1: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
2: %\documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
3: \usepackage{graphicx}% Include figure files
4: \usepackage{dcolumn}% Align table columns on decimal point
5: \usepackage{bm}% bold math
6: \begin{document}
7: \title{Conductivity of a spin-polarized two-dimensional electron liquid in the ballistic regime}
8: \author{A.~A. Shashkin, E.~V. Deviatov, V.~T. Dolgopolov, and A.~A. Kapustin}
9: \affiliation{Institute of Solid State Physics, Chernogolovka, Moscow District 142432, Russia}
10: \author{S. Anissimova, A. Venkatesan, and S.~V. Kravchenko}
11: \affiliation{Physics Department, Northeastern University, Boston, Massachusetts 02115, U.S.A.}
12: \author{T.~M. Klapwijk}
13: \affiliation{Kavli Institute of Nanoscience, Delft University of Technology, 2628 CJ Delft, The Netherlands}
14: \begin{abstract}
15: In the ballistic regime, the metallic temperature dependence of the
16: conductivity in a two-dimensional electron system in silicon is found
17: to change non-monotonically with the degree of spin polarization. In
18: particular, it fades away just before the onset of complete spin
19: polarization but reappears again in the fully spin-polarized state,
20: being, however, suppressed relative to the zero-field case. Analysis
21: of the degree of the suppression allows one to distinguish between
22: the screening and the interaction-based theories.
23: \end{abstract}
24: \pacs{71.30.+h, 73.40.Qv}
25: \maketitle
26:
27: \section{INTRODUCTION}
28:
29: Much interest has been attracted recently to the anomalous properties
30: of low-disordered, strongly correlated two-dimensional (2D) electron
31: systems. The effects of electron-electron interactions are especially
32: strong in silicon metal-oxide-semiconductor field-effect transistors
33: (MOSFETs) (for recent reviews, see
34: Refs.~\cite{kravchenko04,shashkin05}), but they are also pronounced
35: in other systems like GaAs/AlGaAs \cite{tan05} and Si/SiGe
36: heterostructures \cite{lai05}. Interactions lead, in particular, to
37: critical behavior of the Pauli spin susceptibility
38: \cite{shashkin01,vitkalov01} and sharply increasing effective mass at
39: low electron densities \cite{shashkin02,shashkin03,anissimova06}.
40: These phenomena (at least in Si MOSFETs) are not dominated by spin
41: exchange effects, since the Land\'e $g$ factor is found to be close
42: to its value in a bulk semiconductor and the effective mass is
43: insensitive to the degree of spin polarization. At the same time,
44: spin effects are the origin of the strong positive magnetoresistance
45: in parallel magnetic fields (see, e.g.,
46: Refs.~\cite{simonian97,yoon00}). Therefore, one can probe the spin
47: effects by studying peculiarities of a spin-polarized 2D electron
48: system. The case of complete spin polarization of the electron system
49: is especially interesting because it is the simplest from the
50: theoretical point of view.
51:
52: It is known that application of a parallel magnetic field causes
53: giant (orders of magnitude) positive magnetoresistance and fully
54: suppresses the metallic state near the 2D metal-insulator transition
55: \cite{simonian97,yoon00,dolgopolov92}. However, if the electron
56: density is not too low (ballistic regime, $k_BT\gtrsim\hbar/\tau$
57: \cite{zala01}), the metallic temperature dependence of conductivity
58: has been found to persist in the fully spin-polarized state
59: \cite{okamoto00,mertes01,tsui05}. (Note that in silicon-based devices
60: studied in these papers, electrons possess the valley degree of
61: freedom, which survives in the fully spin-polarized state.)
62: Conductivity of silicon MOSFETs in this regime was studied in
63: Ref.~\cite{tsui05}. However, for much of their data (particularly at
64: relatively high temperatures and/or electron densities), the complete
65: spin polarization was in fact not reached as a result of the
66: insufficiently high magnetic fields used.
67:
68: Theoretically, linear-in-temperature corrections to the zero-field
69: conductivity in the ballistic regime were calculated in
70: Ref.~\cite{gold86}. In the newer theory \cite{zala01}, the exchange
71: interaction terms were treated more carefully. However, it turned out
72: that at $B=0$, both the screening \cite{gold86} and the
73: interaction-based \cite{zala01} theories describe the
74: temperature-dependent conductivity equally well \cite{shashkin04a}.
75: To distinguish between them, studies of the effect of the parallel
76: magnetic field on conductivity may be helpful.
77:
78: Here we experimentally study the transport properties of a 2D
79: electron system in silicon in parallel magnetic fields at different
80: degrees of spin polarization in the ballistic regime. We show that in
81: a completely spin-polarized state, disorder effects are dominant when
82: approaching the regime of strong localization, which is in contrast
83: to the behavior of the unpolarized state in low-disordered 2D
84: electron systems. The temperature-dependent correction to the elastic
85: relaxation time is found to change strongly with the degree of spin
86: polarization, reaching a minimum just below the onset of full spin
87: polarization, where the conductivity is practically independent of
88: temperature. In the fully spin-polarized state, the correction
89: mentioned above is about two times weaker than that in $B=0$ at the
90: same electron density. This is consistent with what one expects
91: according to the simple version of screening theory
92: \cite{dolgopolov00}.
93:
94: \section{EXPERIMENTAL TECHNIQUE AND SAMPLES}
95:
96: Measurements were made in an Oxford dilution refrigerator on
97: (100)-silicon MOSFETs with peak electron mobilities of about
98: 3~m$^2$/Vs at 0.1~K. The resistance was measured with a standard
99: 4-terminal technique at a low frequency (1~Hz) to minimize the
100: out-of-phase signal. Excitation current was kept low enough (below
101: 1~nA) to ensure that measurements were taken in the linear regime of
102: response. Contact resistances in these samples were minimized by
103: using a split-gate technique that allows one to maintain a high
104: electron density in the vicinity of the contacts (about $1.5\times
105: 10^{12}$~cm$^{-2}$), regardless of its value in the main part of the
106: sample.
107:
108: \begin{figure}
109: \scalebox{0.42}{\includegraphics[clip]{fig1.eps}}
110: \caption{\label{fig1} Conductivity, mobility, and elastic scattering
111: time vs.\ electron density at a temperature of 0.1~K for
112: spin-unpolarized (triangles) and fully spin-polarized (circles)
113: states in two slightly different samples A at $B=0$ and 9.5~T (a) and
114: B at $B=0$ and 14~T (b).}
115: \end{figure}
116:
117: In Fig.~\ref{fig1}, we show low-temperature conductivity as a
118: function of electron density, $n_s$, in the metallic regime (i.e.,
119: when the conductivity $\sigma>e^2/h$) for fully spin-polarized and
120: spin-unpolarized states in two slightly different samples. As the
121: electron density is decreased, the conductivity significantly drops
122: as a result of a decrease in the electron mobility $\mu$ (insets to
123: Fig.~\ref{fig1}). This mobility decrease originates from the effects
124: of both disorder and electron-electron interactions. The former
125: determines the elastic relaxation time, $\tau$, while the latter is
126: responsible for the enhanced effective mass
127: \cite{shashkin02,shashkin03,anissimova06}. Taking the effective mass
128: values from Ref.~\cite{shashkin03}, we have calculated $\tau$ as
129: shown in the insets to Fig.~\ref{fig1}. For the spin-unpolarized
130: state, the behavior of $\tau$ for the two samples is qualitatively
131: different, although the mobilities are very similar. In sample A, the
132: decrease in $\mu$ at low densities is dominated by electron-electron
133: interactions (increasing effective mass), while in sample B, disorder
134: effects are more pronounced and lead to $\tau$ decreasing at low
135: $n_s$. In the fully spin-polarized state, the disorder effects
136: prevail in both samples.
137:
138: \section{RESULTS}
139:
140: Experimental traces of the parallel-field magnetoresistance at
141: different temperatures are displayed in Fig.~\ref{fig2}. The
142: low-temperature resistivity, $\rho$, rises with $B$ and saturates
143: above a certain $n_s$-dependent magnetic field, $B_{\text{sat}}(0)$,
144: corresponding to the onset of complete spin polarization of the 2D
145: electrons \cite{okamoto99}. Increasing the temperature leads to
146: smearing the dependences so that the resistance saturation occurs at
147: higher magnetic fields. In other words, the saturation field
148: increases as the temperature is increased \cite{vitkalov01}. The
149: resistivity rises appreciably with increasing temperature in both
150: $B=0$ and $B>B_{\text{sat}}$, here the saturation field
151: $B_{\text{sat}}$ corresponds to the highest temperature used in the
152: experiment. The magnetoresistance at two electron densities measured
153: at the highest temperature used, $T\approx1.2$~K, is shown in the
154: inset to Fig.~\ref{fig3}. The fact that the magnetoresistance
155: saturates at sufficiently high magnetic fields confirms that the full
156: spin polarization is reached in our experiment even at this
157: temperature. As seen from Fig.~\ref{fig2}, just below
158: $B_{\text{sat}}(0)$ the resistivity practically does not depend on
159: temperature up to the highest temperatures used. The validity of this
160: effect, which has also been observed in Refs.~\cite{mertes01,tsui05},
161: has been verified at ten electron densities in the range between
162: $1.38\times10^{11}$ and $2.42\times10^{11}$~cm$^{-2}$. We would like
163: to emphasize that the flattening of $\sigma(T)$ just below the onset
164: of complete spin polarization makes it difficult to analyze the data
165: for $\sigma(T)$ obtained in a fixed magnetic field or in a narrow
166: field region \cite{mertes01,tsui05} as the complete spin polarization
167: may have not been reached at higher temperatures and/or electron
168: densities.
169:
170: \begin{figure}
171: \scalebox{0.42}{\includegraphics[clip]{fig2.eps}}
172: \caption{\label{fig2} Magnetoresistance at temperatures 0.5 (solid
173: line), 0.8 (dashed line), and 1.2~K (dotted line) on sample B. The
174: inset shows a detailed view of the magnetoresistance just before the
175: onset of complete spin polarization.}
176: \end{figure}
177:
178: The low-temperature ratio $\rho(B_{\text{sat}})/\rho(0)$ vs.\
179: electron density is shown in Fig.~\ref{fig3}. In agreement with the
180: previously obtained data, it increases weakly with decreasing $n_s$,
181: being close to the value $\rho(B_{\text{sat}})/\rho(0)=4$ predicted
182: by the theory of the spin-polarization-dependent screening of a
183: random potential \cite{dolgopolov00}. As seen from the inset, the
184: ratio $\rho(B_{\text{sat}})/\rho(0)$ diminishes somewhat at higher
185: temperatures.
186:
187: The normalized conductivity as a function of temperature in fully
188: spin-polarized and spin-unpolarized states is depicted in the inset
189: to Fig.~\ref{fig4}. The correction to $\sigma/\sigma(0)$ is linear in
190: temperature with the slope given by $A^*=-\sigma_0^{-1}d\sigma/dT$,
191: where $\sigma_0=\sigma(0)$ is obtained by linear extrapolation of the
192: data to $T=0$. We emphasize that the so-defined slopes do not depend
193: on $\tau$ and are therefore different from the slopes defined in
194: Ref.~\cite{tsui05} as $d\sigma/dT$: the ratio of the slopes
195: $r\equiv[d\sigma(B_{\text{sat}})/dT]/[d\sigma(0)/dT]$ used there is
196: smaller by a factor of $\tau(0)/\tau(B_{\text{sat}})\approx4$ than
197: the ratio $A^*(B_{\text{sat}})/A^*(0)$ used in this paper.
198:
199: In Fig.~\ref{fig4}, we show how the ratio of the slopes
200: $A^*(B_{\text{sat}})/A^*(0)$ for completely spin-polarized and
201: unpolarized states changes with electron density. Being approximately
202: equal to 0.5 at low $n_s$, the slope ratio increases weakly with
203: increasing $n_s$ but remains less than one in the range of electron
204: densities studied. Thus, the metallic behavior of the normalized
205: conductivity is always suppressed in the fully spin-polarized state.
206:
207: \begin{figure}
208: \scalebox{0.42}{\includegraphics[clip]{fig3.eps}}
209: \caption{\label{fig3} Change of the resistance ratio,
210: $\rho(B_{\text{sat}})/\rho(0)$, with electron density in sample B.
211: The inset shows the normalized magnetoresistance, measured at the
212: highest temperature used in this experiment, at electron densities
213: $1.46\times10^{11}$~cm$^{-2}$ (dotted line) and
214: $2.13\times10^{11}$~cm$^{-2}$ (solid line). The field of resistance
215: saturation is marked by arrows.}
216: \end{figure}
217:
218: \section{DISCUSSION}
219:
220: We give a qualitative account of the absence of the $\sigma(T)$
221: dependence just below the onset of complete spin polarization. In the
222: magnetic field $B=B_{\text{sat}}(0)$, the degree of spin polarization
223: decreases linearly with temperature: $\xi=1-\gamma k_BT/E_F$ (where
224: the factor $\gamma\sim1$ and $E_F$ is the Fermi energy of the
225: spin-polarized 2D electrons). The increase in the number of electrons
226: with opposite spin direction naturally leads to increasing
227: conductivity. Therefore, near the onset of complete spin polarization
228: there exists another contribution to the temperature-dependent
229: conductivity, whose sign is opposite compared to the conventional
230: screening behavior of $\sigma(T)$. In the simple version of the
231: screening theory \cite{dolgopolov00}, the derivative $d\rho/d\xi$ at
232: $T=0$ tends to infinity as one approaches the field
233: $B_{\text{sat}}(0)$ from below. This feature will obviously be
234: smeared out at finite temperatures and/or due to the disorder present
235: in real electron systems. It is clear that depending on disorder
236: strength, two opposite contributions to the linear-in-$T$ correction
237: to conductivity can in principle balance each other \cite{rem}.
238:
239: It is worth comparing the behavior of 2D electron system in Si
240: MOSFETs to that in another two-valley system, Si/SiGe quantum wells.
241: Transport properties of the latter system have been found to be very
242: similar to those of silicon MOSFETs
243: \cite{lai05,okamoto00,dolgopolov03}, although the disordered
244: potential in both cases is different resulting, particularly, from
245: the presence/absence of a spacer. However, the peculiarities near the
246: onset of complete spin polarization are less pronounced in Si/SiGe
247: quantum wells than in MOSFETs: only a weakening, but not absence, of
248: the temperature dependence of the resistance has been observed in the
249: metallic regime in a partially spin-polarized state \cite{lai05}.
250: Theoretically, the effect of the weakening of the $\sigma(T)$
251: dependence near the onset of complete spin polarization has been
252: found for the 2D electrons in Si/SiGe quantum wells in the frames of
253: screening approach \cite{hwang05}.
254:
255: \begin{figure}
256: \scalebox{0.42}{\includegraphics[clip]{fig4.eps}}
257: \caption{\label{fig4} The slope ratio as a function of electron
258: density for samples A (circles) and B (squares). The inset shows the
259: temperature dependence of the conductivity in both the fully
260: spin-polarized state for $n_s=1.85\times10^{11}$~cm$^{-2}$ and the
261: unpolarized state for $n_s=1.7\times10^{11}$~cm$^{-2}$ on sample A.
262: The dashed lines are fits of the linear interval of the dependence.}
263: \end{figure}
264:
265: We now compare the experimental ratio of the slopes
266: $A^*(B_{\text{sat}})/A^*(0)$ with theoretical predictions. As we have
267: already mentioned, in zero magnetic field, both the
268: temperature-dependent screening theory \cite{gold86} and the
269: interaction-based theory \cite{zala01} describe reasonably well the
270: available experimental data for $\sigma(T)$ in silicon MOSFETs
271: \cite{shashkin05,shashkin04a,sarma04}. For the fully spin-polarized
272: state, however, their predictions are very different. In theory
273: \cite{zala01}, the ratio $A^*(B_{\text{sat}})/A^*(0)$ (for a
274: two-valley 2D system) is formally equal to $(1+4F_0^\sigma)/(1+\alpha
275: F_0^\sigma)$ \cite{remark}, once the effective mass, as well as the
276: $g$ factor, are independent of the degree of spin polarization
277: \cite{kravchenko04,shashkin05}. Here the interaction parameter
278: $F_0^\sigma$ is responsible for the renormalization of the $g$ factor
279: through $g=2/(1+F_0^\sigma)$, the coefficient $\alpha=8$ if
280: $T<\Delta_v$ and $\alpha=16$ if $T>\Delta_v$, and $\Delta_v$ is the
281: valley splitting. For negative $F_0^\sigma$, the observed slope ratio
282: (Fig.~\ref{fig4}) cannot at all be attained within the approach
283: \cite{zala01}: based on the $B=0$ data for $F_0^\sigma$
284: \cite{shashkin02}, considerably smaller values of the slope ratio are
285: expected compared to the experiment. On the other hand, according to
286: the screening theory in its simple form (ignoring the local field
287: corrections), the ratio $A^*(B_{\text{sat}})/A^*(0)$ is equal to 0.5,
288: as inferred from doubling the Fermi energy due to the lifting of the
289: spin degeneracy \cite{dolgopolov00}. This value is close to the
290: experimental finding. The observed decrease of the slope ratio at low
291: electron densities is likely to be similar to the behavior of the
292: resistance ratio mentioned above (see Fig.~\ref{fig3}). Concerning
293: the data for Si/SiGe quantum wells \cite{lai05}, one can evaluate the
294: slope ratio for $n_s=0.515\times10^{11}$~cm$^{-2}$ at about 0.45,
295: which is consistent with our results.
296:
297: \section{CONCLUSION}
298:
299: In summary, we have found that in the ballistic regime, the metallic
300: temperature dependence of the conductivity in a two-dimensional
301: electron system in silicon changes non-monotonically with the degree
302: of spin polarization. It fades away just below the onset of complete
303: spin polarization but reappears again, being suppressed, in the fully
304: spin-polarized state. A qualitative account of the effect of the
305: disappearance of the $\sigma(T)$ dependence near the onset of
306: complete spin polarization is given. While in zero magnetic field
307: both the temperature-dependent screening theory and the
308: interaction-based theory provide a reasonably good description of
309: experimental data for the temperature-dependent conductivity in the
310: ballistic regime, the results obtained in the fully spin-polarized
311: state favor the screening theory in its simple form.
312:
313: \acknowledgments
314:
315: We gratefully acknowledge discussions with I.~L. Aleiner, A. Gold,
316: and V.~S. Khrapai. This work was supported by the RFBR, the Programme
317: ``The State Support of Leading Scientific Schools'', the National
318: Science Foundation grant DMR-0403026, and the ACS Petroleum Research
319: Fund grant 41867-AC10. EVD acknowledges the Russian Science Support
320: Foundation.
321:
322: \begin{thebibliography}{apssamp}
323: \bibitem{kravchenko04} S.~V. Kravchenko and M.~P. Sarachik, Rep.\ Prog.\ Phys.\ {\bf 67}, 1 (2004).
324: \bibitem{shashkin05} A.~A. Shashkin, Physics-Uspekhi {\bf 48}, 129 (2005).
325: \bibitem{tan05} Y.-W. Tan, J. Zhu, H.~L. Stormer, L.~N. Pfeiffer, K.~W. Baldwin, and K.~W. West, Phys.\ Rev.\ Lett.\ {\bf 94}, 016405 (2005) and references therein.
326: \bibitem{lai05} K. Lai, W. Pan, D.~C. Tsui, S.~A. Lyon, M. M\"uhlberger, and F. Sch\"affler, Phys.\ Rev.\ B\ {\bf 72}, 081313(R) (2005) and references therein.
327: \bibitem{shashkin01} A.~A. Shashkin, S.~V. Kravchenko, V.~T. Dolgopolov, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 87}, 086801 (2001); V.~M. Pudalov, M.~E. Gershenson, H. Kojima, N. Butch, E.~M. Dizhur, G. Brunthaler, A. Prinz, and G. Bauer, Phys.\ Rev.\ Lett.\ {\bf 88}, 196404 (2002); S.~V. Kravchenko, A.~A. Shashkin, and V.~T. Dolgopolov, Phys.\ Rev.\ Lett.\ {\bf 89}, 219701 (2002); A.~A. Shashkin, S. Anissimova, M.~R. Sakr, S.~V. Kravchenko, V.~T. Dolgopolov, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 96}, 036403 (2006).
328: \bibitem{vitkalov01} S.~A. Vitkalov, H. Zheng, K.~M. Mertes, M.~P. Sarachik, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 87}, 086401 (2001).
329: \bibitem{shashkin02} A.~A. Shashkin, S.~V. Kravchenko, V.~T. Dolgopolov, and T.~M. Klapwijk, Phys.\ Rev.\ B\ {\bf 66}, 073303 (2002).
330: \bibitem{shashkin03} A.~A. Shashkin, M. Rahimi, S. Anissimova, S.~V. Kravchenko, V.~T. Dolgopolov, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 91}, 046403 (2003).
331: \bibitem{anissimova06} S. Anissimova, A. Venkatesan, A.~A. Shashkin, M.~R. Sakr, S.~V. Kravchenko, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 96}, 046409 (2006).
332: \bibitem{simonian97} D. Simonian, S.~V. Kravchenko, M.~P. Sarachik, and V.~M. Pudalov, Phys.\ Rev.\ Lett.\ {\bf 79}, 2304 (1997).
333: \bibitem{yoon00} J. Yoon, C.~C. Li, D. Shahar, D.~C. Tsui, and M. Shayegan, Phys.\ Rev.\ Lett.\ {\bf 84}, 4421 (2000).
334: \bibitem{dolgopolov92} V.~T. Dolgopolov, G.~V. Kravchenko, A.~A. Shashkin, and S.~V. Kravchenko, JETP\ Lett.\ {\bf 55}, 733 (1992); A.~A. Shashkin, S.~V. Kravchenko, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 87}, 266402 (2001).
335: \bibitem{zala01} G. Zala, B.~N. Narozhny, and I.~L. Aleiner, Phys.\ Rev.\ B\ {\bf 64}, 214204 (2001).
336: \bibitem{okamoto00} T. Okamoto, K. Hosoya, S. Kawaji, A. Yagi, A. Yutani, and Y. Shiraki, Physica E {\bf 6}, 260 (2000); T. Okamoto, M. Ooya, K. Hosoya, and S. Kawaji, Phys.\ Rev.\ B\ {\bf 69}, 041202(R) (2004).
337: \bibitem{mertes01} K.~M. Mertes, H. Zheng, S.~A. Vitkalov, M.~P. Sarachik, and T.~M. Klapwijk, Phys.\ Rev.\ B {\bf 63}, 041101(R) (2001).
338: \bibitem{tsui05} Y. Tsui, S.~A. Vitkalov, M.~P. Sarachik, and T.~M. Klapwijk, Phys.\ Rev.\ B\ {\bf 71}, 113308(R) (2005).
339: \bibitem{gold86} A. Gold and V.~T. Dolgopolov, Phys.\ Rev.\ B\ {\bf 33}, 1076 (1986); S. Das Sarma, Phys.\ Rev.\ B\ {\bf 33}, R5401 (1986).
340: \bibitem{shashkin04a} A.~A. Shashkin, V.~T. Dolgopolov, and S.~V. Kravchenko, Phys.\ Rev.\ Lett.\ {\bf 93}, 269705 (2004).
341: \bibitem{dolgopolov00} V.~T. Dolgopolov and A. Gold, JETP\ Lett.\ {\bf 71}, 27 (2000).
342: \bibitem{okamoto99} T. Okamoto, K. Hosoya, S. Kawaji, and A. Yagi, Phys.\ Rev.\ Lett.\ {\bf 82}, 3875 (1999); S.~A. Vitkalov, H. Zheng, K.~M. Mertes, M.~P. Sarachik, and T.~M. Klapwijk, Phys.\ Rev.\ Lett.\ {\bf 85}, 2164 (2000).
343: \bibitem{rem} These arguments have been further developed by S. Das Sarma and E.~H. Hwang, Phys.\ Rev.\ B\ {\bf 72}, 205303 (2005).
344: \bibitem{dolgopolov03} V.~T. Dolgopolov, E.~V. Deviatov, A.~A. Shashkin, U. Wieser, U. Kunze, G. Abstreiter, and K. Brunner, Superlattices\ Microstruct.\ {\bf 33}, 271 (2003).
345: \bibitem{hwang05} E.~H. Hwang and S. Das Sarma, Phys.\ Rev.\ B\ {\bf 72}, 085455(R) (2005).
346: \bibitem{sarma04} S. Das Sarma and E.~H. Hwang, Phys.\ Rev.\ Lett.\ {\bf 93}, 269703 (2004).
347: \bibitem{remark} I.~L. Aleiner, private communication. Strictly
348: speaking, theory \cite{zala01} may be insufficient to have predictive
349: power for the fully spin-polarized state because theoretically, the
350: Fermi-liquid parameters are expected to depend on the degree of spin
351: polarization (G. Zala, B.~N. Narozhny, I.~L. Aleiner, and V.~I.
352: Fal'ko, Phys.\ Rev.\ B\ {\bf 69}, 075306 (2004)).
353: \end{thebibliography}
354:
355: \end{document}
356: