Abstract
With favourable error thresholds and requiring only nearestneighbour interactions on a lattice, the surface code is an errorcorrecting code that has garnered considerable attention. At the heart of this code is the ability to perform a lowweight parity measurement of local code qubits. Here we demonstrate highfidelity parity detection of two code qubits via measurement of a third syndrome qubit. With highfidelity gates, we generate entanglement distributed across three superconducting qubits in a lattice where each code qubit is coupled to two bus resonators. Via highfidelity measurement of the syndrome qubit, we deterministically entangle the code qubits in either an even or odd parity Bell state, conditioned on the syndrome qubit state. Finally, to fully characterize this parity readout, we develop a measurement tomography protocol. The lattice presented naturally extends to larger networks of qubits, outlining a path towards faulttolerant quantum computing.
Introduction
Quantum error correction is an essential step towards realizing scalable quantum computers. Theoretically, it is possible to achieve arbitrarily long protection of quantum information from corruption due to decoherence or imperfect controls, so long as the error rate is below a threshold value^{1,2}. However, most of the initial faulttolerant quantum computing proposals were purely theoretical studies that would be impractical to implement in a physical system. Knill’s C4/C6 architecture^{3} showed that it was possible to have a pseudothreshold as high as 3% but with very longrange interactions. The twodimensional surface code (SC) is a topological errorcorrecting code^{4,5} that achieves high error thresholds ~\n1% in a twodimensional lattice of qubits supporting nearestneighbour interactions^{6,7}.
With ongoing improvements to coherence times^{8,9,10,11}, and with gate^{12,13} and readout fidelities^{14,15,16} at or approaching relevant threshold values, superconducting qubits are prime candidates for scaling up towards a faulttolerant architecture. Our scheme for building a network of superconducting qubits employs microwave resonators as the links, as it has been shown that resonators can be used as quantum buses to mediate interactions between qubits^{17}, and that multiple resonators can be coupled to a single qubit^{18}. In the future, larger superconducting qubit systems and networks will also leverage circuit integration techniques that come along with a solidstate architecture.
The original SC layout comprises qubits arranged in a square lattice^{5} with the qubits in the lattice coming in two distinct flavours, either code qubits, which carry logical information, or syndrome qubits, which are used to measure stabilizer operators of surrounding code qubits. Performing a round of error correction in the SC consists of applying controlledNOT (CNOT) gates to map the parity of the surrounding code qubits into the states of the syndrome qubits, which are then measured to determine the plaquette parities. Using superconducting resonators as the links of the lattice, it is possible to construct a SC architecture using superconducting qubits at each vertex along with the ability to couple a single qubit to four resonators. As it is also important to be able to read out and address the qubits individually, this may require an additional fifth resonator per qubit. A layout proposed by DiVincenzo^{19} allows for the number of resonators per qubit to be reduced to two (three with readout). However, each code and syndrome qubit in this scheme is replaced by four qubits, which results in a fourfold increase in the number of twoqubit gates.
Here we introduce an optimal layout for implementing the SC on a twodimensional lattice that does not carry additional overhead as in the original DiVincenzo proposal. In this work, we demonstrate a subsection of this optimized lattice using three transmon qubits, in which two independent outer code qubits are joined to a central syndrome qubit via two independent bus resonators. By combining allmicrowave highfidelity single and twoqubit nearestneighbour entangling gates, with highfidelity measurement of the syndrome qubit, we deterministically measure the parity of the code qubits. This protocol is tested in the case of a superposition of the code qubits, and we are able to detect the outer code qubits in either an even or odd parity Bell state, conditioned on the syndrome state. As a full characterization of this parity detection, we implement a measurement tomography protocol to obtain a fidelity metric (90% for odd and 91% for even). Our results reveal a straightforward path for expanding superconducting circuits towards larger networks for the SC and eventually a primitive logical qubit implementation.
Results
Modified lattice for SC with superconducting qubits
We introduce a direct onetoone mapping of the SC to a twodimensional lattice (Fig. 1) that does not carry the additional overhead of qubits and gates as in the original proposal by DiVincenzo^{19}. Rather than requiring 12 qubits to demonstrate a halfplaquette strand operation on the surface fabric, this arrangement allows such an achievement with just 3 qubits (details about the mapping are given in the Methods). Syndrome qubits can be used as either Xparity checks (red circles) or Zparity checks (green circles) of four surrounding code qubits (blue circles). The fivequbit block consisting of X (Z)syndrome and four surrounding code qubits defines a unit cell X (Z)plaquette.
An advantage of our lattice architecture is that it is commensurate with experiments already demonstrated, where a single qubit can be coupled to two separate resonators^{18}, a single bus has been used to couple up to three qubits^{20}, and an independent readout can be used to measure a single qubit that has been entangled with a separate qubit via a bus resonator^{21}. The experiments presented in this article are performed effectively on a ‘halfplaquette’ subsection, consisting of three qubits (all with individual readout resonators) and two bus resonators, depicted by the pink square of Fig. 1, where we demonstrate all necessary gate operations and measurements that constitute the SC protocol.
Physical subsection and control characterization
The halfplaquette device (Fig. 2a) contains three singlejunction transmon qubits connected by two coplanar waveguide (CPW) resonators serving as the buses, and each qubit is coupled to its own separate CPW resonator for independent readout and control. Here we label the code qubits Q1 and Q3, and the middle qubit Q2 serves as the syndrome. A simplified control and readout diagram is also shown in Fig. 2a, with all microwave sources for single and twoqubit gates and independent readout indicated. The syndrome Q2 is read out with the assistance of a Josephson parametric amplifier (JPA)^{16,22} for highfidelity singleshot state discrimination, which is a critical component for demonstrating the SC parity check protocol. All device parameters, relevant coherence times and a complete schematic are given in the Methods. All singlequbit controls are performed in 40 ns and characterized to >99.7% average gate fidelity via simultaneous randomized benchmarking^{23} (see Methods).
At the crux of the SC protocol is the detection of the parity of the code qubits. Figure 2b shows the parity check protocol (PCP) in a circuit depiction on an arbitrary state of Q1 and Q3, with the parity being indicated through the classical detection of an indication bit, b_{2}. The PCP is realized in a system of three qubits via the quantum circuit shown in Fig. 2c, where the parity of the Q1 and Q3 state is mapped on to the syndrome Q2 via two CNOT gates, between Q1 and Q2, Q3 and Q2, and then subsequently the classical indication bit b_{2} is obtained via a quantum measurement of Q2.
Highfidelity CNOT entangling gates are critical for the PCP. To realize these CNOTs in our device, we implement entangling ZX_{90} gates using the crossresonance interaction^{24,25}. By driving Q1 (Q3) at the Q2 transition frequency ω_{2}, we use a simple decoupling sequence^{12} to implement the twoqubit Clifford generator ZX_{90} gate between Q2 and Q1 (Q3). The ZX_{90} gate is equivalent to a CNOT up to singlequbit rotations, as illustrated in Fig. 2d, and thus can be used interchangeably in the PCP.
Both pairs of ZX_{90} entangling gates are characterized with twoqubit Clifford randomized benchmarking. With a total gate time of 350 ns, the ZX_{90} gate between Q2 and Q1 (Q3) is experimentally shown to have a gate fidelity of 0.962±0.002 (0.957±0.001). The fidelities are consistent with the gate and coherence times in the system. Further details about the gate tuneup, calibration and timescales can be found in the Methods.
To determine the collective state of all qubits in the system, we can perform independent readouts of each qubit through the individually coupled resonators. The syndrome qubit, Q2, is read out dispersively with the JPA, pumped −4 MHz from the readout frequency (ω_{M2}=2π·6.584 GHz) and an optimized singleshot assignment fidelity (with no preparation corrections, and defined in Methods) of 91% is achieved, although fluctuations on the order of 2–3% are observed. The code qubits are read out using the highpowered Josephson nonlinearity of the readout cavities^{26}. Further detail and parameters about the readouts are given in the Methods.
Distributed entanglement and parity measurement
To observe the action of the gate portion of the PCP, in which CNOT (in our case ZX_{90}) gates are performed between the syndrome and the code qubits, it is insightful to perform tomographic reconstruction of the complete threequbit system. State tomography in our system is achieved by correlating individual single shots of all three individual readouts^{27}, Mi, iε[1,3]. Figure 3 shows reconstructed threequbit Pauli state vectors for the entanglement processes necessary for the PCP. In Fig. 3a, a ZX_{90} entangling operation between the code qubit Q3 and the syndrome Q2 is implemented giving a state fidelity _{state}=0.949±0.002 (SDP) 0.954±0.002 (raw), where SDP refers to a semidefinite program reconstruction of the state and raw reflects unconstrained inversion^{28}. The difference between the SDP and raw estimates exceeds the statistical error estimated from the readout signaltonoise and suggests that the main source of error in our experiment is systematic. We estimate that our systematic errors are of order ~\n2–3%. A rigorous error bound on fidelity in the presence of systematic noise is still the subject of future study^{29}. In Fig. 3b, a ZX_{90} entangling operation between Q1 and Q2 is implemented (_{state}=0.951±0.002 (SDP) 0.953±0.003 (raw)) and finally, in Fig. 3c both ZX_{90} gates are applied simultaneously, generating a maximally entangled GHZ state of all three qubits (_{state}=0.935±0.002 (SDP) 0.942±0.003 (raw)). We thus show the ability to distribute entanglement across the full network, first between nearestneighbour qubits Q3 and Q2 or Q1 and Q2, and then across the entire system with the GHZ state, spanning both bus resonators.
With the demonstrated highfidelity ZX_{90} gate primitives for the PCP, the next step is to observe the singleshot readout of the syndrome Q2 to signal the parity of the code qubits Q1 and Q3. We begin by obtaining a syndrome readout calibration histogram (a typical one is shown in Fig. 4a). Then, we prepare the simple computational basis states, , , and , as inputs for the code qubits and observe the proper parity assignment via the PCP, shown via the M2 histograms in Fig. 4b. By thresholding the measurement outcomes of M2 based on the readout calibration traces, we reconstruct the state of Q1 and Q3 conditioned on M2 using standard quantum state tomography techniques (see Methods). In the case of the four computational states, we obtain fidelity of _{SDP}=0.984, 0.987, 0.989, 0.909 and _{raw}=0.975, 0.989, 0.999, 0.905, respectively.
A more complete stress test of the PCP is to observe its function on the maximal superposition state of the code qubits Q1 and Q3. The gate protocol now mimics that of the GHZ state generation from Fig. 3c, and over repeated state preparations and measurements of the syndrome, M2, we obtain a bimodal histogram, indicating that instances of both parities exist (Fig. 4c). We observe the probabilistic entanglement of either the odd or even Bell states or conditioned on M2. For these conditioned entangled states, we find state fidelities of _{odd}=0.891 (raw) 0.891 (SDP) and _{even}=0.970 (raw) 0.948 (SDP). In an errorfree situation, these fidelities should not exceed the assignment fidelity; however, due to systematic errors it is possible for these to deviate. The observed deviation between the assignment and state fidelity is within our estimate of systematic error. Nonetheless, the ability to generate these entangled states of nonnearestneighbour qubits Q1 and Q3 in two separate bus resonators is a crucial element for scaling up towards larger quantum networks.
Finally, to characterize the complete ideal projective nature of the PCP, we perform measurement tomography. This is accomplished via quantum process tomography of the code qubits, for which further details are given in the Methods. Conditioned on the measurement of the syndrome M2, we obtain the two maps for the odd and even parity projection operators shown in Fig. 4f,h (ideal maps are shown in Fig. 4i,k). To quantify the performance of the PCP, we introduce a measurement fidelity metric that takes into account the full quantum dynamics of the measurement, including projection and backaction (see Methods for more details). We obtain a measurement fidelity of and . The loss in measurement fidelity corresponds mostly to the 91% assignment fidelity of M2, best illustrated by the unconditional map (shown in Fig. 4g) having a measurement fidelity _{meas}=0.968.
It is important to note that all the gates used are calibrated and run to achieve these results without any Hamiltonian corrections for either singlequbit or twoqubit errors. An Xparity check is a simple extension through the appropriate application of singlequbit Hadamard pulses on the code qubits.
Discussion
The experiment described implements a subsection of the SC faulttolerant architecture. By combining highcoherence transmon qubits, highfidelity nearestneighbour twoqubit gates and highfidelity quantum nondemolition singleshot readout, we use a syndrome qubit to determine the parity of its neighbouring qubits. With this device, we demonstrate the versatility of superconducting qubits in the extended quantum bus architecture for application towards a larger faulttolerant quantum computing device. Looking ahead, direct extensions to simple error detection and error correction demonstrations will be feasible with existing integration techniques and coherence levels. Overcoming integrated circuit engineering hurdles while preserving long coherence times should pave the way for larger surfaces for quantum error correction.
During the completion of this manuscript, we became aware of similar work by Saira et al.^{33}
Methods
Mapping to SC
The SC requires CNOT gates to be performed between the code and syndrome qubits. To optimize the fidelity of these gates, it is important to use a connective element (connector) to mediate a strong interaction between these qubits. For superconducting transmon qubits, the connections are typically achieved by capacitively coupling two (or more) qubits to a resonator. Since the SC is a twodimensional geometry, connectors should not cross.
The most straightforward layout of the SC is to have the qubits arranged on a square lattice, each connected to its nearest neighbours (see Fig. 5a). This arrangement has considerable overhead, in that each qubit is attached to four connectors. For a realization in a transmonbased architecture it will require each qubit having at least four capacitive couplers (not including readout and control lines). As a result, a significant amount of the qubit charging energy will come from the presence of the couplers.
Previously, DiVincenzo^{19} has shown that it is possible to have a lattice that realizes the SC with each qubit attached only to two connectors. The DiVincenzo lattice is shown in Fig. 6. In this lattice, each small square contains four qubits and is used to represent a single code or syndrome qubit. It is apparent that this arrangement has connectivity between the small squares that is equivalent to the standard square lattice. However, the mapping is inefficient in that it requires four qubits to represent a single qubit in the original SC and furthermore it requires as many as four CNOT operations spanning four connections (resonators) for the furthest reaches of adjacent squares to communicate.
Figure 5b shows three additional arrangements of connections in the square lattice. Connectors are located in every other square of the checkerboard, shown as pink, and each pink square represents either four, two or one connector. In the first case, the pink square is broken up into four connectors, each attached to two qubits and each qubit to four connectors, recovering Fig. 5a. We call this the [2,4] arrangement. The second arrangement is to divide the pink regions into two, with each connector joined to three qubits and each qubit attached to three couplers ([3,3] arrangement).
Finally, the connector can be the entire pink square connecting four qubits with each qubit connecting only to two connectors ([4,2] arrangement). This final arrangement can easily been seen to achieve the desired simultaneous goals of having each qubit attached to just two connections while retaining the connectivity of the square lattice. As shown in Fig. 1, the connector can be realized by using a resonator that connects to four qubits. The alternating horizontal and vertical lines inside the pink squares correspond precisely to the bus resonators.
Two final points are worth discussing. First, note that the [4,2] arrangement is actually more connected than the square lattice, that is, diagonal neighbouring qubits can be coupled. This extra connectivity may be of some advantage, although it comes at the cost of reducing the isolation between these qubits. This is an unavoidable consequence of having more than two qubits per connector. Lastly, DiVincenzo’s lattice is a simple distortion, a skewing, of the [4,2] lattice. Our key observation can be thought of as seeing that his lattice can be directly mapped to the square lattice without having four qubits being used to represent just one qubit from the SC.
Device fabrication
The device is fabricated on a 720μm thick silicon substrate. All superconducting CPW resonators are defined via optical lithography and subtractive reactive ion etching of a sputtered niobium film (200 nm thick). The three singlejunction transmon qubits are patterned using electronbeam lithography, followed by doubleangle deposition of aluminium, with layer thicknesses of 35 and 85 nm. Liftoff process is used to form the final junction structure.
Device parameters
The three transmon qubits (iε[1,3]) have transition frequencies {ω_{i}}/2π={5.0388, 5.0080, 5.2286} GHz, with readout resonators at {ω_{Ri}}/2π={6.698, 6.585, 6.695} GHz, relaxation times {T_{1(i)}}={24, 29, 20} μs, μs. The bus resonators are unmeasured but R_{12} (R_{23}) is designed to resonate at 8 (8.5) GHz. The dispersive cavity shifts of the readout resonators are measured to be {χ_{i}}/π={−2.0, −2.0, −2.3} MHz and the readout resonators have line widths {κ_{Ri}}/2π={443, 976, 793} kHz. All qubits have measured anharmonicities of ~\n−340 MHz. From the above, we calculate coupling strengths {g_{i}}/2π={70, 67, 67} MHz to the readout resonators, which is consistent with electromagnetic simulations.
Experimental setup
The halfplaquette device is cooled to 15 mK in an Oxford Triton dilution refrigerator. A full schematic of the wiring and experimental control hardware is depicted in Fig. 7. Each qubit has its own dedicated readout line with an associated set of isolators and Caltech HEMT (noise temp ~\n6 K) amplifiers. Q2 is unique in that its readout signal is reflected off of a UC Berkeley JPA before going on to the isolator and highelectron mobility transistor (HEMT) chain. The device is housed in a lighttight Ammuneal cryopermshield that is coated throughout with a layer of lossy eccosorb (Emerson & Cuming CR124). Besides explicit cryogenic attenuators at the different stages of the cryostat, all qubits are also attenuated at the lowest temperature stage with inhouse eccosorb coaxial filters.
Outside the cryostat, all microwave qubit control signals are generated via vector modulation combining offtheshelf electronics. The microwave readout signals are pulse modulated using Arbitrary Pulse Sequencers built by Raytheon BBN Technologies. The readout signals are processed via two Alazartech ATS9870 acquisition cards. All singleshot readout traces are processed with various filters, downsampling and an optimal quadrature rotation filter. The details of this full process, which is currently performed explicitly in the acquisition computer, is described in detail in parallel work^{27}. As the devices scale up in the near term towards plaquette and logical qubit demonstrations, the multiple readout signals will need to be processed directly on FieldProgrammable Gate Arrays and then correlated in post processing.
Calibration sequences
Complete tuneup of all microwave gates is accomplished using sets of automated repeated sequences. For singlequbit gates, the repeated calibration sequences are described in a previous publication^{28}.
The crossresonance pulse amplitude is calibrated in close analogy to singlequbit amplitude calibrations. An odd number 2N−1 of ZX_{90} pulses are applied and the amplitude is adjusted so that for each N the expected signal is halfway between 0 and 1. Any amplitude miscalibrations lead to departures from this expected signal and are amplified for increasing N.
In addition to amplitude, we must also calibrate the phase of the ZX_{90} pulse between Q3 and Q2 (as well as Q1 and Q2). In our experiment, we use a separate microwave generator to supply the crossresonance pulse on Q3 at the frequency of Q2. The phase of this microwave signal must be calibrated to match that of the microwave generator supplying the singlequbit pulses on Q2. This is done by applying the pulse sequence IY_{90}(ZU_{180}IX)^{N}IX_{90}. The U denotes the rotation axis defined by the second generator and the goal is to calibrate for an Xrotation. In the case of an Xrotation, we expect the signal to be halfway between 0 and 1 for each N and miscalibration of the phase leads to deviations that are amplified with increasing N. These methods provide a routine for automated calibration with high precision. In the experiments, all crossresonance pulses were calibrated on a regular basis because of phase drift between the two microwave generators.
Randomized benchmarking
All singlequbit gates are 40ns Gaussianshaped microwave pulses (Gaussian width σ=10 ns) resonant with the transition frequencies of the qubits, with scaled derivativeofGaussian shapes applied on the quadrature channel to minimize leakage effects^{30}. The gates are all autonomously calibrated with a set of repeated pulse experiments, correcting for: amplitude of X_{90} and X gates, amplitude imbalance between X and Yrotations, mixer skew and derivativeofGaussianshape parameter. Singlequbit gates are all independently characterized via Clifford^{31} randomized benchmarking (RB) and summarized in Table 1. To characterize the addressability error of the system, we perform simultaneous^{23} RB, applying different sets of randomized singlequbit Clifford gates to all three qubits at the same time. These results are also summarized in Table 1 and essentially indicate that addressability errors are at the 0.1% error level.
The twoqubit ZX_{90} gates for both pairs of qubits are composed as a twoqubit refocusing sequence previously described^{32} and include shaped Gaussian turnon (3σ, σ=24 ns), a flat section and then a Gaussian turnoff, for a total gate time of 350 ns. The ZX_{90} gates are tunedup also using repeated pulse experiments (described in previous section). It is also important to note that the pair of twoqubit gates can be applied simultaneously, as they commute with one another. To characterize the gates, we generate twoqubit Clifford operations^{12} and perform RB. The results for the two cases are shown in Fig. 8, where we show the average fidelity decay over 35 different randomized twoqubit Clifford sequences. Analysing the decay curves gives us error per twoqubit Clifford gate of 0.058±0.003 for Q1 and Q2, and 0.065±0.002 for Q3 and Q2. We find that the reduced chisquare for these fits are 0.583 and 0.385, respectively. This demonstrates that the model is a faithful representation of the data. As each twoqubit Clifford gate is composed of 1.5 ZX_{90} generators, we estimate the twoqubit ZX_{90} gate errors to be 3.8 and 4.3%.
Readout characterization
For this experiment, each qubit has its own measurement resonator. On Q1 and Q3 highpower readout was used and for Q2 a dispersive linear readout with a JPA was used. The readout was performed by using an integrating kernel that takes into account the response of the cavity (see ref. 27 for more details). This is important when most of the information is in the initial transients of the signal. The integration time for the experiment was 4 μs for the highpower readout and 2 μs for the dispersive readout with the JPA.
Shown in Fig. 9 are typical histograms for the three readout channels averaged over all computational basis for the qubits not measured. Here we see that the assignment fidelity, defined by
for the three channels is 0.84, 0.91 and 0.89, respectively. These are typical values and we see about a 2–3% fluctuation over the course of a typical experiment. By fitting a doubleGaussian model to the data, we find that the ratio of the undesired state to the desired state for Q1 prepared in the ground (excited) is 9.9% (22%), for Q2 5.7% (8.3%) and for Q3 13.7% (6.0%). We believe most of the error is due to the highpower nonlinear readout of Q1 and Q3 and is not due to the qubits being initialized in the wrong state. With no power applied to the Q1 and Q3 resonator, the assignment fidelity is 0.95 and the ratios of the two Gaussians are 5.6% when Q2 is prepared in the excited state and negligible when Q2 is prepared in the ground state.
State tomography
For state tomography we used the correlation method as described in ref. 27. The single shots for each measurement resonator are correlated and from a set of complete post rotations we can use either linear inversion or an SDP (with constraints ρ≤ and tr(ρ)=1) to reconstruct the state. The complete set of rotations used are {I, X, X_{+90}, X_{−90}, Y_{+90}, Y_{−90}}^{⊗n}.
Typically, 20,000 shots for each post rotation are used and we find that the statistical error in the measured voltages has signaltonoise ratio (SNR) ~\n1 × 10^{4}, 2 × 10^{4}, 2 × 10^{4} for the three measurement channels M1, M2, M3, respectively. The secondorder correlators range in SNR from ~\n2 to 5 × 10^{3} and the third order has SNR ~\n1,000. Using these and a bootstrapping method^{28} we can estimate the state fidelity and the statistical error. The state fidelity is given by
where ρ_{ideal} is the ideal state and ρ_{noisy} is the reconstructed state.
We find that in all cases the fluctuations in the state fidelity from statistics is much smaller than the difference between the linear reconstruction and the SDP. Furthermore, we find typically the sum of all the negative eigenvalues in the threequbit space to be <0.03.
Measurement tomography
An ideal Zparity check can be described by the quantum operation
where
and the extra system is used to label the outcome of measurement of the syndrome qubit. In the noisy case this is represented by the operation
and the goal of measurement tomography is to determine the conditional maps _{even}(ρ) and _{odd}(ρ). These quantum operations are completely positive but not trace preserving.
By binning the results of the measurement on the syndrome qubit, tomography on the twoqubit subspace is performed by preparing a complete set of different input states and measurement bases via pre and post rotations, and reconstructing the operations from the measurement results. The complete set of rotations that we use are the same as those used in state tomography. We use both a linear reconstruction and a minimization to make the maps physical. For more details on how quantum process tomography can be performed, see ref. 28.
We use the Pauli transfer matrix^{28} defined by
to represent the measurement operations where P_{j}s are the standard Pauli operators {I, X, Y, Z}^{⊗2}.
To quantify the measurement, we define the measurement fidelity by a generalization of the average fidelity. Since the measurement maps are not trace preserving, we need to use normalized outputs , and , where , tr refers to the trace norm and x={even, odd}. Doing this gives
Since the nullspace of a projection operation has measure zero and the noisy realization typically will also have a nullspace of zero measure, this integral is well defined. To compute this, we draw 150,000 different random states from the Fubini–Study measure and compute the average.
One could also define a process fidelity by computing the state fidelity between normalized Choi matrices of the ideal and noisy operations
however, for nonunitary processes there is no simple relationship between them.
The unconditional map can be defined by tracing equation (5) over the syndrome qubit giving
Since this is a quantum operation, the standard fidelity between quantum operations can be used.
Additional information
How to cite this article: Chow, J. M. et al. Implementing a strand of a scalable faulttolerant quantum computing fabric. Nat. Commun. 5:4015 doi: 10.1038/ncomms5015 (2014).
References
 1.
Aharonov, D. & BenOr, M. Faulttolerant quantum computation with constant error. InSTOC '97 176–188ACM, New York, NY (1997).
 2.
Preskill, J. Reliable quantum computers. Proc. R. Soc. Lond. A 454, 385–410 (1998).
 3.
Knill, E. Quantum computing with realistically noisy devices. Nature 434, 39–44 (2005).
 4.
Bravyi, S. & Kitaev, A. Quantum codes on a lattice with boundary. Preprint at http://arXiv.org/quantph/9811052 (1998).
 5.
Dennis, E., Kitaev, A., Landahl, A. & Preskill, J. Topological quantum memory. J. Math. Phys. 43, 4452–4505 (2002).
 6.
Raussendorf, R. & Harrington, J. Faulttolerant quantum computation with high threshold in two dimensions. Phys. Rev. Lett. 98, 190504 (2007).
 7.
Fowler, A. G., Mariantoni, M., Martinis, J. M. & Cleland, A. N. Surface codes: towards practical largescale quantum computation. Phys. Rev. A. 86, 032324 (2012).
 8.
Paik, H. et al. Observation of high coherence in Josephson junction qubits measured in a threedimensional circuit QED architecture. Phys. Rev. Lett. 107, 240501 (2011).
 9.
Rigetti, C. et al. Superconducting qubit in a waveguide cavity with a coherence time approaching 0.1ms. Phys. Rev. B 86, 100506 (2012).
 10.
Chang, J. B. et al. Improved superconducting qubit coherence using titanium nitride. Appl. Phys. Lett. 103, 012602 (2013).
 11.
Barends, R. et al. Coherent Josephson qubit suitable for scalable quantum integrated circuits. Phys. Rev. Lett. 111, 080502 (2013).
 12.
Corcoles, A. D. et al. Process verification of twoqubit quantum gates by randomized benchmarking. Phys. Rev. A 87, 030301 (2013).
 13.
Barends, R. et al. Superconducting quantum circuits at the surface code threshold for fault tolerance. Nature 508, 500–503 (2014).
 14.
Bergeal, N. et al. Phasepreserving amplification near the quantum limit with a Josephson ring modulator. Nature 465, 64–68 (2010).
 15.
Riste, D., van Leeuwen, J. G., Ku, H.S., Lehnert, K. W. & DiCarlo, L. Initialization by measurement of a superconducting quantum bit circuit. Phys. Rev. Lett. 109, 050507 (2012).
 16.
Johnson, J. E. et al. Heralded state preparation in a superconducting qubit. Phys. Rev. Lett. 109, 050506 (2012).
 17.
Majer, J. et al. Coupling superconducting qubits via a cavity bus. Nature 449, 443–447 (2007).
 18.
Johnson, B. R. et al. Quantum nondemolition detection of single microwave photons in a circuit. Nat. Phys. 6, 663–667 (2010).
 19.
DiVincenzo, D. P. Faulttolerant architectures for superconducting qubits. Phys. Scr. T137, 014020 (2009).
 20.
DiCarlo, L. et al. Preparation and measurement of threequbit entanglement in a superconducting circuit. Nature 467, 574–578 (2010).
 21.
Groen, J. P. et al. Partialmeasurement backaction and nonclassical weak values in a superconducting circuit. Phys. Rev. Lett. 111, 090506 (2013).
 22.
Hatridge, M., Vijay, R., Slichter, D. H., Clarke, J. & Siddiqi, I. Dispersive magnetometry with a quantum limited SQUID parametric amplifier. Phys. Rev. B 83, 134501 (2011).
 23.
Gambetta, J. M. et al. Characterization of addressability by simultaneous randomized benchmarking. Phys. Rev. Lett. 109, 240504 (2012).
 24.
Paraoanu, G. S. Microwaveinduced coupling of superconducting qubits. Phys. Rev. B 74, 140504 (2006).
 25.
Rigetti, C. & Devoret, M. Fully microwavetunable universal gates in superconducting qubits with linear couplings and fixed transition frequencies. Phys. Rev. B 81, 134507 (2010).
 26.
Reed, M. D. et al. Highfidelity readout in circuit quantum electrodynamics using the JaynesCummings nonlinearity. Phys. Rev. Lett. 105, 173601 (2010).
 27.
Ryan, C. A. et al. Tomography via Correlation of Noisy Measurement Records. Preprint at http://arxiv.org/abs/1310.6448 (2013).
 28.
Chow, J. M. et al. Universal quantum gate set approaching faulttolerant thresholds with superconducting qubits. Phys. Rev. Lett. 109, 060501 (2012).
 29.
Merkel, S. T. et al. Selfconsistent quantum process tomography. Phys. Rev. A 87, 062119 (2013).
 30.
Motzoi, F., Gambetta, J. M., Rebentrost, P. & Wilhelm, F. K. Simple pulses for elimination of leakage in weakly nonlinear qubits. Phys. Rev. Lett. 103, 110501 (2009).
 31.
Magesan, E., Gambetta, J. M. & Emerson, J. Scalable and robust randomized benchmarking of quantum processes. Phys. Rev. Lett. 106, 180504 (2011).
 32.
Corcoles, A. D. et al. Protecting superconducting qubits from radiation. Appl. Phys. Lett. 99, 181906 (2011).
 33.
Saira, O.P. et al. Entanglement genesis by ancillabased parity measurement in 2D circuit QED. Phys. Rev. Lett. 112, 070502 (2014).
Acknowledgements
We thank M.B. Rothwell and G.A. Keefe for fabricating devices. We thank J.R. Rozen and J. Rohrs for experimental contributions, and D. DiVincenzo and G. Smith for engaging theoretical discussions. We thank I. Siddiqi for providing the Josephson Parametric Amplifier. We acknowledge Caltech for HEMT amplifiers. We acknowledge support from IARPA under contract W911NF1010324. All statements of fact, opinion or conclusions contained herein are those of the authors and should not be construed as representing the official views or policies of the US Government.
Author information
Affiliations
IBM T.J. Watson Research Center, Yorktown Heights, New York 10598, USA
 Jerry M. Chow
 , Jay M. Gambetta
 , Easwar Magesan
 , David W. Abraham
 , Andrew W. Cross
 , Nicholas A. Masluk
 , John A. Smolin
 , Srikanth J. Srinivasan
 & M Steffen
Raytheon BBN Technologies, Cambridge, Massachusetts 02138, USA
 B R Johnson
 & Colm A. Ryan
Authors
Search for Jerry M. Chow in:
Search for Jay M. Gambetta in:
Search for Easwar Magesan in:
Search for David W. Abraham in:
Search for Andrew W. Cross in:
Search for B R Johnson in:
Search for Nicholas A. Masluk in:
Search for Colm A. Ryan in:
Search for John A. Smolin in:
Search for Srikanth J. Srinivasan in:
Search for M Steffen in:
Contributions
J.M.C. and J.M.G. designed the experiments. J.M.G. and J.A.S. came up with the proposed lattice. D.W.A. and J.M.G. performed simulations of devices. J.M.C., D.W.A. and S.J.S. characterized devices. J.M.C., J.M.G., A.W.C. and E.M. interpreted and analysed the experimental data. N.A.M. setup the microwave control hardware. E.M. and J.M.G. developed the measurement tomography formalism. B.R.J. and C.A.R. developed readout hardware and analysis for the correlated readouts. All authors contributed to the composition of the manuscript.
Competing interests
The authors declare no competing financial interests.
Corresponding author
Correspondence to Jerry M. Chow.
Rights and permissions
To obtain permission to reuse content from this article visit RightsLink.
About this article
Further reading

Bellstate generation on remote superconducting qubits with dark photons
Quantum Information Processing (2018)

Onestep implementation of a multitargetqubit controlled phase gate in a multiresonator circuit QED system
Quantum Information Processing (2018)

Building logical qubits in a superconducting quantum computing system
npj Quantum Information (2017)

Fast nonAbelian geometric gates via transitionless quantum driving
Scientific Reports (2016)

Controllable highfidelity quantum state transfer and entanglement generation in circuit QED
Scientific Reports (2016)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.