1: %% LyX 1.3 created this file. For more info, see http://www.lyx.org/.
2: %% Do not edit unless you really know what you are doing.
3: \documentclass[twocolumn,english,aps,prb,floatfix]{revtex4}
4: \usepackage[T1]{fontenc}
5: \usepackage[latin2]{inputenc}
6: \usepackage{amssymb}
7:
8: \makeatletter
9:
10: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% LyX specific LaTeX commands.
11: %% Bold symbol macro for standard LaTeX users
12: \newcommand{\boldsymbol}[1]{\mbox{\boldmath $#1$}}
13:
14:
15: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% User specified LaTeX commands.
16: \usepackage{graphicx}
17: \usepackage{bm}
18:
19: \usepackage{babel}
20: \makeatother
21: \begin{document}
22:
23: \title{Shot noise reduction in quantum wires with ''0.7 structure''}
24:
25:
26: \author{A. Ram\v{s}ak$^{1,2}$ and J. H. Jefferson$^{3}$}
27:
28:
29: \affiliation{$^{1}$Faculty of Mathematics and Physics, University of Ljubljana,
30: Slovenia}
31:
32:
33: \affiliation{$^{2}$J. Stefan Institute, Ljubljana, Slovenia,}
34:
35:
36: \affiliation{$^{3}$QinetiQ, Sensors and Electronics Division, St. Andrews Road,
37: Great Malvern, England}
38:
39: \begin{abstract}
40: Shot noise reduction in quantum wires is interpreted within the model
41: for the ''0.7 structure'' in the conductance of near perfect quantum
42: wires {[}T. Rejec, A. Ram{\v s}ak, and J. H. Jefferson, Phys. Rev.
43: B \textbf{62}, 12985 (2000){]}. It is shown how the Fano factor structure
44: is related to the specific structure of the conductance as a consequence
45: of the singlet--triplet nature of the resonances with the probability
46: ratio 1:3. An additional feature in the Fano factor, related to the
47: ''0.25 structure'' in conductance, is predicted.
48: \end{abstract}
49:
50: \pacs{73.23.-b, 73.63.-b}
51:
52: \maketitle
53: Conductance in various types of quantum wires and quantum point contacts
54: is quantized in units of $G_{0}=2e^{2}/h$. Since the first experimental
55: evidence for this effect\cite{wees88,wharam88} many subsequent experiments
56: have supported the idea of ballistic conductance in clean quantum
57: wires. However, certain anomalies remain, some of which are believed
58: to be related to electron-electron interactions and appear to be spin-dependent.
59: In the rising edge to the first conductance plateau, a structure appears
60: around $0.7G_{0}$, merging into the plateau at higher energies.\cite{thomas96}
61: Traces of such an anomaly are present already in the early measurements.\cite{wees88,patel91}
62: In quantum point contacts an additional structure appears around $0.25G_{0}$
63: with increasing source-drain bias \cite{thomas96} and this structure
64: is also seen at low bias in hard-confined quantum wires.\cite{patel91,philmag98,kaufman99,cronenphd,picciotto04}
65:
66: Under increasing magnetic field the $0.7$ structure moves down and
67: merges with the $\frac{1}{2}G_{0}$ plateau at very high fields.\cite{patel91,philmag98,kaufman99,cronenphd,picciotto04}
68: At elevated temperatures the structures are eventually washed out
69: but surprisingly this also happens at very low temperatures where
70: the disappearance of the 0.7 structure signals the formation of a
71: Kondo-like correlated spin state.\cite{cronenphd,cronenprl} Related
72: anomalies in thermopower measurements may also be explained though
73: a violation of the Mott-law emerges at very low temperatures. This
74: additionally suggests a many-body nature of electron transport in
75: ballistic quantum wires.\cite{appleyard}
76:
77: Recently the non-equilibrium current noise in a one-dimensional quantum
78: wire was measured\cite{will,roche} and this new class of measurements
79: opens up a range of possibilities for gaining new insight into the
80: problem of the transport anomalies in clean quantum wires. In particular,
81: careful measurements of the Fano factor in the expression for shot
82: noise for this system appears to be correlated with anomalies in conductance,
83: signaling different transmission probabilities for spin sub-channels.
84:
85: In this paper we present a theoretical explanation of the peculiar
86: dependence of the Fano factor on conductance and Fermi energy in ballistic
87: quantum wires. We follow the idea developed in the study of conductance
88: of a clean one-dimensional quantum wire with cylindrical\cite{rrj00}
89: or rectangular cross section\cite{jphys00} and a very weak bulge.
90: Several other scenarios for the 0.7 structure were recently proposed,
91: from a phenomenological model involving enhanced spin correlations,\cite{bruus01}
92: Kondo-like physics due to a localised moment,\cite{meir02} fluctuation
93: of local electron density\cite{sushkov03} to corrections due to the
94: backscattering of electrons due to phonons or Wigner crystal state
95: formation.\cite{matveev03} The relevance of these models to the description
96: of the 0.7 anomaly in shot noise measurements has not yet been investigated.
97:
98: Here we do not limit the investigation to a particular geometry of
99: the wire but consider the general case of a slightly imperfect quantum
100: wire. Such structures occur naturally, e.g. from a two-dimensional
101: electron gas (2DEG) in which a surface split-gate, which depletes
102: the 2DEG below it, gives rise to a quasi one-dimensional conducting
103: channel at low temperatures. Slight deviation from a perfect one-dimensional
104: confining potential, either accidental or deliberate, can give rise
105: to a localized potential well with single-electron bound states. This
106: occurs in a wire with a weak symmetric bulge\cite{rrj00} or in the
107: presence of remote gates, impurities or even electric polarization
108: due to the electron itself. The lowest bound state with energy $\epsilon_{b}<0$
109: relative to the bottom of the first conductance channel is shown schematically
110: in Fig.~1(a). Provided the confining potential is sufficiently weak,
111: only a single electron will be bound since the energy of the second
112: electron will be in the continuum due to Coulomb repulsion and the
113: system behaves like an open quantum dot. From the numerically exact
114: solution of two-electron scattering problem, one can extract the transmission
115: probabilities for particular spin configurations.\cite{rrj00} The
116: problem is analogous to treating the collision of an electron with
117: a hydrogen atom, e.g., as studied by J. R. Oppenheimer and N. F. Mott.\cite{oppenheimer}
118: This can be visualized as the scattering of the second electron in
119: an effective potential $V_{eff}$ arising from the combined effect
120: of the Coulomb repulsion from the first electron and the initial confining
121: potential {[}Fig.~1(b){]}. In the absence of a magnetic field the
122: appropriate spin sub-channels are singlet and triplet and, as illustrated
123: in Fig.~1(b), the two-electron system exhibits singlet and triplet
124: quasi-bound-state resonances in transmission. Summing over all electrons
125: in the leads gives the current,\cite{landauer,rrj00}
126:
127: \begin{equation}
128: I=\frac{2e}{h}\int(\frac{1}{4}T_{0}+\frac{3}{4}T_{1})(f_{L}-f_{R})d\epsilon,\label{eq:lbg}\end{equation}
129: where $T_{S}=T_{S}(\epsilon)$ is energy dependent singlet or triplet
130: transmission probability for $S=0$ and $S=1$, respectively. $f_{L,R}=\{1+\exp[(\epsilon-\mu_{L,R})/k_{B}T]\}^{-1}$
131: is the usual Fermi distribution function corresponding to left and
132: right lead, respectively, with temperature $T$ and the Boltzmann
133: constant $k_{B}$. In the linear regime, $\Delta\mu=\mu_{L}-\mu_{R}\to0$,
134: this reduces to a generalized Landauer-B\"{ u}ttiker formula\cite{landauer,rrj00}
135: for conductance, $G=eI/\Delta\mu$. The many-electron problem can
136: be mapped onto an extended Anderson model\cite{pecs,rrj03} for which
137: we have an \emph{open} quantum dot with Coulomb blockade except near
138: the resonances which are analogue to ''mixed-valence'' single-electron
139: tunneling regime. At very low temperatures the effects of Kondo physics
140: are expected and indeed signaled experimentally\cite{cronenprl} and
141: studied theoretically.\cite{meir02} However, at higher temperatures
142: these Kondo effects are suppressed and the extended Anderson model
143: yields a conductance in agreement with Eq.~(\ref{eq:lbg}).
144:
145: In accordance with the Lieb-Mattis theorem\cite{lieb} the singlet
146: resonance is always at lower energies than the triplet, $\epsilon_{0}<\epsilon_{1}$,
147: and consequently the quasi-bound state has longer lifetime (the resonance
148: is sharper) than the triplet. This is clearly seen from the results
149: obtained for the case of a cylindrical (or rectangular) quantum wire
150: with a symmetric bulge,\cite{rrj00,jphys00} presented in Fig~2(${\textrm{a}}_{1}$)
151: and Fig~2(${\textrm{b}}_{1}$).
152:
153: Recent high accuracy shot noise measurements enabled the extraction
154: of the Fano factor in ballistic quantum wires.\cite{roche} The Fano
155: factor $F$ is a convenient measure of the deviation from Poissonian
156: shot noise. It is the ratio of the actual shot noise and the Poisson
157: noise that would be measured in an independent-electron system.\cite{blanter}
158: This factor is, in our model,
159:
160: \begin{equation}
161: F=\frac{\int[T_{0}(1-T_{0})+3T_{1}(1-T_{1})](f_{L}-f_{R})^{2}d\epsilon}{\int(T_{0}+3T_{1})(f_{L}-f_{R})^{2}d\epsilon}.\label{eq:f}\end{equation}
162: This expression and Eq.~(\ref{eq:lbg}), are based on the results
163: of a two-electron scattering between a single bound electron and a
164: propagating conduction electron with a summation over all conduction
165: electrons near the Fermi energy. This approximation is only valid
166: at temperatures above the Kondo scale in this system,\cite{meir02}
167: as discussed in Ref.~\onlinecite{rrj03}. Eq.~(\ref{eq:f}) directly
168: reflects the fact that singlet and triplet modes do not mix in this
169: pairwise interaction approximation, resulting in contributions to
170: the noise that add incoherently with the probability ratio 1:3 for
171: singlet and triplet scattering.
172:
173: The conductance $G(\mu)$ and Fano factor $F(\mu)$ are plotted vs
174: $\mu$ in Fig.~2(${\textrm{a}}_{2}$) and Fig.~2(${\mathrm{b}}_{2}$),
175: with Fig.~2(${\textrm{a}}_{3}$) and Fig.~2(${\mathrm{b}}_{3}$)
176: showing the Fano factor $F$ vs $G$ for various temperatures and
177: in the linear response regime, $\mu=\mu_{L}\sim\mu_{R}$. The dotted
178: lines show zero temperature boundaries for the allowed values of $F$,
179: under the assumption of validity of Eq.~(\ref{eq:f}) in the limit
180: $T\to0$ and the unitarity condition for the transmission probabilities,
181: $0\leq T_{S}\leq1$. Eq.~(\ref{eq:f}) is not strictly valid in the
182: limit $T\to0$ due to many-electron effects which start becoming important
183: at low temperatures. Thus this zero-temperature limit should be regarded
184: as a limiting behavior that would occur in the absence of such many-body
185: effects. The Fano factor exhibits two distinctive features. Firstly,
186: there is a structure for $G/G_{0}<0.5$ corresponding to the sharp
187: $0.25$ singlet conductance anomaly. The second distinctive feature
188: is in the region $0.5<G/G_{0}<1$ and corresponds to the dip in singlet
189: channel just above the singlet resonance and also partially to the
190: triplet channel resonance. In our previous work we assumed a symmetric
191: confining potential fluctuation, giving perfect transmission probabilities
192: at resonance energies. However, in real systems left-right symmetry
193: will not be perfect, especially if the fluctuation is of random origin,
194: and also under finite source-drain bias. In these cases $T_{S}<1$
195: even on resonance. Such an example is presented in Fig.~2(${\textrm{c}}_{1-3}$).
196: In this case the structure of $F(G)$ is less pronounced, consisting
197: of kinks at $G/G_{0}\lesssim0.5$ and $G/G_{0}\lesssim0.75$. This
198: behavior is a consequence of the absence of a pronounced triplet resonance
199: and a dip in the singlet channel as mentioned above. Such a situation
200: is typical for very weak confining potential fluctuations, where the
201: triplet resonance is far in the continuum.
202:
203: The structure at $G/G_{0}\lesssim0.5$ has the same origin as the
204: ''0.25 structure'' in conductance, a direct consequence of a sharp
205: singlet resonance. In Fig.~1(c) we show a schematic representation
206: of the non-linear regime with finite source-drain voltage. In this
207: case the double peak potential barrier is asymmetric and shallower,
208: giving rise to broader singlet and triplet resonances. The triplet
209: resonance can even become over-damped while the singlet becomes more
210: robust to temperature as it broadens. Hence a pronounced ''0.25 structure''
211: in the conductance is expected, surviving to higher source-drain voltage
212: than the triplet (0.7 structure). This is indeed seen in experiments.\cite{patel91,philmag98,picciotto04}
213: If Eq.~(\ref{eq:f}) at least qualitatively holds also in this non-linear
214: regime, a distinctive feature should appear in the Fano factor, as
215: presented in Fig.~2(${\textrm{a}}_{3}$,${\mathrm{b}}_{3}$,${\textrm{c}}_{3}$).
216:
217: Thus far we have calculated conductance and Fano factor from singlet
218: and triplet resonances and found good semi-quantitative agreement
219: with experiment. In the regime of linear conductance and low temperature
220: Eq.~(\ref{eq:lbg}) and Eq.~(\ref{eq:f}) simplify, with $G$ and
221: $F$ determined by $T_{s}(\mu)$ taken at the Fermi energy. We can
222: then invert the procedure and use the experimentally determined conductance
223: and Fano factor to determine the transmission probabilities $T_{0}$
224: and $T_{1}$ using Eq.~(\ref{eq:lbg}) and Eq.~(\ref{eq:f}) in
225: this regime. Unitarity for $F$ requires that $F\leq1-G/G_{0}$ and
226: that $F$ is above some lower limit, $F_{min}(G)$, dotted line in
227: inset of Fig.~3(a). Unfortunately experimental values of $F$ rise
228: above this limit, possibly due to uncertain temperature corrections,
229: and therefore cannot be used to determine the probabilities unambiguously.
230: However, some estimates can be done. First we approximate $F(G)=\min(F_{I},1-G/G_{0}$),
231: where $F_{I}$ corresponds to the line connecting the experimental
232: points for $B=0$. Such a ''fit I'' is presented in the inset of
233: Fig.~3(a) (full line). Another choice is $F(G)=F_{II}$, where $F_{II}$
234: is some ''minimal assumption'' linear approximation for experimental
235: data at low $G$ and presented in Fig.~3(a) inset with a dashed line,
236: ''fit II''. Line-shaded areas between the two choices correspond to
237: the experimentally undetermined regime.
238:
239: In Fig.~3(a) are presented singlet and triplet transmission probabilities
240: extracted from such $F(G)$ and the corresponding experimental values
241: for $G(\mu)$ for $T=$515~mK. In spite of the uncertainty in $F$,
242: the structure of both, $T_{0}$ and $T_{1}$ are relatively well determined
243: and remarkably similar to the theoretically predicted cases from Fig.~2,
244: with a much larger triplet transmission probability at lower energies,
245: where the singlet is just above resonance and only slowly approaches
246: unity. A small resonance in the singlet transmission probability corresponds
247: to the kink structure in $F$ at $G\sim0.5G_{0}$. However, more accurate
248: measurements of $F$ are necessary in order to reduce the error bars
249: in the estimates of singlet and triplet transmission probabilities.
250: In Fig.~3(b) these probabilities are extracted for the case of the
251: lower temperature $T=$273~mK. The most striking observation is the
252: more rapid increase of the singlet transmission probability, a possible
253: signature of Kondo behavior.\cite{cronenprl,meir02}
254:
255: The experimentally measured Fano factor in a strong magnetic field
256: clearly suggests spin-up and spin-down structure of spin sub-channels.
257: Here the singlet--triplet concept is not relevant, $T_{\uparrow}$
258: and $T_{\downarrow}$ being the appropriate sub-channel division.
259: Eq.~(\ref{eq:f}) is therefore not valid in this limit. However,
260: the theoretical results for the conductance of near perfect quantum
261: wires in a magnetic field\cite{jphys00,rrj03} predict that, due to
262: the Zeeman sub-band splitting at finite magnetic field, only one spin
263: channel is open at lower energies and the conductance reaches $G\sim G_{0}$
264: only at higher energies, consistent with experiment. The corresponding
265: Fano factor then follows the unitarity limit for this case, $F\sim1-2G/G_{0}$
266: for $G<0.5G_{0}$ forming a bow with a maximum at $G\sim0.75G_{0}$
267: shown in Ref.~\onlinecite{roche}. This is qualitatively consistent
268: with the experimentally determined high field results for $F$, though
269: a more quantitative description of the transition between the two
270: regimes is still lacking.
271:
272: To conclude, we have shown that anomalous structures in shot noise
273: Fano factor measurements can be understood within the framework of
274: the theory of 0.7 conductance anomalies in near perfect quantum wires.\cite{rrj00}
275: The analysis of temperature dependence indicates stronger \emph{singlet}-channel
276: temperature dependence, which could be related to the Kondo-like behavior
277: at lower temperature.\cite{cronenprl,meir02} High magnetic field
278: measurements are in qualitative agreement with the results of the
279: theory. Finally, our results for a weak asymmetric confining potential
280: in an otherwise perfect quantum wire predict that in finite source-drain
281: voltage measurements a strong structure in Fano factor should appear
282: for $G\lesssim0.5G_{0}$. This structure corresponds to the recently
283: measured 0.25 conductance anomalies.\cite{picciotto04} Additional
284: refined measurements of conductance and the Fano factor could more
285: precisely resolve the singlet and triplet transmission probabilities
286: and test the predictions of the theory based on the singlet--triplet
287: resonant scattering.
288:
289: We thank M. B{\" u}ttiker for drawing our attention to the shot-noise
290: problem and for helpful suggestions. We thank also T. Rejec for comments
291: and B. Bu{\l}ka for stimulating discussions. Support is gratefully
292: acknowledged from the Ministry of Education and Science of Slovenia
293: under grant Pl-0044, the EU and the UK MoD.
294:
295: \begin{thebibliography}{10}
296: \bibitem{wees88}B.~J. van Wees \emph{et al}., Phys. Rev. Lett. \textbf{60}, 848 (1988).
297: \bibitem{wharam88}D.~A. Wharam \emph{et al}., J. Phys. C \textbf{21}, L209 (1988).
298: \bibitem{thomas96}K.~J. Thomas \emph{et al}., Phys. Rev. Lett. \textbf{77}, 135 (1996).
299: \bibitem{patel91}N. K. Patel, J. T. Nicholls, L. Martin-Moreno, M. Pepper, J. E. F.
300: Frost, D. A. Ritchie, and G. A. C. Jones, Phys. Rev. B \textbf{44},
301: 13549 (1991)
302: \bibitem{philmag98}K. J. Thomas \emph{et al}., Phil. Mag. B \textbf{77}, 1213 (1998).
303: \bibitem{kaufman99}D. Kaufman \emph{et al}., Phys. Rev. B \textbf{59}, R10433 (1999).
304: \bibitem{cronenphd}S. M. Cronenwett, Ph. D. Thesis (Stanford, 2001).
305: \bibitem{picciotto04}R. de Picciotto, L. N. Pfeiffer, K. W. Baldwin, and K. W. West, Phys.
306: Rev. Lett. \textbf{92}, 036805 (2004).
307: \bibitem{cronenprl}S. M. Cronenwett \emph{et al}., Phys. Rev. Lett. \textbf{88}, 226805
308: (2002).
309: \bibitem{appleyard}N. J. Appleyard, J. T. Nicholls, M. Pepper, W. R. Tribe, M. Y. Simmons,
310: and D. A. Ritchie, Phys. Rev. B \textbf{62}, R16275 (2000); T. Rejec,
311: A. Ram{\v s}ak, and J. H. Jefferson, Phys. Rev. B \textbf{65}, 235301
312: (2002).
313: \bibitem{will}W.~D.~Oliver, Ph.~D. Thesis (Stanford, 2002); N,~Y,~Kim \emph{et
314: al.,} cond-mat/0311435.
315: \bibitem{roche}P. Roche \emph{et al}., Phys. Rev. Lett. \textbf{93}, 116602 (2004).
316: \bibitem{rrj00}T. Rejec, A. Ram{\v s}ak, and J. H. Jefferson, Phys. Rev. B \textbf{62},
317: 12985 (2000).
318: \bibitem{jphys00}T. Rejec, A. Ram{\v s}ak, and J. H. Jefferson, J. Phys.: Condens.
319: Matter \textbf{12}, L233 (2000).
320: \bibitem{bruus01}H.~Bruus, V.~V.~Cheianov, and K. Flensberg, Physica E (Amsterdam)
321: \textbf{10}, 97 (2001).
322: \bibitem{meir02}Y. Meir, K. Hirose, and N. S. Wingreen, Phys. Rev. Lett. \textbf{89},
323: 196802 (2002).
324: \bibitem{sushkov03}O.~P.~Sushkov, Phys. Rev. B \textbf{67}, 195318 (2003); A.~A.~Starikov,
325: I.~I.~Yakimenko, and K.-F.~Berggren, Phys. Rev. B \textbf{67},
326: 235319 (2003).
327: \bibitem{matveev03}G.~Seelig and K.~A.~Matveev, Phys. Rev. Lett. \textbf{90}, 176804-1
328: (2003); K.~A.~Matveev, Phys. Rev. B \textbf{70}, 245319 (2004).
329: \bibitem{oppenheimer}J. R. Oppenheimer, Phys. Rev. \textbf{32}, 361 (1928); N. F. Mott,
330: Proc. Roy. Soc. A \textbf{126}, 259 (1930).
331: \bibitem{landauer}R. Landauer, IBM J. Res. Dev. \textbf{1}, 223 (1957); \emph{ibid.}
332: \textbf{32}, 306 (1988); M. B{\" u}ttiker, Phys. Rev. Lett. \textbf{57},
333: 1761 (1986).
334: \bibitem{pecs}T. Rejec, A. Ram{\v s}ak, and J. H. Jefferson, in \emph{Kondo Effect
335: and Dephasing in Low-Dimensional Metallic Systems}, eds.V. Chandrasekhar
336: \emph{et al.}, (Kluwer, Dordrecht, 2001); cond-mat/0007420.
337: \bibitem{rrj03}T. Rejec, A. Ram{\v s}ak, and J. H. Jefferson, Phys. Rev. B \textbf{67},
338: 075311 (2003).
339: \bibitem{lieb}Here we deal with quasi-bound states, which do not necesarilly respect
340: the Lieb-Mattis theorem {[}E. Lieb and D. Mattis, Phys. Rev. \textbf{125},
341: 164 (1962){]}. However, our comprehensive numerical analysis shows
342: that singlet resonance is always at lower energy than the triplet,
343: therefore supporting the theorem also for continuum two-body states
344: relevant for the present study.\cite{rrj03}
345: \bibitem{blanter}Ya. M. Blanter and M. B{\" u}ttiker, Phys. Rep. \textbf{336,} 1 (2000).
346: \end{thebibliography}
347: \widetext
348:
349: %
350: \begin{figure}[h]
351: \begin{center}\includegraphics[%
352: clip,
353: width=9cm,
354: keepaspectratio]{Fig1a.eps}\end{center}
355:
356: \begin{center}\includegraphics[%
357: clip,
358: width=9cm,
359: keepaspectratio]{Fig1b.eps}\end{center}
360:
361: \begin{center}\vskip -.5cm\includegraphics[%
362: clip,
363: width=9cm,
364: keepaspectratio]{Fig1c.eps}\end{center}
365:
366:
367: \caption{\label{cap:Fig1}(color online) (a) A weak negative potential fluctuation
368: binds one electron (electron density shown with dashed line) with
369: the energy below first channel minimum, $\epsilon_{b}<0$ (dotted
370: line) and chemical potential $\mu_{L}\sim\mu_{R}$. (b) Finite conduction
371: with electron energy, $\epsilon>0$, and linear regime, $\mu_{L}\gtrsim\mu_{R}$.
372: Two-electron scattering (quasi-bound) state is singlet or triplet.
373: $V_{eff}$ is an effective double barrier tunneling potential for
374: the scattered electron. (c) Larger source-drain voltage where the
375: triplet ($\epsilon_{1}$) resonance becomes over-damped with a broader,
376: more robust, singlet ($\epsilon_{0}$) becoming visible in transport.}
377: \end{figure}
378:
379:
380: %
381: \begin{figure}
382: \begin{center}\includegraphics[%
383: clip,
384: width=15cm,
385: keepaspectratio]{Fig2.eps}\end{center}
386:
387:
388: \caption{\label{cap:Fig2}(color online) Panels (${\textrm{a}}_{1}$) and
389: (${\textrm{b}}_{1}$): singlet and triplet transmission probabilities
390: as a function of Fermi energy $\mu$ measured relative to the bottom
391: of the first channel. Results for cylindrical quantum wires with a
392: symmetric bulge for all parameters as in Fig.~3(b) and Fig.~3(c)
393: of Ref.~\onlinecite{rrj00}. In panels (${\textrm{a}}_{2}$) and
394: (${\textrm{b}}_{2}$) the conductance (full lines) and the Fano factor
395: are presented. Panels (${\textrm{a}}_{3}$) and (${\textrm{b}}_{3}$):
396: Fano factor as a function of $G$ (full line) and unitarity limits
397: (dotted line). (${\textrm{c}}_{1}$) transmission probabilities as
398: would arise, e.g., for a left-right asymmetric confining potential
399: (in resonance $T_{S}<1$). The corresponding energy dependence of
400: the Fano factor and conductance, (${\textrm{c}}_{2}$), and Fano factor
401: as a function of conductance, (${\textrm{c}}_{3}$). Energy and temperature
402: scale is here in the units of single-triplet energy difference $J$.}
403: \end{figure}
404:
405:
406: %
407: \begin{figure}
408: \begin{center}\includegraphics[%
409: clip,
410: width=12cm,
411: keepaspectratio]{Fig3a.eps}\end{center}
412:
413: \begin{center}\includegraphics[%
414: clip,
415: width=12cm,
416: keepaspectratio]{Fig3b.eps}\end{center}
417:
418:
419: \caption{\label{cap:Fig3}(color online) (a) Inset: experimental values from
420: Ref.~\onlinecite{roche} for $F(G)$ for $B=0$ (bullets) and two
421: different interpolating (fitting) forms of $F$. Main figure: conductance
422: for $T=515$ mK and $B=0$ (dashed line), singlet (full line) and
423: triplet (dotted) transmission probabilities extracted from experimental
424: values of $G(\mu)$ vs $\mu$ and using two interpolating forms for
425: $F$. Shaded area corresponds to the shaded area in the inset. (b)
426: As in (a), for $T=273$ mK.}
427: \end{figure}
428:
429: \end{document}
430: