cond-mat0310636/Ahn.tex
1: %\documentclass[aps,prl,preprint,groupedaddress]{revtex4}
2: %\documentclass[aps,prl,preprint,superscriptaddress]{revtex4}
3: %\documentclass[aps,prl,twocolumn,showpacs,groupedaddress]{revtex4}
4: \documentclass[aps,prl,twocolumn,showpacs,superscriptaddress]{revtex4}
5: %\documentclass[aps,prl,draft,twocolumn,superscriptaddress]{revtex4}
6: \usepackage{graphicx}
7: 
8: \begin{document}
9: 
10: \title{Elementary Excitations in One-Dimensional Electromechanical Systems; \\
11: Transport with Back-Reaction}
12: 
13: \author{Kang-Hun Ahn}
14: \affiliation{ Department of Physics, Chungnam National University, Daejon 305-764, Korea}
15: \author{Hangmo Yi}
16: \affiliation{ School of Physics, Korea Institute for Advanced Study, Seoul 130-722, Korea }
17: \date{ \today}
18: 
19: \begin{abstract}
20: Using an exactly solvable model, we study low-energy properties of a one-dimensional
21:  spinless electron fluid contained in a quantum-mechanically moving wire located in
22: a static magnetic field.
23: The phonon and electric current are coupled via Lorentz force and 
24: the eigenmodes are described by two independent boson fluids. 
25: At low energies, the two boson modes are charged while 
26: one of them has excitation gap due to back-reaction of the Lorentz force.
27: The theory is illustrated by evaluating optical absorption spectra.
28: Our results are exact and show a non-perturbative regime of electron transport.
29: \end{abstract}
30: \pacs{73.23.-b, 71.10.Pm, 72.15.-v } 
31: \maketitle
32: 
33: Electronic and mechanical properties of nano-electromechanical
34: systems (NEMS) are recently of fundamental and technical interest\cite{bishop}.  
35: Perpetual trend of miniaturization of electronic devices 
36: has already led NEMS smaller down to molecular scale, e.g., nanotube electromechanical systems\cite{kim}
37: which are
38: likely the bases of new novel devices in NEMS application\cite{rueckes}.
39: These systems are close to the regime of quantum-mechanical operation
40: which is the ultimate limit of nanoelectromechanical devices.
41: For the design of quantum mechanically operating NEMS devices, it is crucial 
42: to understand first the quantum effects of mechanical motion coupled to electric current.
43: Here we ask ourselves;
44: Are the electric currents influenced by the quantum nature of mechanical motion?
45: If so, what is the signature of the quantum fluctuation?
46: 
47: To answer these, we study one-dimensional spinless electron fluids coupled to their quantum-mechanical motion
48: in a static magnetic field. 
49: This model is exactly solvable as will be shown here.
50: In this way, we could describe both the electrical and mechanical degrees of freedom 
51: in  a purely quantum mechanical manner on an equal footing.
52: The harmonic-fluid approach used for electron fluids in a static wire, which 
53: are referred to as (Tomonaga-)Luttinger liquid theory\cite{tomonaga,luttinger,haldane}, 
54: could be successfully applied to
55:  describe the electric currents coupled to mechanical motions through Lorentz force. 
56: This magneto-electro-mechanical coupling is essential in so-called electromotive-force technique which 
57: is popularly used in NEMS-based force sensors\cite{cleland,mohanty}. 
58: 
59: In this paper, we show that electric currents in 
60: nano-electromechanical wires are strongly 
61: influenced by the quantum fluctuations of mechanical motion through the magneto-electro-mechanical coupling.
62: This coupling causes interesting and peculiar phenomena  because of
63: the feedback reaction between electric and mechanical movement.
64: As we shall see, 
65: an important frequency scale related to a gap structure arises due to this feedback reaction. 
66: The feature we found could be observed
67: generally in other vibrating one-dimensional electronic systems 
68: such as atomic wires\cite{chain} and quasi-one-dimensional metals\cite{density}.
69: 
70: We model the electrons with an interacting spinless Fermion fluid with a periodic boundary condition $\Psi(x+L)=\Psi(x)$.
71: These Fermions are contained in a wire of mass density $\rho_{m}$ and
72: tension ${\mathcal T}$, and
73: a magnetic field ${\bf B}$ is applied perpendicular to x-axis. 
74: The displacement $u(x)$ is in the direction of  $\hat{\bf x} \times {\bf B}$, and
75: assumed small ($\partial_{x}u \ll 1 $) and periodic ($u(x+L)=u(x)$).
76: The electrons and the wire are assumed strongly tightened so any force exerted on the electrons in u-direction is 
77: delivered to the wire immediately. 
78: The Hamiltonian  reads ($\hbar=1$, $e < 0$)
79: \begin{eqnarray}
80: \nonumber
81: H&=&\frac{1}{2m}\int dx \Psi^{\dagger}(x)\big( \partial_{x} -i e A_{x} \big)^{2}\Psi(x)
82: \\ 
83: &+&\frac{1}{2} \int dx dx^{\prime}V(x-x^{\prime}) \rho (x) \rho (x^{\prime})+H_{u}
84: \\ \nonumber
85: H_{u}&=&\frac{1}{2}\int dx \big[ \frac{1}{\rho_{m}}\pi_{u}(x)^{2}+{\mathcal T}(\partial_{x} u(x))^{2}  \big].
86: \label{hamiltonian}
87: \end{eqnarray}
88: Here $\pi_{u}$ is the conjugate momentum of $u(x)$; $[u(x),\pi_{u}(x^{\prime})]=i\delta(x-x^{\prime})$.
89: $A_{x}=Bu(x)+\Phi_{B}/L$ is the gauge potential for the 
90: applied magnetic field where $\Phi_{B}$ is a constant magnetic flux 
91: enclosed by the circle of radius $L/2\pi$.
92: For the sake of convenience, let us redefine $u(x)+\Phi_{B}/BL\rightarrow u(x)$ so that $A_{x}=Bu(x)$.
93: 
94: Following the approach of Haldane\cite{haldane},
95: we use  a Boson field $\theta$ 
96: defined from the Fermion density
97:  $\rho(x)=  (k_{F}+\partial_{x} \theta)/\pi$  
98: and its conjugate field $\phi$ ( $[\phi(x) , \theta(x^{\prime})]=i\frac{1}{2}\pi {\rm sgn}(x-x^{\prime}) )$.
99: The boson fields describe the long wavelength fluctuations of the Fermion field $\Psi^{\dagger}(x)$.
100: It proves convenient to define new Fermion field
101:  $\psi^{\dagger}=\exp(i\Lambda)\Psi^{\dagger}$, where
102: $\Lambda(x)=eB\int^{x}u(x^{\prime})dx^{\prime}$.
103: In terms of $\theta$,$\phi$ and $\Lambda$, the new Fermion field $\psi^{\dagger}$ is written as
104: \begin{eqnarray}
105: \psi^{\dagger}(x)\sim \sum_{\rm n~ odd}\exp[ i n [\theta(x)+k_{F}x ]  ] \exp[ i(\phi(x)+\Lambda(x))]
106: \label{newfield}
107: \end{eqnarray}
108: The boundary conditions for the Boson fields are $\theta(x+L)=\theta(x)+\pi {\rm N}$, 
109: $\phi(x+L)=\phi(x)+\pi {\rm J}$, and $\Lambda(x+L)=\Lambda(x)+eBu_{0}L$, where 
110: $u_{0} = \frac{1}{L}\int_{0}^{L}dx u(x) $, and
111: the sum of the topological numbers N and J 
112: is an even integer. 
113: Note that the new Fermion field now has a twisted boundary condition $\psi(x+L)=\exp(-ieBu_{0}L)\psi(x)$.
114: 
115: The effective Hamiltonian describing long-distance physics is given by
116: taking terms of $n=\pm 1$ in Eq.(\ref{newfield}) with a proper normalization constant;
117: \begin{eqnarray}
118: H=\frac{v}{2\pi}\int dx \big( g \big(\partial_{x}\phi +eBu \big)^{2}
119: +g^{-1}\big(\partial_{x}\theta\big)^{2}\big)+H_{u}.
120: \label{hamiltonian-boson}
121: \end{eqnarray}
122: Here the dimensionless constant $g$ is governed by 
123: electron-electron interaction strength and $v=v_{F}/g$ is the electronic wave velocity, where
124: $v_{F}$ is the Fermi velocity for the non-interacting case.
125: For short-ranged electron interactions $V(x-x^{\prime})=V\delta(x-x^{\prime})$, 
126: $g$ is given by $1/\sqrt{1+V/\pi v_{F}}$.
127: In terms of the boson fields, the electrical current density is given by 
128: \begin{eqnarray}
129: J(x)=-ev_{F}\partial_{x}\phi(x)-v_{F}e^{2}Bu(x).
130: \label{current}
131: \end{eqnarray}
132: 
133: 
134: Note that the term of $u^{2}$ in Eq.(\ref{hamiltonian-boson}) and $H_{u}$ describe phonons
135: of finite mass at small momenta;
136: \begin{eqnarray}
137: \frac{gv}{2\pi} e^{2}B^{2}\int dx u^{2}(x)+H_{u} 
138: =\sum_{q}\tilde{\omega}_{s}(q)a^{\dagger}_{q}a_{q},
139: \label{Hph}
140: \end{eqnarray}
141: Here 
142: \begin{eqnarray}
143: \tilde{\omega}_{s}(q)=\sqrt{v^{2}_{s}q^{2}+\omega^{2}_{B}},
144: \end{eqnarray}
145: where
146: $v_{s}=\sqrt{ {\mathcal T}/\rho_{m}  }$ is the bare sound velocity and 
147:  $\omega_{B}=\sqrt{ \frac{v_{F}e^{2}} {\pi\rho_{m}}   }B$.
148: The first term in Eq.(\ref{Hph}) essentially gives mass to phonon and originates from
149: {\it back-reaction process} in the system. 
150: Note that when the wire moves, 
151: an induced electric field ${\mathcal E}_{\rm emf}=-B\partial_{t}u$ changes 
152: the momentum of electrons by 
153: $\int dt e{\mathcal E}_{\rm emf}=-eBu$,
154: which causes the change in electric current
155: $-e^{2}Bu\rho_{0}/m$ where  $\rho_{0}=k_{F}/\pi$ is the mean Fermion density.
156: The Lorentz force $ -v_{F}e^{2}B^{2}u/\pi$   
157: exerted on the induced electric current  again acts on the movement of the
158: wire, which we call here the back-reaction process.
159: 
160:  Now let us use the following Fourier-transformed field operators;
161: \begin{eqnarray}
162: \nonumber
163: \theta(x)&=&\theta_{0}+\frac{\pi{\rm N}x}{L}-i\sum_{q\neq 0}
164: \sqrt{\frac{g\pi}{2|q|L} } (b^{\dagger}_{q}+b_{-q}){\rm sgn}(q) e^{iqx} \\
165: \nonumber
166: \phi(x)&=&\phi_{0}+\frac{\pi {\rm J}x}{L}-i\sum_{q\neq 0}
167: \sqrt{\frac{\pi}{2g|q|L} } (b^{\dagger}_{q}-b_{-q}) e^{iqx} \\
168: \nonumber
169: u(x)&=&u_{0}+\sum_{q\neq 0}
170: \sqrt{\frac{1}{2\rho_{m}\tilde{\omega}_{s}(q)L} } (a^{\dagger}_{-q}+a_{q}) e^{iqx} \\
171: \pi_{u}(x)&=&\frac{\pi_{u 0}}{L}+i\sum_{q\neq 0}
172: \sqrt{\frac{\rho_{m}\tilde{\omega}_{s}(q)}{2L}  } (a^{\dagger}_{-q}-a_{q}) e^{iqx}. 
173: \end{eqnarray}
174: Here ($\theta_{0}$,J) and ($\phi_{0}$,N) are pairs of conjugate action-angle variables\cite{haldane} and
175: $[u_{0},\pi_{u0}]=i$.
176: 
177: 
178: Using these operators, the Hamiltonian in Eq.(\ref{hamiltonian-boson}) is written as
179: \begin{eqnarray}
180: \nonumber
181: H&=&\sum_{q}v|q|b^{\dagger}_{q}b_{q}+\sum_{q}\tilde{\omega}_{s}(q)a^{\dagger}_{q}a_{q}+H_{ZM}\\
182: &+&\frac{\omega_{B}}{2}\sum_{q}\sqrt{\frac{v|q|}{\tilde{\omega}_{s}(q)}}
183: (b_{-q}^{\dagger}-b_{q})(a_{q}^{\dagger}+a_{-q}){\rm sgn}(q)
184: \label{hamilt-op}
185: \end{eqnarray}
186: Here $H_{ZM}$ describes the zero-mode;
187: $H_{ZM}=\frac{\pi v_{F}}{2L}({\rm J}+\frac{eBu_{0}L}{\pi})^{2}+ \frac{\pi_{u 0}^{2}}{2\rho_{m}L} 
188: +\frac{v\pi}{2gL}{\rm N}^{2}$,
189: according to which the center-of-mass of the wire executes 
190: a simple harmonic oscillation with resonance frequency $\omega_{B}$.
191: 
192: \begin{figure}
193: \includegraphics[width=7cm,height=5.5cm]{dispersion}
194:  \caption{ 
195: The dispersion relation of the eigenmodes $\omega_{\pm}(q)$ (solid lines).
196: $\omega$ is in unit of  $\omega_{B}$ and $q$ is in unit of $\omega_{B}/v$.
197: The dotted lines denote the boson mode for electron $\omega=vq$, phonon $\omega=v_{s}q$,
198:  and the massive phonon $\tilde{\omega}_{s}(q)$.
199: We chose $v_{s}=0.3v$ in this figure.  } 
200: \end{figure}
201: 
202: 
203: The Hamiltonian in Eq.(\ref{hamilt-op} )is now diagonalized using new Boson operators
204: $\gamma_{q, \pm}$, 
205: which are linear combinations of the four boson operators $b_{q},b_{-q}^{\dagger},a_{q}$, and $a_{-q}^{\dagger}$;
206: \begin{eqnarray}
207: H=\sum_{q,\pm}\omega_{\pm}(q)\gamma^{\dagger}_{q, \pm}\gamma_{q, \pm}+H_{ZM}
208: \end{eqnarray}
209: where the dispersion relations for the two eigenmodes are 
210: \begin{eqnarray}
211: &&\omega_{\pm}(q)= \\
212: \nonumber
213: &&\sqrt{ \frac{ (v^{2}+v_{s}^{2})q^{2}+\omega_{B}^{2} \pm
214: \sqrt{ ((v^{2}-v_{s}^{2})q^{2}-\omega_{B}^{2} )^{2}+4\omega_{B}^{2}v^{2}q^{2} } } {2} }.
215: \end{eqnarray}
216: Now the ground state is the vacuum of new bosons, $\gamma_{q \pm}|0>=0$.
217: 
218: 
219: For detailed discussion, let us assume $v > v_{s} $ which is usually the case.
220: In Fig. 1, we plot dispersion relations for the two modes. 
221: As $q$ increases or $B$ decreases, the system goes to a weak coupling regime where
222:  the $`+'$ mode  becomes electronic ($\gamma_{q +} \sim b_{q}$,  $\omega_{+}(q) \sim vq$) and 
223: the $'-'$ mode becomes phononic ($\gamma_{q -} \sim a_{q}$, $\omega_{-}(q) \sim \tilde{\omega}_{s}(q)$).
224: In the opposite limit where $q$ is small and $B\neq 0$, the system is in a
225: strong coupling regime where 
226:  the two uncoupled modes undergo level crossing as shown in Fig. 1 and eventually completely exchange their identity 
227: in $q\rightarrow 0$ limit,
228: i.e., $\gamma_{q,+} \rightarrow a_{q}$ and $\gamma_{q,-}\rightarrow b_{q}$.
229: It is instructive to note that electric current can flow 
230: through the '+' modes even in $q\rightarrow 0$ limit, where the '+' modes become phononic.
231: This is because of the 'diamagnetic' contribution of
232: electric current  $-v_{F}e^{2}Bu(x)$ in Eq.(\ref{current}).
233: 
234: In the strong coupling regime, the dispersion relation of the '+' mode is in the same form with
235:  plasmon dispersion relation, 
236: $\omega_{+}(q) \sim \sqrt{v^{2}q^{2}+\omega_{B}^{2}}$ ($q \ll \omega_{B}/v$).
237: In contrast to the plasmon\cite{mahan}, however, the frequency gap $\omega_{B}$  originates from
238: the back-reaction which gives mass to phonon at long-wavelength limit,
239: $\tilde{\omega}_{s}(q) \rightarrow \omega_{B}$. 
240: The '$-$' mode is gapless  $\omega_{-}(q)\sim v_{s}vq^{2}/\omega_{B}$ ($q \ll \omega_{B}/v_{s}$). 
241: This mode is also charged at small $q$, 
242: where electronic and phononic excitation are entangled with each other. 
243: These excitations at long wavelengths are non-perturbative and can not be captured by perturbative
244: expansions in $B$.
245: 
246: 
247: Now let us consider experiments where the mechanical quantum fluctuations can be observed. 
248: In Fig. 2, we depict a system we consider.
249: An external electrical field $V_{e}(t)/d$ is applied along the wire in $0<x<d$ by nearby external metallic
250: gates with oscillating voltage difference $V_{e}(t)=V_{e}(\omega)e^{i\omega t}$.
251: If the current $I(\omega)$ is measured far away from the regime $0<x<d$, 
252: the current will not be sensitive to the position of the contacts so can be given by its averaged value 
253:  $I(\omega) \approx \frac{1}{d}\int_{0}^{d}dx J(x,\omega)$. 
254: From the linear response theory, one gets  
255: \begin{eqnarray}
256: \alpha(\omega)&\equiv&\frac{I(\omega)}{V_{e}(\omega)}=\frac{1}{i\omega d^{2}}\int_{0}^{d}dx\int_{0}^{d}dx^{\prime} \Pi(x,x^{\prime},\omega),
257: \label{gw}
258: \end{eqnarray}
259: where
260:   $\Pi(x,x^{\prime}\omega)=i\int_{0}^{\infty}ds < [ J(x,0),J(x^{\prime},-s)]>e^{i\omega s}$ is the current-current
261: correlation function.
262: 
263: 
264: This form is equivalent to the conductance formula $G(\omega)$ used by Kane and Fisher\cite{kane} 
265: but caution is needed here.
266: In our case,  additional electromotive force $V_{\rm emf}(\omega)$ is induced along the wire when $B \neq 0$.
267: In usual two-terminal conductance measurements, since the measured voltage $V(\omega)$ is given by
268: $V_{\rm emf}(\omega)+V_{e}(\omega)$, the two-terminal conductance is related to our $\alpha(\omega)$
269: as $I(\omega)=G(\omega)(V_{e}(\omega)+V_{\rm emf}(\omega))=\alpha(\omega)V_{e}(\omega)$.
270: 
271: The real part of $\alpha(\omega)$ should be understood as an optical absorption spectrum.
272: After averaging over a time period $2\pi/\omega$,
273: the power $P(\omega)$ 
274: absorbed from the external electric field to the wire
275:  is given by $P(\omega)=\frac{1}{2}{\rm Re}\int_{0}^{d}dx (V^{*}_{e}(\omega)/d)J(x,\omega) 
276: =\frac{1}{2} {\rm Re}\alpha(\omega)|V_{e}(\omega)|^{2}$.
277: 
278: \begin{figure}
279: \includegraphics[width=8.0cm,height=5cm]{setup}
280:  \caption{ 
281: Schematic figure for the experimental set-up in consideration. 
282: The wire is on an insulating layer above two metallic gates denoted by M.  } 
283: \end{figure}
284: 
285: To evaluate conveniently the current correlation function, we switch to a Lagrangian formalism
286: following Kane and Fisher\cite{kane}.
287: The Hamiltonian in Eq.(\ref{hamiltonian-boson}) is equivalent to the following Euclidean action,
288: \begin{eqnarray}
289: \nonumber
290: S&=& 
291: \frac{v}{2g} \int d\tau dx 
292: \left( (\partial_{x} \theta)^{2} 
293: +\frac{1}{v^{2}}(\partial_{\tau} \theta )^{2} \right)
294: \\
295: \label{action}
296: &+& \frac{\mathcal T}{2} \int d\tau dx
297: \left( (\partial_{x} u)^{2} 
298: +\frac{1}{v_{s}^{2}}(\partial_{\tau} u )^{2} \right)
299: \\ \nonumber
300: &-& B
301: \int d\tau dx J(x,\tau) u(x,\tau),
302: \end{eqnarray}
303: where $J(x,\tau)=ie\partial_{\tau}\theta(x,\tau)/\pi$.
304: Evaluation of the current correlation function 
305: $\Pi(x,x^{\prime},i\omega_{n})$
306: at Matsubara frequencies $\omega_{n}=2\pi n/\beta$
307: can be easily done using the above Euclidean action;
308: \begin{eqnarray} 
309: \nonumber
310: &&\Pi(x,x^{\prime},i\omega_{n})=\int_{0}^{\beta}d\tau e^{-i\omega_{n}\tau}<{\rm T}_{\tau}J(x,\tau)J(x^{\prime},0)> \\
311:             &=&
312: -\frac{e^{2}v_{F}}{2\pi^{2}}\int_{-\infty}^{\infty}dq e^{iq(x-x^{\prime})}
313: \frac{\omega_{n}^{2}} 
314: { q^{2}v^{2} +\omega_{n}^{2} 
315: +\omega_{B}^{2} \frac{\omega_{n}^{2}}{q^{2}v_{s}^{2} +\omega_{n}^{2}} }.
316: \end{eqnarray}
317: Now, by performing analytic continuation
318:   $\Pi(x,x^{\prime},\omega)=\Pi(x,x^{\prime},i\omega_{n} \rightarrow \omega +0^{+})$,
319: we get ${\rm Re} \alpha(\omega)$ from Eq.(\ref{gw});
320: \begin{eqnarray}
321: \nonumber
322: &&{\rm Re}\alpha(\omega)
323: \\ \nonumber
324: &=&
325: \frac{e^{2}}{4\pi} v_{F}  |\omega|
326: \int_{-\infty}^{\infty}dq \frac{\sin^{2}(qd/2)}{(qd/2)^{2}}
327: \frac{ \omega^{2}-(qv_{s})^{2}} {\omega^{2}-(\omega_{+}^{2}(q)+\omega_{-}^{2}(q))/2}
328: \\ &\times&
329: \left\{ 
330: \delta(\omega^{2}-\omega_{+}^{2}(q))
331: +\delta(\omega^{2}-\omega_{-}^{2}(q))
332: \right\}.
333: \end{eqnarray}
334: The result of the above integration is written
335: in terms of dimensionless parameters  $\tilde{\omega}=|\omega|d/v_{s}$,
336: $\tilde{B}=L \omega_{B}/v_{s}$, and $\eta=v_{s}/v$; 
337: ${\rm Re} \alpha(\omega)=g\frac{e^{2}}{h}\sum_{\pm}A_{\pm}(\tilde{\omega},\tilde{B},\eta)$.
338:  (Here we revived the Planck constant.)
339: The dimensionless function $A_{\pm}$  are 
340: \begin{eqnarray}
341: A_{\pm}=\pm \frac{\sin^{2}(\kappa_{\pm}/2)}{(\kappa_{\pm}/2)^{2}}
342: \frac{\eta(\tilde{\omega}^{2}-\kappa_{\pm}^{2})/\kappa_{\pm}}{\sqrt{(1-\eta^{2})^{2}\tilde{\omega}^{2}+4\tilde{B}^{2}} },
343: \end{eqnarray}
344: where $
345: \kappa_{\pm}=\sqrt{
346: \left(
347: (1+\eta^{2})\tilde{\omega}^{2}
348: \mp\tilde{\omega}\sqrt{(1-\eta^{2})^{2}\tilde{\omega}^{2}+4\tilde{B}^{2}}
349: \right)/2 }.$
350: 
351: 
352: \begin{figure}
353: \includegraphics[width=7.5cm,height=6cm]{absorpa01bm4}
354:  \caption{ 
355: The optical absorption spectrum  for $v_{s}/v=0.1$
356:  and $\omega_{B}d/v=10^{-4}$.
357: The inset shows the details at low frequencies.
358: In the low frequency regime (inset), the  maximum arises near 
359: $\omega \approx (v_{s}/v)\omega_{B}$. } 
360: \end{figure}
361: 
362: In Fig. 3, we plot the field absorption power ${\rm Re} \alpha (\omega)$ using the exact formula we obtained.
363: Since usually $v_{s} \ll v $ 
364: (for metallic carbon nanotubes, 
365:  $v_{s}\sim 15$ km/s $\ll v$ and $v \sim 10^{5}$ km/s ), let us assume $\eta \ll 1$.
366: In this case, the induced electric current
367: is strongly suppressed when $|\omega| < \omega_{B} $. 
368: %Considering experimentally relevant regime, we will also assume not too stong magnetic field
369: %i.e., $d\omega_{B}/v < 1$. 
370: There are three interesting frequency ranges.
371: (i)~ $\omega \ll  (v_{s}/v)\omega_{B}$. 
372: Here only '-' modes are excited which are mainly electronic density
373: fluctuations.
374: In this regime, using
375: $A_{-} \approx \frac{\eta}{2}\sqrt{\frac{\tilde{\omega}}{\tilde{B}}} \ll 1$ and $A_{+}=0$,
376: we get
377: \begin{eqnarray}
378: {\rm Re}\alpha(\omega) \approx g\frac{e^{2}}{h}\frac{\sqrt{v_{s}/v}}{2}
379: \sqrt{\frac{\omega}{\omega_{B}}}
380: % \ll g\frac{e^{2}}{h}
381: \end{eqnarray}
382: It is interesting to note the anomalous frequency and magnetic-field dependence 
383: ${\rm Re}\alpha(\omega)\propto \sqrt{\omega/B}$
384: which can not be obtained within a perturbation analysis.
385: (ii)~  $(v_{s}/v) \omega_{B} \ll \omega < \omega_{B}$. 
386: Here also only the '$-$' modes are excited while their nature becomes partially phonon-like
387: as the frequency increases. 
388:  In this regime, $A_{-} \approx 4\eta \tilde{B}^{2} \sin^{2}(\tilde{\omega}/2)$
389: and $A_{+}=0$.
390: In particular for $\tilde{\omega} < 1 $ (i.e., $\omega< v_{s}/L$), we get
391: \begin{eqnarray}
392: {\rm Re}\alpha(\omega) \approx g\frac{e^{2}}{h}(\frac{v_{s}}{v})^{3}(\frac{\omega_{B}}{\omega})^{2}
393: %\ll g\frac{e^{2}}{h} 
394: \end{eqnarray}
395: (iii)~ Finally when $\omega_{B} < \omega $,
396: the '+' modes play a significant role and $A_{+}$ has singularity near $\omega=\omega_{B}$.
397: Near this singular point, in spite of the phononic nature of the '+' modes, 
398: significant electric current may flow through its 'diamagnetic' contribution;
399: \begin{eqnarray}
400: {\rm Re}\alpha(\omega) \approx g\frac{e^{2}}{h}\frac{|\omega|}
401: {\sqrt{\omega^{2}-\omega_{B}^{2}}} 
402: \end{eqnarray}
403: The singularity of $\alpha(\omega)$ originates from the density of states of the 
404: massive boson mode $\omega_{+}(q)$ near $\omega=\omega_{B}$.
405: Similar to Landauer conductance quantization,
406: a quantization rule ${\rm Re}\alpha(\omega) \approx ge^{2}/h$ appears in weak coupling regime 
407: $\omega_{B} \ll \omega < v/d $. 
408: 
409: Important to all the physical consequences discussed in this work is
410: a characteristic frequency $\omega_{B}$ associated with the back-reaction 
411: \begin{eqnarray}
412: \omega_{B}=\sqrt{\frac{v_{F} e^{2} } {\rho_{m} \hbar \pi}} B
413: =0.88\times 10^{-2}B[{\rm T}]\sqrt{\frac{v_{F}[m/s]}{\rho_{m}[kg/m]}}
414: ~{\rm Hz},
415: \label{AYfrequency}
416: \end{eqnarray}
417: where we recovered $\hbar$.
418: As clear in its form, electron-electron interactions do not affect this characteristic frequency.
419: For single-walled carbon nanotubes, 
420: the linear mass density is estimated to be $\rho_{m}\sim 4 \times 10^{-13}$ kg/m when its radius is $\sim$ 4 nm .
421: Using a typical Fermi velocity $v_{F}=10^{5} m/s$, 
422: we estimate $ \omega_{B} \approx  4.4 ~B[{\rm T}] ~{\rm MHz} $, which is
423: in the experimentally accessible range. 
424: 
425: While we consider electro-mechanical coupling via Lorentz force,  
426: capacitive coupling to external gates\cite{kim, sapmaz} might be of interest in other contexts. 
427: In this case, the coupled charge density and phonon displacement is 
428: described by $\sim \int dx u(x)\partial_{x}\theta$. 
429: We assumed no external source of mechanical energy 
430: dissipation of the wire,
431: which is another subtle issue of electromechanical system\cite{ahn} 
432: and important to clarify 
433: the detailed operating conditions of quantum NEMS.
434: 
435: In summary, we outlined a theory for electrons in one-dimensional wires in
436: quantum mechanical motion.
437: Quantum mechanical displacement coupled to electric current
438: and the back reaction from electrons 
439: give rise to a current-carrying boson mode with an energy gap.
440: The induced  electric current 
441: by oscillating  electric field might reveal the elementary excitations in the system
442: through its frequency dependence. 
443: We presented here a non-perturbative regime where the induced electric current
444: has non-trivial frequency dependence. 
445: 
446: 
447: 
448: \begin{acknowledgements}
449: This work was supported by Korea Reseacrh Foundation (KRF-2003-070-C00020).
450: We thank the partial financial support of BK21 physics research division 
451: in Seoul National University where a part of this work was done (K.H.A.) and the
452: Swiss-Korean Outstanding Research Efforts Award Program (H.Y.). 
453: \end{acknowledgements}
454: \vspace*{-0.22 truein} 
455: \begin{thebibliography}{99} 
456: \bibitem{bishop} D. Bishop {\it et al.}, Physics Today {\bf 54},No 10, 38 (2001).
457: \bibitem{kim} P. Kim and C. M. Lieber, Science {\bf 286}, 2148 (1999).
458: \bibitem{rueckes} T. Rueckes et. al., Science {\bf 289}, 94 (2000).
459: \bibitem{tomonaga} S. Tomonaga, Prog. Theor. Phys. (Kyoto) {\bf 5}, 544 (1950).
460: \bibitem{luttinger} J.M. Luttinger, J. Math. Phys. N.Y. {\bf 4}, 1154 (1963).
461: \bibitem{haldane} F. D. M. Haldane, Phys. Rev. Lett. {\bf 47}, 1840 (1981).
462: \bibitem{cleland} A.N. Cleland and M.L. Roukes, Appl. Phys. Lett. {\bf 69}, 2653 (1996) 
463: ; Sens. Actuators A {\bf 72}, 256 (1999). 
464: \bibitem{mohanty} P. Mohanty et. al., Phys. Rev. B {\bf 66}, 085416 (2002).
465: \bibitem{chain} H. Onishi et. al.,Nature (London) {\bf 395}, 780 (1998);
466: A.Y. Yanson et. al., Nature (London) {\bf 395}, 783 (1998).
467: \bibitem{density} See e.g., G. Gr\"uner, {\it Density Waves in Solids}, 
468: Addison-Wesley Publishing Company (1994).
469: \bibitem{mahan} See e.g., G. D. Mahan, {\it Many-Particle Physics}, 
470: Plenum Press, New York (1990).
471: \bibitem{kane} C.L. Kane and M.P.A. Fisher, Phys. Rev. Lett. {\bf 68}, 1220 (1992).
472: \bibitem{sapmaz} S. Sapmaz {\it et al.}, Phys. Rev. B {\bf 67}, 235414 (2003).
473: \bibitem{ahn} K.-H. Ahn and P. Mohanty, Phys. Rev. Lett. {\bf 90}. 085504 (2003).
474: 
475: 
476: \end{thebibliography}
477: 
478: 
479: 
480: \end{document}
481: 
482: