cond-mat0610296/cm.tex
1: \documentclass[twocolumn,amsmath,
2:   amssymb,english,aps,prb,floatfix,showpacs]{revtex4}
3: \usepackage[english]{babel}
4: \usepackage{times}
5: \usepackage{graphicx}
6: \makeatletter
7: \makeatother
8: \def \mb{\begin{displaymath}} 
9: \def \me{\end{displaymath}} 
10: \def \eb{\begin{equation}} 
11: \def \ee{\end{equation}}   
12: \def\expect#1{\mathinner{\langle{#1}\rangle}}
13: \def\bra#1{\left<#1\right|}
14: \def\ket#1{\left|#1\right>}
15: {\catcode`\|=\active 
16:   \gdef\expect#1{\left<\mathcode`\|"8000\let|\bravert {#1}\right>}}
17: \def\bravert{\egroup\,\vrule\,\bgroup}
18: \def\braket#1#2{\left<#1|#2\right>}
19: \def\Tr{\mathrm{Tr}}
20: \def\dag{\dagger}
21: 
22: 
23: \begin{document}
24: 
25: \title{Conductance of a molecule with a center of mass motion}
26: \author{J. Mravlje$^{1}$, A. Ram\v{s}ak$^{2,1}$, and T. Rejec$^{1,2,3}$}
27: \affiliation{$^{1}$Jo\v{z}ef Stefan Institute, Ljubljana, Slovenia}
28: \affiliation{$^{2}$Faculty of Mathematics and Physics, University of Ljubljana,
29: Slovenia}
30: \affiliation{$^{3}$Department of Physics, Ben-Gurion University, Beer-Sheva,
31: Israel}
32: 
33: \begin{abstract}
34: We calculate the zero temperature conductance and characteristic
35: correlation functions of a molecule with a center of mass (CM)
36: motion which modulates couplings to the leads. In the first model studied,
37: the CM vibrational mode is 
38: simultaneously coupled to the electron density on the molecule. The
39: conductance is suppressed in regimes corresponding
40: to non-integer occupancy of the molecule.  In the second model, where
41: the CM mode  is not directly coupled to the electron density, the
42: suppression of conductance is related to the dynamic breaking of the inversion symmetry.
43: \end{abstract}
44: 
45: \pacs{72.15.Qm,73.23.-b,73.22.-f}
46: %71.27.+a       Strongly correlated electron systems; heavy fermions
47: %73.23.-b       Electronic transport in mesoscopic systems
48: %73.23.Hk       Coulomb blockade; single-electron tunneling
49: %72.15.Qm       Scattering mechanisms and Kondo effect (see also
50: %               75.20.Hr local moments in compounds and alloys; Kondo effect, valence
51: %               fluctuations, heavy fermions in magnetic properties and materials) 
52: %73.22.-f       Electronic structure of nanoscale materials: clusters,
53: %               nanoparticles, nanotubes, and nanocrystals
54: 
55: \maketitle
56: %
57: \section{INTRODUCTION}
58: Interest in systems with electron-phonon interaction
59: has recently reemerged, as a consequence of advances in 
60: experimental techniques enabling the investigation of the subtle interplay of
61: electronic and vibrational degrees of freedom by means of measuring
62: the conductance of the
63: molecules \cite{madhavan98, park00,park02,
64:   liang02,zhitenev02,yu04,yu04_2,pasupathy05,zhao05,zhao05_2,
65:   wahl05,yu05,natelson06}. % In particular, the influence of the
66: %center-of-mass vibrations has recently recieved some theoretical
67: %attention\cite{alhassanieh05,balseiro06}.
68: Consequently,
69: a considerable amount of work was done in pursuit of understanding the equilibrium
70:  \cite{cornaglia04, cornaglia05a, cornaglia05b,mravlje05} and
71: non-equilibrium \cite{mitra04, paaske05, takahiro05} properties of 
72: these systems. The analysis of
73: low-temperature properties of such systems is challenging as strong
74: repulsive interaction among electrons
75: confined to molecular orbitals leads to surprising results such as the
76: interplay of the Kondo physics and
77: molecular vibrations, which are intractable to conventional
78: perturbational approaches. Investigation of the 
79: molecules which, when attached to the leads,
80: may posses both internal and center-of-mass vibrational
81: modes, makes the calculation of various properties even more
82: involved. 
83: 
84: Here we consider a molecule undergoing a linear transport measurement -- a 
85: system which consists of a molecule and
86: attached leads. The shape of the molecule and its position with
87: respect to the leads oscillate (the setup is sketched in 
88: Fig.\ref{Fig1}). Hence the orbital level of the molecule -- for
89: transport through 
90: molecules it is sufficient to consider a
91: single orbital due to wide inter-level spacings \cite{yu04_2} -- and the
92: tunneling amplitude are both modulated.
93: Here we treat two possibilities: (i) in case I (proposed in
94: Ref.~\onlinecite{alhassanieh05}) the vibrational mode
95: that modulates the tunneling amplitude 
96: is also coupled to the electron density at the molecule; (ii) 
97: in case II (treated also in Ref.~\onlinecite
98: {balseiro06})   the
99: molecule possesses in  addition to the CM mode a breathing
100: mode which modulates the energy level of the molecular orbital.
101: 
102: Experimentally, CM  modes are discerned in
103: nonlinear transport measurements as side-bands in conductance which
104: correspond to frequencies which do 
105: not match any of the eigenmodes of a given molecule. 
106: While the relevance of the separate couplings to two vibrational modes
107: in model~II is
108: intuitively clear, the relevance of the simultaneous modulation of
109: tunneling and coupling to the electron density in model~I is
110: less clear and deserves further exploration. We  note
111: that the model~I could be relevant for the nonlinear transport
112: measurement where the electric 
113: field would directly couple to the electron density at the molecule,
114: or in the presence of some impurities with residual electrical fields
115: even in the (near) linear transport measurements.
116: 
117: The case of a molecule without the CM oscillations described by the
118: Anderson-Holstein Hamiltonian has already been analyzed by several
119: authors (see, \emph{e.g.} Ref.~\onlinecite{mravlje05} and references
120: therein). The dominant effect of the coupling to the
121: breathing mode is the reduction of the repulsive
122: interaction among electron pairs occupying the molecular orbital. When
123: the reduced repulsion $U_\mathrm{eff}$ is positive, a gate-voltage
124: sweep reveals the area (of width of order $U_\mathrm{eff}$) of enhanced
125: linear conductance due to the Kondo effect. With increased electron-phonon coupling
126: and thereby decreased $U_\mathrm{eff}$ this
127: area is diminished. In this paper we show that the CM
128: oscillations suppress the conductance in a new way which cannot be
129: described by an effective model with reduced repulsion. 
130: 
131: This paper is organised as follows. Section II introduces two related
132: electron-phonon models. In Sec. III numerical methods based on the variational wave function are
133: presented and some symmetry properties of the models are considered. In Sec. IV results
134: for linear conductance for both models are presented and some ground state properties of the 
135: molecular system are explained. Results are summarised in Sec. V.
136: \begin{figure}
137: \begin{center}\includegraphics[%
138:   width=70mm,
139:   keepaspectratio]{Fig1.eps}\end{center}
140: 
141: 
142: \caption{\label{Fig1} (Color online) Schematic  plot of the
143:   model devices. In model I the CM vibrational mode which modulates the
144:   tunneling amplitude  $t'$ is coupled also to the
145:   local charge density $n$. In model II the CM displacement modulates
146:   only the tunneling amplitude whereas another (breathing) vibrational mode is
147:   coupled to the local charge density.} 
148: \end{figure}
149: 
150: 
151: \section{Models}
152: Hamiltonians for both models consist of the electron, the phonon and the electron-phonon
153: coupling terms, $H=H_\mathrm{e}+H_\mathrm{p}+H_\mathrm{e-p}$. In the 
154: electronic part of the Hamiltonian,
155: \begin{equation}
156: H_{\mathrm{e}}=H_\mathrm{L}+H_\mathrm{R}+\epsilon_{d}n_{d}+Un_{d\uparrow}n_{d\downarrow}-t'(\widehat{v}_\mathrm{L}
157: + \widehat{v}_\mathrm{R}),
158: \label{eq:he}\end{equation}
159: $H_{\mathrm{L,R}}=-t\sum_{i\in\mathrm{L,R}}c_i^\dagger c_{i+1} +
160:   h.c.$ describe the noninteracting tight-binding left and right leads, respectively, and $\epsilon_d$ is the energy level of the molecular orbital
161:   occupied by $n=\langle n_d\rangle$ electrons where $n_{d}=\sum_{\sigma}n_{d\sigma}$ with
162: $n_{d\sigma}=d_{\sigma}^{\dagger}d_{\sigma}$. The characteristic scale
163:   of the repulsive interaction is $U$ and the tunneling part with 
164:   characteristic tunneling rate $t'$ 
165:   consists of operators
166:   $\widehat{v}_l=\sum_\sigma c_{1l\sigma}^\dagger d_\sigma + h.c.$ for $l=\mathrm{L,R}$ which
167:   couple the molecular orbital to the first site in the left and right
168:   leads, respectively. In $H_\mathrm{e}$ only the coupling between the molecular
169: orbital and even combinations of the left and the right lead orbitals (i.e., the symmetric channel) is present. 
170: 
171: 
172: Vibrational modes are described with $H^\mathrm{I}_{\mathrm{p}}=\Omega_1
173: a_1^\dag a_1$ and $H^\mathrm{II}_{\mathrm{p}}=\Omega_1 a_1^\dag a_1 + \Omega_2 a_2^\dag
174: a_2$ for models I and II, respectively, and  $a_{1,2}^\dagger$ are the
175: phonon creation operators. The corresponding electron-phonon interaction is given by
176: \begin{eqnarray} \mathrm{I:}\quad H^\mathrm{I}_{\mathrm{e-p}}=t'g
177: (\widehat{v}_\mathrm{L}-\widehat{v}_\mathrm{R})x_1 + M (n_d-1)x_1   \:{} \nonumber \\
178:  \mathrm{II:} \quad H^\mathrm{II}_{\mathrm{e-p}}= t'g
179: (\widehat{v}_\mathrm{L}-\widehat{v}_\mathrm{R})x_1+M (n_d-1)x_2 , \end{eqnarray} where
180: $x_{1,2}=a_{1,2}^\dag+a_{1,2}$ are the displacement operators. 
181: \section{METHODS}
182: 
183: The ground state properties are determined using
184: the Gunnarsson and Sch\"{o}nhammer
185: projection-operator method \cite{schonhammer76, gunnarsson85, mravlje05}. Here the Hamiltonian is diagonalized
186: in the basis
187: \begin{equation}
188: |\Psi_{\lambda\left\{ \! m_{\alpha}\!\right\}}\rangle=P_{\lambda}\prod_{\alpha}a_{\alpha}^{\dagger m_{\alpha}}
189: \left|\tilde{0}\right\rangle ,\label{eq:psi}\end{equation}
190: which consists of projectors $P_{\lambda}$ , $P_{0}=\left(1-n_{d\uparrow}\right)\left(1-n_{d\downarrow}\right)$,
191: $P_{1}=\sum_{\sigma}n_{d\sigma}\left(1-n_{d\bar{\sigma}}\right)$,
192: and $P_{2}=n_{d\uparrow}n_{d\downarrow}$ and additional operators
193: involving the operators in leads (for example,
194: $P_3=P_0\widehat{v}_\mathrm{L} P_1$), which are applied to the
195: state $\left|\tilde{0}\right\rangle$ corresponding to the phonon
196: vacuum and the ground state of the auxiliary noninteracting Hamiltonian
197: $\tilde{H}=\tilde{\epsilon}_d n_d +
198: \tilde{t}'_\mathrm{L} \widehat{v}_\mathrm{L} + \tilde{t}'_\mathrm{R}
199: \widehat{v}_\mathrm{R} +H_\mathrm{L}+H_\mathrm{R}$ with renormalized local 
200: energy $\tilde{\epsilon}_d$ and hopping parameters $\tilde{t}'_{\mathrm{L,R}}$.
201: In the model I  $\alpha=1$ and $\alpha=1,2$ for the model II.  Due to
202: the CM displacement which induces the coupling to the asymmetric
203: channel, the renormalized couplings to the left and right lead are
204: not necessarily 
205: equal. The chemical potential was set to the middle of the band. A
206: sufficient number of variational  operators   
207: (up to $\sim 40$) 
208: and excited phonon states (up to $\sim 40$) were used in order to
209: obtain converged results in the parameter regimes presented here.
210: 
211: Zero temperature conductance was calculated by two related methods, both
212: based on the ground state wave function expressed in the basis Eq.~(\ref{eq:psi}). In the
213: first method it was calculated from the sine formula (SF) \cite{rejec03b,rejec03a},
214: $G=G_{0}\sin^{2}[(E_{+}-E_{-})N/4t]$, 
215: where $G_{0}=2e^{2}/h$ and $E_{\pm}$ are the ground state energies
216: of a large auxiliary ring consisted of $N$ non-interacting sites
217: and an embedded interacting system (molecule), with periodic and anti-periodic
218: boundary conditions, respectively. In the second method we use the
219: effective parameters $\tilde{t}'_\mathrm{L,R},\tilde{\epsilon}_d$ of the
220: auxiliary effective noninteracting Hamiltonian $\tilde{H}$ and the conductance
221: is then given from the corresponding Green's function (GF)
222: \cite{rejec03b}.  The advantage of the GF method is
223: that the corresponding  GF can be evaluated for infinite leads so the
224: difficulties with convergence of SF with number of sites are
225: avoided.  However, the SF method is robust and it depends only on the
226: accuracy of the ground-state energy which 
227: improves with the size of the basis in a transparent way. By comparing
228: results of both 
229: methods we checked for the consistency and the convergence. 
230: %In the third method we use the
231: %Friedel sum rule to relate the phase shift and the occupancy.
232: From the effective parameters the conductance can  also be calculated
233: using the Friedel sum rule valid in the limit of large
234: bandwidth ($t \gg U, M, \Omega_i, ...$). When
235: the coupling of the system between left and right lead differ,
236: conveniently parameterized as
237: $\tilde{t}'_\mathrm{L}=(1+b)\tilde{t}'$, $\tilde{t}'_\mathrm{R}=(1-b)\tilde{t}'$,
238: the conductance is given by
239: \begin{equation} \frac{G}{G_0}=\frac{(1-b^2)^2}{(1+b^2)^2}\sin^2{\frac{\pi}{2}}
240:   n. \label{friedel}\end{equation} 
241: Note that conductance may be less than unity when both the symmetric
242: and the antisymmetric channel participate in the transport (\emph{i.e.}, when
243: $b\neq 0 \land b \neq \pm\infty$). 
244: 
245: % \paragraph{Symmetry considerations}
246: The asymmetry stems from the coupling
247: to the CM displacement, hence the expectation values of
248: the CM displacement $x=\langle x_1 \rangle$, and the calculated asymmetry factor $b$  
249: are related. More precisely, the
250: following exact  \cite{schonhammer84, cornaglia05a} relations hold: \begin{eqnarray} 
251: \mathrm{I:} \, \langle x_1\rangle=
252: -\frac{2gt'}{\Omega_1}\langle\widehat{v}_\mathrm{L}-\widehat{v}_\mathrm{R}\rangle-\frac{2M}{\Omega_1}(n-1)
253: \qquad \qquad  {} \nonumber \\ 
254: \mathrm{II:} \, \langle x_1
255: \rangle=-\frac{2gt'}{\Omega_1}\langle\widehat{v}_\mathrm{L}-\widehat{v}_\mathrm{R}\rangle,
256: \,\, \langle x_2 \rangle=-\frac{2M}{\Omega_2}(n-1) \label{eq:disp}.
257: \end{eqnarray}
258:  When
259: $n=1$ the displacement $x$ is proportional to 
260: the activity of the antisymmetric channel $\langle
261: \widehat{v}_\mathrm{L}-\widehat{v}_\mathrm{R}\rangle$ for both models.    
262: 
263: Some general conclusions   may already be
264: drawn  from considering the particle-hole symmetry. Model I
265: is invariant with respect to the  
266: particle-hole transformation
267: $c_{i\sigma}^\dagger \to c_{i\sigma},d^{\dagger} \to d$ with
268: $(\epsilon+U/2) \to - (\epsilon+U/2)$, $t\to -t$, $t'\to -t'$ and
269: $x_1\to-x_1$. Hence at the point of particle-hole
270: symmetry (where $n=1$), $x=0$, therefore $b=0$ and conductance is
271: unity according to the Friedel sum rule Eq.~(\ref{friedel}). 
272: The particle-hole transformation of model II yields $x_2 \to -x_2$
273: with no bounds on $x_1$. The conductance of model II in the symmetric
274: point therefore cannot be deduced from considering the particle-hole
275: symmetry alone and may be less than unity there.
276: 
277: 
278: \begin{figure}
279: \begin{center}\includegraphics[%
280:   width=70mm,
281:   keepaspectratio]{Fig2.eps}\end{center}
282: 
283: 
284: \caption{\label{Fig2} (Color online) (a) Conductance as
285:   calculated from the Green's function corresponding to 
286: renormalised parameters (GF, full line),
287:   sine formula
288:   method with $N=3000$ (squares), and Friedel sum rule (circles) for molecules described by model I with
289:   $g=0.4$. For comparison, conductance
290:   for a molecule without coupling to the CM mode, $g=0$ (dashed).
291:   (b) Local charge occupancy $n-1$ (full), the displacement
292:   $x/2$ (short-dashed) the asymmetry parameter
293:   $b$ (dotted) for a molecule
294:   described by model I with $g=0.4$, and the occupancy
295:   $n-1$ for $g=0$ (long-dashed).
296: }
297: \end{figure}
298:  \section{NUMERICAL RESULTS}
299: Throughout this paper we analyze the system with positive
300: $U_\mathrm{eff}=U-2M^2/\Omega_\alpha$ and show the results for
301: $U=10\Gamma$, $M=\Omega_\alpha=2.5\Gamma$, 
302: $U_\mathrm{eff}=5\Gamma,\Gamma/t=2t'^2/t^2=0.08$, where $\alpha=1,2$ for model I and
303: II, respectively. The scales of the values used correspond to typical
304: molecular transistors \cite{yu04_2}. 
305: 
306:  \subsection{Model I}
307: 
308: 
309: In Fig.~\ref{Fig2}(a) we show that all
310: methods yield the same result for conductance within the
311: model I (the same is true for model II discussed later).  The area of
312: enhanced conductance is diminished in comparison to the 
313: $g=0$ case.  In general,
314: the conductance curves calculated using different methods are 
315: very close.
316: %Obtaining correct conductance from Eq.~(\ref{friedel}) signals 
317: %that the Friedel sum rule is fulfilled which
318: %reflects the fact that the ground state of the system is a Fermi
319: %liquid.
320: However,  due to the finite
321: bandwidth used here minor discrepancies in results applying the Friedel sum rule, 
322: Eq.~(\ref{friedel}), might occur when compared to
323: other methods in the charge transfer 
324: and empty orbital regimes, $\epsilon_d+U/2\gtrsim t$. 
325: 
326: \begin{figure}
327: \includegraphics[%
328:   width=75mm]{Fig3.eps}
329: \caption{(Color online) \label{Fig3}(a) Conductance of a molecule
330:   described by model I for various CM couplings $g$ and fixed Holstein
331:   coupling $M$, $2M^2/\Omega_1=U/2$, $U=10\Gamma$, $M=\Omega_1=2.5\Gamma$. (b)
332:   The occupancy $n-1$ for various $g$. For $g^2=0.25$ also the average displacement $x/2$ (dashed-double-dotted)
333:   plotted. (c) Fluctuations of charge $\Delta
334:   n^2=\langle(n_d-n)^2\rangle$ and
335:  (d) displacement $\Delta x^2=\langle (x_1-x)^2
336:   \rangle$. % as a function  of $(\epsilon_d+U/2)/\Gamma$.
337: }
338: \end{figure}
339: 
340: The results for the asymmetry factor $b$, the 
341: occupancy $n-1$ and the displacement $x$ are shown
342: Fig.~\ref{Fig2}(b).  As discussed above, the 
343: displacement at the symmetric point has to vanish. Correspondingly,
344: the asymmetry $b=0$ there. Near the
345: particle-hole symmetric point we find the relation between the
346: displacement and the asymmetry factor to be approximately linear.
347: 
348: We now systematically explain the properties of the ground state
349: for model I starting
350: from a Hamiltonian with $g=0$. In Fig.~\ref{Fig3}(a) we plot the conductance. 
351: Surprisingly, the influence of increasing $g$ on the conductance is similar to 
352: simple reduction of $U$ as is the case with the
353: Anderson-Holstein model where the electron-phonon coupling term $M$ 
354: effectively reduces $U$.
355: However, as shown in Fig.~\ref{Fig3}(b) the occupancy $n$ does not exhibit the sharp
356: transition from $n\sim 2$ to $n\sim 0$ as seen in the Anderson model with
357: reduced $U_\mathrm{eff}$, where electron (or hole) pairs are preferred in the ground
358: state (see \emph{e.g.} Ref.~\onlinecite{mravlje05}) and the transport
359: is dominated by the pair tunneling \cite{koch06}. In the strong coupling
360: regime the occupancy preferentially takes on values near half integer instead.
361:  This is because to the lowest order only the average
362: displacement couples the electronic and 
363: phononic part, and the average displacement vanishes when $n=1$. Also,
364: for states with well defined occupancy $n\sim 0,1,2$, the hopping
365: matrix element has a vanishing weight.
366: 
367: 
368: To illustrate this further we determined the charge fluctuations on
369: the molecule. In general the range corresponding
370: to increased charge fluctuations $\Delta n^2$ for $g>0$ 
371: [Fig.~\ref{Fig3}(c)] is extended compared to the 
372: $g=0$ case. For large $g$ the charge fluctuations still exhibit a
373: minimum at the
374: particle hole symmetric point,  in contrast to the negative-$U$ (corresponding to strong
375: coupling to the breathing mode) case where the charge analog of the
376: Kondo effect   with
377: increased charge fluctuations evolves \cite{mravlje05}. The straightforward application
378: of the fluctuation-dissipation theorem relates $\Delta
379: n^2\sim -\partial n/ \partial \epsilon_d$, which does not apply to this
380: system in some parameter regimes.
381: On the other hand, the displacement fluctuations shown Fig.~\ref{Fig3}(d) are increased
382: for large $g$ at the symmetric point as in
383: the Anderson-Holstein model  \cite{mravlje05}.
384: 
385: In Fig.~\ref{Fig3}(b) we also plot
386: the average displacement $x/2$ of the molecule for $g=0.5$. According to
387: the relation in Eq.~(\ref{eq:disp}), the difference between the
388: occupancy (times $2M/\Omega_1$) and
389: the displacement $x$ is determined by the difference between
390: expectation values of the tunneling into left and right leads. At the
391: symmetric point, this difference vanishes.
392: 
393: To summarize, in the particle-hole symmetric point the conductance is
394: unity, contrary to the corresponding results presented in
395: Ref.~\onlinecite{alhassanieh05} where a dip in conductance was reported
396: and related  to the non-applicability of the Friedel sum rule due to
397: the breakdown of the Fermi liquid. Our results confirm that the conductance can be
398: expressed with the Friedel sum rule, if left-right asymmetry is
399: correctly taken into account, Eq.~(\ref{friedel}). The
400: displacement of the oscillator in model I is determined by the
401: occupancy. The displacement increases coupling to the anti-symmetric
402: channel and thereby the charge fluctuations which destroy the Kondo
403: correlations with well defined charge and enhanced
404: conductance. Further, the
405: displacement gives rise to asymmetry, which
406: means conductance is additionally suppressed for general $\epsilon_d \neq -U/2$.
407: The width of the conductance peak is seen to decrease approximately as
408: $U-2M^2/\Omega_1-f(M,\Omega_1,t')g^2$ until this quantity is
409: reduced to zero. 
410: 
411: \begin{figure}
412: \includegraphics[%
413:   width=75mm,
414:   keepaspectratio]{Fig4.eps}
415: \caption{(Color online) \label{Fig4}
416: (a) Conductance of a molecule described by model II for 
417:   $g<g_c$ (dashed) and $g_c<g<g'_c$ (full). $\Omega_1=0.2t$. Inset: The
418:   occupancy for the two cases. For $g>g_c$ also the CM displacement
419:   $x$ is shown (dotted). For $g<g_c$ the CM displacement vanishes, $x=0$. (b)
420:   Fluctuations of charge and of CM displacement. The curve of charge
421:   fluctuations for $g=0.5$ is valid for all $g<g_c$. (c) The
422:   conductance in the symmetric point as a function of $g$. }
423: \end{figure}
424: 
425:  \subsection{Model II}
426: We now turn to the results within model II. For small enough values (to be
427: quantified below) of the CM coupling $g < g_c$ the conductance is not
428: significantly affected as unlike in model I the 
429: asymmetry is not determined (directly) by the occupancy of 
430: the molecule. The average displacement vanishes, correspondingly there
431: is no asymmetry in the effective model; the width as well as the shape
432: of the conductance peak and also the correlations
433: exhibit only minor renormalisation compared to the $g=0$
434: case. Conversely, for 
435: $g > g_c$ a soft mode emerges and  
436: the model has a transition reminiscent of dynamical Jahn-Teller distortion
437: \cite{balseiro06} as within a simplified treatment of the related model 
438:  \cite{alascio88}. This is signaled in our approach with non-vanishing
439:  $x$ and  asymmetry $b$ in the effective model $\tilde{H}$ and
440:  correspondingly with
441:  qualitatively modified correlations. For this regime
442:  additional variational basis functions might be required\cite{hewson80}. For CM coupling
443:  large enough, $g > g_c'$, only the asymmetric channel is 
444: active. This is a consequence of the renormalization: 
445: only the channel with the strongest Kondo coupling remains active at low
446: temperatures, in accord with Ref.~\onlinecite{balseiro06}, and with corresponding unity conductance.
447: 
448: 
449: 
450: In Fig.~\ref{Fig4} (a) we plot the conductance and
451: in Fig.~\ref{Fig4} (b) the fluctuations of occupancy and the
452: displacement as a
453: function of $\epsilon_d$ for two distinct $g_c<g<g_c'$ and $g<g_c$ and
454: fixed $\Omega_1/t=0.2$. Whereas
455: for $g<g_c$ the conductance and correlations are almost unaffected by
456: the CM coupling (compare with the $g=0$ curves in Figs.~\ref{Fig2},~\ref{Fig3}),
457: the $g>g_c$ result is markedly different. In the latter, the
458: conductance is less than unity at
459: the symmetric point and is zero at the point corresponding to
460: $b=1$. In the inset of Fig.~\ref{Fig4} (b) we also plot the occupancy
461: and for $g>g_c$ the CM displacement. For $\epsilon_d$
462: large enough, the CM displacement (and correspondingly the asymmetry) is
463: zero. The precise value for $g_c$ hence varies with the gate voltage
464: and is related to the charge fluctuations of the bare model: increased
465: charge fluctuations induce a transition to the state of 
466: spontaneously broken symmetry. 
467: 
468: 
469: 
470: 
471: In Fig.~\ref{Fig4} (c) we plot the conductance as a function of
472: coupling to the CM mode. For small $g$ the
473: conductance is unity as only the even channel is active, $b=0$. By
474: increasing $g$, the conductance drops sharply at $g_c$ which is found to be set by
475: the relation $g_c^2 
476: t/\Omega_1=c(t'),$  where $c(t')\sim 2.5$ for the results presented
477: here \cite{rel}. Note labels A and B indicating the values of $g$
478: for which the results are shown in
479: Figs.~\ref{Fig4}(a,b). For large $\Omega_1$ both channels are active with corresponding
480: suppressed conductance only for a narrow range of $g>g_c$  while for
481: progressively decreasing $\Omega_1$ the range of $g$ with suppressed
482: conductance widens. As in typical devices considered in
483: experiments\cite{yu04_2} $\Omega_1<\Gamma<U$ the anomalies in conductance
484: of the type considered here would would not be sensitive to
485: the precise value of the coupling to the CM mode provided that it is
486: large enough.
487: 
488: 
489: 
490: 
491: 
492: 
493: \section{CONCLUSIONS}
494: In conclusion, we analyzed the properties of molecules with
495: a center-of-mass vibrational mode in linear transport measurements
496: described by two distinct Hamiltonians. In the first case, where the
497: coupling to the CM mode 
498: simultaneously modulates the molecular energy level, the result of the
499: coupling to the CM mode is a reduced range of gate voltages
500: with enhanced conductance. As a direct consequence of the
501: particle-hole symmetry the conductance in the symmetric point remains unity.
502: 
503: In the second case where the coupling to the CM mode modulates only the
504: tunneling matrix elements and additional breathing mode couples to the
505: electron density, the CM mode is of minor relevance for the
506: transport  until a critical value of
507: this coupling $g_c$ is reached. For values of the coupling near 
508: the critical value anomalies due to the development of the soft mode arise
509: signaled by the conductance which is less than unity.
510: The conductance  for these values is within the effective model suppressed for
511: a wide range of gate voltages.
512: 
513: We explained the relation of CM displacement to 
514: the transport of electrons for both models and substantiated the explanation by the
515: numerical data for occupancy, occupancy fluctuations and also related
516: phonon expectation values.
517: 
518: Finally, we may make some remarks regarding the experimental
519: relevance of results presented here. 
520: For linear regime without residual electric fields it does
521: not seem likely that the model I would provide an accurate description
522: of the experimental situation. By analyzing model II which applies
523: then, we have shown that for small couplings to the CM mode there are
524: no discernible effects on the linear conductance (we should note here
525: that this statement would not be changed if additional breathing modes
526: were introduced). At a sufficient coupling strength this picture
527: changes: the displacement of the system from the point of inversion
528: symmetry becomes increasingly sensitive to external perturbations
529: which would result in a displacement of molecule and strong
530: suppression of conductance.
531: 
532: 
533: To check for this behaviour in experiments would require the following
534: steps. First devices in which CM side-peaks are seen would have to be
535: produced, possibly by using different molecules, which would give in
536: other respects identical devices besides having different frequencies
537: of the CM motion. The fingerprint of the physics considered here would
538: then be the suppression of conductance which would predominantly occur
539: for the devices with lower frequencies of the CM motion.  As this
540: procedure would require near-atomic similar lead configuration near
541: the molecule the systematic experimental confirmation of the analyzed
542: effects seems currently hard to achieve. We speculate nevertheless
543: that this effects might have contributed to the anomalously suppressed
544: conductance in some of the devices considered in, \emph{e.g.},
545: Ref.~\onlinecite{pasupathy05}.
546: 
547: 
548: \acknowledgements{
549:  We thank R. \v{Z}itko for inspiring suggestions and S.~El~Shawish for
550:  discussions regarding the Jahn-Teller effect. The work was supported
551:  by SRA under grant Pl-0044.   
552: }
553: %\bibliography{cm}
554: \begin{thebibliography}{31}
555: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
556: \expandafter\ifx\csname bibnamefont\endcsname\relax
557:   \def\bibnamefont#1{#1}\fi
558: \expandafter\ifx\csname bibfnamefont\endcsname\relax
559:   \def\bibfnamefont#1{#1}\fi
560: \expandafter\ifx\csname citenamefont\endcsname\relax
561:   \def\citenamefont#1{#1}\fi
562: \expandafter\ifx\csname url\endcsname\relax
563:   \def\url#1{\texttt{#1}}\fi
564: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi
565: \providecommand{\bibinfo}[2]{#2}
566: \providecommand{\eprint}[2][]{\url{#2}}
567: 
568: \bibitem[{\citenamefont{Madhavan et~al.}(1998)\citenamefont{Madhavan, Chen,
569:   Jamneala, Crommie, and Wingreen}}]{madhavan98}
570: \bibinfo{author}{\bibfnamefont{V.}~\bibnamefont{Madhavan}},
571:   \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Chen}},
572:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Jamneala}},
573:   \bibinfo{author}{\bibfnamefont{M.~F.} \bibnamefont{Crommie}},
574:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{N.~S.}
575:   \bibnamefont{Wingreen}}, \bibinfo{journal}{Science}
576:   \textbf{\bibinfo{volume}{280}}, \bibinfo{pages}{567} (\bibinfo{year}{1998}).
577: 
578: \bibitem[{\citenamefont{Park et~al.}(2000)\citenamefont{Park, Park, Lim,
579:   Anderson, Alivistatos, and McEuen}}]{park00}
580: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Park}},
581:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Park}},
582:   \bibinfo{author}{\bibfnamefont{A.~K.~L.} \bibnamefont{Lim}},
583:   \bibinfo{author}{\bibfnamefont{E.~H.} \bibnamefont{Anderson}},
584:   \bibinfo{author}{\bibfnamefont{A.~P.} \bibnamefont{Alivistatos}},
585:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{P.~L.}
586:   \bibnamefont{McEuen}}, \bibinfo{journal}{Science}
587:   \textbf{\bibinfo{volume}{407}}, \bibinfo{pages}{57} (\bibinfo{year}{2000}).
588: 
589: \bibitem[{\citenamefont{Park et~al.}(2002)\citenamefont{Park, Pasupathy,
590:   I.Goldsmith, Chang, Yaish, Petta, Rinkoski, Sethna, Abrunas, McEuen
591:   et~al.}}]{park02}
592: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Park}},
593:   \bibinfo{author}{\bibfnamefont{A.~N.} \bibnamefont{Pasupathy}},
594:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{I.Goldsmith}},
595:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Chang}},
596:   \bibinfo{author}{\bibfnamefont{Y.}~\bibnamefont{Yaish}},
597:   \bibinfo{author}{\bibfnamefont{J.~R.} \bibnamefont{Petta}},
598:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Rinkoski}},
599:   \bibinfo{author}{\bibfnamefont{J.~P.} \bibnamefont{Sethna}},
600:   \bibinfo{author}{\bibfnamefont{H.~D.} \bibnamefont{Abrunas}},
601:   \bibinfo{author}{\bibfnamefont{P.~L.} \bibnamefont{McEuen}},
602:   \bibnamefont{et~al.}, \bibinfo{journal}{Nature}
603:   \textbf{\bibinfo{volume}{417}}, \bibinfo{pages}{722} (\bibinfo{year}{2002}).
604: 
605: \bibitem[{\citenamefont{Liang et~al.}(2002)\citenamefont{Liang, Shores,
606:   Bockrath, Long, and Park}}]{liang02}
607: \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Liang}},
608:   \bibinfo{author}{\bibfnamefont{M.~P.} \bibnamefont{Shores}},
609:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Bockrath}},
610:   \bibinfo{author}{\bibfnamefont{J.~R.} \bibnamefont{Long}}, \bibnamefont{and}
611:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Park}},
612:   \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{417}},
613:   \bibinfo{pages}{725} (\bibinfo{year}{2002}).
614: 
615: \bibitem[{\citenamefont{Zhitenev et~al.}(2002)\citenamefont{Zhitenev, Meng, and
616:   Bao}}]{zhitenev02}
617: \bibinfo{author}{\bibfnamefont{N.~B.} \bibnamefont{Zhitenev}},
618:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Meng}}, \bibnamefont{and}
619:   \bibinfo{author}{\bibfnamefont{Z.}~\bibnamefont{Bao}},
620:   \bibinfo{journal}{Phys. Rev. Lett} \textbf{\bibinfo{volume}{88}},
621:   \bibinfo{pages}{226801} (\bibinfo{year}{2002}).
622: 
623: \bibitem[{\citenamefont{Yu et~al.}(2004)\citenamefont{Yu, Keane, Ciszek, Cheng,
624:   Stewart, Tour, and Natelson}}]{yu04}
625: \bibinfo{author}{\bibfnamefont{L.~H.} \bibnamefont{Yu}},
626:   \bibinfo{author}{\bibfnamefont{Z.~K.} \bibnamefont{Keane}},
627:   \bibinfo{author}{\bibfnamefont{J.~W.} \bibnamefont{Ciszek}},
628:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Cheng}},
629:   \bibinfo{author}{\bibfnamefont{M.~P.} \bibnamefont{Stewart}},
630:   \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{Tour}}, \bibnamefont{and}
631:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Natelson}},
632:   \bibinfo{journal}{Phys. Rev. Lett} \textbf{\bibinfo{volume}{93}},
633:   \bibinfo{pages}{266802} (\bibinfo{year}{2004}).
634: 
635: \bibitem[{\citenamefont{Yu and Natelson}(2005)}]{yu04_2}
636: \bibinfo{author}{\bibfnamefont{L.~H.} \bibnamefont{Yu}} \bibnamefont{and}
637:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Natelson}},
638:   \bibinfo{journal}{Nanoletters} \textbf{\bibinfo{volume}{4}},
639:   \bibinfo{pages}{79} (\bibinfo{year}{2005}).
640: 
641: \bibitem[{\citenamefont{Pasupathy et~al.}(2005)\citenamefont{Pasupathy, Park,
642:   Chang, Soldatov, Lebedkin, Bialczak, Grose, Donev, Sethna, Ralph
643:   et~al.}}]{pasupathy05}
644: \bibinfo{author}{\bibfnamefont{A.~N.} \bibnamefont{Pasupathy}},
645:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Park}},
646:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Chang}},
647:   \bibinfo{author}{\bibfnamefont{A.~V.} \bibnamefont{Soldatov}},
648:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Lebedkin}},
649:   \bibinfo{author}{\bibfnamefont{R.~C.} \bibnamefont{Bialczak}},
650:   \bibinfo{author}{\bibfnamefont{J.~E.} \bibnamefont{Grose}},
651:   \bibinfo{author}{\bibfnamefont{L.~A.~K.} \bibnamefont{Donev}},
652:   \bibinfo{author}{\bibfnamefont{J.~P.} \bibnamefont{Sethna}},
653:   \bibinfo{author}{\bibfnamefont{D.~C.} \bibnamefont{Ralph}},
654:   \bibnamefont{et~al.}, \bibinfo{journal}{Nanoletters}
655:   \textbf{\bibinfo{volume}{5}}, \bibinfo{pages}{203} (\bibinfo{year}{2005}).
656: 
657: \bibitem[{\citenamefont{Zhao et~al.}(2005{\natexlab{a}})\citenamefont{Zhao, Li,
658:   Chen, Xiang, Wang, Pan, Wang, Xiao, Yang, Hou et~al.}}]{zhao05}
659: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Zhao}},
660:   \bibinfo{author}{\bibfnamefont{Q.}~\bibnamefont{Li}},
661:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Chen}},
662:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Xiang}},
663:   \bibinfo{author}{\bibfnamefont{W.}~\bibnamefont{Wang}},
664:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Pan}},
665:   \bibinfo{author}{\bibfnamefont{B.}~\bibnamefont{Wang}},
666:   \bibinfo{author}{\bibfnamefont{X.}~\bibnamefont{Xiao}},
667:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Yang}},
668:   \bibinfo{author}{\bibfnamefont{J.~G.} \bibnamefont{Hou}},
669:   \bibnamefont{et~al.}, \bibinfo{journal}{Science}
670:   \textbf{\bibinfo{volume}{309}}, \bibinfo{pages}{1542}
671:   (\bibinfo{year}{2005}{\natexlab{a}}).
672: 
673: \bibitem[{\citenamefont{Zhao et~al.}(2005{\natexlab{b}})\citenamefont{Zhao,
674:   Zeng, Cheng, Wang, Wang, Yang, Hou, and Zhu}}]{zhao05_2}
675: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Zhao}},
676:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Zeng}},
677:   \bibinfo{author}{\bibfnamefont{X.}~\bibnamefont{Cheng}},
678:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Wang}},
679:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Wang}},
680:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Yang}},
681:   \bibinfo{author}{\bibfnamefont{J.~G.} \bibnamefont{Hou}}, \bibnamefont{and}
682:   \bibinfo{author}{\bibfnamefont{Q.}~\bibnamefont{Zhu}},
683:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{95}},
684:   \bibinfo{pages}{045502} (\bibinfo{year}{2005}{\natexlab{b}}).
685: 
686: \bibitem[{\citenamefont{Wahl et~al.}(2005)\citenamefont{Wahl, Diekhoner,
687:   Wittich, Vitali, Schneider, and Kern}}]{wahl05}
688: \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Wahl}},
689:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Diekhoner}},
690:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Wittich}},
691:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Vitali}},
692:   \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Schneider}},
693:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Kern}},
694:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{95}},
695:   \bibinfo{eid}{166601} (\bibinfo{year}{2005}).
696: 
697: \bibitem[{\citenamefont{Yu et~al.}(2005)\citenamefont{Yu, Keane, Ciszek, Cheng,
698:   Tour, Baruah, Pederson, and Natelson}}]{yu05}
699: \bibinfo{author}{\bibfnamefont{L.~H.} \bibnamefont{Yu}},
700:   \bibinfo{author}{\bibfnamefont{Z.~K.} \bibnamefont{Keane}},
701:   \bibinfo{author}{\bibfnamefont{J.~W.} \bibnamefont{Ciszek}},
702:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Cheng}},
703:   \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{Tour}},
704:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Baruah}},
705:   \bibinfo{author}{\bibfnamefont{M.~R.} \bibnamefont{Pederson}},
706:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Natelson}},
707:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{95}},
708:   \bibinfo{pages}{256803} (\bibinfo{year}{2005}).
709: 
710: \bibitem[{\citenamefont{Natelson et~al.}(2006)\citenamefont{Natelson, Yu,
711:   Ciszek, Keane, and Tour}}]{natelson06}
712: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Natelson}},
713:   \bibinfo{author}{\bibfnamefont{L.~H.} \bibnamefont{Yu}},
714:   \bibinfo{author}{\bibfnamefont{J.~W.} \bibnamefont{Ciszek}},
715:   \bibinfo{author}{\bibfnamefont{Z.~K.} \bibnamefont{Keane}}, \bibnamefont{and}
716:   \bibinfo{author}{\bibfnamefont{J.~M.} \bibnamefont{Tour}},
717:   \bibinfo{journal}{Chem. Phys.} \textbf{\bibinfo{volume}{324}},
718:   \bibinfo{pages}{267} (\bibinfo{year}{2006}).
719: 
720: \bibitem[{\citenamefont{Cornaglia et~al.}(2004)\citenamefont{Cornaglia, Ness,
721:   and Grempel}}]{cornaglia04}
722: \bibinfo{author}{\bibfnamefont{P.~S.} \bibnamefont{Cornaglia}},
723:   \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Ness}}, \bibnamefont{and}
724:   \bibinfo{author}{\bibfnamefont{D.~R.} \bibnamefont{Grempel}},
725:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{93}},
726:   \bibinfo{eid}{147201} (\bibinfo{year}{2004}).
727: 
728: \bibitem[{\citenamefont{Cornaglia et~al.}(2005)\citenamefont{Cornaglia,
729:   Grempel, and Ness}}]{cornaglia05a}
730: \bibinfo{author}{\bibfnamefont{P.~S.} \bibnamefont{Cornaglia}},
731:   \bibinfo{author}{\bibfnamefont{D.~R.} \bibnamefont{Grempel}},
732:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Ness}},
733:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{71}},
734:   \bibinfo{eid}{075320} (\bibinfo{year}{2005}).
735: 
736: \bibitem[{\citenamefont{Cornaglia and Grempel}(2005)}]{cornaglia05b}
737: \bibinfo{author}{\bibfnamefont{P.~S.} \bibnamefont{Cornaglia}}
738:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{D.~R.}
739:   \bibnamefont{Grempel}}, \bibinfo{journal}{Phys. Rev. B}
740:   \textbf{\bibinfo{volume}{71}}, \bibinfo{eid}{245326} (\bibinfo{year}{2005}).
741: 
742: \bibitem[{\citenamefont{Mravlje et~al.}(2005)\citenamefont{Mravlje, Ram\v{s}ak,
743:   and Rejec}}]{mravlje05}
744: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Mravlje}},
745:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ram\v{s}ak}},
746:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Rejec}},
747:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{72}},
748:   \bibinfo{pages}{121403(R)} (\bibinfo{year}{2005}).
749: 
750: \bibitem[{\citenamefont{Mitra et~al.}(2004)\citenamefont{Mitra, Aleiner, and
751:   Millis}}]{mitra04}
752: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Mitra}},
753:   \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Aleiner}}, \bibnamefont{and}
754:   \bibinfo{author}{\bibfnamefont{A.~J.} \bibnamefont{Millis}},
755:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{69}},
756:   \bibinfo{eid}{245302} (\bibinfo{year}{2004}).
757: 
758: \bibitem[{\citenamefont{Paaske and Flensberg}(2005)}]{paaske05}
759: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Paaske}} \bibnamefont{and}
760:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Flensberg}},
761:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{94}},
762:   \bibinfo{pages}{176801} (\bibinfo{year}{2005}).
763: 
764: \bibitem[{\citenamefont{Yamamoto et~al.}(2005)\citenamefont{Yamamoto, Watanabe,
765:   and Watanabe}}]{takahiro05}
766: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Yamamoto}},
767:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Watanabe}}, \bibnamefont{and}
768:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Watanabe}},
769:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{95}},
770:   \bibinfo{pages}{065501} (\bibinfo{year}{2005}).
771: 
772: \bibitem[{\citenamefont{Al-Hassanieh et~al.}(2005)\citenamefont{Al-Hassanieh,
773:   B{\"u}sser, Martins, and Dagotto}}]{alhassanieh05}
774: \bibinfo{author}{\bibfnamefont{K.~A.} \bibnamefont{Al-Hassanieh}},
775:   \bibinfo{author}{\bibfnamefont{C.~A.} \bibnamefont{B{\"u}sser}},
776:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Martins}}, \bibnamefont{and}
777:   \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Dagotto}},
778:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{95}},
779:   \bibinfo{pages}{256807} (\bibinfo{year}{2005}).
780: 
781: \bibitem[{\citenamefont{Balseiro et~al.}(2006)\citenamefont{Balseiro,
782:   Cornaglia, and Grempel}}]{balseiro06}
783: \bibinfo{author}{\bibfnamefont{C.~A.} \bibnamefont{Balseiro}},
784:   \bibinfo{author}{\bibfnamefont{P.~S.} \bibnamefont{Cornaglia}},
785:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{D.~R.}
786:   \bibnamefont{Grempel}}, \bibinfo{journal}{cond-mat/0605064}
787:   (\bibinfo{year}{2006}).
788: 
789: \bibitem[{\citenamefont{Sch\"onhammer}(1976)}]{schonhammer76}
790: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Sch\"onhammer}},
791:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{13}},
792:   \bibinfo{pages}{4336} (\bibinfo{year}{1976}).
793: 
794: \bibitem[{\citenamefont{Gunnarsson and Sch{\"o}nhammer}(1985)}]{gunnarsson85}
795: \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Gunnarsson}} \bibnamefont{and}
796:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Sch{\"o}nhammer}},
797:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{31}},
798:   \bibinfo{pages}{4815} (\bibinfo{year}{1985}).
799: 
800: \bibitem[{\citenamefont{Rejec and Ram\v{s}ak}(2003{\natexlab{a}})}]{rejec03b}
801: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Rejec}} \bibnamefont{and}
802:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ram\v{s}ak}},
803:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{68}},
804:   \bibinfo{eid}{035342} (\bibinfo{year}{2003}{\natexlab{a}}).
805: 
806: \bibitem[{\citenamefont{Rejec and Ram\v{s}ak}(2003{\natexlab{b}})}]{rejec03a}
807: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Rejec}} \bibnamefont{and}
808:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ram\v{s}ak}},
809:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{68}},
810:   \bibinfo{pages}{033306} (\bibinfo{year}{2003}{\natexlab{b}}).
811: 
812: \bibitem[{\citenamefont{Sch\"onhammer and Gunnarsson}(1984)}]{schonhammer84}
813: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Sch\"onhammer}}
814:   \bibnamefont{and}
815:   \bibinfo{author}{\bibfnamefont{O.}~\bibnamefont{Gunnarsson}},
816:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{30}},
817:   \bibinfo{pages}{3141} (\bibinfo{year}{1984}).
818: 
819: \bibitem[{\citenamefont{Koch et~al.}(2006)\citenamefont{Koch, Raikh, and von
820:   Oppen}}]{koch06}
821: \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Koch}},
822:   \bibinfo{author}{\bibfnamefont{M.~E.} \bibnamefont{Raikh}}, \bibnamefont{and}
823:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{von Oppen}},
824:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{96}},
825:   \bibinfo{eid}{056803} (\bibinfo{year}{2006}).
826: 
827: \bibitem[{\citenamefont{Alascio et~al.}(1988)\citenamefont{Alascio, Balseiro,
828:   Ort\'iz, Kiwi, and Lagos}}]{alascio88}
829: \bibinfo{author}{\bibfnamefont{B.}~\bibnamefont{Alascio}},
830:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Balseiro}},
831:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Ort\'iz}},
832:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Kiwi}}, \bibnamefont{and}
833:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Lagos}},
834:   \bibinfo{journal}{Phys. Rev. B} \textbf{\bibinfo{volume}{38}},
835:   \bibinfo{pages}{4698} (\bibinfo{year}{1988}).
836: 
837: \bibitem[{\citenamefont{Hewson and Newns}(1980)}]{hewson80}
838: \bibinfo{author}{\bibfnamefont{A.~C.} \bibnamefont{Hewson}} \bibnamefont{and}
839:   \bibinfo{author}{\bibfnamefont{D.~M.} \bibnamefont{Newns}},
840:   \bibinfo{journal}{J. Phys. C: Solid St. Phys.,}
841:   \textbf{\bibinfo{volume}{13}}, \bibinfo{pages}{4477} (\bibinfo{year}{1980}).
842: 
843: \bibitem[{rel()}]{rel}
844: \bibinfo{howpublished}{In the symmetric point
845:   $g_c\sim0.45\sqrt{\Omega_1/t}/(t'/t)^{3/4}$ irrespective of $U, M$ and
846:   $\Omega_2$.}
847: 
848: \end{thebibliography}
849: 
850: 
851: \end{document}
852: