betaFeedback

WorksPhysics and Astronomy

Universal quantum computation based on Nano-Electro-Mechanical Systems

Abstract

We propose to use a buckled plate as a qubit, where a double-well potential is mechanically produced by pushing the plate from both the sides. The right and left positions of the plate are assigned to be quantum states ∣0⟩|0\rangle and ∣1⟩|1\rangle . Quantum effects emerge when the displacement is of the order of picometers, although the size of a buckled plate is of the order of 1μm1μm. The NOT gate is executed by changing the buckling force acting on the plate, while the Pauli-Z gate and the phase-shift gate are executed by applying electric field. A two-qubit phase shift gate is materialized with the use of an electrostatic potential. They constitute a set of universal quantum gates. An examination of material parameters leads to a feasibility of a NEMS(Nano-Electro-Mechanical System)-based quantum computer.

Introduction

Introduction. According to Moor’s law, elements of integrated circuits become exponentially small as a function of year. The size will become the order of nanometers within 10 years, where quantum mechanical effects are inevitable. For example, the superposition of states and the entanglement occur, which are absent in classical mechanics. It is impossible to decrease the size of elements smaller than 1nm, which is a typical scale of atoms. This is the end of Moore’s law. Quantum computation[1, 2, 3] is a candidate of "More than Moore", which resolves the limit of Moor’s law. It gives an exponential speed up for some algorithms. The problem is how to materialize a qubit based on actual materials. Various proposals are made such as superconductors [4], photonic systems [5], quantum dots [6], trapped ions [7], and nuclear magnetic resonance [8, 9]. Recently, nanoscale-skyrmion-based qubits[10, 11] and meron-based qubits[12] are also proposed.

Micro-Electro-Mechanical System (MEMS) is one of the basic elements in the current technology[13, 14, 15]. They use electrostatic energy to induce mechanical motions. If the size becomes of the order of nanometer, they are called Nano-Electro-Mechanical System (NEMS)[16, 17]. It has been demonstrated[18, 19, 20] that quantum effects emerge in the oscillation modes of a cantilever when its sample size is of the order of 100 nm∼1 μm100\,\mathrm{nm}\sim1\,\mu\mathrm{m} but with the displacement being of the order of picometers. It is described by a quantum harmonic oscillator. Carbon nanotubes, DNAs or biomolecules are used to compose elements in NEMS. Quantum effects have also been observed for a buckled beam made of a carbon nanotube as in the case of a cantilever .

The buckled plate has two stable positions. It can be used as a classical bit. It has been proposed that an Ising annealing machine is executable by a series of the buckled plates[21]. Its mechanism is based on the electrostatic potential inducing the Ising interaction between two adjacent plates.

A buckled plate would also reveal quantum effects when the displacement is of the order of picometers. In this paper, we propose to use it as a quantum bit, which is well described by a double-well potential. We then propose how to construct a set of universal quantum gates based on buckled plate MEMS, which consists of the phase-shift gate, the Hadamard gate and the CNOT gate. They are constructed by tuning the tension, applying electric field and voltage. A merit is that it is not necessary to use external magnetic field.

Buckled plate MEMS

In the field of MEMS, the bistable structure has been studied with a typical application to memories[1, 21, 23], where a plate is buckled. It was proposed[21] to use this buckled plate as the classical bit information 11 (0), when it is buckled rightward (leftward).

We push a plate from both the ends. The position along the xx axis is determined as x=ax=a by minimizing the double-well potential[21, 24] VDW(x)V_{\mathrm{DW}}(x),

VDW(x)=λ(x2−a2)2.(1)V_{\mathrm{DW}}(x)=\lambda\left(x^2-a^2\right)^2 . \tag*{(1)}

Explicit representations of λ\lambda and aa are given in terms of material parameters in Supplemental Material I.

This buckled plate is an example of a MEMS (NEMS), when its size is of the order of micrometers (nanometers).

Quantum NEMS. The dynamics of a buckled NEMS is described by the Schrödinger equation

iℏddtψ(x,t)=Hψ(x,t),(2)i\hbar\frac{d}{dt}\psi(x,t)=H\psi(x,t), \tag*{(2)}

where the Hamiltonian is

H=−ℏ22md2dx2+VDW(x),(3)H=-\frac{\hbar^2}{2m}\frac{d^2}{dx^2}+V_{\mathrm{DW}}(x), \tag*{(3)}

together with the double-well potential (1). In what follows, we use tu=(m2/ℏλ)1/3t_u=(m^2/\hbar\lambda)^{1/3} and xu=ℏ1/3/(mλ)1/6x_u=\hbar^{1/3}/(m\lambda)^{1/6} as the units of time and space, respectively, about which we explain in Fig.S2 in Supplemental Material II.

Qubit. We numerically solve the eigenenergies of the double-well system[25, 26], and obtain the energy spectrum as in Figure 1(a). As is well known, it consists of undegenerated levels for small aa (a/xu≪1a/x_u\ll1) and two-fold degenerated levels for large aa (a/xu≫1a/x_u\gg1). What is unexpected is a sharp transition of the spectrum at a≈1.5xua\approx1.5x_u as a function of aa. We propose to use the lowest two-fold degenerated states at a≳1.5xua\gtrsim1.5x_u as a qubit. Corresponding wavefunctions are shown in Supplemental Material III.

Energy spectrum, energy difference, absolute wave-function value, and phase difference

Figure 1. (a1) The energy spectrum as a function of aa. The lowest six energy levels are shown. (a2) The logarithm plot log⁡10(E1−E0)\log_{10}(E_1-E_0) of the energy difference between the ground state and the first excited state. (b1) The absolute value ∣ψ(±a0,t3)∣|\psi(\pm a_0,t_3)| and (b2) The phase difference arg⁡ψ(a0,t3)−arg⁡ψ(−a0,t3)\arg\psi(a_0,t_3)-\arg\psi(-a_0,t_3). The horizontal axis is the position aa in units of xux_u in (a1) and (a2), and the time tt in units of tut_u in (b1) and (b2). We have set T=tu/5T=t_u/5.

We represent the state ∣0⟩\lvert0\rangle by the wave function ψ+(x)\psi_+(x) localized at the right bottom and the state ∣1⟩\lvert1\rangle by the wave function ψ−(x)\psi_-(x) localized at the left bottom. Their degeneracy is resolved for a≲1.5xua\lesssim1.5x_u, where the ground state is well described by the symmetric state (ψ++ψ−)/2(\psi_+ + \psi_-)/\sqrt{2}. We propose to use this transition of the level splitting at a≈1.5xua\approx1.5x_u for a NOT gate operation.

Actually, the two-fold degeneracy is slightly broken for finite aa. The energy difference between the ground state and the first-excited state is calculated. The log⁡10(E1−E0)\log_{10}(E_1-E_0) is plotted in Figure 1(a2). It is found that the energy difference is as tiny as 10−14ℏλ/m10^{-14}\hbar\sqrt{\lambda/m} at a≈2.5xua\approx2.5x_u.

The potential is expanded by the harmonic potential

VDW(x)≃4a2λ(x∓a)2+o((x∓a)3)(4)V_{\mathrm{DW}}(x)\simeq4a^2\lambda(x\mp a)^2+o\left((x\mp a)^3\right) \tag*{(4)}

in the vicinity of x=±ax=\pm a with a≳2xua\gtrsim2x_u, where the ground-state wave function is given by

ψ±(x)=(mωπℏ)1/4exp⁡(−mω2ℏ(x∓a)2),(5)\psi_{\pm}(x)=\left(\frac{m\omega}{\pi\hbar}\right)^{1/4}\exp\left(-\frac{m\omega}{2\hbar}(x\mp a)^2\right), \tag*{(5)}

with the characteristic frequency ω=2a2λ/m\omega=2a\sqrt{2\lambda/m} and the ground state energy E0=ℏω/2E_0=\hbar\omega/2.

Universal quantum gates. It is known that a set of the π/4\pi/4 phase-shift gate, the Hadamard gate and the CNOT gate is enough for constructing any quantum circuits. It is known as the Solovay-Kitaev theorem of universal quantum computation[27, 28, 29]. We explicitly show that they are actually constructed in buckled NEMS.

Construction of NOT\sqrt{\mathrm{NOT}} and NOT gates. Our scenario reads as follows. Let us start with either the state ∣0⟩|0\rangle or ∣1⟩|1\rangle at a=a0≳2xua=a_0\gtrsim2x_u. For definiteness we take a0=3xua_0=3x_u. When we change adiabatically the stable position aa from a0a_0 to a=0a=0, the state is moved to the symmetric state. Then, we change adiabatically the position aa back to the point a0a_0. More explicitly, by pushing the plate from the both ends, we temporally control the stable position aa according to a smooth function,

a(t)=a02[tanh⁡t−t2T−tanh⁡t−t1T+2],(6)a(t)=\frac{a_0}{2}\left[\tanh\frac{t-t_2}{T}-\tanh\frac{t-t_1}{T}+2\right], \tag*{(6)}

with three parameters t1t_1, t2t_2 and TT. The resultant state needs not be the state ∣0⟩|0\rangle or ∣1⟩|1\rangle but can be a combination of ∣0⟩|0\rangle and ∣1⟩|1\rangle in general.

We study the dynamics of the wave packet by numerically solving the Schrödinger equation (2) with a time-dependent Hamiltonian (3), where the double-well potential (1) is time-dependent with the use of the time-dependent position (6).

We start from the initial state ψ+\psi_+ given by Eq. (5) localized at the right-hand side. In order to see the result of the gate operation, we focus on the state ψ(x,t)\psi(x,t) after enough time of the gate operation at t=t3≫t2t=t_3\gg t_2. We show the amplitudes ∣ψ(±a0,t3)∣|\psi(\pm a_0,t_3)| and the phase shifts arg⁡ψ(±a0,t3)\arg\psi(\pm a_0,t_3) as a function of t2t_2 in Figure 1(b1) and (b2), respectively. We have found that the amplitude changes as a function of t2t_2 significantly. We also find that the phase difference is π/2\pi/2 and 3π/23\pi/2 shown in Figure 1(a2). The jumps occur where ∣ψ(±a0,t3)∣=0|\psi(\pm a_0,t_3)|=0.

NOT\sqrt{\mathrm{NOT}} gate: We first construct the square-root NOT gate,

UNOT±=12(eiπ/4I2±e−iπ/4σx),(7)U_{\sqrt{\mathrm{NOT}}}^{\pm}=\frac{1}{\sqrt{2}}\left(e^{i\pi/4}I_2\pm e^{-i\pi/4}\sigma_x\right), \tag*{(7)}

which satisfies (UNOT±)2=±σx\left(U_{\sqrt{\mathrm{NOT}}}^{\pm}\right)^2=\pm\sigma_x. We observe in Figure 1(b1) that the amplitude ∣ψ(a0,t)∣|\psi(a_0,t)| at the initial position colored in magenta decreases for t2>26tut_2>26t_u, and it becomes identical to ∣ψ(−a0)∣|\psi(-a_0)| colored in cyan at t2NOT=27.02tut_2^{\mathrm{NOT}}=27.02t_u. This value of the parameter is special, where we study the time evolution of the spatial profile. The result is given in Figure 2(aFigure 2, where the wave packet is split equally to the right and left positions. The time evolution of the amplitude at x=±a0x=\pm a_0 is shown in Figure 2(aFigure 3. We find that the wave function becomes stationary after the gate operation for t≳40tut\gtrsim40t_u. Precisely in the same way, the equal splitting occurs when we start from the initial state ψ−\psi_- given by Eq. (5) localized at the left-hand side. This gate operation at t2NOTt_2^{\mathrm{NOT}} is summarized as UNOT±U_{\sqrt{\mathrm{NOT}}}^{\pm}.

Square-root NOT and NOT gate dynamics

Figure 2. (a) The NOT\sqrt{\mathrm{NOT}} gate, where we have set t2NOT=27.02tut_2^{\sqrt{\mathrm{NOT}}}=27.02t_u. (b) The NOT gate, where we have set t2NOT=28.87tut_2^{\mathrm{NOT}}=28.87t_u. (a1) and (b1) The time evolution of a(t)a(t) given in Eq. (6) in units of xux_u. (a2) and (b2) The time evolution of the spatial profile of the absolute value of the wave function ψ(x,t)\psi(x,t) starting from the localized state at the right hand side. (a3) and (b3) The time evolution of ∣ψ(±a0,t)∣|\psi(\pm a_0,t)|, where ∣ψ(a0,t)∣|\psi(a_0,t)| colored in magenta and ∣ψ(−a0,t)∣|\psi(-a_0,t)| is colored in cyan. We have set a0=3xua_0=3x_u, t1=20tut_1=20t_u and T=tu/5T=t_u/5. The horizontal axis is time ranging 0<t<60tu0<t<60t_u.

Figure 3. The wave functions and the energy spectrum under electric field. (a) a=xua=x_u and (b) a=2xua=2x_u. The ground state wave function is colored in red, while the first-excited states wave function is colored in cyan. We have set Ex=0.5E_x=0.5 in units of Eu≡ℏω/a0E_u\equiv\hbar\omega/a_0.

NOT gate: Next, we construct the NOT gate UNOT≡σxU_{\mathrm{NOT}}\equiv\sigma_x. We observe in Figure 1(bFigure 1 that the amplitude ∣ψ(a0,t)∣|\psi(a_0,t)| at the initial position colored in magenta becomes zero at t2NOT=28.87tut_2^{\mathrm{NOT}}=28.87t_u. This value of the parameter is also special, where the wave packet moves to the left position as shown in Figure 2(bFigure 2. The corresponding time evolution of the amplitude at x=±a0x=\pm a_0 is shown in Figure 2(bFigure 3. This is the NOT gate σx\sigma_x. We find that the wave function becomes stationary after the gate operation for t≳40tut\gtrsim40t_u.

Construction of phase-shift and Pauli-Z gates. We proceed to construct the phase-shift gate. We apply an electric field along the xx direction to the buckled plate, where the potential is given by ExxE_xx. The potential and eigenfunctions under electric field, VDW(x)+ExxV_{\mathrm{DW}}(x)+E_xx, are shown in Figure 3. We numerically evaluate the energy spectrum as a function of the electric field, which is shown in Figure 4(a). The energy difference between the ground state and the first-excited state monotonically increases as the increase of electric field as shown in Figure 4(b). For a0=3xua_0=3x_u, the energy spectrum changes linearly as a function of the electric field as shown in Figure 4(c). In the first-order perturbation theory, the energy is

Figure 4. The energy spectrum as a function of the applied electric field ExE_x in the case of (a) a=xua=x_u and (c) a=3xua=3x_u. (b) The energy difference between the ground state and the first excited state in the case of a=xua=x_u. The horizontal axis is ExE_x in units of Eu≡ℏω/a0E_u\equiv\hbar\omega/a_0

estimated as

⟨ψ+∣Exx∣ψ+⟩=±Exa0,(8)\langle\psi_+|E_xx|\psi_+\rangle=\pm E_xa_0, \tag*{(8)}

which is consistent with the numerical results shown in Figure 4(c). Namely, the energy-shift (8) is represented by the effective Hamiltonian HEx=a0ExσzH_{E_x}=a_0E_x\sigma_z.

We have numerically evaluated the dynamics of the wave packet starting from the Gaussian distribution (5) under the temporally controlled electric field according to the formula

Ex(t)=E02[tanh⁡t−t1T−tanh⁡t−t2T],(9)E_x(t)=\frac{E_0}{2}\left[\tanh\frac{t-t_1}{T}-\tanh\frac{t-t_2}{T}\right], \tag*{(9)}

as shown in Figure 5(a). We found that the absolute value ∣ψ∣|\psi| does not change. However, the phase is modulated as in Figure 5(b) for the state ∣ψ+⟩|\psi_+\rangle. The phase modulation for the state ∣ψ−⟩|\psi_-\rangle is precisely opposite of that of the state ∣ψ+⟩|\psi_+\rangle. Hence, the result is summarized as the unitary operator

Time dependence of the applied electric field and phase difference

Figure 5. (a) The time dependence of the applied electric field according to (9) ExE_x. (b) The time evolution of the phase difference compared with the phase evolution without electric field. The color indicates the time t2t_2 ranging from t2=2tut_2=2t_u colored in red to t2=3tut_2=3t_u colored in cyan. We have set t1=2tut_1=2t_u. The horizontal axis is time ranging 0<t<5tu0<t<5t_u. We have set T=tu/5T=t_u/5.

UZ(θ)=diag.⁡(e−iθ/2,eiθ/2)=exp⁡[−iθ2σz],(10)U_Z(\theta)=\operatorname{diag.}\left(e^{-i\theta/2},e^{i\theta/2}\right)=\exp\left[-\frac{i\theta}{2}\sigma_z\right], \tag*{(10)}

where θ\theta is determined as a function of t2t_2 as illustrated in Figure 5(b). This is the phase-shift gate by angle θ\theta.

π/4\pi/4 phase-shift gate: The π/4\pi/4 phase-shift gate UT≡diag.⁡(1,eiπ/4)U_T\equiv\operatorname{diag.}(1,e^{i\pi/4}) is realized by the zz rotation (10) with the angle θ=π/4\theta=\pi/4 as UT=e−iπ/8UZ(π4)U_T=e^{-i\pi/8}U_Z\left(\frac{\pi}{4}\right) up to the overall phase factor eiπ/8e^{i\pi/8}.

Pauli-Z gate. The Pauli-Z gate is realized by the zz rotation with the angle π\pi as UZ=−iUZ(π)U_Z=-iU_Z(\pi) in a similar way.

The Hadamard gate. The Hadamard gate is defined by UH≡(σz+σx)/2U_H\equiv(\sigma_z+\sigma_x)/\sqrt{2}. It is realized by a sequential application of the zz rotation and the xx rotation [30] as

UH=−iUZUNOTUZ.(11)U_H=-iU_ZU_{\mathrm{NOT}}U_Z. \tag*{(11)}

Two-qubit phase-shift gate. Next, we construct two-qubit gates made of two buckled plates. When we apply the voltage V1V_1 between the two plates, the potential energy is given by

V(x1,x2)≡VDW(x1)+VDW(x2)+Cpara(x1,x2)2V12,(12)V(x_1,x_2)\equiv V_{\mathrm{DW}}(x_1)+V_{\mathrm{DW}}(x_2)+\frac{C_{\mathrm{para}}(x_1,x_2)}{2}V_1^2, \tag*{(12)}

with Cpara(x1,x2)C_{\mathrm{para}}(x_1,x_2) the capacitance between the plates,

Cpara(x1,x2)=ε0SXcap+x1−x2,(13)C_{\mathrm{para}}(x_1,x_2)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}+x_1-x_2}, \tag*{(13)}

where ε0\varepsilon_0 and SS are the permittivity and the plate area, while XcapX_{\mathrm{cap}} is the distance between the two plates. We assume that the plate distance XcapX_{\mathrm{cap}} is very large compared with the displacement aa. We calculate

V(a,a)=V(−a,−a)=ε0SXcapV122≡E0,(14)V(a,a)=V(-a,-a)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}}\frac{V_1^2}{2}\equiv E_0, \tag*{(14)}
V(a,−a)=ε0SXcap+2aV122≡E+,(15)V(a,-a)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}+2a}\frac{V_1^2}{2}\equiv E_+, \tag*{(15)}
V(−a,a)=ε0SXcap−2aV122≡E−.(16)V(-a,a)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}-2a}\frac{V_1^2}{2}\equiv E_-. \tag*{(16)}

Then, the potential differences are given by

E+−E0≃−aXcap2ε0SV12=−EX,(17)E_+-E_0\simeq-\frac{a}{X_{\mathrm{cap}}^2}\varepsilon_0 S V_1^2=-E_X, \tag*{(17)}
E−−E0≃aXcap2ε0SV12=EX.(18)E_--E_0\simeq\frac{a}{X_{\mathrm{cap}}^2}\varepsilon_0 S V_1^2=E_X. \tag*{(18)}

The detailed derivation is shown in Supplemental Material IV. We start with the Gaussian state Ψσ1σ2(x1,x2)≡ψσ1(x1)ψσ2(x2)\Psi_{\sigma_1\sigma_2}(x_1,x_2)\equiv\psi_{\sigma_1}(x_1)\psi_{\sigma_2}(x_2) with Eq. (5) localized at four points x1=σ1ax_1=\sigma_1a and x2=σ2ax_2=\sigma_2a, where σ1=±\sigma_1=\pm, σ2=±\sigma_2=\pm. The absolute value of this wave function almost remains as it is for a potential at a/xu≫1a/x_u\gg1, but a phase shift occurs. The unitary evolution is given by

U(t)=exp⁡[−i(E0/ℏ+ω)t](19)U(t)=\exp[-i(E_0/\hbar+\omega)t] \tag*{(19)}

for σ1=σ2=+\sigma_1=\sigma_2=+ and σ1=σ2=−\sigma_1=\sigma_2=-,

U(t)=exp⁡[−i(E+/ℏ+ω)t](20)U(t)=\exp[-i(E_+/\hbar+\omega)t] \tag*{(20)}

for σ1=+\sigma_1=+ and σ2=−\sigma_2=-,

U(t)=exp⁡[−i(E−/ℏ+ω)t](21)U(t)=\exp[-i(E_-/\hbar+\omega)t] \tag*{(21)}

for σ1=−\sigma_1=- and σ2=+\sigma_2=+, where we have added the zero-point energy.

It corresponds to the two-qubit phase-shift gate

U2-phase(t)=diag.⁡(e−iE0ℏt,e−iE+ℏt,e−iE−ℏt,e−iE0ℏt)=e−iE0ℏtdiag.⁡(1,e−iEXℏt,eiEXℏt,1),(22)U_{\mathrm{2\text{-}phase}}(t)=\operatorname{diag.}\left(e^{-i\frac{E_0}{\hbar}t},e^{-i\frac{E_+}{\hbar}t},e^{-i\frac{E_-}{\hbar}t},e^{-i\frac{E_0}{\hbar}t}\right) =e^{-i\frac{E_0}{\hbar}t}\operatorname{diag.}\left(1,e^{-i\frac{E_X}{\hbar}t},e^{i\frac{E_X}{\hbar}t},1\right), \tag*{(22)}

by identifying the qubit state (∣00⟩,∣01⟩,∣10⟩,∣11⟩)t=(∣++⟩,∣+−⟩,∣−+⟩,∣−−⟩)t(|00\rangle,|01\rangle,|10\rangle,|11\rangle)^t=(|++\rangle,|+-\rangle,|-+\rangle,|--\rangle)^t.

Ising gate: The Ising gate UZZ≡diag.⁡(1,−1,−1,1)U_{ZZ}\equiv\operatorname{diag.}(1,-1,-1,1), by setting EXt/ℏ=πE_Xt/\hbar=\pi up to the global phase exp⁡[−iE0t/ℏ]\exp[-iE_0t/\hbar].

CZ gate: The controlled-Z (CZ) gate UCZU_{\mathrm{CZ}} is a unitary operation acting on two adjacent qubits defined by UCZ=diag.⁡(1,1,1,−1)U_{\mathrm{CZ}}=\operatorname{diag.}(1,1,1,-1). We construct it by a sequential application of the Ising gate and the one-qubit phase-shift gates as[31]

UCZ=eiπ/4UZ(π2)UZ(π2)UZZ,(23)U_{\mathrm{CZ}}=e^{i\pi/4}U_Z\left(\frac{\pi}{2}\right)U_Z\left(\frac{\pi}{2}\right)U_{ZZ}, \tag*{(23)}

By using the results (S19) and (18), it is enough to set ϕ1=π−ϕ2=EXt/ℏ\phi_1=\pi-\phi_2=E_Xt/\hbar to construct the CZ gate.

CNOT gate: The CNOT is constructed by a sequential application of the CZ gate (23) and the Hadamard gate (11) as UCNOT1→2=UH(2)UCZUH(2)U_{\mathrm{CNOT}}^{1\to2}=U_H^{(2)}U_{\mathrm{CZ}}U_H^{(2)}. See the definition of the CNOT gate in Supplemental Material V.

Discussions. We would like to address the feasibility of a NEMS-based quantum computer by examining typical material parameters[19]. The length of an element is of the order of 1μm∼100 μm1\mu\mathrm{m}\sim100\ \mu\mathrm{m}. The displacement is of the order of 0.1pm∼10fm0.1\mathrm{pm}\sim10\mathrm{fm}. The mass mm is of the order of 10−21kg∼10−14kg10^{-21}\mathrm{kg}\sim10^{-14}\mathrm{kg}. The characteristic frequency is of the order of κ/m\sqrt{\kappa/m}, or 1MHz∼1GHz1\mathrm{MHz}\sim1\mathrm{GHz}. Then, such a NEMS-based quantum computer would be realizable by improving the present technology.

M.E. is grateful to N. Nagaosa for helpful discussions on the subject. This work is supported by CREST, JST (Grants No. JPMJCR20T2).

Buckled plate MEMS

We consider a plate with length 2L02L_0 and the spring constant κ\kappa placed along the yy axis. The form of the buckled plate is determined by the Euler-Bernoulli equation[1, 2],

d4wdy4+m2d2wdy2=0,(24)\frac{d^4 w}{dy^4}+m^2\frac{d^2w}{dy^2}=0, \tag*{(24)}

with m2≡P/EIm^2 \equiv P/EI, where PP is the axial load, EE is Young’s modulus of the beam material and II is the second moment of area of the beam. In solving the Euler-Bernoulli equation for a clamped-clamped beam, we may use the fixed boundary condition or the free boundary condition.

First, we impose the fixed boundary condition, which reads

w(−y0)=w(y0)=0,dwdy∣x=−y0=dwdy∣x=y0=0,(25)w(-y_0)=w(y_0)=0,\qquad\left.\frac{dw}{dy}\right|_{x=-y_0}=\left.\frac{dw}{dy}\right|_{x=y_0}=0, \tag*{(25)}

where y0y_0 is the position of the supporting point along the yy axis.

The lowest energy solution is given by[3]

w=x02(1+cos⁡πyy0),(26)w=\frac{x_0}{2}\left(1+\cos\frac{\pi y}{y_0}\right), \tag*{(26)}

where x0x_0 is the position along the xx axis. The length of the buckled plate is given by L(x0)=∫−y0y0dy1+(dw/dy)2L(x_0)=\int_{-y_0}^{y_0}dy\sqrt{1+(dw/dy)^2}. It is calculated as

L(x0)=∫−y0y0dy1+π2x024y02sin⁡2πyy0=y0π∫−ππdθ1+π2x024y02sin⁡2θ=4y0π∫0π/2dθ1+π2x024y02sin⁡2θ=4y0πE(−iπx02y0),(27)\begin{aligned} L(x_0)&=\int_{-y_0}^{y_0}dy\sqrt{1+\frac{\pi^2x_0^2}{4y_0^2}\sin^2\frac{\pi y}{y_0}}=\frac{y_0}{\pi}\int_{-\pi}^{\pi}d\theta\sqrt{1+\frac{\pi^2x_0^2}{4y_0^2}\sin^2\theta}\\ &=\frac{4y_0}{\pi}\int_0^{\pi/2}d\theta\sqrt{1+\frac{\pi^2x_0^2}{4y_0^2}\sin^2\theta}=\frac{4y_0}{\pi}E\left(-i\frac{\pi x_0}{2y_0}\right), \tag*{(27)} \end{aligned}

where EE is the complete elliptic integral of the second kind defined by

E(k)≡∫0π/21−k2sin⁡2θ dθ.(28)E(k) \equiv\int_{0}^{\pi/2} \sqrt{1-k^2\sin^2\theta}\,\mathrm{d}\theta. \tag*{(28)}

The potential energy is given by the Hooke law

V(x0)=κ2(2L(x0)−2L0)2,(29)V(x_0)=\frac{\kappa}{2}\left(2L(x_0)-2L_0\right)^2, \tag*{(29)}

which we show for various y0y_0 in Figure S1. It is a double-well potential. The potential minimum is shown as a function of y0y_0 in Figure S1.

Illustrations of fixed and free boundary conditions, electric-field-induced potential, and electrostatic potential

Figure S1. Illustration of a buckled plate with (a1) fixed boundary condition and (a2) free boundary condition. (b) Electric field is applied along the xx direction in order to construct the Pauli-Z gate. (c) The electrostatic potential is induced by applying voltage in order to construct the two-qubit phase-shift gate.

We expand Eq. (29) in terms of x0x_0 by fixing L0L_0 and y0y_0 as

V(x0)=κ2(2L0−2y0)2+14y0(L0−y0)κπ2x02+(L0−3y0)256y03κπ4x04+o(x06).(30)V(x_0)=\frac{\kappa}{2}(2L_0-2y_0)^2+\frac{1}{4y_0}(L_0-y_0)\kappa\pi^2x_0^2+\frac{(L_0-3y_0)}{256y_0^3}\kappa\pi^4x_0^4+o(x_0^6). \tag*{(30)}

It is summarized as V(x0)=λ(x02−a2)2+V0+o(x06)V(x_0)=\lambda(x_0^2-a^2)^2+V_0+o(x_0^6) with

λ≡(L0−3y0)256y03κπ4,a≡y042πL0−y03L0−y0.(31)\lambda\equiv\frac{(L_0-3y_0)}{256y_0^3}\kappa\pi^4,\qquad a\equiv y_0\frac{4\sqrt{2}}{\pi}\sqrt{\frac{L_0-y_0}{3L_0-y_0}}. \tag*{(31)}

The parameter x0x_0 is determined as x0=ax_0=a by minimizing the potential V(x0)V(x_0). This buckled plate is an example of a MEMS (NEMS), when its size is of the order of micrometers (nanometers).

Second, we impose the fixed boundary condition, which reads

d2wdy2∣x=−y0=d2wdy2∣x=y0=0,(32)\left.\frac{d^2w}{dy^2}\right|_{x=-y_0}=\left.\frac{d^2w}{dy^2}\right|_{x=y_0}=0, \tag*{(32)}

whose solution is [4]

w=x0cos⁡πy2y0.(33)w=x_0\cos\frac{\pi y}{2y_0}. \tag*{(33)}

The length L(x0)L(x_0) is given by Eq. (27) precisely as in the case of the fixed boundary condition. Hence, the potential energy is given by the same Hooke law as Eq. (29).

Units of time and space

It is straightforward to rewrite the Schrödinger equation (2) and the Hamiltonian (3) in the main text as

iddτψ(X,τ)=Hψ(X,τ)(34)i\frac{d}{d\tau}\psi(X,\tau)=H\psi(X,\tau) \tag*{(34)}

and

H=−12(d/dX)2+(X2−A2)2(35)H=-\frac{1}{2}(d/dX)^2+(X^2-A^2)^2 \tag*{(35)}

in terms of the dimensionless time τ=t/tu\tau=t/t_u and the dimensionless parameters X≡x/xuX\equiv x/x_u and A≡a/xuA\equiv a/x_u, where tu=(m2/ℏλ)1/3t_u=(m^2/\hbar\lambda)^{1/3} and xu=ℏ1/3/(mλ)1/6x_u=\hbar^{1/3}/(m\lambda)^{1/6} give the units of time and space, respectively. We have carried out numerical analysis based on the dimensionless formulas.

Wavefunction

We also numerically determine the wave functions for a double-well potential, which are shown in Figure S2. They are well described by those of the harmonic potential for large aa.

Figure S2. Wave functions and energy spectrum with varying stress. (a) a=xua=x_{\mathrm{u}}, (b) a=2xua=2x_{\mathrm{u}} and (c) a=3xua=3x_{\mathrm{u}}. The horizontal axis is xx. The ground state wave function is colored in red, while the first-excited states wave function is colored in cyan.

Electrostatic energy

In the parallel-plate NEMS, the capacitance between the plates x1x_1 and x2x_2 is well described by

Cpara(x1,x2)=ε0SXcap+x1−x2,(36)C_{\mathrm{para}}(x_1,x_2)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}+x_1-x_2}, \tag*{(36)}

where XcapX_{\mathrm{cap}} is the distance between the adjacent plates without buckling, SS is the area of the plate, and ε0\varepsilon_0 is the permittivity. The electrostatic potential is given by

Ucap=Cpara(x1,x2)2V12,(37)U_{\mathrm{cap}}=\frac{C_{\mathrm{para}}(x_1,x_2)}{2}V_1^2, \tag*{(37)}

when we control the voltage V1V_1 between the plates.

We consider a set of two adjacent plates, where the potential energy is given by

V(x1,x2)≡V(x1)+V(x2)+Cpara(x1,x2)2V12.(38)V(x_1,x_2)\equiv V(x_1)+V(x_2)+\frac{C_{\mathrm{para}}(x_1,x_2)}{2}V_1^2. \tag*{(38)}

We show the potential in the x1x_1-x2x_2 plane in Figure S3. In the absence of the applied voltage V1=0V_1=0, there are four-fold degenerated bottoms at u1=±au_1=\pm a and u2=±au_2=\pm a as shown in Figure S3(a), where the ground state energy is ℏω\hbar\omega. Under applied voltage, they are split as

Density plot of energy spectrum in the $x_1$-$x_2$ plane

Figure S3. Density plot of energy spectrum (S15) in the x1x_1-x2x_2 plane. (a) Zero voltage V1=0V_1=0. (b) Nonzero voltage ε0SV12/2=100\varepsilon_0 S V_1^2/2=100. We have set Xcap=10xuX_{\mathrm{cap}}=10x_{\mathrm{u}} and a=2xua=2x_{\mathrm{u}}.

V(a,a)=V(−a,−a)=ε0SXcapV122≡E0,(39)V(a,a)=V(-a,-a)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}}\frac{V_1^2}{2}\equiv E_0, \tag*{(39)}
V(a,−a)=ε0SXcap+2aV122≡E+,(40)V(a,-a)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}+2a}\frac{V_1^2}{2}\equiv E_+, \tag*{(40)}
V(−a,a)=ε0SXcap−2aV122≡E−.(41)V(-a,a)=\frac{\varepsilon_0 S}{X_{\mathrm{cap}}-2a}\frac{V_1^2}{2}\equiv E_-. \tag*{(41)}

We also plot the potential profile with nonzero voltage in Figure S3. We assume that the plate distance XcapX_{\mathrm{cap}} is very large compared with the deviation aa and obtain

E+−E0≃−aXcap2ε0SV12,E−−E0≃aXcap2ε0SV12.(42)E_{+}-E_{0} \simeq-\frac{a}{X_{\mathrm{cap}}^{2}}\varepsilon_{0}SV_{1}^{2}, \qquad E_{-}-E_{0} \simeq\frac{a}{X_{\mathrm{cap}}^{2}}\varepsilon_{0}SV_{1}^{2}. \tag*{(42)}

CNOT gates

The CNOT gate UCNOT1→2U_{\mathrm{CNOT}}^{1\to2} is defined by

UCNOT1→2≡(1000010000010010),(43)U_{\mathrm{CNOT}}^{1\to2} \equiv\begin{pmatrix} 1 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 \\ 0 & 0 & 0 & 1 \\ 0 & 0 & 1 & 0 \end{pmatrix}, \tag*{(43)}

where the first qubit is the controlled qubit and the second qubit is the target qubit.

References

References

  1. [1]R. Feynman, Simulating physics with computers, Int. J. Theor. Phys. 21, 467 (1982).
  2. [2]P. DiVincenzo, Quantum Computation, Science 270, 255 (1995).
  3. [3]M. Nielsen and I. Chuang, "Quantum Computation and Quantum Information", Cambridge University Press, (2016); ISBN 978-1-107-00217-3.
  4. [4]Y. Nakamura; Yu. A. Pashkin; J. S. Tsai, Coherent control of macroscopic quantum states in a single-Cooper-pair box, Nature 398, 786 (1999).
  5. [5]E. Knill, R. Laflamme and G. J. Milburn, A scheme for efficient quantum computation with linear optics, Nature, 409, 46 (2001).pubmed.ncbi.nlm.nih.gov/11343107
  6. [6]D. Loss and P. DiVincenzo, Quantum computation with quantum dots, Phys. Rev. A 57, 120 (1998).DOI
  7. [7]J. I. Cirac and P. Zoller, Quantum Computations with Cold Trapped Ions, Phys. Rev. Lett. 74, 4091 (1995).pubmed.ncbi.nlm.nih.gov/10058410
  8. [8]L. M.K. Vandersypen, M. Steffen, G. Breyta, C. S. Yannoni, M. H. Sherwood, I. L. Chuang, Experimental realization of Shor’s quantum factoring algorithm using nuclear magnetic resonance, Nature 414, 883 (2001).pubmed.ncbi.nlm.nih.gov/11780055
  9. [9]B. E. Kane, A silicon-based nuclear spin quantum computer, Nature 393, 133 (1998).DOI
  10. [10]C. Psaroudaki and C. Panagopoulos, Skyrmion Qubits: A New Class of Quantum Logic Elements Based on Nanoscale Magnetization, Phys. Rev. Lett. 127, 06720 (2021).DOI
  11. [11]J. Xia, X. Zhang, X. Liu, Y. Zhou, M. Ezawa, Universal quantum computation based on nanoscale skyrmion helicity qubits in frustrated magnets arXiv:2204.04589.DOI
  12. [12]J. Xia, X. Zhang, X. Liu, Y. Zhou, M. Ezawa, Qubits based on merons in magnetic nanodisks, arXiv:2205.04716.DOI
  13. [13]Y. Mita, A. Hirakawa, B. Stefanelli, I. Mori, Y. Okamoto, S. Morishita, M. Kubota, E. Lebrasseur, A. Kaiser, Japanese Journal of Applied Physics 57, 04FA05 (2018).
  14. [14]R. Reddy, K. Komeda, Y. Okamoto, E. Lebrasseur, A. Higo, Y. Mita, Sensors and Actuators A: Physical, 295, 1 (2019).
  15. [15]H. Toshiyoshi, Electrostatic Actuation. In: Yogesh B Gianchandani, Osamu Tabata, Hans Zappe, (Editors in Chief) Comprehensive Microsystems, 2, 1 Amsterdam: Elsevier. (2008).
  16. [16]H. G. Craighead, Science 290, 1532 (2000).
  17. [17]K. L. Ekinci and M. L. Roukes, Rev. Sci. Instruments 76, 6 (2005).
  18. [18]M. Blencowe, Physics Reports 395, 159 (2004)
  19. [19]M. Poot, H. S.J. van der Zant, Physics Reports 511, 273 (2012).
  20. [20]O. Slowik, K. Orlowska, D. Kopec, P. Janus, P. Grabiec and T. Gotzalk, Measurement Automation Monitoring, Mar., 62, 2450 (2016).
  21. [21]M. Ezawa, Eric Lebrasseur and Yoshio Mita, Nonvolatile bistable memory and Ising machine based on Micro-Electro-Mechanical Systems, app-ph/arXiv:2106.09931.
  22. [23]Y. Intraparsonk and S. Fan, Appl. Phys. Lett. 98, 241104 (2011).
  23. [24]J. Martinez-Rincon and Y. V. Pershin, IEEE Transactions on Electron Devices 58, 1809 (2011).
  24. [25]H. J. Korsch and M. Gluck Eur. J. Phys. 23 413 (2002).
  25. [26]Solving the eigenvalue problem for a double well potential using a 1D particle in a box as a basis set https://mathematica.stackexchange.com/questions/187471/solving-the-eigenvalue-problem-for-a-double-well-potential-using-a-1d-particle-i
  26. [27]D. Deutsch, Quantum theory, the Church Turing principle and the universal quantum computer, Proceedings of the Royal Society A. 400, 97 (1985).DOI
  27. [28]C. M. Dawson and M. A. Nielsen, The Solovay-Kitaev algorithm, quant-ph/arXiv:0505030.
  28. [29]M. Nielsen and I. Chuang, "Quantum Computation and Quantum Information", Cambridge University Press, Cambridge, UK (2010).
  29. [30]N. Schuch and J. Seiwert, Natural two-qubit gate for quantum computation using the XY interaction, Phys. Rev. A 67, 032301 (2003).DOI
  30. [31]Y. Makhlin, Quant. Info. Proc. 1, 243 (2002).

Paper details

vlnc:W_g9vxmn34
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents