1: \documentclass{jetpl}
2: \twocolumn
3:
4: %%% article in English
5: \lat
6:
7: \title{Superconducting energy gap distribution in c-axis oriented MgB$_{2}$
8: thin film from point contact study}
9:
10: %%% article title - for colontitle (at the top of the page)
11: \rtitle{Superconducting energy gap distribution in MgB$_{2}$}
12:
13: %%% article title - for table of contents (usualy identical with \title)
14: \sodtitle{Superconducting energy gap distribution in c-axis oriented MgB$_{2}$
15: thin film from point contact study}
16:
17:
18: %%% author(s) ( + e-mail)
19: \author{Yu.\,G.\,Naidyuk\/\thanks{e-mail: naidyuk@ilt.kharkov.ua }, I.\,K.\,Yanson,
20: L.\,V.\,Tyutrina, N.\,L.\,Bobrov, P.\,N.\,Chubov,
21: W.\,N.\,Kang$^{*}$, Hyeong-Jin Kim$^{*}$, Eun-Mi Choi$^{*}$, and
22: Sung-Ik Lee$^{*}$ }
23:
24: %%% author(s) - for colontitle (at the top of the page)
25: \rauthor{Yu.\,G.\,Naidyuk, I.\,K.\,Yanson, L.\,V.\,Tyutrina et
26: al.}
27:
28: %%% author(s) - for table of contents
29: \sodauthor{Yu.\,G.\,Naidyuk$^{+}$, I.\,K.\,Yanson$^{+}$, L.\,V.\,Tyutrina$^{+}$
30: et. al.}
31:
32:
33: %%% author's address(es)
34: \address{B.Verkin Institute for Low Temperature Physics and
35: Engineering, National Academy of Sciences of Ukraine, 47 Lenin
36: Ave., 61103, Kharkiv, Ukraine\\~\\
37: $^*$National Creative Research Initiative Center for
38: Superconductivity, Department of Physics, Pohang University of
39: Science and Technology, Pohang 790-784, South Korea}
40:
41: %%% dates of submition & resubmition (if submitted once, second argument is *)
42: \dates{15 January 2002}{*}
43:
44:
45: \abstract{We have analyzed about hundred voltage-dependent
46: differential resistance $dV/dI(V)$ curves of metallic point
47: contacts between $c$-axis oriented MgB$_{2}$ thin film and Ag,
48: which exhibit clear Andreev reflection features connected with the
49: superconducting gap. About one half of the curves show the
50: presence of a second larger gap. The histogram of the double gap
51: distribution reveals distinct maxima at 2.4 and 7 meV, while
52: curves with a single-gap features result in more broad maximum at
53: 3.5 meV. The double-gap distribution is in qualitative agreement
54: with the distribution of gap values over the Fermi surface
55: calculated by H. J. Choi et al. (cond-mat/0111183). The data
56: unequivocally show the presence of two gaps $\Delta _{S}=2.45\pm
57: 0.15$ meV and $\Delta _{L}=7.0\pm 0.45$ meV in MgB$_{2}$ with gap
58: ratio $\Delta _{L}/\Delta _{S}=2.85\pm 0.15$. Our observations
59: prove further a widely discussed multigap scenario for MgB$_{2}$,
60: when two distinct gaps are seen in the clean limit, where a
61: single averaged gap is present in the dirty one.}
62:
63: \PACS{73.40.Jn, 74.76Db, 74.80.Fp }
64:
65: \begin{document}
66:
67: \maketitle
68:
69: \textit{Introduction}.
70: Direct spectroscopic investigations of the superconducting
71: order parameter in recently discovered \cite{Nagam} superconductor
72: MgB$_2$ with T$_c\simeq$40\,K by tunneling
73: \cite{Karapet,Rubio,Sharoni,Chen,Giubileo1,Giubileo2,Badr,Zhang,Zasad}
74: and point-contact spectroscopy
75: \cite{Zasad,Deuts,Plec01,Laube,Gonelli,Szabo,Bugos,Bobrov} show
76: unambiguously an energy gap $\Delta $ in the quasiparticle density
77: of states (DOS). However, the experimental results are
78: controversial as to the gap width $\Delta $, whose variation from
79: 1.5 to 8 meV (see, e. g., review \cite{Buzea}) is unexpectedly
80: large, pointing to the possibility of multiphase or
81: nonhomogeneous samples, degraded surface, or anisotropic energy
82: gap. Another way to solve this puzzle is to consider two
83: superconducting gaps in MgB$_{2}$, as proposed by Liu {\it et al.}
84: \cite{Liu}, accounting complex electronic structure of MgB$_{2}$
85: with both quasi-2D and 3D Fermi surface \cite{Kortus}. Indeed,
86: several papers \cite{Giubileo1,Giubileo2,Badr,Szabo,Bugos,Bobrov}
87: have reported double gap structure in the differential conductance
88: (resistance) with the smaller gap being far below weak-coupling
89: BCS value $\Delta $=1.76k$_{\rm B}$T$_c\simeq $6\,meV and the
90: larger gap slightly above the standard BCS one, in accordance with
91: theory \cite{Liu}.
92:
93: Therefore, one of the intriguing key issues of superconducting
94: state of MgB$_2$ is whether the double gap state is intrinsic or
95: the spread of the gap values is a result of anisotropy,
96: nonhomogeneity, surface effect, etc. In other words, before
97: macroscopic high quality single crystals will be available for
98: thorough investigations, the sample imperfection may raise doubts
99: about the final conclusion. However, in our mind, good
100: reproducibility of the double-gap values given by different
101: authors \cite{Giubileo1,Giubileo2,Badr,Szabo,Bugos,Bobrov} by
102: different, in their physical background, methods such as tunneling
103: and point-contact spectroscopy carried out on different samples
104: such as pellets, films, grains, all this with a great degree of
105: probability supports intrinsic nature of the double gap in
106: MgB$_2$.
107:
108: In this paper we will give further confirmation of double gap scenario
109: in MgB$_2$ based on analysis of about hundred point-contact
110: spectra of c-axis oriented thin films.
111:
112: \textit{Experimental and calculation details}.
113: We have measured the high-quality c-axis oriented 0.4\,$\mu$m
114: thick MgB$_2$ film \cite{Lee01} grown by a PLD technique on Al$_2$O$_3$
115: substrate. The resistivity of the film exhibits a
116: sharp transition at 39 K with a width of $\sim$ 0.2 K from 90\% to
117: 10\% of the normal state resistivity \cite{Lee01}. The residual
118: resistivity $\rho_0$ at 40\,K is $\sim$ 6 $\mu\Omega$\,cm
119: \footnote{There is a scattering by factor of 4 in $\rho_0$ for the
120: different films.} and RRR=2.3.
121:
122: Different point contacts (PCs) were established in situ directly
123: in liquid $^{4}$He by touching as-prepared surface (sometimes
124: etched by 1\% HCl solution in ethanol) of the MgB$_2$ film by a
125: sharpened edge of an Ag counterelectrode, which were cleaned by
126: chemical polishing in HNO$_3$. This geometry corresponds to the
127: current flowing preferably along the c axis. A number of contacts
128: were measured by touching of the film edge after breaking
129: Al$_2$O$_3$ substrate. By this means, the current flows preferably
130: along the ab plane. The differential resistance d$V/$d$I$ vs $V$
131: was recorded using a standard lock-in technique. The normal
132: resistance $R_N$ (at $V\gg\Delta$) of investigated contacts ranged
133: mainly between 10 and 1000 $\Omega$ at 4.2\,K.
134:
135: The important characteristic of PC is their size or diameter $d$,
136: which can be determined from the simple formula derived by Wexler
137: \cite{Wexler} for contact resistance:
138: \begin{equation}
139: \label{Rwex} R_{\rm PC}(T) \simeq \frac {16 \rho l}{3\pi d^2} +
140: \frac{\rho (T)}{d},
141: \end{equation}
142: where two terms represent ballistic Sharvin \footnote{In the case
143: of interface scattering Sharvin resistance should be multiple by
144: factor (1+Z$^2$)\cite{BTK82}.} and diffusive Maxwell resistance,
145: correspondingly. Here $\rho l = p_{\rm F}/n$e$^2$, where $p_{\rm
146: F}$ is the Fermi momentum and $n$ is the density of charge
147: carriers. The latter for MgB$_2$ is estimated at $n\simeq
148: 6.7\times 10^{22}$ \cite{Canfield}, which results in $\rho l\simeq
149: 2\times 10^{-12}\Omega\cdot $cm$^2$ using $v_{\rm F}\simeq 5\times
150: 10^{7}$cm/s \cite{Kortus}. Hence, the upper limit for elastic mean
151: free path $l=\rho l/\rho_0$ for our film is about 3\,nm. In this
152: case, according to Eq.\,(1), the condition $d<l$ is fulfilled for
153: PC with $R>40\,\Omega $ or for lower resistance supposing multiple
154: contacts in parallel.
155:
156: We have utilized generally used Blonder-Tinkham-Klapwijk equations
157: \cite{BTK82} describing $I-V$ characteristic of ballistic N-c-S
158: metallic junctions (here N is normal metal, c is constriction and
159: S is superconductor) by accounting for the processes of Andreev
160: reflection. At finite barrier strength at the N-S interface
161: characterized by parameter $Z\neq0$ and $T\ll T_c$, the theory
162: gives the d$V/$d$I$ curves with minima at $V\simeq \pm\Delta /$e.
163: To get the correct $\Delta $, the fit of the measured curves to
164: the theory should be done. The additional smearing of d$V/$d$I$
165: curves due to, e. g., broadening of the quasiparticle DOS in the
166: superconductor can be taken into account by including parameter
167: $\Gamma$ \cite{Dynes}.
168:
169: In the case of curves with double gap structure we calculated,
170: according to the theory \cite{BTK82}, the sum of two differential
171: conductances d$I/$d$V$ with the weight $w$ for the larger gap and,
172: correspondingly, with $(1-w)$ for the smaller one. After this, we
173: have transformed d$I/$d$V$ into d$V$/d$I$ to compare with the
174: measured dependences. The best fit was achieved, as a rule, by
175: using its own values of $Z$ and $\Gamma$ for large and small gap.
176: It is acceptable if we suppose that we have a number of
177: microconstrictions with various $Z$ in the region of mechanical
178: touch. It is worthy to note that, with increasing of weight
179: factor, the difference between $Z$ and $\Gamma$ values for large
180: and small gap becomes smaller or even vanishes for some PCs.
181:
182: \textit{Results and discussion}. Approximately one half (44 of
183: total 91) of analyzed raw d$V$/d$I$ vs. $V$ curves show visible
184: two-gap structure, although, in most cases with shallow features
185: corresponding to a larger gap. The samples of some d$V/$d$I$
186: curves taken at 4.2\,K $\ll T_c$ with double-gap structure, along
187: with calculated curves, are shown in Fig.\,1. In spite of a number
188: of fitting parameters ($\Delta$, $\Gamma$, $Z$, $w$ ) for curves
189: with pronounced (or at least visible as shown in Fig.\,1) double
190: gap features determined $\Delta_L$ and $\Delta_S$ are robust as to
191: fitting procedure.
192:
193: It turns out that histogram of gaps distribution built on the
194: basis of fitting of 44 spectra (see Fig.\,2a) has two
195: well-separated and quite narrow (especially for the small gap)
196: maxima.
197:
198: The double-gap distribution is in qualitative agreement with the
199: distribution of gap values over the Fermi surface recently
200: calculated in \cite{Choi} (see Fig.\,2c). The main difference is
201: that theoretical distribution for lower gap is wider and has a
202: dominant maxima around 1.6 meV. This discrepancy can be resolved
203: when we draw attention that we have measured curves with double
204: gap structure for contacts that is predominantly along the c-axis.
205: In this case, according to \cite{Choi}, gap values along c-axis
206: spreads between 2 and 3 meV. The c-axis directionality of our
207: measurements is, apparently, the main reason of a shallow large
208: gap structure in d$V$/d$I$, because large gap dominates in the
209: "a-b" plane \cite{Choi}.
210:
211: It should be mentioned that two very different order parameters
212: exist only in the clean limit $l\gg 2\pi\xi$. Since in our case
213: $l$ has upper limit in 3 nm and the coherence length $\xi\sim$5 nm
214: \cite{Finnemore}, the observation of two gaps is in line with our
215: supposition that in the PC area there are small grains with a much
216: larger mean free path. Indeed, SEM image of MgB$_2$ films
217: \cite{Bugos} shows that the film is granular with 100-500 nm large
218: grains. Therefore, in the area of mechanical contact there are
219: some amount of small metallic bridges, perhaps, with slightly
220: different crystallographic orientation being in parallel.
221:
222: The single gap $\Delta$ is seen for the dirty limit \footnote{On
223: d$V/$d$I$ of "edge" contacts (a total of 11 curves) only single
224: gap structure was observed, probably due to the deterioration of
225: the film edge after breaking.} and is average of small and large
226: gaps with some weights. If we assume that this weight has some
227: relation to the weight $w$ used in the fitting procedure, then,
228: admittedly, $\Delta \simeq w\,\Delta_L +(1-w)\Delta_S$=3.4\,meV by
229: using upper limit $w\simeq$0.2 (see Fig.\,3). This agrees with the
230: position of the maximum of single-gap distribution at 3.5\,meV
231: (see. Fig.\,2b). By the way, according to the calculation in
232: \cite{Brink}, a large amount of impurity scattering will cause the
233: gaps to converge to $\Delta \simeq$4.1 meV.
234:
235: Therefore the superconducting properties of this compound can be
236: strongly influenced by nonmagnetic defects and impurities, which
237: seem to have a great impact also on the scattering of gap value(s)
238: given by different authors.
239:
240: As to $w$ factor it is hardly to see in Fig.\,3 its dependence on
241: $R_N$ or PC size, which one would expect if small gap reflects a
242: degraded surface or large gap is a result of surface states
243: \cite{Servedio}.
244:
245: Table II shows double gap values given by different authors. A
246: quite good correspondence between our results and data of other
247: authors carried out on different types of MgB$_2$ samples is
248: evident. In our case averaged over 44 curves, the ratio of the
249: larger gap to the lower one 2.85$\pm$0.15 is close to the
250: theoretical value 3:1 \cite{Liu}.
251:
252: %\section{Conclusion}
253: \textit{In conclusion}. We have analyzed d$V$/d$I$ point-contact
254: spectra of MgB$_{2}$ with clear single- and double-gap structure.
255: The observed distinct maxima in the double gap distribution which
256: is consistent with theoretical calculations \cite{Choi} ruled out
257: surface or multiphase origin of gap structure and testify about
258: intrinsic superconducting double-gap state in MgB$_2$. The
259: averaged gap value ratio turned out to be in accordance with the
260: theoretically predicted ratio 1:3 \cite{Liu}.
261:
262:
263: %\section*{Acknowledgements}
264: \textit{Acknowledgments} The work in Ukraine was supported by the
265: State Foundation of Fundamental Research, Grant $\Phi$7/528-2001.
266: The work at Postech was supported by the Ministry of Science and
267: Technology of Korea through the Creative Research Initiative
268: Program. IKY is grateful to Forschungszentrum Karlsruhe for
269: hospitality.
270:
271:
272: \begin{thebibliography}{99}
273:
274: \bibitem{Nagam}
275: J.\,Nagamatsu, N.\,Nakagawa, T.\,Muranaka et al., Nature {\bf410} 63
276: (2001).
277:
278: \bibitem{Karapet} G.\,Karapetrov, M.\,Iavarone, W.\,K.\,Kwok et al.,
279: Phys. Rev. Lett. {\bf86} 4374 (2001).
280:
281: \bibitem{Rubio} G.\,Rubio-Bollinger, H.\,Suderow, S.\,Vieira,
282: Phys. Rev. Lett. {\bf 86} 5582 (2001).
283:
284: \bibitem{Sharoni} A.\,Sharoni, I.\,Felner, D.\,Millo, Phys. Rev. B {\bf 63}
285: 220508R (2001).
286:
287: \bibitem{Chen} P.\,Seneor, C.-T.\,Chen, N.-C.\,Yeh et al.,
288: Phys. Rev. B {\bf 65} 012505 (2002).
289:
290: \bibitem{Giubileo1} F.\,Giubileo, D.\,Roditchev, W.\,Sachs et al.,
291: cond-mat/0105146.
292:
293: \bibitem{Giubileo2}F.\,Giubileo, D.\,Roditchev, W.\,Sacks et al.,
294: Phys. Rev. Lett. {\bf 87} 177008 (2001).
295:
296: \bibitem{Badr} Mohamed H. Badr, Mario Freamat, Yuri Sushko, and K.-W. Ng,
297: cond-mat/0110421.
298:
299: \bibitem{Zhang} Y.\,Zhang, D.\,Kinion, J.\,Chen et al., cond-mat/0107478.
300:
301: \bibitem{Zasad} H.\,Schmidt, J.\,F.\,Zasadzinski, K.\,E.\,Gray et al.,
302: Phys. Rev. B {\bf 63} 220504(R) (2001).
303:
304: \bibitem{Deuts} A.\,Kohen and G.\,Deutscher, Phys. Rev. B {\bf 64}
305: 060506(R) (2001).
306:
307: \bibitem{Plec01} A.\,Plecenik, \v S.\,Be\v na\v cka, P.\,K\'u\v s et al.,
308: Physica C, {\bf 368} 251 (2002). %cond-mat/0104038v2.
309:
310: \bibitem{Laube} F.\,Laube, G.\,Goll, J.\,Hagel et al., Europhysics Lett.
311: {\bf 56} 296 (2001).
312:
313: \bibitem{Gonelli}R.\,S.\,Gonnelli, A.\,Calzolari, D.\,Daghero et al.,
314: cond-mat/0107239.
315:
316: \bibitem{Szabo}
317: P.\,Szab\'o, P.\,Samuely, J.\,Kacmar\'cik et al.,
318: Phys. Rev. Lett. {\bf 87} 137005 (2001).
319:
320: \bibitem{Bugos} Y.\,Bugoslavsky, Y.\,Miyoshi, G.\,K.\,Perkins et al.,
321: cond-mat/0110296.
322:
323: \bibitem{Bobrov} N.\,L.\,Bobrov, P.\,N.\,Chubov, Yu.\,G.\,Naidyuk et al.,
324: cond-mat/0110006.
325:
326: \bibitem{Buzea} C.\,Buzea and T.\,Yamashita, Superconductors, Science \&
327: Technology, {\bf 14} N11, R115-R146, (2001).
328:
329: \bibitem{Liu} A.\,Y.\,Liu, I.\,I.\,Mazin, and J.\,Kortus, Phys. Rev. Lett.
330: {\bf 87} 087005 (2001).
331:
332: \bibitem{Kortus} J.\,Kortus, I.\,I.\,Mazin, K.\,D.\,Belashchenko et al.,
333: Phys. Rev. Lett. {\bf 86} 4656 (2001).
334:
335: \bibitem{Lee01}W.\,N.\,Kang, Hyeong-Jin Kim, Eun-Mi Choi et al., Phys. Rev.
336: Lett. {\bf 87} 087002 (2001).
337:
338: \bibitem{Wexler} A.\,Wexler, Proc. Phys. Soc. {\bf 89} 927
339: (1966).
340:
341: \bibitem{Canfield} P.\,C.\,Canfield, D.\,K.\,Finnemore, S.\,L.\,Bud'ko et al.,
342: Phys. Rev. Lett. {\bf 86} 2423 (2001).
343:
344: \bibitem{BTK82} G.\,E.\,Blonder, M.\,Tinkham, T.\,M.\,Klapwijk, Phys. Rev. B
345: {\bf 25} 4515 (1982).
346:
347: \bibitem{Dynes} R.\,C.\,Dynes, V.\,Narayanamurti and J.\,P.\,Garno, Phys. Rev.
348: Lett. {\bf 41} 1509 (1978).
349:
350: \bibitem{Choi} Hyoung Joon Choi, David Roundy, Hong Sun et al.,
351: cond-mat/0111183.
352:
353: \bibitem{Finnemore} D.\,K.\,Finnemore, J.\,E.\,Ostenson, S.\,L.\,Bud'ko et al.,
354: Phys. Rev. Lett. {\bf 86} 2420 (2001).
355:
356: \bibitem{Brink} A.\,Brinkman, A.\,A.\,Golubov, H.\,Rogalla et al.,
357: cond-mat/0111115.
358:
359: \bibitem{Servedio} V.\,D.\,P.\,Servedio, S.-L.\,Drechsler, T.\,Mishonov,
360: cond-mat/0111434.
361:
362: \end{thebibliography}
363:
364: \newpage
365:
366: %FIGURE CAPTIONS
367:
368: \begin{figure}
369: \includegraphics[width=7.5cm,angle=0]{f1mgb.eps}
370: \caption[] {Figure 1.\\ Reduced differential resistance
371: $R_N^{-1}$d$V$/d$I$ vs. $V$ measured at T = 4.2 K for four
372: MgB$_2$-Ag contacts with double gap structure (symbols). Thin
373: lines are theoretical dependences calculated with parameters
374: given in the Table I. The curves (1-3) are vertically offset for
375: clarity. Vertical dashed lines show approximately position of
376: large $\Delta_L$ and small $\Delta_S$ gaps. Experimental curves
377: are taken nominally in c-directions} \label{fig1}
378: \end{figure}
379:
380: \begin{figure}
381: \includegraphics[width=7cm,angle=0]{f2mgb.eps}
382: \caption[] {Figure 2.\\ The superconducting energy-gap
383: distribution in c-axis-oriented MgB$_{2}$ thin film in the case:
384: (a) double gap and (b) single gap. Thin lines show Gaussian fit
385: with maxima at (a) 2.45 and 7 meV and (b) 3.5 meV. The histogram
386: window of 0.25 meV for (a) and 0.5 meV for (b) is chosen to get
387: the most close to normal (Gaussian) distribution. (c) Distribution
388: of gap values over the Fermi surface calculated in \cite{Choi}}
389: \label{fig2}
390: \end{figure}
391:
392: \begin{figure}
393: \includegraphics[width=7cm,angle=0]{f3mgb.eps}
394: \caption[] {Figure 3.\\ Dependence of the weight factor $w$ on
395: the point-contact resistance} \label{fig3}
396: \end{figure}
397:
398: \newpage
399:
400: %TABLES
401:
402:
403: \begin{table}
404: \caption{Table 1. Fitting parameters for curves presented in
405: Fig.\,1.}
406: %\begin{ruledtabular}
407: \begin{tabular}{ccccc}
408: \hline
409: Parameters& Curve 1 & Curve 2 & Curve3 & Curve 4\\
410: \hline
411: R$_N$, $\Omega$ & 47 & 35 & 20 & 34\\
412: $\Delta_L$, meV & 7.4 & 6.25 & 7.35& 7.3\\
413: $\Delta_S$, meV & 2.6 & 2.54 & 2.4& 2.6\\
414: $w$-factor& 0.11 & 0.08 & 0.07& 0.06\\
415: Z$_L$ & 0.7 & 0.71 & 0.63& 0.21\\
416: Z$_S$ & 0.75 & 0.55 & 0.56& 0.76\\
417: $\Gamma_L$, meV & 0.4 & 0.1 & 0.55& 0\\
418: $\Gamma_S$, meV & 0.5 & 0.54 & 0.38& 0.3\\
419: \hline
420: \end{tabular}\label{tab1}
421: %\end{ruledtabular}
422: \end{table}
423:
424:
425: \begin{table}
426: \caption{Table 2. Gap values in MgB$_2$ measured by point-contact
427: (PCS) or tunneling spectroscopy (TS). }
428: %\begin{ruledtabular}
429: \begin{tabular}{ccccc}
430: \hline
431: Method&Sample& $\Delta_S$, meV & $\Delta_L$, meV & Refs.\\
432: \hline PCS & Film & $2.45\pm 0.15$ & $7.0\pm 0.4$ & This pap.\\
433: PCS & Film & $2.3\pm 0.3$ & $6.2\pm 0.7$& \cite{Bugos}\\ PCS &
434: Grain & 2.8 & 7& \cite{Szabo}\\ TS & Granular & 3.9 & 7.5&\cite{Giubileo1}\\
435: TS & 50$\mu$ crys. & 3.8 & 7.8&\cite{Giubileo2}\\
436: TS & polycrys. & 1.75 & 8.2&\cite{Badr}\\
437: \hline
438: \end{tabular}\label{tab2}
439: %\end{ruledtabular}
440: \end{table}
441:
442: \end{document}
443: