1: %\documentstyle{article}
2: %\documentstyle[preprint,aps]{revtex}
3: %\def\mathcal{\cal}
4: \documentstyle[prb,aps,multicol,epsfig]{revtex}
5: %\input psfig
6: \def \be{\begin{equation}}
7: \def \ee{\end{equation}}
8: \def \r{{\bf r}}
9: \def \k{{\bf k}}
10: \def \o{\omega}
11: \def \p{{\bf p}}
12: \def \ve{\varepsilon}
13: \def \q{{\bf q}}
14: \def \e{\epsilon}
15: \def \dzeta{\zeta}
16: \def \H{{\cal H}}
17: \def \K{G^{(K)}}
18: \def \R{G^{(R)}}
19: \def \A{G^{(A)}}
20: \def \t{\tau_{imp}}
21: \def \x{{\bf x}}
22: \def \D{{\cal D}}
23: \def \C{{\cal C}}
24: \def \Nch{{N_{\rm ch}}}
25: \def \varQ{{\overline{Q^2}}}
26: \def \mls{\delta_1}
27: \renewcommand{\narrowtext}{\begin{multicols}{2} \global\columnwidth20.5pc}
28: \renewcommand{\widetext}{\end{multicols} \global\columnwidth42.5pc}
29: \newcommand{\Lrule}{\vspace*{-0.2in}\noindent\vrule width3.5in height.2pt
30: depth.2pt \vrule depth0em height1em}
31: \newcommand{\Rrule}{\vspace{-0.1in}\hfill\vrule depth1em height0pt \vrule
32: width3.5in height.2pt depth.2pt\vspace*{-0.1in}}
33:
34:
35: \begin{document}%*******************************************************
36:
37: \bibliographystyle{simpl1}
38:
39: \title{Conductance Fluctuations of Open Quantum Dots under Microwave
40: Radiation.}
41: \date{\today}
42:
43: \author{Maxim~G.~Vavilov$^a$ and Igor
44: L.~Aleiner$^b$}
45:
46: \address{ $^a$ Laboratory of Atomic and Solid State
47: Physics, Cornell University, Ithaca, NY 14853\\ $^b$ Department of
48: Physics and Astronomy, SUNY at Stony Brook, Stony Brook, NY 11794 }
49: \maketitle
50: \begin{abstract}
51: We develop a time dependent random matrix theory describing the
52: influence of a time-dependent perturbation on mesoscopic conductance
53: fluctuations in open quantum dots. The effect of external field is
54: taken into account to all orders of perturbation theory, and our
55: results are applicable to both weak and strong fields. We obtain
56: temperature and magnetic field dependences of conductance
57: fluctuations. The amplitude of
58: conductance fluctuations is determined by electron temperature in
59: the leads rather than by the width of electron distribution function
60: in the dot. The asymmetry of conductance with respect to inversion of
61: applied magnetic field is the main feature allowing to distinguish
62: the effect of direct suppression of quantum interference from the simple
63: heating if the frequency of external radiation is larger than
64: the temperature of the leads $\hbar\omega \gg T$.
65: \end{abstract}
66: \draft
67: \pacs{PACS numbers: 73.23.Ad, 72.15.Rn, 72.70.+m}
68:
69: \narrowtext
70:
71: {\em Introduction} --
72: Transport coefficients of disordered and chaotic
73: electron systems fluctuate from sample to
74: sample \cite{A1,LS,A,B,AK,H1,Huibers}.
75: These fluctuations are commonly called mesoscopic
76: fluctuations. Mesoscopic fluctuations of conductance of
77: non-interacting systems are universal. The universality means
78: that the variance of the conductance $\langle \delta g^2\rangle$ is of
79: the order $G_0^2$, where $G_0=e^2/\pi\hbar$ is the quantum of
80: conductance and is weakly dependent of the sample geometry, see \cite{B,AK}.
81:
82: The fluctuations of transport properties of electron systems is a
83: quantum mechanical phenomenon based on the interference of quantum
84: states. As any other interference phenomena, conductance fluctuations
85: are very sensitive to inelastic processes, commonly referred to as
86: dephasing \cite{AAKL}. The dephasing processes in open quantum dots
87: were considered on the purely phenomenological basis \cite{BM1}.
88: First microscopic consideration of the microwave radiation on the weak
89: localization in quantum dots was performed in Ref.~\cite{VA}.
90: In this reference, the concept of the time dependent random matrix
91: theory (TRMT) was used.
92:
93: The purpose of the present paper is to extend the results of
94: Ref.~\cite{VA} to describe the effect of the external microwave
95: radiation on the mesoscopic fluctuations of the conductance. The
96: ultimate goal is to identify observable features which allow one to
97: distinguish the effect of the external field to the dot itself from a
98: simple heating \cite{Huibers}.
99:
100:
101: However, there is a significant difference in the calculation of the
102: mesoscopic conductance fluctuations and the averaged conductance
103: \cite{VA}. Because, the dot is subjected to the external classical
104: radiation which produces non-equilibrium in the dot, the
105: d.c.-current $I_0$ through the dot is finite (though randomly changing
106: from one configuration to another) even if the {\em d.c.}-voltage
107: $V=V_{\rm l}-V_{\rm r}$ across the
108: dot is zero, see Fig.~1.
109: This current $I_0$ is due either to the photovoltaic effect (see
110: \cite{VAA} and references therein) or to rectification of a.c.-bias across the
111: dot, see \cite{Brouwer-ac}. We are interested in the linear response to
112: the applied {\em d.c.}-voltage across the dot:
113: \begin{equation}
114: \label{16}
115: I_{\rm dc} = I_{0}+gV+{\cal O}(V^2).
116: \end{equation}
117: In principal, the linear in $V$ contribution to the current comes
118: comes from two sources:
119: ({\it i}) the non-equilibrium of the distribution functions in the
120: leads, ({\it ii}) change in the photovoltaic current,
121: correspondent to a different realization of the dot due to the finite
122: bias.
123: Nevertheless, we will show that due to the electro-neutrality condition
124: the non-equilibrium current prevails \cite{Y}.
125:
126: Closing the introductory part, we note that the
127: present paper has a certain overlap with the recent preprint by Wang
128: and Kravtsov \cite{WK}, where the conductance fluctuations were
129: calculated for open quantum dots subjected to a periodic $ac$
130: pumping. Our treatment is different in several aspects. Firstly, our
131: results are applicable for the frequencies of the external radiation
132: $\omega$ smaller than the Thouless energy of the dot, $E_T$, whereas
133: treatment of Ref.~\cite{WK} is valid in the opposite regime.
134: Secondly, unlike Ref.~\cite{WK}, we will restrict ourselves to the
135: case of the monochromatic radiation acting on the dot. Finally, we
136: will highlight the role of the electro-neutrality requirement in a
137: separability of the photovoltaic effect and the mesoscopic conductance
138: fluctuations, which was not done in Ref.~\cite{WK}.
139:
140:
141:
142: {\em Model}---
143: We apply the random matrix theory (RMT)
144: to study the conductance of open quantum dots, see Ref.~\cite{B}. All
145: corrections to the RMT are governed by a small parameter
146: $N_{\rm ch}/g_{\rm dot}$, where
147: $g_{\rm dot}=E_{\rm T}/\delta_1$ and $\delta_1$ is
148: the mean level spacing, see Ref.~\cite{ABG} for the detailed discussion.
149: We consider the conductance fluctuations of quantum
150: dots with a large number of open
151: channels $N_{\rm ch}\gg 1$. In this approximation, we neglect the effects
152: of the electron--electron interaction on the conductance which are as
153: small
154: as $1/N_{\rm ch}^2$\cite{ABG,Brouwer}, while the
155: conductance fluctuations are proportional
156: to $1/N_{\rm ch}$. The same condition also allows us to use a
157: conventional diagrammatic technique\cite{AGD} to
158: take the ensemble average. External microwave radiation is
159: modeled as time dependent random part of the Hamiltonian of the dot.
160:
161:
162:
163:
164: \begin{figure}
165: \centerline{\psfig{figure=figure1.eps,width=7cm}}
166: {
167: \caption{
168: An open quantum dot is connected to two leads with applied voltages
169: $V_{\rm l,r}$. The measured current through the
170: dot has an offset $I_0$ at zero bias and the linear
171: response to small applied voltage $V=V_{\rm l}-V_{\rm r}$. }
172: }
173: \end{figure}
174:
175:
176:
177:
178: The Hamiltonian of the system is
179: \cite{ABG}:
180: \be
181: \label{1}
182: {\cal H}(t)={\cal H}_{\rm D}(t)+{\cal H}_{\rm L}+{\cal H}_{\rm LD},
183: \ee
184: where $\hat H_{\rm D}$ is the Hamiltonian of the electrons in the dot,
185: which is determined by the $M\times M$ matrix $H_{nm}$:
186: %\begin{mathletters}
187: \be
188: \label{2}
189: {\cal H}_{\rm D}(t)=\sum_{n,m=1}^M \psi^\dag_n H_{{\rm D}nm}(t) \psi_m
190: +E_c{n}^2,
191: \ee
192: $\psi_n$ corresponds to the states of the dot and
193: the thermodynamic limit $M \to\infty$ is assumed, $E_c$ is the
194: charging energy, and ${n}=\sum_{m=1}^M \psi^\dag_m \psi_m$,
195: the last term in Eq.~(\ref{2}) is the largest contribution to
196: the interaction effects in quantum dot, see Ref.~\cite{ABG}
197: for the discussion of the status of this approximation.
198: Matrix $\hat H_{\rm D}(t)$ is given by
199: \be
200: \label{03}
201: \hat H_{\rm D}(t)=\hat H+\hat V \varphi(t).
202: \ee
203: Here the time independent part of the Hamiltonian $\hat H$ is a
204: random realization of a $M\times M$ matrix, which obeys the correlation
205: function
206: \begin{eqnarray}
207: \label{04}
208: &&\langle H_{nm}(\Phi_1) H^*_{n'm'}(\Phi_2)\rangle=
209: M\left(\frac{\mls}{\pi}\right)^2 \\
210: &&\quad
211: \times
212: \left[\left(1-\frac{N_{\rm d}}{4M}\right)
213: \delta_{nn'}\delta_{mm'}
214: +\left(1-\frac{N_{\rm c}}{4M}\right)\delta_{mn'}\delta_{nm'}\right],
215: \nonumber
216: \end{eqnarray}
217: where $\mls$ is the mean level spacing of the dot and
218: parameters $N_{d,c}$ describe the effect of the magnetic field on the
219: dot\cite{ABG}.
220: These parameters can be estimated as $N_{\rm d,c} \simeq g_{\rm dot}
221: \left(\Phi_1\mp\Phi_2\right)^2/\Phi_0^2$
222: where $\Phi_{1,2}$ is the magnetic flux through the dot and $\Phi_0=hc/e$ is
223: the flux quantum. The time dependent perturbation is described by
224: symmetric $M\times M$ matrices $V_{nm}$ and function $\varphi(t)$
225: of time $t$.
226: We assume that the perturbation is harmonic with single frequency
227: $\omega$, $\varphi (t)=\cos \omega t$, even though most of the
228: consideration [up to Eq.~(\ref{24})] is valid for an arbitrary function.
229: The effect of the perturbation on
230: the system is totally determined by two parameters, \cite{fn1}
231: \be
232: Z= \frac{1}{M}{\rm Tr} \hat{V}, \quad
233: C_0 = \frac{\pi}{M^2\mls}{\rm Tr} \hat{V}^2.
234: \label{05}
235: \ee
236: Parameter $Z$ has a meaning of the average velocity of the energy
237: levels of the dot under the external perturbation and can be omitted
238: from our consideration due to screening [see below]. The
239: parameter $C_0$ characterizes its typical deviation, \cite{SA}.
240: Since all the physical responses of the system are characterized by
241: the same parameters, the value of $C_0$ can be eliminated by an independent
242: measurement.
243:
244:
245: The electron spectrum in the leads near Fermi
246: surface can be linearized:
247: \be
248: \label{9}
249: {\cal H}_L=\hbar v_{\rm F}\sum_{\alpha, k} k \psi^\dag_\alpha(k)\psi_\alpha(k),
250: \ee
251: where $\psi_{\alpha}(k)$ denotes different electron states in the
252: leads, $k$ labels the continuum of momentum states in each
253: channel $\alpha$, $\hbar v_{\rm F}=1/2\pi \nu$ is the
254: Fermi velocity and $\nu$ is the density of states per channel at the
255: Fermi surface. We put $\hbar =1$ in all intermediate formulas below.
256:
257: The coupling between the dot and the leads is
258: \be
259: \label{7}
260: {\cal H}_{\rm LD}=\sum_{\alpha, n, k}\left( W_{n \alpha}\psi^\dag_\alpha
261: (k)
262: \psi_n+
263: {\rm H.c.}\right).
264: \ee
265: For the reflectionless point contacts,
266: the coupling constants, $W_{n \alpha}$,
267: in Eq.(\ref{7}) are given by \cite{B,ABG}:
268: \be
269: \label{8}
270: W_{n \alpha}=\cases{\displaystyle
271: \sqrt{\frac{M\mls}{\pi^2\nu}}, & if $n=\alpha\leq N_{\rm ch}$,\cr
272: 0,& otherwise.
273: }
274: \ee
275:
276:
277: For open dots with a large number of open channels $N_{\rm ch}\gg 1$
278: the interaction term can be
279: treated within mean field approximation, so that the Hamiltonian
280: (\ref{2}) takes the form
281: \begin{mathletters}
282: \begin{eqnarray}
283: {\cal H}_{\rm D}^{\rm mf}(t)& = & \sum_{n,m=1}^M \psi^\dag_n
284: \left[H_{{\rm D}nm}(t)+ e V_d(t) \delta_{nm}\right]
285: \psi_m,
286: \label{99}
287: \\
288: e V_d & = & 2 E_c\langle {n} \rangle_{q},
289: \label{100}
290: \end{eqnarray}
291: \end{mathletters}
292: where $\langle {n} \rangle_{q}$ stands for the quantum mechanical
293: (but not ensemble) of the number of electrons in the dot.
294: Corrections to mean-field treatment (\ref{99}) were calculated in
295: Ref.~\cite{Brouwer} and shown to be small as $1/N_{\rm ch}^2$.
296:
297: In the mean field approximation (\ref{99}), one can introduce one-particle
298: $S$ - matrix ${\cal S}_{\alpha\beta}(t_1,t_2)$ as
299: \begin{eqnarray}
300: \label{13}
301: \label{S}
302: {\cal S}_{\alpha\beta}(t,t')=
303: \delta_{\alpha\beta}\delta(t-t')-2\pi i\nu
304: W^\dag_{\alpha n} G^{({\rm R})}_{nm}(t,t') W_{m\beta},
305: \end{eqnarray}
306: and the Green functions $\hat G^{({\rm R,A})}(t,t')$ are the solutions of:
307: \begin{eqnarray}
308: \nonumber
309: & &
310: \left(i\frac{\partial}{\partial t} - {\hat H}_{\rm D}(t)-eV_{\rm d}(t)
311: \pm i\pi\nu \hat{W}\hat{W}^\dag \right)\hat{G}^{({\rm R,A})}(t,t')
312: \\
313: & & = \delta(t-t'),
314: \label{14}
315: \end{eqnarray}
316: where the matrices $\hat{H}_{\rm D}$ and $\hat{W}$ are
317: defined by Eqs.~(\ref{03}) and (\ref{8}).
318:
319: The d.c.-current through the dot is given by, see \cite{VAA}:
320: \begin{eqnarray}
321: I_{\rm dc} & = & {e}\int_0^{T_{\rm p}}\!\!
322: \frac{dt}{T_{\rm p}} \int\! dt_1dt_2
323: {\rm Tr}\left\{ \hat f(t_1-t_2)
324: \right.
325: \nonumber
326: \\
327: & \times &
328: \left.
329: \left(
330: \hat{\cal S}^{\dagger}(t_2,t) \hat \Lambda
331: \hat{\cal S}(t,t_1) - \hat \Lambda \delta(t_2-t_1)
332: \right)
333: \right\}
334: \label{10}
335: \end{eqnarray}
336: % the other form of the equation.
337: %\begin{eqnarray}
338: %I_{\rm dc} & = & {e}\int_0^{T_{\rm p}}\!\!
339: %\frac{dt}{T_{\rm p}}
340: %{\rm Tr}\left[\hat\Lambda
341: %\int\! dt_1dt_2
342: %\hat{\cal S}(t,t_1)\hat f(t_1-t_2)
343: %\hat{\cal S}^{\dagger}(t_2,t)\right]
344: %\nonumber \\
345: %&-&
346: %e{\rm Tr} \hat\Lambda\hat f(+0),
347: %\label{10}
348: %\end{eqnarray}
349: where $T_p$ is the period of the external perturbation,
350: $\hat{f}(t)$ is related to the
351: Fourier transform of the electron distribution
352: function in the $\alpha$th channel as
353: \be
354: \label{11}
355: f_{\alpha\beta}(t)=\delta_{\alpha\beta}
356: \frac{i Te^{-ieV_\alpha t} }{\sinh\pi T t}
357: \ee
358: and
359: \begin{equation}
360: \label{12}
361: \Lambda_{\alpha\beta}=\delta_{\alpha\beta}\cases{\displaystyle
362: \frac{N_{\rm r}}{N_{\rm ch}}, &
363: for $1\leq \alpha\leq N_{\rm l}$; \cr
364: \displaystyle -\frac{N_{\rm l}}{N_{\rm ch}}, &
365: for $N_{\rm l}+1\leq \alpha\leq N_{\rm ch}$. \cr
366: }
367: \end{equation}
368: The spin degeneracy is taken into account in Eq.~(\ref{10}). We assume
369: that the degeneracy is not lifted by magnetic field.
370:
371: To complete the theory one needs an equation for the averaged
372: number of particles $\langle {n} \rangle_{q}$,
373: see Eq.~(\ref{100}). It is found from the continuity
374: relation as
375: \begin{eqnarray}
376: \frac{d \langle {n}(t) \rangle_{q}}{dt} &=&
377: - \int\! dt_1dt_2 {\rm Tr}\left\{ \hat f(t_1-t_2) \right.
378: \label{110}
379: \\
380: & \times &
381: \left.
382: \left(
383: \hat{\cal S}^{\dagger}(t_2,t) \hat{\cal S}(t,t_1) - \delta(t_2-t_1)
384: \right)
385: \right\}
386: \nonumber
387: \end{eqnarray}
388:
389:
390: Equations (\ref{10}) --- (\ref{110}) are similar to those used in
391: Ref.~\cite{Buttikker} for studying the frequency dependence of the
392: conductance of mesoscopic system.
393:
394:
395: {\em Ensemble averaging} ---
396: Our goal now is to perform calculations of the conductance correlation
397: function
398: \be
399: \label{R}
400: {\cal R}(\Phi_1, \Phi_2)=\langle g(\Phi_1) g(\Phi_2)\rangle - \langle
401: g(\Phi_1) \rangle\langle g(\Phi_2)\rangle,
402: \ee
403: using the model outlined above.
404:
405: \begin{figure}
406: \centerline{\psfig{figure=figure2.eps,width=7cm}}
407: {
408: \caption{
409: Two diagrams, which contribute to the conductance correlation function
410: $R$. }}
411: \end{figure}
412:
413: We use the leading approximation in small parameter $1/N_{\rm ch}$.
414: The fluctuations of the conductance are smaller than its average value
415: and we can use instead of sample specific Eq.~(\ref{110}) its ensemble
416: averaged counterpart:
417: \begin{eqnarray}
418: &&\frac{d \langle {n} (t) \rangle_{q}}{dt}
419: = - \frac{\Nch\mls}{2\pi}\langle {n}(t) \rangle_{q}
420: \label{110a} \\
421: && + \frac{eN_{\rm l}}{\pi}\left(V_{\rm l} - V_{\rm d} -
422: Z\varphi(t)\right)
423: + \frac{eN_{\rm r}}{\pi}\left(V_{\rm r} - V_{\rm d} - Z\varphi(t)\right)
424: .
425: \nonumber
426: \end{eqnarray}
427: Equation (\ref{110a}) is nothing but a discrete form of the
428: diffusion equation for the bulk system and the last two terms correspond to
429: the divergence of the drift current. Substituting Eq.~(\ref{100})
430: into Eq.~(\ref{110a}), solving the resulting differential equation,
431: we find
432: \begin{eqnarray}
433: e V_{\rm d}(t) + Z\varphi(t) &=&
434: \frac{4e E_{\rm c}}{4 E_{\rm c} + \mls}
435: \frac{N_{\rm l} V_{\rm l}+N_{\rm r} V_{\rm r}}
436: {\Nch}
437: \label{200}\\
438: &+& Z
439: \frac{\mls+ (2\pi/N_{\rm ch}) \partial_t}
440: {\mls+4 E_{\rm c} + (2\pi/N_{\rm ch})\partial_t }\varphi(t).
441: \nonumber
442: \end{eqnarray}
443: We notice from Eq.~(\ref{200}) that the characteristic energy
444: scale governing charge dynamics is $E_{\rm c}N_{\rm ch}/2\pi$. Usually,
445: this scale is of the order of the Thouless energy, $E_T$. Because
446: all the random matrix theory is capable to describe the energy scale
447: only smaller than $E_T$, we can consider only $\omega \ll
448: E_T \simeq E_{\rm c}N_{\rm ch}/2\pi$. Moreover, for the small quantum
449: dot $E_{\rm c} \gg \delta_1$, so that Eq.~(\ref{200}) gives
450: \be
451: V_{\rm d} = \frac{N_{\rm l} V_{\rm l}+N_{\rm r} V_{\rm r}}{\Nch}
452: \label{Vd}
453: \ee
454: and the time dependent perturbation (\ref{05})
455: can be considered as traceless, $Z=0$.
456: This constant in time component of the bias of the dot
457: can be
458: removed from Eq.~(\ref{14}) for the Green function
459: by the following gauge transformation:
460: \begin{equation}
461: \label{gaugetransform}
462: \label{202}
463: \hat G(t,t')=\left.
464: \hat G\right|_{V_{\rm d}=0}(t,t') e^{- i e V_{\rm d}(t-t')}.
465: \end{equation}
466:
467: Substituting Eq.~(\ref{202}) into Eq.~(\ref{10}) and
468: expanding up to the first power in $V=V_{\rm l}-V_{\rm r}$,
469: we find
470: \begin{eqnarray}
471: \label{17}
472: g = \frac{\partial I_{\rm dc}}{\partial V} = g_{\rm cl} & +&
473: \displaystyle
474: G_0
475: \int\limits_0^{T_{\rm p}}\!\!\! dt\!\! \int\limits_{-\infty}^{+\infty}
476: \!\!\!dt_1dt_2
477: F(T(t_1-t_2))
478: \\
479: \nonumber
480: &\times & {\rm Tr}
481: \Big{\{}
482: \hat {\cal S}(t,t_1)\hat \Lambda {\cal S}^\dagger(t_2,t)\hat \Lambda
483: \Big{\}}.
484: \end{eqnarray}
485: Here
486: \[
487: g_{\rm cl}= G_0 \frac{N_{\rm l}N_{\rm r}}{N_{\rm ch}}
488: \]
489: is the classical conductance of the dot, $G_0=e^2/\pi\hbar$ is the quantum
490: conductance and $F(x)$ is the Fourier transform of the derivative of
491: electron distribution function:
492: \begin{equation}
493: \label{18}
494: F(x)=
495: \frac{\pi x}{\sinh\pi x}.
496: \end{equation}
497:
498:
499:
500: The correlation function of the conductance
501: (\ref{R}) is given by the diagrams
502: shown in Fig. 2 and can be found from the following analytical
503: expression
504: %\end{multicols}
505: \widetext
506: \Lrule
507: \begin{eqnarray}
508: \label{19}
509: R & =&2 \frac{g_{cl}^2\mls^2}{4\pi^2}
510: \int\limits_0^{T_{\rm p}}\!\!
511: \frac{dtdt'}{T_{\rm p}^2}\int\limits_0^{\infty}\!\!
512: d\tau F^2(T\tau)
513: \int\limits_{\tau/2}^{\infty}\!\!d\theta
514: \left[
515: \D\left(\frac{t+t'}{2}+\theta,\frac{t+t'}{2}-\frac{\tau}{2},t-t' \right)
516: \D\left(\frac{t+t'}{2}+\theta,\frac{t+t'}{2}+\frac{\tau}{2},t-t'\right)
517: \right.
518: \nonumber
519: \\
520: & &+ \left.
521: \C \left(
522: t-t'+\theta+\frac{\tau}{2},t-t'-\theta-\frac{\tau}{2},\frac{t+t'}{2}+
523: \frac{\tau}{4}
524: \right)
525: \C
526: \left(
527: t'-t+\theta-\frac{\tau}{2},t'-t-\theta+\frac{\tau}{2},\frac{t+t'}{2}-
528: \frac{\tau}{4}
529: \right)
530: \right].
531: \end{eqnarray}
532: \Rrule
533: \narrowtext
534: %\begin{multicols}{2}
535: The diffuson and the Cooperon are defined by the following equations:
536: \begin{mathletters}
537: \begin{eqnarray}
538: \label{20}
539: {\cal C}(\tau_1,\tau_2,t) & = &
540: \Theta(\tau_1-\tau_2)
541: \exp\left(-\frac{1}{2}\int_{\tau_2}^{\tau_1}
542: K_{\rm c}(\tau,t)
543: d\tau \right), \\
544: \label{21}
545: {\cal D}(t_1,t_2,\tau) & = &\Theta(t_1-t_2) \exp\left(-\int_{t_2}^{t_1}
546: K_{\rm d}(\tau,t) d t \right),
547: \end{eqnarray}
548: \end{mathletters}
549: where
550: \begin{mathletters}
551: \begin{eqnarray}
552: \label{22}
553: K_{\rm d,c} & = & \gamma_{\rm d,c} +
554: C_0\left(\varphi(t+\tau/2)-\varphi(t-\tau/2)\right)^2
555: \\
556: \displaystyle
557: \gamma_{\rm d, c} & = & \frac{\mls}{2\pi}
558: \left( N_{\rm ch}+N_{\rm d,c}\right),
559: \label{22.b}
560: \end{eqnarray}
561: \end{mathletters}
562: and parameters $N_{\rm d,c}$ describe the effect of the magnetic field,
563: see Eq.~(\ref{04}).
564:
565: \begin{figure}
566: \centerline{\psfig{figure=figure3.eps,width=7cm }}
567: {
568: \caption{
569: The elements of the diagrammatic technique. }}
570: \end{figure}
571:
572:
573: Derivation of Eq.~(\ref{22.b}) deserves a little bit of additional
574: discussion. One notices, that the diagrams of Fig.~3 does not contain
575: any piece corresponding to the classical distribution function in the
576: dot. We can trace it into the expression for conductance (\ref{17})
577: which contains traceless vertices $\hat \Lambda$, which can not be
578: dressed by the dashed line. On the other hand, any vertex with
579: finite trace corresponds the modified distribution function of
580: electrons in the dot and represents the effect of heating ~\cite{VAA}.
581: Since the distribution function is not dressed in the expression for
582: conductance fluctuations, we conclude, that the effect of heating is
583: not relevant for the conductance fluctuations and the temperature
584: dependence of the conductance fluctuations is uniquely determined by
585: {\em the electron temperature in the leads\cite{Y}}. That means that
586: contrary to the common believes, see e.g. Ref.~\cite{Huibers} the
587: amplitude of the mesoscopic fluctuations can {\em not} be used for the
588: study of the distribution function of electrons in the dot. From the
589: theoretical side, it is important to emphasize, that the appearance of the
590: traceless vertices is determined solely by the electro-neutrality
591: condition (\ref{Vd}), any other choice of the dot bias would lead to
592: the change in the photovoltaic current ~\cite{VAA}.
593:
594:
595:
596: {\em Limiting cases} ---
597: Below we consider the limit of high ($\hbar\omega\gg C $) and low
598: ($\hbar\omega\ll \gamma_{\rm d,c}$) frequencies.
599: For the high frequencies, $\omega\gg C $, we
600: obtain
601: \begin{equation}
602: \label{24}
603: R = \frac{\mls^2 g_{\rm cl}^2}{4\pi^2}\left(\frac{1}{\gamma_{\rm d}^2}
604: Q_{\rm d}
605: \left(
606: \frac{C_0}{\gamma_{\rm d}},
607: \frac{T}{\gamma_{\rm d}}
608: \right)
609: +
610: \frac{1}{\gamma_{\rm c}^2} Q_{\rm c}
611: \left(
612: \frac{C_0}{\gamma_{\rm c}},
613: \frac{T}{\gamma_{\rm c}},
614: \frac{\hbar\omega}{\gamma_{\rm c}}
615: \right)
616: \right),
617: \end{equation}
618: where dimensionless $Q-$ functions are given by
619: \begin{mathletters}
620: \begin{eqnarray}
621: \label{25}
622: Q_{\rm d}(x,y) & = &
623: \int\limits_0^\infty d\tau F^2(y\tau) \\
624: &\times &
625: \int\limits_0^{2\pi} \frac{\exp\left(-\tau
626: \left(1+2x\sin^2\zeta/2\right)\right)}{1+2x\sin^2\zeta/2}\frac{d\zeta}{2\pi},
627: \nonumber
628: \\
629: \label{26}
630: Q_{\rm c}(x,y,z) & = & \int\limits_0^\infty d\tau F^2(y\tau) \\
631: & \times &
632: \int\limits_0^{2\pi}
633: \frac{\exp\left(-\tau \left(1+2x\sin^2\zeta/2\right)\right)}
634: {1+x(\sin^2\zeta/2+\sin^2(\zeta/2+z\tau/2 ))}\frac{d\zeta}{2\pi}.
635: \nonumber
636: \end{eqnarray}
637: \end{mathletters}
638:
639: Let us now consider the dependence of the functions $Q_{\rm d,c}$ and
640: $Q_{\rm c}(x,y,z)$ from Eq.~(\ref{26})
641: on temperature $y$. For the limit of high
642: temperature, $y \gg 1$, we obtain
643: \begin{equation}
644: \label{27}
645: Q_{\rm c}(x,y,z)\approx Q_{\rm d}(x,y)\approx
646: \frac{\pi^2}{3 y}\frac{1}{\sqrt{1+2x}}.
647: \end{equation}
648: The equality between functions $Q_{\rm c}$ and $Q_{\rm d}$ means
649: the magnetic field symmetry of the conductance \cite{Onsager}. Indeed, using
650: Eqs.(\ref{22}), (\ref{24})
651: and (\ref{27}) we observe, that
652: \be
653: R (\Phi_1,\Phi_2) =
654: R(\Phi_1,-\Phi_2).
655: \label{ons1}
656: \ee
657: However, in low temperature limit $y \ll 1$, we obtain for $x\gg 1$
658: \begin{mathletters}
659: \begin{eqnarray}
660: \label{28}
661: Q_{\rm d}(x,0)=\frac{1+x}{(1+2x)^{3/2}} & \approx & \frac{1}{2\sqrt{2x}}.\\
662: \label{28a}
663: Q_{\rm c}(x,0,z)& \approx & \frac{1}{2x}.
664: \end{eqnarray}
665: \end{mathletters}
666:
667: \begin{figure}
668: \centerline{\psfig{figure=figure4.eps,width=8cm,height=7cm}}
669: \caption{
670: Functions $Q_c(x,y,z)$ and $Q_d(x,y)$ computed for $x=1$ and
671: $z=10$. As temperature $y=T/\gamma$ increases, function $Q_{\rm c}(x,y,z)$
672: approaches frequency independent function $Q_{\rm d}(x,y)$. }
673: \end{figure}
674:
675:
676: Comparison of Eqs.(\ref{22}), (\ref{24}) and (\ref{28}) reveals
677: an important fingerprint of the dephasing by the external radiation
678: --- violation of the
679: Onsager relation
680: \be
681: \frac{R (\Phi_1,-\Phi_2)}{R(\Phi_1,\Phi_2)}
682: =
683: \sqrt{\frac{2\tilde \gamma}{C_0}}
684: ,
685: \label{ons2}
686: \ee
687: where $\tilde\gamma=\gamma_{\rm d}(\Phi_1,\Phi_2)=
688: \gamma_{\rm c}(\Phi_1,-\Phi_2)$, provided that
689: $\gamma_{\rm c}(\Phi_1,\Phi_2)=\gamma_{\rm d}(\Phi_1,-\Phi_2)\gg
690: \tilde \gamma$.
691: This breakdown of the Onsager relation is a simple manifestation of
692: the lifting of the time reversal symmetry in the system with time dependent
693: Hamiltonian.
694:
695: In the limit of low frequency $\hbar\omega\ll \gamma_{\rm d, c}$, the
696: contribution from the Cooperon and diffuson parts are described by the
697: same function, so that the conductance correlation function can be
698: represented in the form
699: \begin{equation}
700: \label{29}
701: R= \frac{\mls^2 g_{\rm cl}^2}{4\pi^2}\left[\frac{1}{\gamma_{\rm d}^2}
702: Q\left(\frac{C_0}{\gamma_{\rm d}}, \frac{T}{\gamma_{\rm d}}\right)
703: +\frac{1}{\gamma_{\rm c}^2}
704: Q\left(\frac{C_0}{\gamma_{\rm c}}, \frac{T}{\gamma_{\rm c}}\right)
705: \right],
706: \end{equation}
707: so the Onsager relation (\ref{ons1}) holds.
708: Here,
709: \begin{eqnarray}
710: Q(x,y)& =& \int_0^{2\pi}\frac{d\xi d\zeta}{4\pi^2}\int_0^{\infty}
711: F^2(y\tau)
712: \\
713: \nonumber
714: & \times &
715: \frac{\exp(-(1+4x\sin^2\xi/2\sin^2\zeta/2)\tau)}
716: {1+4x\sin^2\xi/2\sin^2\zeta/2} d\tau.
717: \end{eqnarray}
718: This expression in the limit of high temperature $T\gg
719: \gamma_{\rm d, c}$ has an asymptotic behavior
720: \begin{equation}
721: \label{30}
722: Q(x,y)=\frac{\pi}{3 y}K\left(-4x\right).
723: \end{equation}
724: At zero temperature $Q(x,y)$ is given by the expression
725: \begin{equation}
726: \label{31}
727: Q(x,0)=\frac{1}{\pi}\frac{E\left(-4x\right)+(1+4x)K(-4x)}{1+4x},
728: \end{equation}
729: where $K(x)$ and $E(x)$ are the elliptic integrals of the first and
730: second kind respectively
731: \begin{eqnarray*}
732: %\label{32}
733: K(x) & = & \int_0^{\pi/2}\frac{d\varphi}{\sqrt{1-x\sin^2\varphi}}\\
734: %\label{33}
735: E(x) & = & \int_0^{\pi/2}\sqrt{1-x\sin^2\varphi}d\varphi .
736: \end{eqnarray*}
737:
738:
739:
740:
741: We conclude that the conductance fluctuations are suppressed
742: by external radiation
743: even in the limit of the low frequency, see Eq.~(\ref{29}).
744: Indeed, during one period of time, the system
745: goes along a closed loop in the parameter space
746: and the contribution to the d.c.-conductance
747: is effectively determined by the equilibrium conductance, correspondent
748: to each point of the loop.
749: The equilibrium conductance fluctuates along this loop.
750: Thus, the observed d.c.-conductance is already partially
751: averaged over some realizations of the quantum dot and its
752: fluctuations decrease. The perturbation strength is related to the
753: length of the contour in the parameter space and effectively determines
754: how many different dot's configurations contribute to the
755: d.c.-conductance.
756: Consequently, the stronger perturbation, over the larger number
757: of the realizations the d.c.-conductance is averaged and the smaller
758: fluctuations of the d.c.-conductance.
759:
760: This should be contrasted with the suppression of the averaged
761: magnetoresistance \cite{AAKL,VA}. There, the stationary field does not
762: do anything because the result is already ensemble averaged. In order
763: to suppress the average quantum correction, the field should have
764: change on the time scale of the order of $1/\gamma_{\rm esc}$, where
765: $\gamma_{\rm esc}=\mls N_{\rm ch}/2\pi$ is the escape rate from the
766: dot. That is why
767: the effect of the low-frequency radiation on conductance fluctuations
768: and weak localization corrections are significantly different.
769: At high frequency $\hbar\omega\gg \gamma_{\rm esc}$ the d.c.-conductance no
770: longer can be represented in terms of the stationary conductance and
771: the suppression of both the conductance fluctuations and the weak
772: localization correction to the conductivity can be interpreted as
773: dephasing.
774:
775:
776:
777:
778: {\em Comparison with experiment} ---
779: Our results still contains an unknown parameter $C_0$ characterizing
780: the strength of the perturbation. There is a way, however, to present
781: the results in a form not depending on this parameter, thus
782: eliminating a need for additional fitting.
783: Following Ref.~\cite{Huibers}, we represent the parametric dependence
784: of the weak localization correction $\delta g_{\rm wl}$ versus
785: ${\rm var}\ \ g$, where $\delta g_{\rm wl}$ is given by\cite{VA}:
786: \begin{eqnarray}
787: \label{39}
788: \delta g_{\rm wl} & = &-\frac{e^2}{\pi\hbar}
789: \frac{N_{\rm l}N_{\rm r}}{(N_{\rm l}+N_{\rm r})^2}
790: P\left(\frac{C_0}{\gamma_{\rm esc}},
791: \frac{\hbar\omega}{\gamma_{\rm esc}}\right)
792: \\
793: \label{40}
794: P(x,z) & = & \int_0^\infty e^{-\xi-x\phi}I_0[x\phi] d\xi, \ \ \
795: \phi=\xi-\frac{\sin z\xi}{z} .
796: \end{eqnarray}
797: The conductance variance is determined from
798: Eq.~(\ref{19}) for broken time-reversal
799: symmetry $\gamma_{\rm c} \to \infty$.
800: Figure~5 shows the parametric dependence for various values
801: of the parameters $C_0$ and $\omega$ and $T= 10\gamma_{\rm esc}$.
802:
803:
804: We observe that the shape of the curves depends on the frequency of
805: external radiation. Particularly, in the limit of low frequency
806: $\hbar\omega\ll \gamma_{\rm esp}$ the weak localization correction is not
807: changed by the radiation, while the conductance fluctuations may be
808: significantly suppressed. At high frequency, $\hbar\omega \gg C_0,\
809: \gamma_{\rm esc}$ the curves become non-sensitive to the radiation frequency.
810:
811: The authors of Ref.~\cite{Huibers} found that the
812: radiation applied to their device produced curves in ${\rm
813: var} g$ vs. $\delta g_{\rm wl}$ plane identical to the curve
814: produced by increasing temperature of the device for a wide
815: range of frequencies. This observation apparently
816: demonstrates that the radiation produces the heating of electrons and
817: the effect of dephasing without heating, see Ref.~\cite{AGA},
818: is not observed in experiments\cite{Huibers}.
819: It was also suggested that the main mechanism is
820: the increase of the temperature in the dot due to the Joule heat by
821: induced $ac$ source-drain bias.
822:
823: \begin{figure}
824: \centerline{\psfig{figure=figure5.eps,width=7.5cm}}
825: \caption{
826: Weak localization correction $\delta g_{\rm wl}$ versus conductance
827: fluctuations ${\rm var} \ g$ of an open quantum dot with
828: $N_{\rm l}=N_{\rm r}\gg 1$ for three values of frequency $\omega$:
829: $\hbar\omega =0.5\gamma_{\rm esc}$ (o);
830: $\hbar\omega=5\gamma_{\rm esc}$ ($\bigtriangledown$);
831: $\hbar\omega=50 \gamma_{\rm esc}$ ($\bigtriangleup$).
832: The temperature for all lines was taken $T=10 \gamma_{\rm esc}$.
833: The amplitude of the field $C_0$ varies from $10^{-2}\gamma_{\rm esc}$
834: to $10^2\gamma_{\rm esc}$. $G_0=e^2/\pi\hbar$ is the quantum conductance.
835: }
836: \end{figure}
837:
838:
839: Although the present data of Ref.~\cite{Huibers} support the heating
840: mechanism of suppression of the weak localization correction to the
841: conductance and the conductance fluctuations, we believe that more
842: detailed analysis has to be done. According to our theory, see
843: Eq.~(\ref{17}) and the paragraph below Eq.~(\ref{22.b}),
844: ({\em i}) mesoscopic fluctuations are
845: sensitive only to the temperature in the leads, and therefore, the notion
846: of heating of electrons in the dot responsible for $1/T$ dependence of
847: the mesoscopic fluctuations is not relevant: if there is a heating,
848: it manifests itself only through {\em the temperature of the leads};
849: ({\em ii}) high frequency
850: curves of our theoretical Fig.~5 quantitatively agrees with data on
851: Fig.~3 of Ref.~\cite{Huibers}, for frequencies $f=1$, $10$ and $25$ GHz.
852: An exception is the lowest frequency curve ($f=100$ MHz)
853: represented in this plot, for
854: the dot with $\delta_1 = 2.4\mu eV$, $\Nch=2$ corresponds to
855: $h f/\gamma_{esc}\approx 0.5 $, so according to our Fig.~5 it
856: should have observable deviations from the high frequency curves,
857: which is not seen. However, taking into account uncertainty in
858: determination of the level spacing $\mls$ from the geometrical
859: area of the dot, this does not unambiguously rule out the microwave
860: dephasing mechanism.
861:
862: We believe that the ``smoking gun'' evidence for the mechanism considered
863: in the present paper is the violation of the Onsager relation (\ref{ons2})
864: in high frequency regime, $\hbar\omega \gg \gamma_{\rm c,d},\ T$.
865: The dependence
866: of this violation on the amplitude of the field $C_0$ is the main
867: prediction of the theory.
868:
869: {\em Summary}---
870: We constructed the time dependent random matrix theory to describe the
871: effect of the non-equilibrium external radiation on conductance
872: fluctuations of an open quantum dot. The main experimental feature to
873: reveal such a mechanism is the breakdown of the magnetic symmetry of
874: the conductance by high-frequency radiation, $\hbar\omega \gg T$, see
875: Eq.~(\ref{ons2}).
876:
877: We thank V. Ambegaokar, P. W. Brouwer, V. E. Kravtsov and C. M. Marcus
878: for useful discussions. This work was supported by the
879: Cornell Center for Materials Research,
880: under NSF grant no. DMR 0079992 (M.G.V.) and Packard Foundation
881: Fellowship (I.L.A.).
882:
883:
884:
885:
886:
887:
888:
889: \begin{references}
890:
891: \bibitem{A1} B.L. Altshuler, Pis'ma Zh. Eksp. Teor. Fiz. {\bf 51}, 530
892: (1985) [JETP Lett. {\bf 41}, 648 (1985)].
893:
894: \bibitem{LS} P.~A.~Lee and A.~D.~Stone, Phys. Rev. Lett. {\bf 55},
895: 1622 (1985).
896:
897:
898: \bibitem{A} B.~L.~Altshuler and B.~I.~Shklovskii, Sov. Phys. JETP,
899: {\bf 64}, 127, (1986).
900:
901: \bibitem{B} C.W.J. Beenakker, Rev. Mod. Phys, {\bf 69}, 731, (1997).
902:
903: \bibitem{AK} B. L. Altshuler and D. E. Khmelnitskii, JETP Lett. {\bf
904: 42}, 359 (1985).
905: Also see P. A. Lee,
906: Fukuyama, A.D Stone Phys. Rev. B {\bf 35}, 1039 (1985).
907:
908:
909: \bibitem{H1} A.~G.~Huibers {\it et. al.} Phys. Rev. Lett. {\bf 81},
910: 200 (1998).
911:
912: \bibitem{Huibers} A.~G.~Huibers, J.~A.~Folk, S.~R.~Patel,
913: C.~M.~Marcus, C.~I.~Duru\"oz and J.~S.~Harris, Phys.~Rev.~Lett.,
914: {\bf 83}, 5090, (1999).
915:
916: \bibitem{AAKL} B.L. Altshuler, {\em et. al.}, in
917: {\it Quantum Theory of Solids}, (Mir, Moscow, 1982).
918:
919: \bibitem{BM1}
920: H.U.~Baranger, P.A.~Mello, Phys. Rev. B, {\bf 51}, 4703, (1995);
921: P.W.~Brouwer and C.W.J.~Beenakker,
922: {\em ibid}., {\bf 51}, 7739, (1995);
923: {\em ibid}., {\bf 55}, 4695, (1997);
924: I.L.~Aleiner and A.I.~Larkin. {\em ibid}., {\bf 54},
925: 14423 (1996);
926: E. McCann and I.V. Lerner, {\em ibid}.,
927: {\bf 57}, 7219 (1998).
928:
929: \bibitem{VA} M.G.~Vavilov and I.L.~Aleiner, Phys. Rev. B, {\bf 60},
930: R16311, (1999).
931:
932: \bibitem{VAA} M.~G.~Vavilov, V.~Ambegaokar, and I.~L.~Aleiner,
933: cond-mat/0008469.
934:
935: \bibitem{Brouwer-ac} P.~W.~Brouwer, Phys. Rev. B {\bf 63}, 121303 (2001).
936:
937: \bibitem{Y} The same conclusion for the conductance fluctuations of
938: open quantum dots was reached by V.~I.~Yudson,
939: E.~Kanzieper, and V.~E.~Kravtsov, cond-mat/0012200. We do not quite
940: understand reasoning of this paper, since the electroneutrality
941: condition is not invoked there. Nevertheless, the answer is
942: correct.
943:
944: \bibitem{WK} X.-B. Wang and V.~E.~Kravtsov, cond-mat/0008193.
945:
946: \bibitem{ABG} I. L. Aleiner, P. W. Brouwer, and L. I. Glazman,
947: cond-mat/0103008.
948:
949: \bibitem{Brouwer} P.W. Brouwer and I.L. Aleiner, Phys. Rev. Lett.,
950: {\bf 82}, 390 (1999).
951:
952: \bibitem{AGD} A. A.~Abrikosov, L. P.~Gorkov, I. E.~Dzyaloshinskii,
953: {\it Methods
954: of Quantum Field Theory in Statistical Physics}, (Prentice--Hall, Englewood
955: Cliffs, NJ, 1963).
956:
957: \bibitem{fn1} Note that $C_0$ is defined in ref.~\cite{VA} with a
958: different numerical factor.
959:
960: \bibitem{SA} B.D.~Simons and B.L.~Altshuler, Phys. Rev. Lett., {\bf 70}, 4063,
961: (1993); B.L.~Altshuler and B.D.~Simons, in {\it Mesoscopic Quantum Physics},
962: eds. E.~Akkermans {\it et. al}, (Elsevier, 1995).
963:
964: \bibitem{Buttikker} M. B\" uttikker, H. Thomas, and A. Pretre, Z. Phys. B
965: {\bf 94}, 133 (1994).
966:
967: \bibitem{Onsager} L. Onsager, Phys. Rev. {\bf 38}, 2265 (1931);
968: M. B\" uttikker, Phys. Rev. Lett. {\bf 57}, 1761 (1986).
969:
970: \bibitem{AGA} B.~L.~Altshuler, M.E.~Gershenzon, and I.L.~Aleiner,
971: Physica E {\bf 3}, 58 (1999).
972:
973: \end{references}
974: \widetext
975:
976: \end{document}
977:
978:
979:
980:
981:
982:
983:
984:
985:
986:
987:
988:
989:
990:
991:
992:
993:
994:
995:
996:
997:
998:
999:
1000:
1001:
1002:
1003:
1004:
1005:
1006:
1007:
1008:
1009:
1010:
1011:
1012:
1013:
1014:
1015:
1016:
1017:
1018:
1019:
1020:
1021:
1022:
1023:
1024:
1025:
1026:
1027:
1028:
1029:
1030:
1031:
1032:
1033:
1034:
1035:
1036:
1037:
1038:
1039:
1040:
1041:
1042:
1043:
1044:
1045:
1046:
1047:
1048:
1049:
1050:
1051:
1052:
1053:
1054:
1055:
1056:
1057:
1058:
1059:
1060:
1061:
1062:
1063:
1064:
1065:
1066:
1067:
1068:
1069:
1070:
1071:
1072:
1073:
1074:
1075:
1076:
1077:
1078:
1079:
1080:
1081:
1082:
1083:
1084:
1085:
1086:
1087:
1088:
1089:
1090:
1091:
1092:
1093:
1094:
1095:
1096:
1097:
1098:
1099:
1100:
1101:
1102:
1103:
1104:
1105:
1106:
1107:
1108:
1109:
1110:
1111:
1112:
1113:
1114:
1115:
1116: