1: %requires fig1.eps, fig2.eps, fig3.eps,fig4.eps,fig5.eps
2: \documentclass[%preprint,
3: twocolumn,
4: showpacs,superscriptaddress,amssymb,aps,prl]{revtex4}
5:
6: \usepackage{graphicx}
7: \usepackage{bm}
8: %\usepackage{amsmath}
9:
10: \begin{document}
11: \title{Oscillatory Magneto-Thermopower and Resonant Phonon Drag \\in a
12: High-Mobility 2D Electron Gas}
13: \author{Jian Zhang}
14: \affiliation{Department of Physics, University of Utah, Salt Lake City, Utah 84112}
15: \author{S. K. Lyo}
16: \affiliation{Sandia National Laboratories, Albuquerque, New Mexico 87185}
17: \author{R. R. Du}
18: \affiliation{Department of Physics, University of Utah, Salt Lake City, Utah 84112}
19: \author{J. A. Simmons}
20: \affiliation{Sandia National Laboratories, Albuquerque, New Mexico 87185}
21: \author{J. L. Reno}
22: \affiliation{Sandia National Laboratories, Albuquerque, New Mexico 87185}
23:
24:
25: \date{October,10, 2003}
26: \begin{abstract}
27: Experimental and theoretical evidence is presented for new low-magnetic-field
28: ($B<5$ kG) $1/B$-oscillations in the thermoelectric power of a high-mobility
29: GaAs/AlGaAs two-dimensional (2D) electron gas. The oscillations result from
30: inter-Landau-Level resonances of acoustic phonons carrying a momentum
31: equal to twice the Fermi wavenumber at $B = 0$. Numerical calculations show
32: that both 3D and 2D phonons can contribute to this effect.
33: \end{abstract}
34: \pacs{73.50.Rb, 73.40.-c, 72.20.Pa}
35: \maketitle
36:
37:
38: Thermoelectric power (TEP) of a two-dimensional electron gas (2DEG) reflects
39: the electron density of states, scattering dynamics, and
40: electron-phonon interactions. Phonon scattering and consequently the TEP
41: are particularly important in GaAs/Al$_x$Ga$_{1-x}$As heterostructures owing
42: to the strong electron-lattice coupling in these materials. The TEP
43: experiments in the following two limits of the magnetic field ($B$) have been
44: widely pursued
45: \cite{fletcher4,nicholas,davidson,fletcher3,fletcher2,ying,zeitler,bayot,fletcher16,btieke}
46: and the phenomena in these regimes are relatively well-understood
47: \cite{cantrell,lyo38,kuba,lyo40,smith}.
48: At $B = 0$, the TEP ($S_0$) shows a power-law, e.g., $T
49: ^{3-4}$, temperature dependence above $T\sim0.3$ K, indicating that the
50: phonon drag
51: mechanism dominates and the electron-diffusion effect is relatively weak.
52: In the regime of a high field ($B>10$ kG),
53: Shubnikov de-Haas (SdH) oscillation and quantum Hall effect have been
54: observed in TEP \cite{btieke, fletcher16}. The peak
55: amplitude of the TEP are found to be greater than the
56: predicted values of the diffusion TEP by 2 orders of
57: magnitude, but consistent with the phonon-drag TEP. Most
58: of the above experimental data at $B=0$ and at relatively high fields have been
59: successfully explained quantitatively by theories based on the phonon drag
60: TEP.
61:
62: On the contrary, little is known experimentally about the TEP in the
63: regime of a weak magnetic field, where many Landau levels (LLs)
64: are occupied by electrons with large quantum numbers $n\gg1$ at the Fermi
65: level, and electronic transport is generally treated semiclassically. This
66: regime is characteristically distinct for the following reasons. 1)
67: The Fermi wave length is much shorter than the magnetic length
68: $l_B=\sqrt{\hbar/eB}$ (i.e., $2k_F\ell_B\gg1$), and a momentum selection rule
69: governs the scattering of electron guiding centers. In particular, $2k_F$
70: scattering is strongly enhanced, giving rise to the low
71: $B$ resonance phenomena
72: \cite{zudovlip,yang} in this regime. 2) Acoustic phonons in GaAs
73: have suitable energies, and in combination with 1) can participate in
74: inter-LL scattering. This is dramatically different from
75: the regime of higher $B$, where intra-LL scattering dominates
76: and inter-LL scattering is negligible at low temperatures\cite{lyo40}.
77: Elastic intra-LL scattering is important for LL broadening.
78:
79: In this Letter, we report low-$B$ TEP oscillations observed
80: in a high-mobility 2DEG in GaAs/Al$_{x}$Ga$_{1-x}$As
81: heterostructures. The oscillations are periodic in $1/B$ with the
82: peak positions of $B$ proportional to $\sqrt{n_e}$, or to the
83: Fermi wavenumber $k_F=
84: \sqrt {2\pi n_e}$, where $n_e$ is the electron density. Characteristically,
85: such oscillations appear in the temperature range 0.5 K - 1 K, and their
86: amplitude increases with $T$. It will be shown that the TEP oscillations
87: result from inter-LL cyclotron resonance promoted by resonant absorption of
88: phonons carrying a $2k_F$ in-plane momentum and an energy
89: $\ell\hbar \omega_c$ equal
90: to the integer ($\ell=1,2,\cdots$) multiple of $\hbar\omega_c=\hbar
91: eB/m^*$, which is the cyclotron energy
92: with an effective mass $m^*$. 2D phonon modes propagating along the
93: GaAs/Al$_x$Ga$_{1-x}$As interface and 3D phonons are responsible
94: for the TEP oscillations. However, the contribution from 2D phonons is
95: difficult to assess quantitatively because of the lack of electron-2D phonon
96: interaction parameters. On the other hand, numerical
97: calculations show that 3D phonons yield a substantial contribution to the
98: oscillation and behave like 2D phonons for the
99: $2k_F$-oscillation because $q_z$ is restricted to small
100: $q_z\ll2k_F$ at low temperatures due to the phonon occupation factor.
101:
102: The TEP experiments were performed using a sorption pumped $^3$He cryostat
103: and a superconducting solenoid, with the B axis always oriented
104: perpendicular to the 2DEG plane. The samples used here are made from
105: GaAs/Al$_{0.3}$Ga$_{0.7}$As
106: heterostructures grown by molecular beam epitaxy on the (001) GaAs
107: substrate. At low temperature ($T\sim1$K), the $n_e$ and the mobility,
108: $\mu$, can be varied using a red light-emitting diode (LED). Without LED
109: (saturate LED) the $n_e$ (in unit of $10^{11}$cm$^{-2}$ throughout this
110: paper) $\sim$ 1.33 (2.03) and $\mu$ $\sim$ 2 $\times 10^6$ (3 $\times
111: 10^6$) cm$^2$/Vs. To make TEP measurement, a specimen of dimension 8 mm
112: $\times$ 2 mm was first cleaved from the wafer and a Hall bar mesa (width
113: 0.5 mm) was then chemically wet-etched from it by optical lithography. Heat sink
114: to a copper post (0.3 K) was achieved by indium soldering at one end of
115: the specimen.
116: A strain gauge was glued to the
117: other end of the specimen by Ag paint and was used as a heater to create a
118: temperature gradient ($\nabla T$) along the Hall bar direction. Two
119: calibrated RuO$_2$
120: chip sensors were glued by epoxy 2850FT \cite{2850} on the back of
121: the specimen and used to
122: measure the $\nabla T$ along the sample. Electrical leads were made
123: with 38 gauge manganin wires, whose low thermal conductance ensures a
124: negligible heat leak to the $^{3}$He liquid. The whole system was sealed
125: in a
126: vacuum can made of epoxy \cite{1266}. The vacuum can including the copper
127: cold sink were immersed in the $^{3}$He liquid.
128:
129: The TEP $S_{xx}$ is defined by $\nabla V_{xx}=S_{xx}\nabla T$, with
130: the quantities measured in the following manner.
131: A low frequency ($f_{0}$ = 2.7Hz) ac voltage was applied to the heater and
132: the $T_{1}$, $T_{2}$ (see inset, Fig. 1) were measured using the
133: RuO$_{2}$ sensors and a AC bridge. The $V_{xx}$ induced by the
134: thermo-gradient was measured by lock-in method at the frequency of
135: $2f_{0}$ = 5.4 Hz. Both the $T$ and voltage gradients were
136: calculated using the dimensions given by the specimen.
137:
138: \begin{figure}
139: \includegraphics{fig1.eps}
140: \caption{
141: -$S_{xx}$ traces are shown for three densities $n_{e}$ of 1.33, 1.74,
142: and 2.03 in units of 10$^{11}$cm$^{-2}$, respectively; arrows indicate the
143: maxima
144: for $l$ =1, 2, 3, 4 and the shift of the primary ($l$ = 1) peak with
145: increasing $n_{e}$.
146: In the lower figure, the second derivative against $B$ for the
147: high-density
148: trace is shown; the numbers mark the oscillation peaks. Inset shows that
149: oscillations are periodic in $1/B$.}
150: \label{fig1}
151: \end{figure}
152:
153: In Fig.\,1(a), we show the low field magneto-TEP measured at 800
154: mK, for
155: three electron densities. These traces
156: reveal strong new oscillations appearing at $B < 3$ kG, where SdH
157: oscillations are relatively weak. For example, up to four maxima can be
158: clearly seen in the trace of $n_e$ = 2.03. The arrows close to the
159: traces indicate the maxima (indexed as $l$ = 1, 2, 3, 4). A second
160: derivative with respect to $B$, - $d^{2}S_{xx}/ dB^{2}$, for the $n_e$ =
161: 2.03 is plotted in (b), together with a fan diagram (inset) showing the
162: linear relation between the order, $l$, and the inverse $B$. We conclude
163: from all three traces that the TEP oscillations are periodical in
164: $1/B$. Moreover, the peak positions of $B$ scales with $\sqrt{n_e}$, as
165: shown in the inset of
166: Fig.\,\ref{fig1}(a). This behavior is distinct from that of SdH which scales
167: with
168: $n_e$, but consistent with the characteristics of a class
169: \cite{zudovlip,yang} of weak $B$ oscillations originating from cyclotron
170: resonance with a $2k_F$ momentum transfer.
171:
172: \begin{figure}
173: \includegraphics{fig2.eps}
174: \caption{\label{fig2}
175: -$S_{xx}$ (for $n_{e}$ = 2.03$\times$10$^{11}$cm$^{- 2}$) traces at
176: different temperatures show that the low field oscillations become
177: stronger as $T$ increases.
178: }
179: \end{figure}
180:
181: The TEP oscillations exhibit a remarkable $T$ dependence, which,
182: as will be shown later, can be attributed to inter-LL scattering by
183: acoustic phonons. As an example, Fig. 2 shows the TEP data for $n_e =
184: 2.03$ at a $T$ range from 420 mK to 800 mK. Note that the TEP oscillation
185: can be discerned at $T$ as low as 300 mK for our experiment. With increasing
186: $T$, the TEP signal increases dramatically, both at $B$ = 0, and at $B <
187: 5$ kG. The TEP at $B = 0$, $S_0$, is well understood as due to phonon
188: drag. As will be shown, the $S_0$ is power-law dependent in this
189: experiment. As $T$ rises, the TEP oscillation amplitude increases faster
190: than $S_0$. This fact indicates that inter-LL scattering is dominant in
191: the TEP oscillations observed here, since inter-LL scattering mechanism
192: leads to an exponential rather than a power-law $T$
193: dependence. As the cyclotron energy increases, the inter-LL
194: scattering regime with an exponential-law $T$-dependence disappears and only
195: the intra-LL-scattering SdH regime with a power-law
196: $T$-dependence prevails\cite{lyo40}. Indeed, at higher
197: temperatures,
198: $T\ge 800$ mK, the TEP oscillations become the dominating feature in
199: TEP measurement at low field.
200:
201: \begin{figure}
202: \includegraphics{fig3.eps}
203: \caption{\label{fig3}
204: The left figure shows the $T$ dependence of the TEP at zero
205: field ($S_{0}$), for
206: three densities. The solid lines represent $T^3$ temperature
207: dependence. The right figure shows the $T$ dependence of
208: the TEP at primary oscillation maxima ($l=1$), for the three
209: corresponding densities. The E is the fitted value for the phonon energy
210: at $l=1$.}
211: \end{figure}
212:
213: We now turn to a quantitative analysis for $S_0$ and the TEP
214: oscillations. In
215: Fig.\,\ref{fig3}(a), we plot $S_0$ vs. $T$ for three
216: densities. All data show a power-law dependence on $T$, with an exponent
217: between 3 and 4. This observation confirms quantitatively that the TEP at
218: $B$ = 0 is by the phonon drag mechanism \cite{btieke}. The
219: $T$-dependent amplitude of the first peak ($l = 1$) is presented in
220: Fig.\,\ref{fig3}(b). It is worth noting that the data strongly deviates from
221: a power-law, but can be fitted by a modified exponential relation
222: $S_{xx}\sim exp(-E/k_B T)/T^2$, predicted by equation
223: (1). From the slope of the fit we arrive at the activation energies $E \sim$
224: 3.2 K, 3.5 K, 4.0 K, respectively for the three densities, which are
225: somewhat smaller than $\hbar\omega_c\sim$ 4.0 K, 4.4 K, 4.8 K due to the LL
226: broadening. These data strongly support the interpretation of a resonant
227: phonon drag mechanism being the origin of the TEP oscillations.
228:
229: In principle, both 3D and 2D phonons can contribute to the resonant
230: phonon drag mechanism responsible for the oscillations. We start by
231: considering the 3D case, in the following, on a more general
232: ground. Qualitatively, the 2D case can be reduced from the 3D case.
233:
234: We begin with the formula for the phonon drag TEP in a magnetic
235: field derived by Lyo \cite{lyo40}; a similar formulism was given by
236: Kubakaddi
237: and Butcher \cite{cantrell}. The TEP is obtained for a unit volume:
238:
239: $$S_{xx}={{-k_Bh} \over {e\nu(k_BT)^2}}
240: \sum_{s{\bf q}}\sum_{n,n^\prime} u_s\Lambda_{s{\bf q}}q_y^2\ n_{s{\bf
241: q}}|V_{s{\bf
242: q}}|^2 \Delta_z(q_z)\Delta_{n,\ell}(q_\|)$$
243: \begin{equation}\times\int d\varepsilon\rho_n(\varepsilon)\int
244: d\varepsilon^\prime\rho_{n^\prime}(\varepsilon^\prime)
245: f(\varepsilon)[1-f(\varepsilon^\prime)]\delta (\varepsilon+\hbar
246: \omega _{s{\bf q}}-\varepsilon^\prime),\label{e1}
247: \end{equation}
248: where $\nu=\pi n_el_B^2$ is the filling factor for spin-degenerate LLs,
249: $\Lambda_{s{\bf q}}$ the phonon
250: mean-free-path, $n_{s{\bf q}}$ the boson function,
251: $f(\varepsilon)$ the Fermi function, $u_s$ the sound velocity for the mode
252: $s$, $\rho_n$ the spectral function for the LLs, and $\Delta_z(q_z)$ is the
253: conservation factor for $q_z$\cite{lyo40}. The square of the absolute value
254: of the electron-phonon matrix element
255: $|V_{s{\bf q}}|^2$ is proportional
256: to the
257: in-plane-momentum conservation factor:
258: \begin{equation}
259: \Delta_{n,\ell}(q_\|)=\frac{n!}{(n+\ell)!}\chi^\ell
260: e^{-\chi}[L_n^\ell(\chi)]^2,\ \ \chi=\frac{(q_\|l_B)^2}{2},
261: \label{e2}
262: \end{equation}
263: which has a sharp
264: principal maximum near $\chi=4n$, namely, near the in-plane momentum
265: transfer $q_\|\simeq 2k_{\rm F}$ for
266: $n \gg 1,\ \ell $ in view of $\varepsilon_F\simeq
267: n\hbar\omega_c$ \cite{zudovlip,holstein}. In Eq.\,(\ref{e2}),
268: $L_n^\ell(\chi)$ is the associated Laguerre polynomial and
269: $n^\prime=n+\ell$ is the larger of $n$ and $n^\prime$. The phonon occupation
270: factor restricts
271: $q_z$ to small values
272: $q_z\ll 2k_{\rm F}$ for this resonance at low temperatures. There
273: are other secondary peaks below the main peak for $\Delta_{n,\ell}(q_\|)$. To
274: reduce the computing time, we approximate the spectral density function
275: $\rho_n(\varepsilon)$ by a rectangular distribution with a full width
276: $2\Gamma$ centered at the LL energy
277: $\varepsilon_n=\hbar\omega_c(n+1/2)$ and take $\Gamma
278: = 0.2$ meV for numerical evaluation.
279:
280: The calculated TEP is plotted as a function of $B$ in Fig. 4, for
281: $\Lambda_{s{\bf q}}=2$ mm and respectively for three densities, 2.03,
282: 1.74, and 1.33$\times 10^{11}$ cm$^{-2}$ employing field-free screening for
283: the electron-phonon interaction. Other parameters are well-known and are given
284: in Ref.\cite{lyo40}. The TEP is proportional to the phonon mean-free-path
285: $\Lambda_{s{\bf q}}$, which is basically an adjustable parameter. At low
286: temperatures considered here,
287: $\Lambda_{s{\bf q}}$ is determined by boundary scattering and is expected to
288: be of the order of the smallest sample dimension $\sim 2$\ mm. It is seen
289: that the
290: $\ell=1,2,3,\cdots$ peaks appear on top of the regular SdH
291: oscillations which are from the phonons with $q_\|< 2k_F$.
292: Larger
293: $\Gamma$ makes the peaks broader and the valleys shallower compared to the
294: sharper structures obtained for
295: $\rho_n(\varepsilon)=\delta(\varepsilon-\varepsilon_n)$. A, B and C traces
296: display the TEP as a function of
297: $B$ for three densities
298: $n_{e}=1.33,\ 1.74,\ 2.03\times 10^{11}$ cm$^{-2}$ at $0.8$ K. It is seen
299: that there is an approximate scaling relationship between the peak
300: positions of $B$, satisfying
301: $B\propto \sqrt{n_{e}}/\ell$. This
302: relationship is consistent with the inter-LL resonance phonon picture
303: $\hbar\omega_{2k_F}\simeq\ell\hbar eB/m^\ast$ in view of
304: $\omega_{2k_F}\propto k_{\rm F}\propto\sqrt{n_{e}}$, yielding reasonable
305: agreement with experimental data. We also find $S_{xx}\propto
306: \exp(-E/k_BT)/T^2$ in agreement with Fig.\,\ref{fig3}(b) with
307: $E\propto\sqrt{n_e}$ close to the transverse $2k_F$ phonon energy. Transverse
308: phonons yield a dominant contribution ($\sim 70$ \%) through strong
309: piezoelectric scattering at low temperatures. The calculated background TEP
310: is much lower than the peaks compared with the data in Fig.\,\ref{fig2}
311: probably due to the simplistic non-self-consistent density of states employed
312: in the present low-$B$ situation where the LLs are closely separated. Also,
313: the magnitude of the calculated TEP keeps decreasing as
314: $B$ approaches
315: $B=0$ in contrast to the data: Below $B<0.4$ kG, the number of the LLs
316: becomes very large ($n>100$), requiring a zero-$B$ formalism for a more
317: accurate result.
318:
319: The 2D phonon modes relevant in the GaAs/Al$_x$Ga$_{1-x}$As
320: heterostructures are the leaky phonons \cite{zudovlip,efros} with phonon
321: wave vector component $q_{z}$ = 0 and $q=q_\|$. The TEP is again given by
322: Eq.\,(\ref{e1}) with $\Delta(q_z)\equiv 1$ with the summation on $q_z$
323: replaced by the summation over the leaky modes. For a rough estimate, we
324: take same $|V_{s{\bf q}}|$ with the effective sample volume given by
325: $\Omega=S\ell_p$, where $S$ is the cross section of the well and
326: $\ell_p$ is roughly the penetration depth of the mode. The result for
327: the 2D phonons is compared with that of the 3D phonons in Fig.\,\ref{fig4} for
328: $\ell_p=200$ $\AA$ using a pair of longitudinal and transverse modes. The $B$
329: dependence of $S_{xx}$ from the 2D phonons in Fig.\,\ref{fig4} is very similar
330: to that from the 3D phonons except that it is slightly shifted to lower $B$.
331:
332: In Fig.\,\ref{fig3}(a), the linear contribution $S_0\propto T$ from
333: the electron-diffusion TEP is negligible at $B=0$. This situation is
334: similar to the data of Tieke, $et$ $al$.
335: \cite{btieke}, but different from the results of C. Ruf, $et$ $al$. \cite{ruf}, in which the
336: diffusive TEP is visible at 0.6K and the data deviate from the $T^3$ slope at
337: 0.6K.
338:
339: A high-mobility sample is necessary to observe this low-field
340: acoustic-phonon resonance. In high mobility samples, high LLs
341: are not easily smeared out by impurity-scattering. We expect stronger and
342: sharper $2k_F$ oscillations for higher-mobility samples.
343:
344: \begin{figure}
345: \includegraphics{fig4.eps}
346: \caption{\label{fig4}
347: Numerical calculation of the TEP at low fields based on the 3D phonon
348: model (upper three curves). The bottom curve is an estimate from a
349: 2D-phonon model. The primary oscillation maxima increase with
350: $T$, decrease with increasing $n_e$. Their $B$ positions shift to lower field
351: with decreasing $n_e$.}
352: \end{figure}
353:
354: In conclusion, we have reported for the first time an oscillatory
355: TEP in a weak magnetic field, where inter-LL
356: scattering is accessible by acoustic phonons. The observation of such
357: oscillations confirms a generic $2k_F$ momentum selection rule
358: in electronic transport in a weak magnetic field where many LLs
359: are occupied. Finally, it is found that while both 3D and 2D phonons cause
360: qualitatively similar TEP oscillations in this regime, 3D phonons yield a
361: substantial contribution to the oscillations to explain the data.
362:
363: We acknowledge helpful conversations with Prof. R. Fletcher. S. K. L thanks
364: Dr. D. E. Amos for his indispensable help with the computation. The work at
365: the University of Utah was supported by NSF and by a DARPA-QuIST
366: grant. Sandia is a multiprogram
367: laboratory operated by Sandia Corporation, a Lockheed Martin Company, for
368: the U.S. DOE under Contract No.DE-AC04-94AL85000.
369:
370: \bibliographystyle{unsrt}
371: \begin{thebibliography}{10}
372: \bibitem{fletcher4}
373: R. Fletcher, J. C. Maan, and G. Weimann, Phys. Rev. B. {\bf 32}, 8477
374: (1985).
375: \bibitem{nicholas}
376: R. J. Nicholas, J. Phys. C: Solid State Phys. {\bf 18}, L695 (1985).
377: \bibitem{davidson}
378: J. S. Davidson, E. D. Dahlberg, A. J. Valois and G. Y. Robinson,
379: Phys. Rev. B. {\bf 33}, 2941 (1986).
380: \bibitem{fletcher3}
381: R. Fletcher, J. C. Maan, K. Ploog and G. Weimann, Phys. Rev. B. {\bf 33},
382: 7122 (1986).
383: \bibitem{fletcher2}
384: R. Fletcher, M. D'Iorio, A. S. Sachrajda, R. Stoner, C. T. Foxon, and
385: J. J. Harris, Phys. Rev. B. {\bf 37}, 3137 (1988).
386: \bibitem{ying}
387: X. Ying, V. Bayot, M. B. Santos and M. Shayegan, Phys. Rev. B. {\bf 50},
388: 4969 (1994).
389: \bibitem{zeitler}
390: U. Zeitler, R. Fletcher, J. C. Maan, C. T. Foxon, J. J. Harris, and
391: P. Wyder, Surf. Sci., {\bf 305}, 91 (1994).
392: \bibitem{bayot}
393: V. Bayot, E. Grivei, H. C. Manoharan, X. Ying, and M. Shayegan,
394: Phys. Rev. B. {\bf 52}, R8621 (1995).
395: \bibitem{btieke}
396: B. Tieke, R. Fletcher, U. Zeitler, M. Henini, and J. C. Maan,
397: Phys. Rev. B. {\bf 58}, 2017 (1998).
398: \bibitem{fletcher16}
399: R. Fletcher, V. M. Pudalov, Y. Feng, M. Tsaousidou, and P. N. Butcher,
400: Phys. Rev. B. {\bf 56}, 12422 (1997).
401: \bibitem{cantrell}
402: D. G. Cantrell and P. N. Butcher, J. Phys. C. {\bf 20}, 1985 (1987).
403: \bibitem{lyo38}
404: S. K. Lyo, Phys. Rev. B. {\bf 38}, 6345 (1988).
405: \bibitem{kuba}
406: S. S. Kubakaddi, P. N. Butcher, and B. G. Mulimani, Phys. Rev. B. {\bf 40}, 1377 (1989).
407: \bibitem{lyo40}
408: S. K. Lyo, Phys. Rev. B. {\bf 40}, 6458 (1989).
409: \bibitem{smith}
410: M. J. Smith and P. N. Butcher, J. Phys.: Condens.Matter. {\bf 2}, 2375
411: (1990).
412: \bibitem{zudovlip}
413: M. A. Zudov, I. V. Ponomarev, A. L. Efros, R. R. Du, J. A. Simmons, and
414: J. L. Reno, Phys. Rev. Lett. {\bf 86}, 3614 (2001).
415: \bibitem{yang}
416: C. L. Yang, J. Zhang, R. R. Du, J. A. Simmons, and J. L. Reno,
417: Phys. Rev. Lett. 89, 076801 (2002).
418: \bibitem{2850}
419: Stycast 2850FT, Emerson and Cuming Inc., Lexington, MA.
420: \bibitem{1266}
421: Stycast 1266 A/B, Emerson and Cuming Inc., Lexington, MA.
422: \bibitem{holstein}
423: H. Scher and T. Holstein, Phys. Rev. {\bf 148}, 598 (1966).
424: \bibitem{efros}
425: I. V. Ponomarev and A. L. Efros, Phys. Rev. B. {\bf 63}, 165305 (2001).
426: \bibitem{ruf}
427: C. Ruf, H. Obloh, B. Junge, E. Gmelin, K. Ploog, and G. Weimann, Phys. Rev. B. {\bf
428: 37}, 6377 (1988).
429:
430: \end{thebibliography}
431:
432: \end{document}
433:
434:
435:
436:
437: