Common logical operations in a silicon quantum processor

System fabrication

The qubit gadget was fabricated on an isotopically purified ²⁸Si buffer layer (20 nm thick), epitaxially grown at 380 °C by way of ultrahigh-vacuum electron-beam deposition on a p-type pure silicon substrate (50–100 Ω cm). This layer maintains a ²⁹Si focus beneath 130 ppm, successfully isolating qubits from the nuclear spin tub of the pure silicon substrate and enhancing spin coherence instances. Utilizing STM-assisted hydrogen depassivation lithography at 77 Ok (Scienta Omicron INFINITY SPM Lab), we deterministically patterned the hydrogen-passivated silicon floor with atomic precision. The uncovered areas have been then dosed with phosphine (PH₃) at 5 × 10⁻⁷ mbar for five min (room temperature), adopted by thermal incorporation at 330 °C for 1 min to activate phosphorus donors. This course of creates three tunnel-coupled quantum dot constructions (P-atom clusters) positioned 17.0 nm away from a single-electron transistor cost sensor for particular person qubit addressability and read-out. Subsequently, the doped nanostructure was encapsulated with a 20 nm epitaxial ²⁸Si capping layer, grown at 250 °C (0.7 nm h⁻¹ development price). This low-temperature epitaxial course of ensures minimal dopant diffusion whereas sustaining crystalline perfection. After eradicating the gadget from the STM system, a polymethyl methacrylate resist and electron-beam lithography have been used to outline a collection of small holes (200 nm in diameter) aligned with the phosphorus patches. This step was adopted by reactive ion etching and aluminium lead deposition. By way of these holes, the leads penetrated the dopant layer, establishing direct electrical contact between the metallic and the phosphorus patches. The ohmic contacts between the buried P-dopant gadget and aluminium electrodes have been fashioned by silicon vias. Then, an aluminium microwave antenna was fabricated atop the gadget, separated by a ten nm atomic-layer-deposited Al₂O₃ dielectric layer to forestall present leakage and allow high-fidelity spin manipulation.

Estimation of uncertainties

All reported uncertainties correspond to 1 s.d. from the imply, estimated by 2,000 Monte Carlo bootstrap resampling trials. Particularly, we mannequin the measured counts for every spin state (for instance, (left|Downarrow Downarrow Downarrow Downarrow rightrangle) to (left|Uparrow Uparrow Uparrow Uparrow rightrangle)) throughout N measurement repetitions as following a multinomial distribution. To make clear the illustration of the uncertainties, we offer the next illustrative instance: the worth 96.5(20)% corresponds to 96.5 ± 2.0%, the place the quantity in parentheses denotes the s.d.

[[4, 2, 2]] detection code

The logical code phrases (left|10rightrangle _01rightrangle left_leftrightrangle) within the computational foundation are

The stabilizers performing the syndrome extraction of the bodily phase-flip and bit-flip errors are ({widehat{S}}^{X}={X}_{1}{X}_{2}{X}_{3}{X}_{4}) and ({widehat{S}}^{Z}={Z}_{1}{Z}_{2}{Z}_{3}{Z}_{4}), respectively. All logical foundation states in equation (1) have even parity, whereas any single bodily qubit errors result in a state with odd parity.

The correspondence between logical operators and bodily operators is as follows: the logical operation X_LI_L corresponds to making use of X gates on bodily qubits 1 and three (X₁I₂X₃I₄). Equally, I_LX_L maps to X₁X₂I₃I₄. In contrast, the Z_LI_L operation corresponds to Z₁Z₂I₃I₄, whereas I_LZ_L interprets to Z₁I₂Z₃I₄. The logical Hadamard operation H_LH_L requires H gates on all 4 bodily qubits (H₁H₂H₃H₄), adopted by swapping of bodily qubits 2 and three (SWAP₂₃). Lastly, the logical CNOT_L is carried out as a SWAP₁₂ operation, the place the logical L₁ is the management qubit.

In contrast with the logical operations above, the implementation of the S_LI_L gate is much less intuitive. The S_LI_L gate is carried out by making use of bodily S gates to the primary two bodily qubits, adopted by a CZ gate between them (Fig. 3a). The bodily S gates accumulate the goal section on the logical qubit L₁, whereas the CZ gate cancels the additional section collected on bodily qubits 1 and a pair of. For instance, when the 2 bodily S gates are utilized for the primary two bodily qubits, with the preliminary state being ({left|{psi }_{0}rightrangle }_{{rm{L}}},=,00rightrangle _{{rm{L}}}+10rightrangle _{{rm{L}}}), the corresponding bodily qubit state turns into (left|{psi }_{1}rightrangle =left|0000rightrangle +{{rm{e}}}^{i{{uppi }}}left|1111rightrangle +,{{rm{e}}}^{i{{uppi}}/2}left|0101rightrangle +{{rm{e}}}^{i{{uppi }}/2}left|1010rightrangle). Then, the CZ gate is utilized to right the section collected on the (left|1111rightrangle) state. Because of this, the goal logical qubit state is obtained: ({left|{psi }_{2}rightrangle }_{{rm{L}}}=00rightrangle _{{rm{L}}}+{{rm{e}}}^{i{{uppi }}/2}10rightrangle _{{rm{L}}}).

Quantum state tomography

The density matrix of a four-qubit state may be expressed as follows:

the place c_i,j,okay,l denotes the growth coefficients of the density matrix and ({M}_{i},{M}_{j},{M}_{okay},{M}_{l}in left{I,X,Y,Zright}) are the Pauli operators appearing on the qubits i, j, okay and l. There are 4⁴ = 256 linear impartial operators in complete, from I ⊗ I ⊗ I ⊗ I, X ⊗ I ⊗ I ⊗ I, …to Z ⊗ Z ⊗ Z ⊗ Z.

To reconstruct the density matrix, 255 expectation values are to be decided within the experiment (apart from c_1,1,1,1 = 1). For qubit j within the I base measurement, no prerotation is utilized after the goal state (left|psi rightrangle) is ready; for qubit okay within the X base measurement, a ({R}_{y}^{okay}(-{{uppi }}/2)) prerotation is utilized; a ({R}_{x}^{l}({{uppi }}/2)) prerotation is utilized for qubit l within the Y base measurement; and we use I base projection outcomes for the Z base calculation reasonably than performing R_x(π) pretotation to any qubit, to forestall rotation error⁵². All of the above prerotations are realized by NMR pulses, that are conditional on the electron spin being within the (left|downarrow rightrangle) state.

After that, the utmost probability estimation methodology is used to limit the measured density matrix to be bodily (Hermitian, optimistic semidefinite and hint preserving) whereas offering the closest measurement chances. In line with the Cholesky decomposition, a bodily density matrix may be written as (rho ={T}^{dagger }T/{rm{Tr}}({T}^{dagger }T)). The matrix T is a fancy decrease triangle matrix, which has 2^2D (D = 4 for 4 qubits) parameters t₁, t₂, …t₂₅₆. The loss perform is outlined as follows:

the place P_i are the measured chances projected at foundation (left|{psi }_{i}rightrangle). We decrease this loss perform to acquire the ultimate and bodily density matrix. All of the constancy on this paper is calculated by (F={(mathrm{Tr}sqrt{sqrt{{rho }^{mathrm{best}}}rho sqrt{{rho }^{mathrm{best}}}})}^{2}), the place ρ^best is the corresponding best density matrix.

Logical density matrix reconstruction

The entire quantum state tomography reconstructs the bodily density matrix, which usually accommodates elements outdoors the logical subspace outlined by the stabilizers ({widehat{S}}^{X}) and ({widehat{S}}^{Z}). The parity projection is basically enabled by the projection operator (Π + I^⊗4)/2, the place Π represents the stabilizers of the [[4, 2, 2]] code, and I^⊗4 denotes the four-qubit identification operator. We are able to mission the bodily density matrix into three subspaces outlined by the +1 eigenspace of ({widehat{S}}^{X}), the +1 eigenspace of ({widehat{S}}^{Z}) and their simultaneous eigenspace. The logical density matrix is then obtained after making use of the projection.

Quantum course of tomography

A quantum operation may be described by the operator-sum illustration, expressed as

the place the operation parts {α_i} fulfill the completeness situation:

To experimentally decide the operators α_i, we undertake an equal illustration utilizing a set foundation ({{widetilde{alpha }}_{i}}) spanning the area of d² complicated matrices (for N-qubit methods, d = 2^N). This provides the χ-matrix illustration:

the place the d² × d² matrix χ absolutely characterizes the quantum course of. For the single-qubit system, the usual foundation operators are chosen as follows:

Experimentally figuring out an arbitrary single-qubit operation requires measuring 12 parameters by enter states (left|0rightrangle), (left|1rightrangle), (left|+rightrangle =(left|0rightrangle +left|1rightrangle )/sqrt{2}) and (left|+irightrangle =(left|0rightrangle +ileft|1rightrangle )/sqrt{2}). Quantum state tomography on the output states yields 4 matrices:

These correspond to ({rho }_{j}^{{prime} }={mathcal{E}}({rho }_{j})) with ({rho }_{1}=left|0rightrangle leftlangle 0right|), ρ₂ = ρ₁X, ρ₃ = Xρ₁ and ρ₄ = Xρ₁X. The linear decomposition of ({mathcal{E}}({rho }_{j})) within the foundation states ρ_okay results in the next relation:

the place coefficients ({beta }_{jk}^{mn}) are decided by the idea states and operators. For the single-qubit case, these relations are encapsulated within the transformation matrix:

The χ-matrix is then constructed by block matrix operations:

The method constancy definition we used right here is

the place χ^best is the perfect course of matrix.

Classical optimizer

For the classical optimizer, we employed the Nelder–Mead algorithm⁵³ for optimization, which is a simplex-based direct-search methodology appropriate for unconstrained minimization of goal capabilities. In contrast with different direct-search strategies comparable to gradient descent which will exhibit higher convergence on easy capabilities, Nelder–Mead demonstrates robustness for noisy capabilities and may discover close by valleys, thereby avoiding native minima trapping and overcoming non-smoothness of goal capabilities.

We improved the Nelder–Mead algorithm to go well with our particular experimental scenario. To implement the algorithm’s angular area constraint of [−π, π), we applied periodic boundary conditions by computationally wrapping the optimization parameter θ (Fig. 5a) into this interval during each iterative optimization step. Furthermore, we optimized the Nelder–Mead algorithm’s reflection, expansion, contraction and shrink coefficients to achieve convergence to the global minimum with fewer iterations.

Clifford fitting

To reduce the impact of noise in near-term quantum computations, error mitigation techniques process noisy results to more accurately estimate ideal expectation values⁵⁴. Clifford fitting is also referred to as Clifford data regression^50,55. The core idea is to learn a correction function, f^CF, using training data generated from efficiently simulatable Clifford circuits. For these circuits, both the noisy expectation values (X^noisy) measured on hardware and the exact expectation values (X^exact) computed classically are obtained. The function f^CF is determined by fitting these data pairs. Subsequently, this learned function is applied to the noisy result measured for the actual target circuit, for which an exact value may be hard to compute. Specifically, given the noisy measurement ({X}_{psi }^{,mathrm{noisy}}) for a target state (left|psi rightrangle) (prepared by circuit U_ψ) and observable X, Clifford data regression aims to find f^CF to approximate the ideal value ({X}_{psi }^{,mathrm{exact}}=langle psi | X| psi rangle) via the following relation:

where (vec{a}) represents a set of parameters characterizing the function f^CF.

In the VQE experiment, we implement Clifford fitting in a quantum parameterized circuit. First, according to the method in ref. ⁵⁰, we distinguish between observables that vary with parameters and those that remain constant. Constant observables cannot be analysed using the Clifford fitting method. Let M denote the number of parameterized gates. Then, by randomly selecting K gates from M, we independently and uniformly sample their K parameters from the range [−π, π), while the remaining M − K gates are converted into identity operations. This process generates m random circuits ({{{S}_{i}^{(K)}}}_{i=1}^{m}). The number m should be sufficient (typically over 10) to ensure an adequate fitting performance of (vec{a}) in the linear regression. The dataset consists of pairs of noisy and exact expectation values, ({{mathcal{T}}}_{psi }={({X}_{psi }^{,mathrm{noisy}},{X}_{psi }^{mathrm{exact}})}), generated from a set of quantum circuits ({{{S}_{i}^{(K)}}}_{i=1}^{m}). Thus, the parameters (vec{a}) of the correction function f^CF can be fitted by the the linear regression function

The parameters a₁ and a₂ are typically determined by minimizing the least-squares cost function

There is theoretical justification for the linear model, as it can perfectly correct for certain noise models, such as global depolarizing noise.

Finally, after determining the optimal parameters ({vec{a}}^{* }=({a}_{1}^{* },{a}_{2}^{* })), the error-mitigated expectation value for the original target circuit is predicted using equation (13):

Symmetry verification

Molecular Hamiltonians in quantum simulations inherently possess symmetry constraints such as particle number conservation. Symmetry verification through subspace expansion provides an effective error mitigation strategy. Following the methodology in ref. ⁵⁶, we implement this approach for the [[4, 2, 2]] code. For a normal [[n, k, d]] code, the code area is outlined by the projection operator:

the place l = n − okay; ({P}_{i}=frac{1}{2}(I+{u}_{i})) initiatives onto the +1 eigenspace of stabilizer turbines u_i ∈ G = 〈u₁, …, u_l〉; and S denotes the stabilizer group. For a code-space Hamiltonian H_c = ∑_iγ_iΓ_i, the place γ_i denotes the coefficient of the ith time period, and Γ_i represents the corresponding Pauli string operator. the symmetry-verified expectation worth turns into the next:

the place (c={rm{Tr}}[widehat{P}rho ]) normalizes the density matrix ρ projected onto the symmetry-preserving subspace.