ordinary signature of a soft Coulomb gap in the
124: density of states (the Efros-Shklovskii or ES gap)
125: \cite{Efros1,Efros2}, meaning that strong Coulomb
126: (electron-electron) interactions are present. Theory predicts
127: $T^{*} = A e^2/(k_B 4\pi \varepsilon_0 \kappa \xi)$ ($k_B$ is the
128: Boltzmann constant, $\varepsilon_0$ is the permittivity of the
129: air, $\kappa$ the relative dielectric constant, $A= $2.8 and 6.2
130: for respectively 3D \cite{Efros2} and 2D \cite{nguyen} systems,
131: and $\xi$ is the localization length). This approach is a single
132: particle approach and neglects the correlated motion of the
133: charges. It was found from numerical simulations that correlations
134: effects on the carriers motion relies essentially on a decrease of
135: $A$ ($A \approx$ 0.61 was found for the 2D case in a small
136: system). One of the principle merit of this result is to reconcile
137: the introduction of many electron configurations with an unchanged
138: functional form of the ES-VRH \cite{perez}, but this shows that
139: there is a large uncertainty on $A$. From the data at $B=$ 14 $T$,
140: we get $T^{*}=$ 159 $K$ from $T=$ 2 $K$ to about 90 $K$. $\kappa .
141: \xi \approx$ 62 and 640 $nm$ for $A=$ 0.61 and 6.2 respectively
142: can be deduced. $\kappa$ is to our knowledge unknown for
143: $Ca_{3}Co_{4}O_{9}$, but considering $\kappa \approx$ 10 as rough
144: approximation, reasonable $\xi$ values can be deduced. Exactly the
145: same analysis was applied to the zero field curve from $T=$ 90 $K$
146: up to $T = T_0 \approx$ 15 $K$. For $T \leq T_0$, a large negative
147: magnetoresistance is present. This reflects in a slight increase
148: of the exponent $\mu$ (0.56 $\pm$ 0.02 with the Zabroski plot).
149: This indicates a small departure from a pure parabolic form of the
150: Coulomb gap \cite{ra}. However, the most significant feature is
151: the change of $T^{*}$. This is this change, from $T^{*}= $159 $K$
152: for $B=$ 14 $T$ to $T^{*}=$ 267 $K$ for $B=$ 0 $T$, which gives
153: the weight to the observed negative magnetoresistance. Note that
154: since $\mu$ does not change significantly for our large magnetic
155: fields range,
156: this means that the electronic interactions keep dominant with no sign of a magnetic hard gap at least for $T \gtrsim$ 2 $K$. Unlike in the 1D cobaltite $Ca_{3}Co_{2}O_{6}$
157: \cite{raquet}, the Coulomb gap seems to be robust to a high
158: magnetic field. We suggest thus that the magnetoresistance effect
159: comes principally from orbital, rather than spin, effect
160: \cite{sivan}. This agrees also with the linear field dependence of
161: the magnetoconductance at intermediate fields (fig.2)
162: \cite{nguyen2}. Nevertheless, this point deserves clearly more
163: attention, with a systematic study of the magnetoresistance as a
164: function of the magnetic field orientation. Concerning the DC
165: properties, it can be noted that simple activated form, together
166: with transport non linearities usually associated to a spin
167: density waves condensation and to their collective depinning
168: \cite{gruner} have not been observed.
169:
170:
171: We will now focus on the time series of the change in resistance.
172: When $T \lesssim $25 $K$, large telegraphic-like fluctuations
173: appear. They are generally two levels fluctuations (Fig. 3). The
174: square of the Fourier transform of the time traces gives their
175: spectral representation. After a long acquisition (several hours),
176: we observe a Lorentzian form, meaning that a single characteristic
177: time is sufficient to be statistically representative of the
178: switching process (Fig.4). The fluctuations are slow, a switcher
179: being in average about 100 $sec$ in a resistance state. These
180: characteristic times are only slightly temperature dependent. For
181: high field values (several Teslas), some sequences of three and
182: four levels have been also seen. At 25 $K$, the first minutes of
183: the resistance traces are dominated by large switches which
184: contribute to the noise, but after some times (typically 1000
185: $sec$), the noise relaxes to low value. A small perturbation as a
186: 50 $G$ field is enough to destroy the kinetic between the two
187: states (Fig. 3a). Thus, the noisy regime is clearly not really
188: established and the rare events can be thought of precursors of
189: the low temperature state. In strong contrast, at 22.8 $K$, the
190: noisy regime is robust to a 14 $T$ field (fig.3b) and is
191: stabilized at least up to several hours, the typical time scale of
192: our longest recording. This indicates a dramatic change on the
193: charge dynamics. This change is confirmed by a peak in the noise
194: value $\sigma_R$ at a temperature $T=$ 22.9 $\pm$ 0.1 $K$
195: ($\sigma_R$ corresponds to the integrated noise spectrum over a
196: 10$^{-3}$-10$^{-1}$ $Hz$ bandwidth) (fig. 3c). This temperature is
197: reasonably separated from $T_0$, the temperature where the
198: magnetoresistance begins to be measurable. This is a good
199: indication that the noise has not a magnetic origin, in agreement
200: with the small sensitivity to the large magnetic fields that has
201: been observed. Correlated hopping conduction should be rather
202: involved, in agreement with the very long time scales associated
203: to the resistance fluctuations \cite{massey}.
204:
205: At the lowest temperatures, a notable change in the fluctuation dynamics, with the appearance of a large intermittence (Fig.4), is observed.
206: After Fourier transforming and squaring the time traces, the noise
207: spectra are found with a $1/f^{\alpha}$ shape, where $\alpha$ is
208: close to 1.3 (inset of Fig.5). This contrasts to the Lorentzian
209: shape observed near $T=$ 22.9 $K$. We stress also that the
210: observed small temperature dependence of the noise power when the
211: temperature decreases rules out simple thermally activated
212: trapping processes. In addition, it can be realized that the
213: observed intermittence implies that the fluctuations are not
214: stationary anymore. This means that the noise spectrum itself
215: significantly changes with time. This "noise of the noise" can be
216: characterized by the second spectrum $S^{2}(f) = 1/f^{1-\beta}$
217: introduced in ref. \cite{restle}. This corresponds to the fourth
218: order statistic spectrum of the ordinary $1/f^{\alpha}$ noise. A
219: non white $S^{2}(f)$ is typical of a non Gaussian averaging
220: process. If one excludes the case of a dynamic redistribution of
221: the current in an inhomogeneous paths gallery \cite{percol}, this
222: implies interacting fluctuators. In practice, the voltage signal
223: was acquired during a very long time (several hours) then
224: segmented, and finally each segment was Fourier transformed and
225: integrated to obtain time series of noise power. They were taken
226: for different ranges of frequencies and then Fourier transformed,
227: to finally obtain $S_2(f)$. As shown in fig.6, this second
228: spectrum is found nearly white at $T=$ 22.8 $K$ in the telegraphic
229: noise regime. In strong contrast, a large exponent $(1-\beta)
230: \approx$ 0.9 is measured at 8 $K$. We conclude that charges
231: fluctuations are strongly correlated at this moderately low
232: temperature. This behavior is close to what was observed in 2D
233: electronic systems far in the localized regime, at a much lowest
234: temperature \cite{bogda}. In summary, the low temperature phase of
235: $Ca_{3}Co_{4}O_{9}$ exhibits glassy electronic transport
236: properties and a Coulomb gap form of the dc conductivity, i.e.,
237: can be called a Coulomb Glass.
238:
239: Furthermore, when the sample is cooled down from high
240: temperature to 4 $K$, the resistance is measured hysteretic when
241: cycling the magnetic field for the first time (fig.7). It is
242: reversible after this first cycle and then reveals the equilibrium
243: magnetoresistance properties (fig.2). In addition, after a field
244: cooling, the resistance relaxes slowly downwards with time. A
245: simple exponential relaxation gives generally a good fit to the
246: data, and its asymptotic limit after a long time is close to the
247: low resistance state value. Such out of equilibrium features can
248: be observed in very large number of systems which exhibits pinning
249: by disorder, metastability, genuine glassy properties (etc). In
250: the case of $Ca_{3}Co_{4}O_{9}$, the possible frustration of the
251: interactions between spins in the triangular Co lattice
252: degenerates the free energy. This can be a reason for spin
253: glass-like properties as observed for example in Kagome lattices
254: \cite{ladieu}. A discriminating test is to check if the system
255: itself evolves with time, i.e., "ages". For that, a standard
256: procedure is started at a temperature $T_i$ and at a field $B_i$,
257: and then the system evolves with the time $t_w$. The field is
258: quenched and the resistance is measured during the time $t$
259: \cite{vincent}. After this procedure, if aging occurs, the
260: relaxation should differ for different $t_w$. A scaling law in
261: $t/t_w^{\alpha}$ provides generally a good fit to the data
262: \cite{vincent}. In the inset of the fig.7, it can be observed that
263: the system evolves during the waiting time, but with a relevant
264: rescaling in $t-t_w$. This means that the system relaxes but
265: respects time-translation invariance and that the magnetic out of
266: equilibrium state $\textit{is not}$ a spin-glass in its proper
267: sense. Nevertheless, a memory of the relaxation after a small
268: temperature change is seen. It is known that similar experimental
269: signatures, namely slow relaxation and memory without aging, are
270: observed when weakly interacting magnetic nanoparticles are
271: present, without the need of a correlation length of a spin-glass
272: order \cite{sasaki}. Thus, we propose an explanation of the
273: relaxation properties in terms of superparamagnetic-like
274: relaxation. The associated relaxation time $\tau^{-1}$ is maximal
275: at the temperature $T_0$ where the magnetoresistance has been
276: found to disappear (not shown), confirming the magnetic origin of
277: the relaxation. We propose that short range ordered clusters of
278: mostly ferromagnetic Co ions are the relevant entities. It is
279: clear that any probe of magnetic correlations at intermediate
280: scales, such as probed by Small Angle Neutron scattering, should
281: be interesting to give more clues on this magnetic ground state.
282:
283: In summary, we have shown that the low temperature magnetotransport properties of $Ca_{3}Co_{4}O_{9}$ can be
284: described as a localized state with a Coulomb gap. A non
285: Gaussian regime of resistance fluctuations is present at moderately low temperature. This can be attributed to
286: correlated charge fluctuations characteristics of an electron glass. An additional field dependent resistance relaxation is observed
287: but does not show glassy phenomena like aging, and thus is not directly coupled to the pure electronic part.
288:
289:
290:
291: Acknowledgments: A.P would like to acknowledge L.
292: M$\acute{e}$chin (GREYC-ENSICAEN) for patterning the microbridges,
293: S. H$\acute{e}$bert and Ch. Simon (CRISMAT-ENSICAEN) for critical
294: readings of the paper, and F. Ladieu (SPEC-CEA-SACLAY) for very
295: helpful remarks concerning the slow dynamic of magnetic systems.
296: H.W.E. acknowledges the CNRS and the conseil r$\acute{e}$gional de
297: Basse Normandie for funding.
298:
299: %%%%%%%%%%%%%%%%% BIBLIOGRAPHIE %%%%%%%%%%%%%%%%%%%%%%%%%%%%
300:
301: \begin{references}
302: \label{sec:TeXbooks}
303:
304: \bibitem{sylvie1} S. H$\acute{e}$bert, S. Lambert, D. Pelloquin, and A. Maignan , Phys. Rev. B 64, 172101
305: (2001).
306:
307: \bibitem{supra} K. Takada, H. Sakurai, E. Takayama-Muromachi, F. Izumi, R.A. Dilanian, and T. Sasaki, Nature (London) 422, 53 (2003).
308:
309: \bibitem{ando} Y. Ando, N. Miyamoto, K. Segawa, T. Kawata, and I. Terasaki, Phys. Rev. B 60, 10580 (1999).
310:
311: \bibitem{limelette} P. Limelette, V. Hardy, P. Auban-Senzier, D. J$\acute{e}$rome, D. Flahaut, S. H$\acute{e}$bert, R. Fr$\acute{e}$sard, Ch. Simon, J. Noudem, and A. Maignan, Phys. Rev. B 71, 233108 (2005).
312:
313: \bibitem{riri} D. Grebille, S. Lambert, F. Bour$\acute{e}$e, and V. Petricek, J. Appl. Cryst. 37, 823 (2004).
314:
315: \bibitem{koshi} W. Koshibae and S. Maekawa, Phys. Rev. Lett. 91, 257003
316: (2003).
317:
318: \bibitem{sylvie} L. B. Wang, A. Maignan, D. Pelloquin, S. H$\acute{e}$bert, and B. Raveau , J. Appl. Phys. 92, 124 (2002).
319: J. Sugiyama, C. Xia, and T. Tani, Phys. Rev. B 67, 104410 (2003).
320:
321: \bibitem{Sujiyama} J. Sugiyama, J. H. Brewer, E. J. Ansaldo, H. Itahara, K. Dohmae, Y. Seno, C. Xia, and T. Tani, Phys. Rev. B 68, 134423 (2003).
322:
323: \bibitem{mike} M. B. Weissman, Rev. Mod. Phys. 60, 537 (1988).
324:
325: \bibitem{eng} H.W. Eng, W. Prellier, S. H$\acute{e}$bert, D. Grebille, L. M$\acute{e}$chin, and B. Mercey, J. Appl. Phys. 97, 013706
326: (2005).
327:
328: \bibitem{masset} A. C. Masset, C. Michel, A. Maignan, M. Hervieu, O. Toulemonde, F. Studer, B. Raveau, and J. Hejtmanek, Phys. Rev. B 62, 166 (2000).
329:
330: \bibitem{mott} N.F. Mott, Metal-Insulator Transitions (Taylor and
331: Francis, London), 1974.
332:
333: \bibitem{zabro} A.G. Zabrodskii and K.N. Ninov'eva, Zh. Eksp. Teor. Fiz. 86, 727 (1984) [Sov. Phys. JETP 59, 425 (1984)].
334:
335: \bibitem{Efros1} A.L. Efros and B.I. Shklovskii, J. Phys. C8, L49
336: (1975).
337:
338: \bibitem{Efros2} B.I. Shklovskii and A.L. Efros, Electronic
339: Properties of doped semiconductor (Springer, New York, 1984).
340:
341: \bibitem{nguyen} V.L. Nguyen, Sov. Phys. Semicond. 18, 207 (1984).
342:
343: \bibitem{perez} A. P$\acute{e}$rez–Garrido, M. Ortu$\tilde{n}$o, E. Cuevas, J. Ruiz , and M. Pollak, Phys Rev B 55, R8630 (1997).
344:
345: \bibitem{ra} M. E. Raikh and I. M. Ruzin, Sov. Phys. JETP 68, 642 (1989).
346:
347: %\bibitem{yu} K. Shtengel and C. C. Yu, Phys. Rev. B 67, 165106 (2003).
348:
349: %\bibitem{butko} V. Yu. Butko, J.F. DiTusa and P.W. Adams, Phys.
350: %Rev. Lett. 84, 1543 (2000).
351:
352: \bibitem{raquet} B. Raquet, M. N. Baibich, J. M. Broto, H. Rakoto, S. Lambert, and A. Maignan, Phys. Rev. B 65, 104442 (2002).
353:
354:
355: \bibitem{gruner} G. Gr$\ddot{u}$ner, Rev. Mod. Phys. 66, 1 (1994).
356:
357:
358: \bibitem{sivan} U. Sivan, O. Entin–Wohlman, and Y. Imry, Phys. Rev. Lett. 60, 1566 (1988).
359: O. Entin–Wohlman, Y. Imry, and U. Sivan, Phys. Rev. B 40, 8342
360: (1989).
361:
362:
363: \bibitem{nguyen2} V. I. Nguyen, B. Z. Spivak and B. I. Shklovskii, Zh. Eksp. Teor. Fiz. 89, 1770 (1985) [Sov. Phys—JETP 62, 1021 (1985)]
364:
365: \bibitem{massey} J.G. Massey and M. Lee, Phys. Rev. Lett. 79, 3986
366: (1997).
367:
368: \bibitem{restle} P. J. Restle, R. J. Hamilton, M. B. Weissman, and M. S. Love, Phys. Rev. B 31, 2254 (1985).
369:
370: \bibitem{percol} G. T. Seidler, S. A. Solin, and A. C. Marley, Phys. Rev. Lett. 76, 3049
371: (1996).
372:
373: \bibitem{bogda} S. Bogdanovich and D. Popovi$\acute{c}$, Phys. Rev. Lett. 88, 236401
374: (2002). J. Jaroszy$\acute{n}$ski, D. Popovi$\acute{c}$, and T. M.
375: Klapwijk, Phys. Rev. Lett. 89, 276401 (2002).
376:
377: \bibitem{ladieu} F. Ladieu, F. Bert, V. Dupuis, E. Vincent, J. Hammann, J. Phys. Condens. Matter
378: 16, 735 (2004).
379:
380: \bibitem{vincent} E. Vincent, V. Dupuis, M. Alba, J. Hammann, and J.P. Bouchaud, Europhys. Lett. 50, 674 (2000).
381:
382: \bibitem{sasaki}M. Sasaki, P.E. J$\ddot{o}$nsson, H. Takayama, and H. Mamiya, Phys. Rev. B 71, 104405 (2005).
383:
384:
385: \end{references}
386:
387: \newpage
388:
389: \begin{figure}[tbp]
390: \centering \includegraphics*[width=6cm]{fig1a.eps} \caption{Color
391: online. Semilog plot of the resistance as function of $1/\sqrt T$,
392: for $B=0$ $T$ and $14$ $T$. Linearity in this scale relies on a
393: variable range hopping regime with a Coulomb gap. Inset:
394: Resistance as function of the temperature in a semi-log scale for
395: B=0T.} \label{fig.1}
396: \end{figure}
397:
398: \begin{figure}[tbp]
399: \centering \includegraphics*[width=6cm]{magneto.eps}
400: \caption{Normalized magnetoconductance at $T=4.2$ $K$. Note the
401: small negative magnetoconductance for $B \lesssim 3$ $T$ followed
402: by the strong positive magnetoconductance. The dotted line is a
403: guide to the eyes.} \label{fig.2}
404: \end{figure}
405:
406: \begin{figure}[tbp]
407: \centering \includegraphics*[width=8cm]{peak.eps} \caption{Color
408: online. a/ and b/ resistance traces at respectively 25 K and 22.8
409: K for $B=0$ and $5.10^{-3}$ T and for $B=0$ and $14$ $T$. The
410: shift of the time traces in each graph is arbitrary. $R_0$ is the
411: mean value of the resistance averaged during the time of the
412: measurement. c/ The normalized noise versus the temperature. Note
413: the peak of noise at $T=22.9 \pm 0.1$ $K$. The dotted line is a
414: guide to the eyes.} \label{fig.3}
415: \end{figure}
416: \begin{figure}[tbp]
417: \centering \includegraphics*[width=6cm]{lorentz.eps}
418: \caption{Color online. The power spectal density (P.S.D)
419: corresponding to a very long recording of the time trace at
420: $T=22.80$ $K$. The well defined telegraph-like carriers dynamics
421: are evidenced by both the Lorentzian spectrum (corner frequency
422: $f_{c} \approx 10^{-2}$ Hz) and the bimodal histogram.}
423: \label{fig.4}
424: \end{figure}
425:
426: \begin{figure}[tbp]
427: \centering \includegraphics*[width=6cm]{fig5a.eps} \caption{Color
428: online. Long time acquisition of the resistance trace at $T=8$ $K$
429: and $B=14$ $T$. Inset: The associated Fourier transform with a
430: ($1/f^{1.3}$) form.} \label{fig.5}
431: \end{figure}
432:
433: \begin{figure}[tbp]
434: \centering \includegraphics*[width=6cm]{fig6a.eps} \caption{Color
435: online. Second spectra $S^{2}(f)$ of the resistance fluctuations
436: at 22.8 K and 8 K. Correlated fluctuations are evidenced by the
437: non-white second spectrum such as measured at the low
438: temperature.} \label{fig.6}
439: \end{figure}
440:
441: \begin{figure}[tbp]
442: \centering \includegraphics*[width=6cm]{fig7a.eps} \caption{Color
443: online. Variation of the resistance as function of the magnetic
444: field at $T=4$ $K$ after a zero field cooling. Inset: relaxation
445: of the thermo-remanent resistance after a 2T FC for $t_{w}= 10,
446: 100, 1000$ $sec$ as function of $t-t_w$.} \label{fig.7}
447: \end{figure}
448:
449: \end{document}
450: