1: \documentclass{elsart}
2: %\documentstyle{elsart}
3:
4: \usepackage{graphicx}
5:
6: \begin{document}
7: \begin{frontmatter}
8: \title{Electron-electron interaction in multiwall carbon nanotubes\thanksref{talk}}
9: \thanks[talk]{Expanded version ..}
10:
11: \author{A.I. Romanenko, A.V. Okotrub, O.B. Anikeeva, L.G. Bulusheva, N.F. Yudanov}
12: \address{$^1$Institute of Inorganic Chemistry Siberian Branch of Russian Academy of Science,
13: Novosibirsk, Russia}
14: \author{C. Dong, Y. Ni}
15: \address{$^2$National Laboratory for Superconductivity, Institute of Physics Chinese Academy
16: of cience, Beijing, China}
17:
18: \begin{abstract}
19: Magnetic susceptibility $\chi$ of pristine and brominated
20: arc-produced sample of multiwall carbon nanotubes was measured
21: from 4.2 to 400 K. An additional contribution $\Delta \chi(T)$ to
22: diamagnetic susceptibility $\chi(T)$ of carbon nanotubes was found
23: at T $<$ 50 K for both samples. It is shown that $\Delta \chi(T)$
24: are dominated by quantum correction to $\chi$ for interaction
25: electrons (interaction effects-IE). The IE shows a crossover from
26: two-dimensional to three-dimensional at $B$ = 5.5 T. The effective
27: interaction between electrons for interior layers of nanotubes are
28: repulsion and the electron-electron interaction $\lambda$$_c$ was
29: estimated to be $\lambda_c\sim $ 0.26.
30: \end{abstract}
31: \begin{keyword}
32: Electron-electron interaction; Carbon nanotubes; Magnetic
33: susceptibility; Brominated carbon nanotubes. \PACS 72.15.Rn,
34: 75.20.-g, 71.20.Tx
35: \end{keyword}
36: \end{frontmatter}
37:
38:
39:
40: From the time of discovery nanotubes one of the most important
41: problems is the possibility of a superconducting state in them. In
42: a series of experiments were observed the effects which indicate
43: on its existence. There is supercurrents through single-walled
44: carbon nanotubes ~\cite {Kasu99}, persistent currents and magnetic
45: flux trapping ~\cite {Tseb99}. As is known the nature of a
46: superconducting state is the electron-electron interaction (
47: Namely - attraction between electrons). On the other hand
48: electron-electron interaction is exhibited in electronic transport
49: properties of conductors in normal state - so-called quantum
50: interference effects - interaction effects (IE) ~ \cite
51: {Lee85,Al't83}. IE are connected with the correction to density of
52: states of conduction electrons in a results of quantum
53: interferences of electrons at their diffuse motion in random
54: conductors. But in such systems the one-particle processes,
55: so-called weak localization (WL) ~ \cite {Lee85} and weak
56: antilocalization (WAL) ~ \cite {Hika80}, always accompany with IE.
57:
58: The observation of IE corrections to $\chi$ is very important for
59: partition of including of WL, WAL, and IE to different physical properties. From all these
60: corrections only IE contributes to $\chi$.
61:
62: For observation of IE corrections to $ \chi $ it is necessary to
63: divide the contributions connecting with IE and much more on
64: quantity a magnetic susceptibility of sample, and exclude the
65: contribution of paramagnetic impurities. Only in separate cases it
66: is possible. Earlier, with the using of relaxation processes in
67: Mo$_ 2$S $ _ 3 $ ~\cite {Roma85a}, we changed the contribution
68: connected with IE correction to $ \chi $ by quenching of
69: high-temperature metastable state of a sample. In a results, with
70: the using of difference contribution to $ \chi $ in equilibrium
71: and metastable states, we received the IE correction to $ \chi $
72: in the pure state ~\cite {Roma85b}. In this work, with the using
73: of chemical modification of a sample (brominated), we singled out
74: IE correction to $ \chi $ in laminated structures based on
75: multi-layer carbon nanotubes (LS of MWNT).
76:
77: The material which contained MWNTs was synthesized using a set-up
78: for arc discharge graphite evaporation, which was described
79: elsewhere~\cite{Okot96,Okot95}. The arc was maintained with a
80: voltage of 35 V and a current of 1000 A for 15-20 minutes in
81: helium atmosphere of 800 Torr. A nanotube content in the inner
82: part of carbon deposit grown on the cathode was estimated by
83: transmission electron microscopy to reach about
84: $80\%$~\cite{Okot00}. Tubes have from 2 to 30 shells with an outer
85: diameter of 60-150 {\AA }. Scanning electron microscopy (SEM)
86: revealed a predominant orientation of MWNTs was perpendicular to
87: the deposit growth axis (Fig. 1).
88:
89: \begin{figure}
90: \begin{center}
91: \includegraphics[scale=0.8]{figure1.ps}
92: \end{center}
93: \caption{The SEM micrograph of prestine macro-samples LS of MWNT.
94: The deposit axis was perpendicular to picture.} \label{ret}
95: \end{figure}
96:
97: \newpage
98:
99:
100:
101: The brominated material was prepared by exposure of the pristine
102: one to Br$_2$ vapors during 7 days. Its composition CBr$_{0.06} $
103: was determined by x-ray photoelectron spectroscopy.
104:
105: For magnetic measurements a cylindrical sample of diameter 5mm and
106: of length 10 mm was cut out from the pristine or brominated
107: material so that its axis was coincident with the deposit one. As
108: a result, the magnetic field was perpendicular to most of carbon
109: nanotubes in the sample. The weight of samples was about 0.3
110: grams. The temperature dependence of magnetic susceptibility $
111: \chi $ for the samples was measured from 4.5 to 400 K in a field
112: of 0.01, 0.5, and 5.5 T by using a model MPMS-5 SQUID (QUANTUM
113: DESIGN, USA).
114:
115: According to experimental and theoretical data, the basic
116: contribution in $ \chi $ of quasi-two-dimensional graphite (QTDG),
117: including MWNTs, gives orbital magnetic Susceptibility $ \chi _
118: {or} $ connected with extrinsic carriers (EC) ~ \cite
119: {Koto97,Koto87,Koto91}.
120:
121:
122: Figure 2(a) presents the magnetic susceptibility $\chi $ of
123: pristine sample as a function of temperature. The observed
124: behavior was similar to the previously reported measurements on
125: the samples of MWNTs~ \cite{Koto97,Koto87,Koto91,Here94,Tsui00}.
126:
127: \begin{figure}
128: \begin{center}
129: \includegraphics[scale=0.29]{figure2a.ps}
130: \end{center}
131: \label{ret}
132: \end{figure}
133: \begin{figure}
134: \begin{center}
135: \includegraphics[scale=0.28]{figure2b.ps}
136: \end{center}
137: \label{ret}
138: \end{figure}
139: \begin{figure}
140: \begin{center}
141: \includegraphics[scale=0.28]{figure2c.ps}
142: \end{center}
143: \caption{The temperature dependence of magnetic susceptibility
144: $\chi (T)$ ({\bf a}) and $\Delta \chi_{or} (T)/\chi _{or}(T)$ =
145: [$\chi (T) -\chi _{or}(T)]/\chi _{or}(T)$ [({\bf b}) and ({\bf
146: c})] for prestine sample. The solid lines are fits: for ({\bf a})
147: by Eq. (1) in interval 50 - 400 K with parameters; for curve
148: ($\circ$) , $\gamma _0$ = 1.6 eV, $T_0$ = 215 K, $\delta$ = 159
149: K; for ($\bullet$) , $\gamma _0$ = 1.6 eV, $T_0$ = 215 K,
150: $\delta$ = 159 K; for ($\star$) , $\gamma _0$ = 1.7 eV, $T_0$ =
151: 327 K, $\delta$ = 210 K; by Eq. (2) and Eq. (3) for ({\bf b}) and
152: ({\bf c}) respectively in interval 4.5 - 45 K with parameters
153: $T_c$ = 10000 K , $l_{el}/a$ = 0.15.} \label{ret}
154: \end{figure}
155:
156: \newpage
157:
158:
159:
160: Available models well reproduce the temperature dependence of
161: magnetics susceptibility for MWNTs only at T $>$ 50 K~ \cite
162: {Koto97,Koto87,Koto91,Ajik93,Lu95}. In the low-temperature region
163: the experimental data deviate from the theoretical ones that are
164: usually attributed to the paramagnetic impurities contribution. To
165: analyze this anomalous part of the magnetic susceptibility $ \chi
166: $ in detail, that was a goal of the our work, it was necessary
167: to select its high-temperature portion previously. According to
168: theoretical consideration the magnetic susceptibility $\chi $ of
169: quasi-two-dimensional graphite (QTDG) is generally contributed by
170: two components: diamagnetic susceptibility $\chi_D $ ~ \cite
171: {Koto97,Koto87,Koto91,Ajik93,Lu95} and paramagnetic spin
172: susceptibility $\chi_s $. The amount of metallic impurities in the
173: samples under investigation was detected by spectrographic
174: analysis and was less than $10^{-5}$. Furthermore, the signal
175: corresponding to the unpaired spins of transition metals was
176: absents in the EPS spectrum of samples (detection limit equal to
177: $10^{-6}$). Thus the paramagnetic spin contribution to the
178: magnetic susceptibility $ \chi $ of the measured materials was
179: negligible. Hence, the $\chi$ of MWNT sample have been determined
180: by the component $\chi_D $~\cite{Koto97}
181:
182: \begin{equation}
183: \chi _{or}(T) = -\frac{5.45\times 10^{-3}\gamma_0^2}{(T+\delta )[2+exp(\eta )+exp(-\eta )]} ,
184: \end{equation}
185:
186: where $\gamma_0$ is the band parameter for two-dimensional case,
187: $\delta $ is the additional temperature formally taking into
188: account "smearing" the density of states due to electron
189: nonthermal scattering by structure defects, $\eta $ =
190: $E_F/k_B(T+\delta )$ represents reduced Fermi level ($E_F$), $k_B$
191: is the Boltzmann constant. Using an electrical neutrality equation
192: in the 2D graphite model ~\cite{Koto97} $\eta $ can be derived by
193: $\eta $ = sgn($\eta _0$)[0.006$\eta _0^4$ - 0.0958$\eta _0^3$ +
194: 0.532$\eta _0^2$ - 0.08$\eta_0$] ~\cite{Koto87} with an accuracy
195: no less then $1\%$. The $\eta_0$ is determined by $\eta_0 $ =
196: $T_0/(T+\delta )$, where $T_0$ being degeneracy temperature of
197: extrinsic carriers (EC) depends on its concentration $n_0$ only.
198: The value of $\delta $ can be estimated independently
199: ~\cite{Koto91} as $\delta $ = $\hbar $/$\pi k_B\tau _0$ ,
200: where $\hbar $ is the Planck constant, $\tau _0$ is a relaxation time of the carrier
201: nonthermally scattered by defects~\cite{Koto91}. Generally, the
202: number of EC in QTDG is equal to that of scattering centers and
203: $\delta$ depends only on $T_0$, i.e. $\delta = T_0/r$, where $r$
204: is determined by scattering efficiency ~\cite{Koto87}. These
205: parameters were chosen to give the best fit of the experimental
206: data (Fig. 2a). At high field ($B > $ 1 T) the magnetic
207: susceptibility $\chi (T)$ decreases in all interval of
208: temperature. In this field region the magnetic
209: length~\cite{Lee85,Al't83} $l_B$ = $(\hbar c/2eB)^{1/2}$ is much
210: less than a tube length. Therefore, susceptibility probes only
211: small local areas of the graphite plane, and is expected to be the
212: geometrical averaged of that of rolled-up sheets of graphite. At
213: low fields the magnetic length is larger than the dimension of
214: most tubes in the sample, and this geometrical correction may be
215: neglected~\cite{Here94}.
216:
217: The data in Fig. 2 show that at T$ < $50 K there is an additional
218: contribution $\Delta \chi _{or}(T)=\chi (T) -\chi _{or}(T)$ to
219: $\chi (T)$. According to theoretical
220: calculations~\cite{Lee85,Al't83} only electron-electron
221: interactions can contribute to magnetic susceptibility, which may
222: be divided into two parts. The first part $\Delta \chi _{or}$
223: comes from correction to orbital susceptibility $\chi_{or}$. The
224: other part $\Delta \chi _s$ is associated with the correction to
225: spin susceptibility $\chi_s$. The latter contribution is
226: negligible because the carbonaceous material used in sample
227: preparation was indicated by spectrographic analysis and these
228: material contained very small amount of magnetic impurities (less
229: than detection limit). The dominated part $\Delta \chi _{or}$,
230: divided on $\chi _{or}$, was described by~\cite{Lee85,Al't83}
231:
232: \begin{equation}
233: \frac{\Delta \chi_{or} (T)}{\chi _{or}(T)}=
234: -\frac{\frac{4}{3}(\frac{l_{el}}{h})ln[ln(\frac{T_c}{T})]}
235: {ln(\frac{k_B T_c\tau _{el}}{\hbar })} ,(d = 2) ,
236: \end{equation}
237: \begin{equation}
238: \frac{\Delta \chi_{or} (T)}{\chi _{or}(T)}=
239: -\frac{2(\frac{\pi }{6})\xi (\frac{1}{2})
240: (\frac{k_B T\tau _{el}}{\hbar })^{1/2}}{ln(\frac{T_c}{T})} ,(d = 3) ,
241: \end{equation}
242:
243: where value of $\xi (\frac{1}{2})$ $\sim $ 1, $l_{el}$ is the
244: electron mean free path; $\tau _{el}$ represents the elastic
245: relaxation time, which is about 10$^{-13}$ sec for MWNT
246: ~\cite{Baxe97}; $h $ is the thickness of layer in two-dimensional
247: case; $d $ denoted a system dimensionality; $T_c$ = $\theta
248: _Dexp(\lambda _c^{-1}$), where $\theta _D$ is the Debye
249: temperature and $\lambda _c$ is the constant described the
250: electron-electron interaction in Cooper canal ($\lambda _c > 0 $
251: in a case of electron repulsion). The dependence in Eq. (2) is
252: determined by $ln[ln(\frac{T_c}{T})]$ term because at low
253: temperature, in the disordered systems, $\tau _{el}$ is the
254: temperature independent while all other terms are constants. The
255: dependence in Eq. (3) is governed by $\ T^{1/2}$ term as $T_c \gg
256: T $ and, therefore, $ln(\frac{T_c}{T})$ is constant relative to $\
257: T^{1/2}$.
258:
259: The additional contribution to $\chi (T)$ as a function of
260: $ln[ln(\frac{T_c}{T})]$ and $\ T^{1/2}$ is presented in Fig. 2(b)
261: and Fig. 2(c). The $\triangle \chi_{or} (T)/\chi _{or}(T)$ clearly
262: shows the dependence given by Eq. (2) at low magnetic field and
263: one given by Eq. (3) at high magnetic field, while at $B$ = 0.5 T
264: the temperature behavior of $\triangle \chi_{or} (T)/\chi
265: _{or}(T)$ is deviated from these two cases. As seen from Fig. 2,
266: an absolute value of $\triangle \chi_{or} (T)/\chi _{or}(T)$ at
267: all magnetic fields applied to the pristine sample increases with
268: decreasing temperature that has been predicted for IE in the
269: systems characterized by electron-electron repulsion
270: ~\cite{Lee85,Al't83}. Hence, at $B$ = 5.5 T a transfer from
271: two-dimensional IE correction to three-dimensional one takes
272: place. At lower magnetic field the interaction length $L_I(T)$ =
273: $(\hbar D/k_BT)^{1/2}$ is much less than the magnetic length $l_B
274: = (\hbar c/2eB)^{1/2}$, which in turn becomes dominant at high
275: field. An estimation of the characteristic lengths gave
276: respectively the value of $L_I(4.2K)$ = 130 {\AA } (taking into
277: account that the diffusion constant $D$ = 1
278: cm$^2$/s~\cite{Baxe97}) and the value of $l_B$ = 100 {\AA } at $B$
279: = 5.5 T.
280:
281:
282: Taking into account experimentally apparent crossover (Fig. 2)
283: from two-dimensional (equation 2) to three-dimensional (equation
284: 3) of temperature dependence of $ \triangle \chi _ {or} (T) /\chi
285: _ {or} (T) $ we concluded that effective diameter of nanotubes was
286: in an interval between 100 {\AA} and 130 {\AA}.
287:
288: Partitioning of the contributions $ \chi _ {or} (T) $ and $
289: \triangle \chi _ {or} (T) /\chi _ {or} (T) $ can be carried out in
290: a result of changing of extrinsic carriers by the chemical
291: modification of the prestine sample. In this case the $ \chi _
292: {or} (T) $ should be changed ~\cite {Koto97} and the $ \triangle
293: \chi _ {or} (T) /\chi _ {or} (T) $ remain the invariable ~\cite
294: {Lee85,Al't83}.
295:
296: \begin{figure}
297: \begin{center}
298: \includegraphics[scale=0.28]{figure3a.ps}
299: \end{center}
300: \label{ret}
301: \end{figure}
302: \begin{figure}
303: \begin{center}
304: \includegraphics[scale=0.26]{figure3b.ps}
305: \end{center}
306: \label{ret}
307: \end{figure}
308: \begin{figure}
309: \begin{center}
310: \includegraphics[scale=0.26]{figure3c.ps}
311: \end{center}
312: \caption{The temperature dependence of magnetic susceptibility
313: $\chi (T)$ ({\bf a}) and $\Delta \chi_{or} (T)/\chi _{or}(T)$ =
314: [$\chi (T) -\chi _{or}(T)]/\chi _{or}(T)$ [({\bf b}) and ({\bf
315: c})] for bromineted sample. The solid lines are fits: for ({\bf
316: a}) by Eq. (1) in interval 50 - 400 K with parameters; for curve
317: ($\circ$) , $\gamma _0$ = 1.4 eV, $T_0$ = 340 K, $\delta$ = 252
318: K; for ($\bullet$) , $\gamma _0$ = 1.4 eV, $T_0$ = 300 K,
319: $\delta$ = 273 K; for ($\star$) , $\gamma _0$ = 1.5 eV, $T_0$ =
320: 435 K, $\delta$ = 325 K; by Eq. (2) and Eq. (3) for ({\bf b}) and
321: ({\bf c}) respectively in interval 4.5 - 45 K with parameters
322: $T_c$ = 10000 K , $l_{el}/a$ = 0.15.} \label{ret}
323: \end{figure}
324:
325: \newpage
326:
327: The intercalation of single wall carbon nanotubes (SWNT) by
328: bromine lied to increase of $n_0$~\cite{Rao97}. We supposed that
329: in MWNT may be similarly situation and used brominated MWNT for
330: our investigations. Figure 3 show $\chi (T)$ for the brominated
331: sample. The $T_0$ and $\delta $ for these sample are shown in
332: figure 3 caption. We estimated $n_0$ at low temperature and low
333: field in framework of theory of QTDG~\cite{Koto91} $n_0 =
334: 4(k_BT_0)^2/(3\pi a^2\gamma _0^2)$, where $a$ = 0.246 nm -
335: lattice parameter in layer. These estimations gives: $n_{0ini}$
336: $\sim $ 3$\times$10$^{10}$ cm$^{-2}$ - for prestine sample;
337: $n_{0Br} \sim $ 10$^{11}$ cm$^{-2}$ - for brominated sample. The
338: $n_0$ increases in brominated sample about 3 times. According to
339: the Drude formula, the conductivity is proportional to $n_0$ and
340: $\tau _{el}$. We measured the conductivity of pristine and
341: brominated samples and find that conductivity increase in 3 times
342: in a result of bromination. Taking into account that $n_0$ also
343: increase by about 3 times in a result of bromination we conclude
344: that $\tau _{el}$ practically did not change during bromination.
345: The $\Delta \chi_{or} (T)/\chi _{or}(T)$ don't change during
346: intercalation [Fig. 3(b) and Fig. 3(c)]. This result indicates
347: that $\Delta \chi_{or} (T)/\chi _{or}(T)$ doesn't connected with
348: n$_0$, and connected with the IE correction to $\chi$. From
349: experimental data $\Delta \chi_{or} (4.5K)/\chi _{or}(4.5K)$ =
350: 0.027 for three-dimensional IE correction to $\chi$ we estimated
351: the $T_c$ from Eq. (3) and fined $T_c$ = 5$\times $10$^4$ K. We
352: estimated the $\lambda _c$ from $T_c = \theta _Dexp(\lambda
353: _c^{-1})$ with Debye temperature for carbon
354: nanotubes~\cite{Bene96} $\theta _D$ = 1000 K and obtained $\lambda
355: _c \sim $ 0.26. From experimental data $\Delta \chi_{or}
356: (4.5K)/\chi _{or}(4.5K)$ = 0.07 for two-dimensional IE correction
357: to $\chi$ we estimated the $l_{el}/h$ from Eq. (2) and fined
358: $l_{el}/h \sim $ 0.15. If $h \sim d_m$ and 100 {\AA } $\leq d_m
359: \leq$ 130 {\AA } so 20 {\AA } $\leq l_{el} \leq $ 15 {\AA }. This
360: estimation in a quite good agreement with estimation $l_{el} =
361: (D\tau _{el})^{1/2} \sim $ 30 {\AA } for so crude estimation.
362:
363:
364:
365: In summary, we have investigated the additional contribution to
366: temperature dependence of orbital magnetic susceptibility $\Delta
367: \chi_{or} (T)/\chi _{or}(T)$ of lamination structure of multiwall
368: carbon nanotubes at $T < $ 50 K. It is shown that $\Delta
369: \chi_{or} (T)/\chi _{or}(T)$ is connected with quantum correction
370: to magnetic susceptibility for interaction electron. At low field
371: this correction is two-dimensional. At $B$ = 5.5 T was observed
372: three-dimensional correction to magnetic susceptibility. This
373: crossover from $2d$ to $3d$ behavior of $\Delta \chi_{or} (T)/\chi
374: _{or}(T)$ is connected with decreasing of magnetic length up to
375: value less then typical mean diameter of nanotubes. It is shown
376: that brominated of samples lead to increasing of extrinsic
377: carriers $n_0$ from $n_{0ini} \sim $ 3$\times $10$^{10}$ cm$^{-2}$
378: for prestine sample up to $n_{0Br} \sim $ 10$^{11}$ cm$^{-2}$ for
379: brominated samples. But $\Delta \chi_{or} (T)/\chi _{or}(T)$ did
380: not changed when $n_0$ increases, which is in full agreement with
381: theoretical predictions. From $\Delta \chi_{or} (4.5K)/\chi
382: _{or}(4.5K)$, we estimated the constant of electron-electron
383: interaction $\lambda _c \sim $ 0.26. This interaction is repulsion
384: for interior layers, which give the domination contribution to
385: $\Delta \chi_{or} (T)/\chi _{or}(T)$ as a integration value.
386:
387:
388: It is necessary to note, that the correction to a magnetic
389: susceptibility observation by us, and, accordingly estimation of a
390: constant of electron-electron interaction it is integrated values
391: average on all stratums in structural nanotubes. Therefore the
392: made conclusion about repulsion character of interaction between
393: electrons does not eliminate opportunity of an attraction between
394: electrons in high layer nanotubes, which contribution in magnetic
395: susceptibility is small in comparison with sum by the contribution
396: of all remaining stratums of a tube.
397:
398:
399:
400: \section{Acknowledgements}
401: The authors thank Ms. Chaoying WANG for the SEM analysis of the
402: samples, and Dr. V.A. Nadolinny for the EPR measurements. The work
403: was supported by Lu Jiaxi grant for international joint research
404: from Chinese Academy of Sciences, and Russian scientific and
405: technical program "Fullerenes and atomic clusters" (Projects No
406: 5-1-98), INTAS (Grant Nos 97-1700, 00-237), Russian Foundation of Basic Research
407: (Grants No: 00-02-17987; 00-03-32510; 00-03-32463; 01-02-0650),
408: and Interdisciplinary Integral Program of Siberian Branch of
409: Russian Academy of Science (Grant No 61).
410:
411:
412: \begin{thebibliography}{99}
413:
414:
415: \bibitem{Kasu99} A.Yu. Kasumov {\em et al.}, {\em Science\/} {\bf 284} (1999) 1508--1510.
416:
417: \bibitem{Tseb99} V.I. Tsebro, O.E. Omel'yanovski, and A.P. Moravski,
418: {\em Sov. Phys. JETP Lett.\/} {\bf 70} (1999) 462--468.
419:
420: \bibitem{Lee85} P.A. Lee, and T.V. Ramakrishnan, {\em Rev. Modern Physics\/} {\bf 57}
421: (1985) 287--337.
422:
423: \bibitem{Al't83} B.L. Al'tshuler, A.G. Aronov, A.Yu. Zyuzin,
424: {\em Sov. Phys. JETP\/} {\bf 53} (1983) 889--902.
425:
426: \bibitem{Hika80} S. Hikami, A.I. Larkin, Y. Nagaoka,
427: {\em Prog. Theor. Phys.\/} {\bf 63(2)} (1980) 707--727.
428:
429: \bibitem{Roma85a} A.I.Romanenko, A.K.Dzhunusov, I.N. Kuropyatntsk, and E.V. Kholopov,
430: {\em Sov. Phys. JETP Lett.\/} {\bf 41} (1985) 237--239.
431:
432: \bibitem{Roma85b} A.I.Romanenko, F.S. Rakhmenkulov, V.N. Ikorski, P.S. Nikitin,
433: {\em Sov. Phys. JETP Lett.\/} {\bf 42} 377 (1985) 377--380.
434:
435: \bibitem{Okot96} A.V. Okotrub {\em et al.}, {\em Inorganic Materials\/} {\bf 32}
436: (1996) 974--978.
437:
438: \bibitem{Okot95} A.V. Okotrub {\em et al.}, {\em Phys. Low-Dim. Struct.\/} {\bf 8/9}
439: (1995) 139--158.
440:
441: \bibitem{Okot00} A.V. Okotrub {\em et al.}, {\em Appl. Phys. A\/} {\bf 71} 1 (2000) 481--486.
442:
443: \bibitem{Koto97} A.S. Kotosonov, and S.V. Kuvshinnikov,
444: {\em Phis. Lett. A\/} {\bf 229} (1997) 377--380.
445:
446: \bibitem{Koto87} A.S. Kotosonov, {\em Sov. Phys. JETP\/} {\bf 93} 1870 (1987) 1870--1878.
447:
448: \bibitem{Koto91} A.S. Kotosonov, {\em Sov. Phys. Solid State\/} {\bf 33} (1991) 1477--1485.
449:
450: \bibitem{Here94} J. Heremans, C.H. Olk, and D.T. Morelli,
451: {\em Phys. Rev. B\/} {\bf 49} 15122 (1994) 15122--15125.
452:
453: \bibitem{Tsui00} F. Tsui, L. Jin and O. Zhou, {\em Appl. Phys. Lett. B\/} {\bf 76},
454: (2000) 1452--1454.
455:
456: \bibitem{Ajik93} H. Ajiki and Ando, {\em J. Phys. Soc. Jpn.\/} {\bf 62}, (1993) 2470--2480.
457:
458: \bibitem{Lu95} J.P. Lu, {\em Phys. Rev Lett.\/} {\bf 74}, (1995) 1123--1126.
459:
460:
461: \bibitem{Baxe97} M. Baxendale, V.Z. Mordkovich, S. Yoshimura, and R.P.H. Chang,
462: {\em Phys. Rev. B\/} {\bf 56} (1997) 2161--2165.
463:
464: \bibitem{Rao97} A.M. Rao, P.C. Eklund, S. Bandow, A. Thess, and R.E. Smalley,
465: {\em Nature\/} {\bf 388} (1997) 257--259.
466:
467: \bibitem{Bene96} L.X. Benedict, S.G. Louie, and M.L. Cohen,
468: {\em Solid State Commun.\/} {\bf 100} (1996) 177--180.
469:
470:
471:
472: \end{thebibliography}
473:
474: \end{document}
475: