METHOD AND DEVICE FOR QUANTUM COMPUTING WITH MAJORANA MODES

Abstract

A quantum computing device includes: a microwave cavity (CH) having an input port (PE) and an output port (PS) that are separate or coincident;a superconducting electronic device (DS) capacitively coupled to the microwave cavity and configured to support a chain of 2N Majorana modes (MM1-MM4, MM2′, MM3′), N being a positive integer; andcoupling means for applying microwave excitations between each pair of adjacent Majorana modes of the chain.

Description

FIELD OF THE INVENTION

The invention lies in the field of quantum computing, and more generally of quantum information.

BACKGROUND

Majorana fermions are hypothetical spin-½ elementary particles that are their own antiparticle. More recently, this term—or expressions such as “Majorana states”, “Majorana zero modes”—have been used to designate a specific type of quasi-particle in solids, and notably in topological superconductors. A Majorana mode corresponds to a zero-energy excitation (hence the term “zero mode”) that may occur in particular in correspondence with a vortex in a topological superconductor.

By definition, a Majorana mode is represented by a self-adjoint operator:

γ₀^†=γ₀

It will be understood that an isolated Majorana mode does not make it possible to define a creation or annihilation operator, since these two operators would coincide, which is contradictory. The difficulty may however be overcome by considering two Majorana modes, identified by indices (1) and (2). It is then possible to define operators

$\begin{array}{cc}{c}_{12}^{†}=\frac{{\gamma }_0^{\left(1\right)}+i{\gamma }_0^{\left(2\right)}}{2},c_{12}=\frac{{\gamma }_0^{\left(1\right)}+i{\gamma }_0^{\left(2\right)}}{2}& \left(1\right)\end{array}$

that satisfy the usual anti-commutation relation of fermions:

{c₁₂^†, c₁₂}=1

and may therefore be considered, respectively, to be a creation and annihilation operator. Since two Majorana modes are needed to define a creation or annihilation operator, such a mode is generally said to constitute “a half-fermion”.

The fact that a pair of separate modes is necessary to define the creation and annihilation operators of a fermion introduces a non-local quantum correlation between these modes, which drastically modifies their quantum nature. Therefore, the exchange of two Majorana modes (in real space or in an appropriate parameter space) produces a state of the same energy as the initial state that is related to the initial state not by a simple phase factor, as for bosons or fermions, but by a unitary transformation:

$U_{\mathrm{ij}}=\exp \left(-\frac{\pi }{4}{\gamma }_0^{\left(i\right)}{\gamma }_0^{\left(j\right)}\right)$

If consideration is given to three Majorana modes with indices (1), (2) and (3) and consideration is given to the exchanges (1)-(2) and (2)-(3), it is possible to demonstrate that the operators U₁₂ and U₂₃ do not commute. It is then said that the exchange process is non-abelian.

Majorana mode exchanges are also called “braiding” operations because the world lines representing these modes in spacetime form a kind of braid.

It has been demonstrated that a set of four Majorana modes makes it possible to encode a qubit and that the braiding operations make it possible to produce quantum gates called “Clifford gates”, which are useful for performing quantum computations. Other operations on Majorana modes also make it possible to manipulate qubits, notably to “fuse” them (which is equivalent to a projective measurement).

For a more detailed introduction to Majorana modes and to the application thereof to quantum computing, reference may be made to (Sato 2007) and to (Beenakker 2019).

Majorana modes are a particularly promising approach to producing a quantum computer since the braiding operations, given that they perform transformations between locally indistinguishable states, are protected from local sources of decoherence (“topological protection”). However, there are considerable obstacles to overcome to produce a Majorana-mode quantum processor.

It is for example known to implement Majorana modes based on a one-dimensional superconducting structure (nanowire) exhibiting strong spin-orbit coupling, see for example (Cottet 2013). However, the Majorana-mode exchange cannot be performed in a strictly one-dimensional geometry. More complex superconducting circuits, exhibiting a two-dimensional or grid geometry, have been proposed—see for example (You 2014) and (Vijay 2016)—but the implementation of such circuits encounters technological difficulties that are at present insurmountable.

(Kornich 2020) proposes to carry out braiding based on a one-dimensional arrangement of Majorana modes by utilizing transitions between said Majorana modes and the first excited level above them. One drawback of this approach is that it compromises topological protection, since the system is subject to relaxation when it is in the excited state.

SUMMARY OF THE INVENTION

The invention aims to overcome these drawbacks and to propose a Majorana-mode quantum computing device the implementation of which does not pose any major technological difficulties.

According to the invention, this object is achieved by coupling a preferably one-dimensional superconducting electronic device supporting Majorana modes to a microwave cavity and by using a photonic degree of freedom of the cavity to perform the braiding. Coupling a one-dimensional superconducting electronic device supporting Majorana modes to a microwave cavity has already been described in (Cottet 2013), but the use of this coupling to carry out braiding has not.

One subject of the invention is therefore a quantum computing device comprising:

• • a microwave cavity having an input port and an output port that are separate or coincident;
  • a superconducting electronic device capacitively coupled to the microwave cavity and configured to support a chain of 2N Majorana modes, N being a positive integer; and
  • coupling means for applying microwave excitations between each pair of adjacent Majorana modes of the chain.

According to some particular embodiments:

The device may also comprise an electronic measurement circuit suitable for measuring at least one quadrature of a microwave field coming from the output port of the cavity.

The device may also comprise at least one microwave signal generator configured to generate:

• • microwave pulses at a first frequency resonant with the cavity; and
  • microwave pulses at a second frequency not resonant with the cavity and to selectively apply these pulses to said coupling means and to the input port of the cavity.
  • N may notably be a multiple of 3, whereby the superconducting electronic device supports at least one group consisting of four Majorana modes capable of encoding a qubit plus two auxiliary Majorana modes. In this case, the device may also comprise an electronic control circuit configured or programmed to drive the microwave signal generator so as to:
  • a) apply a first microwave excitation at the first frequency between a fourth and a fifth Majorana mode from one end of the chain or from a portion of the chain comprising six Majorana modes; and then
  • b) apply a second microwave excitation at the first frequency between a first and a second Majorana mode from said end; and then
  • c) simultaneously apply a third microwave excitation at the second frequency between the second and a third Majorana mode from said end; a fourth microwave excitation at the second frequency between the third and the fourth Majorana mode, the third and the fourth microwave excitations exhibiting a phase shift; and a fifth microwave excitation at the first frequency to the input port of the cavity; and then
  • d) apply a sixth microwave excitation at the first frequency between the second and the third Majorana mode; and then
  • e) simultaneously apply a seventh microwave excitation at the second frequency between the fourth and the fifth Majorana mode; an eighth microwave excitation at the second frequency between the fifth and a sixth Majorana mode from said end; and a ninth microwave excitation at the first frequency to the input port of the cavity; the order of operations c) and d) being able to be swapped.

The electronic control circuit may be configured or programmed to drive the electronic measurement circuit so as to measure a component of the microwave field coming from the output port of the cavity in phase with the excitations at the first frequency after the application of each excitation or group of excitations.

The electronic control circuit may also be configured or programmed to drive the microwave signal generator so as to: simultaneously apply a ninth and a tenth microwave excitation at the second frequency between two adjacent pairs of adjacent Majorana modes, and to drive the electronic measurement circuit so as to measure two components of the microwave field coming from the output port of the cavity in phase and in quadrature with the excitations at the first frequency.

The superconducting electronic device may comprise a semiconductor nanowire exhibiting spin-orbit coupling, placed in correspondence with an antinode of the electric field of a mode of the cavity and on which a superconducting material is deposited, the device also comprising a magnet generating a magnetic field parallel to the nanowire.

The superconducting electronic device may also comprise 2N-1 electrodes for applying electrostatic potentials in order to generate the Majorana modes, said electrodes also constituting said coupling means for applying microwave excitations.

Another subject of the invention is a method for producing a quantum braiding gate on a qubit by way of such a quantum computing device, the method comprising the following steps:

• • a) applying a first microwave excitation at the first frequency between a fourth and a fifth Majorana mode from one end of the chain or from a portion of the chain comprising six Majorana modes; and then
  • b) applying a second microwave excitation at the first frequency between a first and a second Majorana mode from said end; and then
  • c) simultaneously applying a third microwave excitation at the second frequency between the second and a third Majorana mode from said end; a fourth microwave excitation at the second frequency between the third and the fourth Majorana mode; and a fifth microwave excitation at the first frequency to the input port of the cavity; and then
  • d) applying a sixth microwave excitation at the first frequency between the second and the third Majorana mode; and then
  • e) simultaneously applying a seventh microwave excitation at the second frequency between the fourth and the fifth Majorana mode; an eighth microwave excitation at the second frequency between the fifth and a sixth Majorana mode from said end; and a ninth microwave excitation at the first frequency to the input port of the cavity; the order of steps c) and d) being able to be swapped.

The method may also comprise measuring a component of the microwave field coming from the output port of the cavity in phase with the excitations at the first frequency after the application of each excitation or group of excitations.

A further subject of the invention is a method for producing a quantum T-gate on a qubit by way of a quantum computing device, the method comprising the following steps:

• • simultaneously applying two microwave excitations at the second frequency between two adjacent pairs of adjacent Majorana modes, and measuring two components of the microwave field coming from the output port of the cavity in phase and in quadrature with the excitations at the first frequency.

BRIEF DESCRIPTION OF THE DRAWINGS

Other features, details and advantages of the invention will become apparent on reading the description given with reference to the appended drawings, which are given by way of example and in which, respectively:

FIG. 1 shows a system of four Majorana modes in a microwave cavity;

FIG. 2A shows the output signal from the cavity as a function of time during an operation of fusing two Majorana modes of the system of FIG. 1;

FIG. 2B shows a representation in the I/Q plane of the photon field in the cavity during said fusion operation;

FIG. 2C shows the evolution of the density matrix of the superconducting device as a function of time during said fusion operation;

FIG. 2D and FIG. 2E show two sequences of microwave pulses for carrying out said fusion operation and a parity measurement;

FIG. 2F and FIG. 2G show the states of the photon field of the cavity during the sequences of figures FIG. 2D and FIG. 2E, respectively;

FIG. 3A shows a sequence of microwave pulses for carrying out a braiding operation on the Majorana modes of the system of FIG. 1;

FIG. 3B shows the evolution of the photon field of the cavity during the third pulse of the sequence of figure FIG. 3A;

FIG. 3C shows the evolution of the system during said braiding operation on a Bloch sphere;

FIG. 4 shows the basic diagram of a device according to one embodiment of the invention, using six Majorana modes;

FIG. 5A and FIG. 5B show two sequences of alternating pulses for carrying out respective braiding operations on the Majorana modes of the system of FIG. 4;

FIG. 6 shows one physical exemplary embodiment of the device of FIG. 4; and

FIG. 7 shows a sequence of pulses for carrying out a “T-gate” operation on the Majorana modes of the system of FIG. 4.

Hereinafter, the term “microwave” will denote the frequency range between 300 MHz and 300 GHz, and more particularly between 1 GHz and 100 GHz.

DETAILED DESCRIPTION

FIG. 1 shows a linear chain of four Majorana modes MM₁, MM₂, MM₃, MM₄, which may be implemented physically in a superconducting device, capacitively coupled to a microwave cavity CH, which may be represented by an individual photonic mode âof frequency ω_c. Each Majorana mode is associated with a respective self-adjoint operator {circumflex over (γ)}₁, {circumflex over (γ)}₂, {circumflex over (γ)}₃, {circumflex over (γ)}₄. A superposition between adjacent Majorana modes generates energy shifts ϵ_L, ϵ_M, ϵ_R that decrease exponentially as the distance between Majorana modes increases. These energy shifts in principle break the topological protection but, as will be discussed later, it is possible to make them small enough to preserve the exponential scaling of the topological protection over a wide range of control parameters, thereby making it possible to implement the invention based on various existing technological platforms.

The low-energy Hamiltonian H of the system may be written as the sum of three terms: a term H_cav represents the energy of the photon field of the cavity, a term H_el represents the energy of the Majorana modes and a term H_int represents the coupling between the latter and the cavity:

H=H _el +H _int +H _cav

H _cav=ℏω_câ^†â

H _el=ℏ(iϵ_L{circumflex over (γ)}₁{circumflex over (γ)}₂+iϵ_M{circumflex over (γ)}₂{circumflex over (γ)}₃+iϵ_R{circumflex over (γ)}₃{circumflex over (γ)}₄)

H _int=ℏ(ig _L{circumflex over (γ)}₁{circumflex over (γ)}₂+ig _M{circumflex over (γ)}₂{circumflex over (γ)}₃+ig _R{circumflex over (γ)}₃{circumflex over (γ)}₄) (â+â^†)

where â, â^†are respectively the annihilation and creation operator of the photon field in the cavity and g_L, g_M and g_R are the coupling coefficients between pairs of adjacent Majorana modes and this photon field.

Based on the Majorana operators {circumflex over (γ)}₁, {circumflex over (γ)}₂, {circumflex over (γ)}₃, {circumflex over (γ)}₄ it is possible to define topological charges. For a chain of four Majorana modes, there are three ways of pairing these modes so as to form two topological charges corresponding to Fermionic creation operators:

ĉ _L= 1/2 ({circumflex over (γ)}₁+i {circumflex over (γ)}₂) and ĉ_R= 1/2 ({circumflex over (γ)}₃+i {circumflex over (γ)}₄)

ĉ _m= 1/2 ({circumflex over (γ)}₂+i {circumflex over (γ)}₃) and ĉ_ex= 1/2 ({circumflex over (γ)}₁+i {circumflex over (γ)}₄)

ĉ _o= 1/2 ({circumflex over (γ)}hd 1+i {circumflex over (γ)}₃) and ĉ_e= 1/2 ({circumflex over (γ)}₄+i {circumflex over (γ)}₂)

The electron system may therefore be expressed in the bases {|0_i, 0_j, |1_i, 1_j, |0_i, 1_j, |1_i, 1_j} where (i, j)∈{(L, R), (m, ex), (o, e)}. The parity operator associated with one of these topological charges is given by {circumflex over (P)}_ij=iγ_iγ_j.

It is then considered that the electron-photon coupling may be modulated at a frequency ω_RF in the microwave domain:

g_O(t)=g_O+{tilde over (g)}_Ocos(ω_RFt+ϕ_O) with O=L, M, R and ϕ_O being a phase, for example by way of gates (insulated electrodes) capacitively coupled to respective regions of the superconducting device (left, middle, right). The microwave excitation thereby modulates the coupling go via a modulation of the superposition between the Majorana modes, and therefore of the energy shifts:

ϵ_i(t)=ϵ_i+{tilde over (ϵ)}_icos(ω_RFt+ϕ_i) where ϵ_i is a constant term the value of which may be varied by applying a DC voltage to the gate (i=L, M, R).

An electromagnetic field CF develops in the cavity when a coupling coefficient, for example g_L, is modulated at the resonant frequency of the cavity: ω_RF=ω_c This field is related directly to the parity of the left section of the chain of Majorana modes. The components H_el and H_int of the low-energy Hamiltonian may also be written:

H_el=iℏϵ_L(t){circumflex over (γ)}₁{circumflex over (γ)}₂

H_int=iℏϵ_L(t){circumflex over (γ)}₁{circumflex over (γ)}₂(â+â^†)

The low-energy Hamiltonian may be rewritten in a reference frame rotating at the frequency ω_c neglecting a static term, proportional to g_L according to the rotating-wave approximation:

$\tilde{H}=H_{\mathrm{el}}+\frac{h}{2}{\tilde{g}}_L\left(e^{i{\varphi }_L}{\hat{a}}^{†}+e^{-i{\varphi }_L}\hat{a}\right)i{\hat{\gamma }}_1{\hat{\gamma }}_2$

This induces an_effective coupling between the Majorana modes MM₁ and MM₂, which may be used to measure the parity {circumflex over (P)}₁₂=i{circumflex over (γ)}₁{circumflex over (γ)}₂ thereof through the field CF, as illustrated by figures FIG. 2A to FIG. 2G. It is therefore possible to fuse the Majorana modes MM₁ and MM₂ and to detect this fusion through the photons of the cavity, extracted via an output port PS (an input port PE, coincident with or separate from PS, is also provided in order to directly excite a mode of the cavity).

FIG. 2A shows the evolution over time t of the output signal α_out from the cavity, representative of the electromagnetic field excited therein, assuming that the modulation of the coupling g_L starts at t=0. A transient is observed the duration of which is of the order of magnitude of 1/κ, κ being the line width of the cavity. The amplitude of the signal after the transient is /√{square root over (κ)}. FIG. 2B shows a representation of the output field in the I/Q plane; it is noted that the position of the circular spot representing the coherent mode of the field depends on the parity, thereby allowing the latter to be measured, and on the phase ϕ_L. The contrast, /√{square root over (κ)}, may be much greater than the width of the spots even in a strongly topological regime (→0), provided that the cavity width is small enough.

Measuring the parity through the photons of the cavity, which makes it possible to fuse two Majorana modes, requires taking into account the other two Majorana modes, hitherto considered to be decoupled. By taking into account the other Majorana modes (MM₃, MM₄), the coupling between these modes and the cavity is no longer able to be considered purely longitudinal: there is also a transverse component of the coupling that induces a temporal evolution of the parity operator. However, this evolution may be neglected provided that ϵ_O, g_O«ω_c for O=L, M, R, for example, ϵ_O,g_O≤0.1ω_c.

The consequences of the operation of fusing Majorana modes MM₁ and MM₂ on the additional modes MM₃, MM₄ are not trivial. Indeed, in the base of the two charges of the operators number {circumflex over (n)}_m=½(i{circumflex over (γ)}₂{circumflex over (γ)}₃+1), {circumflex over (n)}_ex=½(i{circumflex over (γ)}₁{circumflex over (γ)}₄+1), the fusion operation projects the two charges into an entangled state:

$❘{\Psi }_{+}\rangle =❘1_L{0}_R\rangle =\frac{1}{\sqrt{2}}(❘0_m{1}_{ex}\rangle +❘1_m{0}_{ex}\rangle )$ $\mathrm{for}{P}_{12}=1\mathrm{and}❘{\Psi }_{-}\rangle =❘0_L{1}_R\rangle =\frac{1}{\sqrt{2}}(❘0_m{1}_{ex}\rangle -❘1_m{0}_{ex}\rangle )$ $\mathrm{for}{P}_{12}=-1.$

Entanglement may be observed by measuring the central charge, which is carried out in the same way as in measurement-based fusion, as it involves measuring the parity associated with two adjacent Majorana modes, MM₂ and MM₃. At the same time, the projection expressed in the base of the charges {circumflex over (n)}_L, {circumflex over (n)}_R rib gives a state vector; the measurement of the parity of the modes MM₃ and MM₄ is therefore deterministic.

FIG. 2C shows the temporal evolution of the four components ρ₀₀, ρ₁₀, ρ₀₁, ρ₁₁ of the density matrix of the chain of Majorana modes expressed in the base L, R. It may be seen that the starting state is an entangled state and that the coherences (elements off the diagonal of the matrix) cancel one another out during the projective measurement.

FIG. 2D and FIG. 2F respectively illustrate the sequence of microwave excitations and the sequence of measurements used to perform and measure the fusion of MM₁ and MM₂. First of all, g_L is modulated at the frequency ω_c for a time greater than 1/κ (and typically less than 10/κ, for example of the order of 3/κ, because the maximum admissible duration of the measurement is limited by the parity lifetime) to perform the fusion, and a projective measurement is performed on the cavity field. Next, the parity {circumflex over (P)}₂₃=i{circumflex over (γ)}₂{circumflex over (γ)}₃ is measured by modulating g_M at the frequency ω_c and by performing a second measurement of the cavity field. This second measurement, which is probabilistic, depends on the state in which the chain of Majorana modes was projected, |Ψ₊ or |Ψ₋.

As an alternative, it is possible to measure the parity {circumflex over (P)}₃₄ by way of the sequences of FIG. 2E and FIG. 2G. It is noted that, in this case, the second measurement of the field is deterministic.

If it is considered that κ=1 MHz, which is easily obtained for example with coplanar-waveguide cavities, the measurement duration is of the order of a few μs, which requires a parity lifetime of a few tens of μs. Charge-cavity couplings of the order of g=100 MHz may be obtained, which is compatible with the abovementioned condition ϵ_O, g_O«ω_c. Assuming that the strength of the coupling may be modulated by 10%, it is therefore possible to perform “single-shot” reading of the cavity without requiring electrical manipulations.

The fusion operation only requires measuring the parity operator between adjacent links of the linear chain of Majorana modes (a “link” being formed by a pair of adjacent Majorana modes) or, equivalently, activating coupling between adjacent pairs of such modes. Braiding is in principle more restrictive because it requires reading the parity corresponding to distant Majorana modes, for example MM₂ and MM₄. According to a widespread prejudice, this would require a two-dimensional or at least grid geometry, since it seems difficult to “jump over” a Majorana mode (MM₃, in this case) in a strictly one-dimensional geometry. One idea on which the present invention is based is that this is in fact made possible through the coupling with the microwave cavity, using two pulsed excitations to modulate the coupling coefficients g_M and g_R. This effectively converts the one-dimensional system to a two-dimensional system through line-star mapping. This makes it possible to create dynamically reconfigurable two-dimensional arrays through the use of cavity photons.

Starting from the Hamiltonian H_el=ℏ(iϵ_L{circumflex over (γ)}₁{circumflex over (γ)}₂+iϵ_M{circumflex over (γ)}₂{circumflex over (γ)}₃+iϵ_R{circumflex over (γ)}₃{circumflex over (γ)}₄), consideration is given to microwave excitations detuned with the cavity (ω_RF≠ω_c) between the Majorana modes MM₂ and MM₃ and the Majorana modes MM₃ and MM₄. These excitations induce the following unitary transformation:

$U=e^{i{\omega }_{\mathrm{RF}}{\hat{a}}^{†}\hat{a}t}{e}^{\left[\frac{{\tilde{g}}_M}{{\omega }_{\mathrm{RF}}-{\omega }_c}\left(e^{i|{\varphi }_M}{\hat{a}}^{†}+e^{-i{\varphi }_M}\hat{a}\right)i{\hat{\gamma }}_2{\hat{\gamma }}_3+\frac{{\tilde{g}}_R}{{\omega }_{\mathrm{RF}}-{\omega }_c}\left(e^{i{\varphi }_R}{\hat{a}}^{†}+e^{-i{\varphi }_R}\hat{a}\right)i{\hat{\gamma }}_3{\hat{\gamma }}_4\right]}$

Thereby giving the following Hamiltonian:

$H=\hslash \left({\omega }_c-{\omega }_{\mathrm{RF}}\right){\hat{a}}^{†}\hat{a}+8i\frac{{\tilde{g}}_M{\tilde{g}}_R}{{\omega }_{\mathrm{RF}}-{\omega }_c}\sin \left({\varphi }_M-{\varphi }_R\right){\hat{\gamma }}_2{\hat{\gamma }}_4\left({\hat{a}}^{†}\hat{a}+1/2\right)$

More generally, considering a chain comprising an arbitrary number of Majorana modes, the use of a resonant cavity coupled longitudinally to said modes and of non-resonant excitations makes it possible to obtain an effective Hamiltonian having the form:

$H_{\mathrm{eff}}=\sum _{n,m}i{\hat{\gamma }}_n{\hat{\gamma }}_m{f}_{\mathrm{nm}}\left({\hat{a}}^{†},\hat{a},{\hat{a}}^{†}\hat{a}\right)+\mathrm{\hslash \delta }{\hat{a}}^{†}\hat{a}$

where f_nm is a linear function and δ=ω_c−ω_RF is the detuning between the excitation and the cavity. This shows that the system constituted by a chain of Majorana modes (regardless of its physical implementation) coupled longitudinally to a microwave cavity, with means (for example gates) for applying microwave excitations between adjacent Majorana modes, is equivalent to a 2D array, thereby allowing manipulations such as T-gates and, above all, braidings.

The protocol for carrying out the braiding {circumflex over (B)}₁₄ of the Majorana modes MM₁ and MM₄ is illustrated in FIG. 3A. First of all, from a state M₀ of the system, modulation, at the frequency ω_c, of the coupling g_M followed by reading of the cavity field makes it possible to read the parity P₂₃, after which the system is in a state M₁. For example, consideration is given to the case where M₀=|1_m0_ex, in which case P₂₃=1 and M₁=M₀. Second of all, modulation, at the frequency ω_c, of the coupling g_L followed by reading of the cavity field makes it possible to read the parity P₁₂, after which the system is in a state M₂. In the example under consideration here, P₁₂=1 and M₂=|1_L0_R. Third of all, a measurement of the parity P₂₄ is performed. This measurement concerns two non-adjacent Majorana modes, and it is more complex: it requires the modulation, at the frequency (ω_d≠ω_c), of the coupling g_M and simultaneously, but with a phase shift (ideally of π/2, in any case other than 0 or an integer multiple of π) of the coupling g_R, and also the application of a resonant excitation α_in (at the frequency ω_c) to the input port of the cavity. Following this measurement, the system is in a state M₃. In the chosen example, M₃=|0_o1_e and P₂₄=1. FIG. 3B shows the evolution of the field in the cavity in both cases P₂₄=1 and P₂₄=−1. Finally, a second measurement of the parity P₂₃ (P₂₃=1 in the example) brings the system into the state M₄=M₀e^iπ/4—in other words, the system accumulates a phase of −π/4.

FIG. 3C shows the evolution of the system on a Bloch sphere for the braiding {circumflex over (B)}₁₄ and for the braiding {circumflex over (B)}₄₁, which travels the same trajectory in the opposite direction and which leads to a π/4 phase accumulation. To sum up:

{circumflex over (B)}₁₄:

|1_m0_ex→^{{circumflex over (Π)}23}|1_m0_ex→^{{circumflex over (Π)}12}i|1_L0_R→^{{circumflex over (Π)}24}−e^iπ/4|0_o1_e→^{{circumflex over (Π)}23}e^−iπ/4|1_m0_ex

{circumflex over (B)}₄₁:

|1_m0_ex→^{{circumflex over (Π)}23}|1_m0_ex→^{{circumflex over (Π)}24}−i|0_o1_e→^{{circumflex over (Π)}12}ie^iπ/4|1_L0_R→^{{circumflex over (Π)}23}e^iπ/4|1_m0_ex

Where {circumflex over (Π)}_ij is the operator that projects the state of the system onto the parity subspace {circumflex over (P)}_ij=1.

However, the protocol of FIG. 3A does not really make it possible to encode and manipulate a qubit, because the two charges {circumflex over (n)}_m and {circumflex over (n)}_ex are not independent. To overcome this difficulty, it is possible to resort to a chain of 6 Majorana modes, as illustrated in FIG. 4. This system is obtained by extending the chain of Majorana modes of FIG. 2 on the side of MM₄ by adding two additional Majorana modes MM_2′, and MM_3′, associated with the self-adjoint operators {circumflex over (γ)}₂′, {circumflex over (γ)}₃′. g′_E denotes the coupling coefficient between MM₄ and MM_2′, and g′_m denotes that between MM_2′and MM_3′.

A qubit is encoded using the four Majorana modes MM₁, MM₂, MM_3′, and MM₄, while the Majorana modes MM₂ and MM₃ are said to be “auxiliary”: they are used to prepare, braid and read the state of the qubit, but cannot be used to encode it because they are projected during the braiding operation.

The following Fermionic operators are defined:

ĉ _o′= 1/2 (γ₁+iγ_3′)

ĉ _m′= 1/2 (γ_2′+iγ_3′)

ĉ _e′= 1/2 (γ₄+iγ_2′)

ĉ _L′= 1/2 (γ₁+iγ_2′)

ĉ _R′= 1/2 (γ_3′+iγ₄ )

and a base |n_i, n_j, n_k with i, j, k∈[e′, o′, m,, m′, ex, m, L′, R′, m]

First of all, the system is initialized in a state

$❘{\Psi }_{\mathrm{init}}\rangle =❘0_{e^{\prime }}{1}_{0^{\prime }}{1}_m\rangle =\frac{1}{\sqrt{2}}(❘0_{m^{\prime }}{1}_{\mathrm{ex}}{1}_m\rangle +❘0_{m^{\prime }}{1}_{\mathrm{ex}}{1}_m\rangle ).$

This thus creates a superposition of two different parities in the subspace associated with the Majorana modes MM₁-MM₄. As |0_m′1_ex1_m and |1_m′0_ex1_mbelong to different subspaces, they evolve independently during the braiding operation and, even if the system undergoes a projection in each step, this ends up with a superposition of the evolutions of these two states, which acquire different phases.

FIG. 5A illustrates the protocol for carrying out the braiding {circumflex over (B)}₁₄.

The initialization is performed by modulating the coupling coefficient g′_E at the frequency ω_c in order to measure the parity P_42′. The braiding itself—represented by an operator {circumflex over (Π)}₂₃{circumflex over (B)}₁₄ comprises the modulation of the coupling coefficient g_L at the frequency ω_c; and then the modulation, at the frequency ω_d≠ω_c, of the coupling g_M and simultaneously, but with a phase shift, ideally of π/2, of the coupling g_R, and also the application of a resonant excitation α_in (at the frequency ω_c) to the input port of the cavity; and then the modulation, at the frequency ω_c, of the coupling coefficient g_M. Finally, reading of the parity P_3′4 completes the protocol; this requires the modulation, at the frequency cod ω_d≠ω_c, of the coupling g_M and simultaneously, but with a phase shift, ideally of π/2, of the coupling g′_E, and also the application of a resonant excitation α_in (at the frequency ω_c) to the input port of the cavity.

Pulses at the frequencies ω_c and ω_d exhibit constant phase relationships. For pulses of different frequencies, the phase relationship is defined at the start of the pulse; for example, carriers at the frequencies we and wei may be in phase with one another at the initial times of the pulses. This is typically achieved by using an atomic clock to synchronize between the various microwave sources.

FIG. 5B illustrates the protocol for carrying out the braiding {circumflex over (B)}₄₁, which differs from the previous one only by the order of two operations and which leads to a result of opposite sign to the measurement of the parity P_3′4.

As in the previous protocol, the initialization is performed by modulating the coupling coefficient g′E at the frequency ω_c in order to measure the parity P_42′. The braiding itself—represented by an operator {circumflex over (Π)}₂₃{circumflex over (B)}₄₁ comprises the modulation of the coupling coefficient g_L at the frequency ω_c; and then the modulation, at the frequency ω_c, of the coupling coefficient g_M; and then the modulation, at the frequency ω_d≠ω_c, of the coupling g_M and simultaneously, but with a phase shift, ideally of π/2, of the coupling g_R, and also the application of a resonant excitation α_in (at the frequency ω_c) to the input port of the cavity. Finally, reading of the parity P_3′4 completes the protocol; this requires the modulation, at the frequency ω_d≠ω_c, of the coupling g_M and simultaneously, but with a phase shift, ideally of π/2, of the coupling g′_E, and also the application of a resonant excitation α_in (at the frequency ω_c) to the input port of the cavity.

FIG. 6 shows, highly schematically, a device according to one embodiment of the invention, implementing the chain of 6 Majorana modes of FIG. 4.

The cavity CH is of coplanar waveguide type, obtained by etching a resonator in a superconducting metal film (Goeppl 2008). The frequency of its fundamental mode is typically of the order of 10 GHz.

The cavity has an input port PE at one end and an output port, or read port, PS at the opposite end (in other embodiments, one and the same port may take on both functions). A microwave signal generator GH_in is connected to the input port, making it possible to excite a mode of the cavity, while a measurement circuit MES connected to the output port makes it possible to measure at least one quadrature, and preferably both quadratures, of this mode.

A magnet AM generates a stationary magnetic field Bo in the axis of the cavity.

A superconducting electronic device DS is arranged at an antinode of the electric field of the mode of the cavity. This device comprises a semiconductor nanowire NF with strong spin-orbit coupling, electrically insulated from the ground plane of the cavity and oriented parallel to the magnetic field B. The spin-orbit coupling may be intrinsic, as for example in the case of InSb or InAs nanowires, or extrinsic, induced by a magnetic texture of a substrate on which the nanowire (for example a carbon nanotube) is deposited. This nanowire is covered with a superconducting film FS of bandgap Δ and contacted by 5 electrodes (gates) GR₁-GR₅ as described in (Cottet 2013); more generally, there are 2N gates to generate 2N Majorana modes. The gates should be spaced enough to ensure relatively good localization of the Majorana modes (that is to say that the energy ϵ₀ of these modes should be much smaller, for example by a factor of 10 or more, than the bandgap Δ).

The magnetic field B should be strong enough to induce a topological electron phase, at the electrochemical equilibrium potential μ (in a simple model, >√{square root over ((gμ_BB)²−Δ²)}), where g is the Landéfactor and μ_B is the Bohr magneton), but not too strong, so that the cavity is able to remain superconducting.

Voltage generators GT_ii=1−5 (only GT₅ is shown) connected to the gates make it possible to locally change the chemical potential of the nanowire, for example to a value of μ′ such that μ′<√{square root over ((gμ_BB)²−Δ²)})as proposed in (Cottet 2013). It is this chemical potential modulation that allows the Majorana modes to appear.

In addition, the gates are connected to microwave signal generators GH_ii=1−5 (only GH₅ is shown) via respective bias tees BT. These generators are driven by a processor PR so as to generate a pulse sequence at the frequencies ω_d and ω_c of the type described above with reference to FIG. 5A and to FIG. 5B.

As explained above, the pulses have a duration of a few times 1/κ, κ being the spectral width of the mode of the cavity, for example 3/κ. Their power is typically of the order of a few milliwatts. The frequency shift between ω_d and ω_c is typically between κ and 10 κ; it may be for example 5κ.

In addition to braiding, the device of FIG. 6 (and more generally the diagram of FIG. 4) makes it possible to produce a T-gate (also called “π/8 gate”). To this end, it is enough to simultaneously apply two off-resonance excitations (typically at the frequency ω_c) to two adjacent links of the chain of Majorana modes and to simultaneously read the two quadratures of the field at the output of the cavity (which may be achieved through a single measurement in a direction forming angles of 45° with the axes of the IQ plane). For example, in the device of FIG. 6, the excitations may be applied to the gates GR₄, GR₅ in order to modulate the couplings g′_E, g′_M of the diagram of FIG. 4. This is illustrated in FIG. 7.

However, it has been demonstrated in (Karzig 2017) that braiding operations are enough to produce all 1-qubit Clifford gates, as well as a two-qubit entanglement gate denoted W. The set of both makes it possible to produce two-qubit gates between neighboring qubits, including the CNOT gate. In addition, CNOT gates and 1-qubit π/8 gates are enough to produce a universal computer (Sau 2010). A multi-qubit extension to the device of FIG. 6 or the diagram of FIG. 4, comprising six Majorana modes for each qubit, therefore makes it possible to produce a universal quantum computer.

To be able to perform two-qubit operations, it is enough to have a chain of 12 Majorana modes—that is to say two subsets of 6 modes each encoding one qubit—and to perform braiding on modes of the two subsets. The invention may therefore serve as a basis for a generic quantum computer.

FIG. 6 constitutes one example of a physical platform for implementing the invention, but it is in no way limiting. Other embodiments are conceivable without departing from the scope of the invention; a strict one-dimensionality of the superconducting electronic device supporting the chain of Majorana modes is highly advantageous technologically, but not essential in principle.

REFERENCES

(Sato 2017): Masatoshi Sato, Yoichi Ando “Topological superconductors: a review” arXiv:1608.03395v3, 4 Apr. 2017.

(Beenakker 2019): C. W. J. Beenakker “Search for non-Abelian Majorana braiding statistics in superconductors” arXiv:1907.06497v1, 15 Jul. 2019.

(Cottet 2013): A. Cottet, T. Kontos, B. Douçot “Squeezing light with Majorana fermions” arXiv:1307.4185v4, 12 Nov. 2013.

(You 2014): J. Q. You et al. “Encoding a qubit with Majorana modes in superconducting circuits” arXiv: 1108.3712v2, 20 Sep. 2014.

(Vijay 2016): Sager Vijay, Liang Fu “Braiding without Braiding: Teleportation-Based Quantum Information Processing with Majorana Zero Modes” arXiv: 1609.00950v1, 4 Sep. 2016.

(Goeppl 2008): M. Göppl et al. “Coplanar Waveguide Resonators for Circuit Quantum Electrodynamics”, arXiv: 0807.4094v1, 25 Jul. 2008

(Karzig 2017): T. Karzig et al. “Scalable Designs for Quasiparticle-Poisoning-Protected Topological Quantum Computation with Majorana Zero Modes” arXiv:1610.05289v4, 21 Jun. 2017.

(Sau 2010): J. D. Sau et al. “Universal quantum computation in a semiconductor quantum wire network. ” arXiv:1007.4204v3, 24 Nov. 2010.

(Kornich 2020): V. Kornich et al. “Braiding and all quantum operations with Majorana modes in 1D”, arXiv 14 Sep. 2020 and Physical Review Letters, 126(11), 117701.

Claims

1. A quantum computing device comprising: a microwave cavity (CH) having an input port (PE) and an output port (PS) that are separate or coincident;a superconducting electronic device (DS) capacitively coupled to the microwave cavity and configured to support a chain of 2N Majorana modes (MM1-MM4, MM2′, MM3′), N being a positive integer; andcoupling means (GR1-GR5) for applying microwave excitations between each pair of adjacent Majorana modes of the chain andan electronic measurement circuit (MES) suitable for measuring at least one quadrature (I, Q) of a microwave field coming from the output port of the cavity, characterized in that it also comprises at least one microwave signal generator (GHin, GH1-GH5) configured to generate: microwave pulses at a first frequency resonant with the cavity; andmicrowave pulses at a second frequency not resonant with the cavity,and to selectively apply these pulses to said coupling means and to the input port of the cavity,wherein N is a multiple of 3, whereby the superconducting electronic device supports at least one group consisting of four Majorana modes capable of encoding a qubit (MM1, MM2, MM2′, MM3′) plus two auxiliary Majorana modes (MM2, MM3).

2. The quantum computing device as claimed in claim 1, further comprising an electronic control circuit (PR) configured or programmed to drive the microwave signal generator so as to: a) apply a first microwave excitation at the first frequency between a fourth and a fifth Majorana mode from one end of the chain or from a portion of the chain comprising six Majorana modes; and thenb) apply a second microwave excitation at the first frequency between a first and a second Majorana mode from said end; and thenc) simultaneously apply a third microwave excitation at the second frequency between the second and a third Majorana mode from said end; a fourth microwave excitation at the second frequency between the third and the fourth Majorana mode, the third and the fourth microwave excitations exhibiting a phase shift; and a fifth microwave excitation at the first frequency to the input port of the cavity; and thend) apply a sixth microwave excitation at the first frequency between the second and the third Majorana mode; and thene) simultaneously apply a seventh microwave excitation at the second frequency between the fourth and the fifth Majorana mode; an eighth microwave excitation at the second frequency between the fifth and a sixth Majorana mode from said end; and a ninth microwave excitation at the first frequency to the input port of the cavity;the order of operations c) and d) being able to be swapped.

3. The quantum computing device as claimed in claim 2, the electronic control circuit being configured or programmed to drive the electronic measurement circuit so as to measure a component of the microwave field coming from the output port of the cavity in phase with the excitations at the first frequency after the application of each excitation or group of excitations.

4. The quantum computing device as claimed in claim 2, wherein the electronic control circuit (PR) is also configured or programmed to drive the microwave signal generator so as to: simultaneously apply a ninth and a tenth microwave excitation at the second frequency between two adjacent pairs of adjacent Majorana modes, and to drive the electronic measurement circuit so as to measure two components of the microwave field coming from the output port of the cavity in phase and in quadrature with the excitations at the first frequency.

5. The quantum computing device as claimed in claim 1, wherein the superconducting electronic device comprises a semiconductor nanowire (NF) exhibiting spin-orbit coupling, placed in correspondence with an antinode of the electric field of a mode (CF) of the cavity and on which a superconducting material (FS) is deposited, the device also comprising a magnet (AM) generating a magnetic field (B) parallel to the nanowire.

6. The quantum computing device as claimed in claim 5, wherein the superconducting electronic device further comprises 2N-1 electrodes (GR1-GR5) for applying electrostatic potentials in order to generate the Majorana modes, said electrodes also constituting said coupling means for applying microwave excitations.

7. A method for producing a quantum braiding gate on a qubit by way of a quantum computing device as claimed in claim 1, the electronic control circuit being configured or programmed to drive the electronic measurement circuit so as to measure a component of the microwave field coming from the output port of the cavity in phase with the excitations at the first frequency after the application of each excitation or group of excitations, the method comprising the following steps: a) applying a first microwave excitation at the first frequency between a fourth and a fifth Majorana mode from one end of the chain or from a portion of the chain comprising six Majorana modes; and thenb) applying a second microwave excitation at the first frequency between a first and a second Majorana mode from said end; and thenc) simultaneously applying a third microwave excitation at the second frequency between the second and a third Majorana mode from said end; a fourth microwave excitation at the second frequency between the third and the fourth Majorana mode; and a fifth microwave excitation at the first frequency to the input port of the cavity; and thend) applying a sixth microwave excitation at the first frequency between the second and the third Majorana mode; and thene) simultaneously applying a seventh microwave excitation at the second frequency between the fourth and the fifth Majorana mode; an eighth microwave excitation at the second frequency between the fifth and a sixth Majorana mode from said end; and a ninth microwave excitation at the first frequency to the input port of the cavity;the order of steps c) and d) being able to be swapped.

8. The method as claimed in claim 7, further comprising measuring a component of the microwave field coming from the output port of the cavity in phase with the excitations at the first frequency after the application of each excitation or group of excitations.

9. A method for producing a quantum T-gate on a qubit by way of a quantum computing device as claimed in claim 1, the method comprising the following steps: simultaneously applying two microwave excitations at the second frequency between two adjacent pairs of adjacent Majorana modes, and measuring two components of the microwave field coming from the output port of the cavity in phase and in quadrature with the excitations at the first frequency.