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 and . Quantum effects emerge when the displacement is of the order of picometers, although the size of a buckled plate is of the order of . 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 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 (0), when it is buckled rightward (leftward).
We push a plate from both the ends. The position along the axis is determined as by minimizing the double-well potential[21, 24] ,
Explicit representations of and 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
where the Hamiltonian is
together with the double-well potential (1). In what follows, we use and 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 () and two-fold degenerated levels for large (). What is unexpected is a sharp transition of the spectrum at as a function of . We propose to use the lowest two-fold degenerated states at as a qubit. Corresponding wavefunctions are shown in Supplemental Material III.

Figure 1. (a1) The energy spectrum as a function of . The lowest six energy levels are shown. (a2) The logarithm plot of the energy difference between the ground state and the first excited state. (b1) The absolute value and (b2) The phase difference . The horizontal axis is the position in units of in (a1) and (a2), and the time in units of in (b1) and (b2). We have set .
We represent the state by the wave function localized at the right bottom and the state by the wave function localized at the left bottom. Their degeneracy is resolved for , where the ground state is well described by the symmetric state . We propose to use this transition of the level splitting at for a NOT gate operation.
Actually, the two-fold degeneracy is slightly broken for finite . The energy difference between the ground state and the first-excited state is calculated. The is plotted in Figure 1(a2). It is found that the energy difference is as tiny as at .
The potential is expanded by the harmonic potential
in the vicinity of with , where the ground-state wave function is given by
with the characteristic frequency and the ground state energy .
Universal quantum gates. It is known that a set of the 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 and NOT gates. Our scenario reads as follows. Let us start with either the state or at . For definiteness we take . When we change adiabatically the stable position from to , the state is moved to the symmetric state. Then, we change adiabatically the position back to the point . More explicitly, by pushing the plate from the both ends, we temporally control the stable position according to a smooth function,
with three parameters , and . The resultant state needs not be the state or but can be a combination of and 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 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 after enough time of the gate operation at . We show the amplitudes and the phase shifts as a function of in Figure 1(b1) and (b2), respectively. We have found that the amplitude changes as a function of significantly. We also find that the phase difference is and shown in Figure 1(a2). The jumps occur where .
gate: We first construct the square-root NOT gate,
which satisfies . We observe in Figure 1(b1) that the amplitude at the initial position colored in magenta decreases for , and it becomes identical to colored in cyan at . 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 is shown in Figure 2(aFigure 3. We find that the wave function becomes stationary after the gate operation for . Precisely in the same way, the equal splitting occurs when we start from the initial state given by Eq. (5) localized at the left-hand side. This gate operation at is summarized as .

Figure 2. (a) The gate, where we have set . (b) The NOT gate, where we have set . (a1) and (b1) The time evolution of given in Eq. (6) in units of . (a2) and (b2) The time evolution of the spatial profile of the absolute value of the wave function starting from the localized state at the right hand side. (a3) and (b3) The time evolution of , where colored in magenta and is colored in cyan. We have set , and . The horizontal axis is time ranging .
Figure 3. The wave functions and the energy spectrum under electric field. (a) and (b) . The ground state wave function is colored in red, while the first-excited states wave function is colored in cyan. We have set in units of .
NOT gate: Next, we construct the NOT gate . We observe in Figure 1(bFigure 1 that the amplitude at the initial position colored in magenta becomes zero at . 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 is shown in Figure 2(bFigure 3. This is the NOT gate . We find that the wave function becomes stationary after the gate operation for .
Construction of phase-shift and Pauli-Z gates. We proceed to construct the phase-shift gate. We apply an electric field along the direction to the buckled plate, where the potential is given by . The potential and eigenfunctions under electric field, , 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 , 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 in the case of (a) and (c) . (b) The energy difference between the ground state and the first excited state in the case of . The horizontal axis is in units of
estimated as
which is consistent with the numerical results shown in Figure 4(c). Namely, the energy-shift (8) is represented by the effective Hamiltonian .
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
as shown in Figure 5(a). We found that the absolute value does not change. However, the phase is modulated as in Figure 5(b) for the state . The phase modulation for the state is precisely opposite of that of the state . Hence, the result is summarized as the unitary operator

Figure 5. (a) The time dependence of the applied electric field according to (9) . (b) The time evolution of the phase difference compared with the phase evolution without electric field. The color indicates the time ranging from colored in red to colored in cyan. We have set . The horizontal axis is time ranging . We have set .
where is determined as a function of as illustrated in Figure 5(b). This is the phase-shift gate by angle .
phase-shift gate: The phase-shift gate is realized by the rotation (10) with the angle as up to the overall phase factor .
Pauli-Z gate. The Pauli-Z gate is realized by the rotation with the angle as in a similar way.
The Hadamard gate. The Hadamard gate is defined by . It is realized by a sequential application of the rotation and the rotation [30] as
Two-qubit phase-shift gate. Next, we construct two-qubit gates made of two buckled plates. When we apply the voltage between the two plates, the potential energy is given by
with the capacitance between the plates,
where and are the permittivity and the plate area, while is the distance between the two plates. We assume that the plate distance is very large compared with the displacement . We calculate
Then, the potential differences are given by
The detailed derivation is shown in Supplemental Material IV. We start with the Gaussian state with Eq. (5) localized at four points and , where , . The absolute value of this wave function almost remains as it is for a potential at , but a phase shift occurs. The unitary evolution is given by
for and ,
for and ,
for and , where we have added the zero-point energy.
It corresponds to the two-qubit phase-shift gate
by identifying the qubit state .
Ising gate: The Ising gate , by setting up to the global phase .
CZ gate: The controlled-Z (CZ) gate is a unitary operation acting on two adjacent qubits defined by . We construct it by a sequential application of the Ising gate and the one-qubit phase-shift gates as[31]
By using the results (S19) and (18), it is enough to set 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 . 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 . The displacement is of the order of . The mass is of the order of . The characteristic frequency is of the order of , or . 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 and the spring constant placed along the axis. The form of the buckled plate is determined by the Euler-Bernoulli equation[1, 2],
with , where is the axial load, is Young’s modulus of the beam material and 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
where is the position of the supporting point along the axis.
The lowest energy solution is given by[3]
where is the position along the axis. The length of the buckled plate is given by . It is calculated as
where is the complete elliptic integral of the second kind defined by
The potential energy is given by the Hooke law
which we show for various in Figure S1. It is a double-well potential. The potential minimum is shown as a function of in Figure S1.

Figure S1. Illustration of a buckled plate with (a1) fixed boundary condition and (a2) free boundary condition. (b) Electric field is applied along the 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 by fixing and as
It is summarized as with
The parameter is determined as by minimizing the potential . 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
whose solution is [4]
The length 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
and
in terms of the dimensionless time and the dimensionless parameters and , where and 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 .
Figure S2. Wave functions and energy spectrum with varying stress. (a) , (b) and (c) . The horizontal axis is . 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 and is well described by
where is the distance between the adjacent plates without buckling, is the area of the plate, and is the permittivity. The electrostatic potential is given by
when we control the voltage between the plates.
We consider a set of two adjacent plates, where the potential energy is given by
We show the potential in the - plane in Figure S3. In the absence of the applied voltage , there are four-fold degenerated bottoms at and as shown in Figure S3(a), where the ground state energy is . Under applied voltage, they are split as

Figure S3. Density plot of energy spectrum (S15) in the - plane. (a) Zero voltage . (b) Nonzero voltage . We have set and .
We also plot the potential profile with nonzero voltage in Figure S3. We assume that the plate distance is very large compared with the deviation and obtain
CNOT gates
The CNOT gate is defined by
where the first qubit is the controlled qubit and the second qubit is the target qubit.
References
References
- [1]R. Feynman, Simulating physics with computers, Int. J. Theor. Phys. 21, 467 (1982).
- [2]P. DiVincenzo, Quantum Computation, Science 270, 255 (1995).
- [3]M. Nielsen and I. Chuang, "Quantum Computation and Quantum Information", Cambridge University Press, (2016); ISBN 978-1-107-00217-3.
- [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]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]D. Loss and P. DiVincenzo, Quantum computation with quantum dots, Phys. Rev. A 57, 120 (1998).DOI
- [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]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]B. E. Kane, A silicon-based nuclear spin quantum computer, Nature 393, 133 (1998).DOI
- [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]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]J. Xia, X. Zhang, X. Liu, Y. Zhou, M. Ezawa, Qubits based on merons in magnetic nanodisks, arXiv:2205.04716.DOI
- [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]R. Reddy, K. Komeda, Y. Okamoto, E. Lebrasseur, A. Higo, Y. Mita, Sensors and Actuators A: Physical, 295, 1 (2019).
- [15]H. Toshiyoshi, Electrostatic Actuation. In: Yogesh B Gianchandani, Osamu Tabata, Hans Zappe, (Editors in Chief) Comprehensive Microsystems, 2, 1 Amsterdam: Elsevier. (2008).
- [16]H. G. Craighead, Science 290, 1532 (2000).
- [17]K. L. Ekinci and M. L. Roukes, Rev. Sci. Instruments 76, 6 (2005).
- [18]M. Blencowe, Physics Reports 395, 159 (2004)
- [19]M. Poot, H. S.J. van der Zant, Physics Reports 511, 273 (2012).
- [20]O. Slowik, K. Orlowska, D. Kopec, P. Janus, P. Grabiec and T. Gotzalk, Measurement Automation Monitoring, Mar., 62, 2450 (2016).
- [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.
- [23]Y. Intraparsonk and S. Fan, Appl. Phys. Lett. 98, 241104 (2011).
- [24]J. Martinez-Rincon and Y. V. Pershin, IEEE Transactions on Electron Devices 58, 1809 (2011).
- [25]H. J. Korsch and M. Gluck Eur. J. Phys. 23 413 (2002).
- [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
- [27]D. Deutsch, Quantum theory, the Church Turing principle and the universal quantum computer, Proceedings of the Royal Society A. 400, 97 (1985).DOI
- [28]C. M. Dawson and M. A. Nielsen, The Solovay-Kitaev algorithm, quant-ph/arXiv:0505030.
- [29]M. Nielsen and I. Chuang, "Quantum Computation and Quantum Information", Cambridge University Press, Cambridge, UK (2010).
- [30]N. Schuch and J. Seiwert, Natural two-qubit gate for quantum computation using the XY interaction, Phys. Rev. A 67, 032301 (2003).DOI
- [31]Y. Makhlin, Quant. Info. Proc. 1, 243 (2002).


