SCIENCE ADVANCES | RESEARCH ARTICLE

QUANTUM INFORMATION Copyright © 2017

The Authors, some

Fault-tolerant quantum error detection rights reserved;

exclusive licensee
American Association
Norbert M. Linke,1* Mauricio Gutierrez,2† Kevin A. Landsman,1 Caroline Figgatt,1 for the Advancement
Shantanu Debnath,1‡ Kenneth R. Brown,2 Christopher Monroe1,3 of Science. No claim to
original U.S. Government
Quantum computers will eventually reach a size at which quantum error correction becomes imperative. Quantum Works. Distributed
information can be protected from qubit imperfections and flawed control operations by encoding a single logical under a Creative
qubit in multiple physical qubits. This redundancy allows the extraction of error syndromes and the subsequent Commons Attribution
detection or correction of errors without destroying the logical state itself through direct measurement. We show the NonCommercial
encoding and syndrome measurement of a fault-tolerantly prepared logical qubit via an error detection protocol on License 4.0 (CC BY-NC).
four physical qubits, represented by trapped atomic ions. This demonstrates the robustness of a logical qubit to
imperfections in the very operations used to encode it. The advantage persists in the face of large added error rates
and experimental calibration errors.

INTRODUCTION pffiffiffi
j01iL ¼ ðj0011i þ j1100〉Þ= 2 ð1BÞ

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

The discovery of quantum error correction codes gave credibility to the idea
of scaling up physical quantum systems to arbitrary sizes (1–3). Showing
that all elements of error correction can be realized in a fault-tolerant way is pffiffiffi
therefore of fundamental interest. Fault tolerance removes the assump- j10iL ¼ ðj0101i þ j1010〉Þ= 2 ð1CÞ
tion of perfect encoding and decoding of logical qubits (4), because the
logical error probability scales as a convex function of the physical error pffiffiffi
probability for small errors (5). Although several experiments have j11iL ¼ ðj0110i þ j1001〉Þ= 2 ð1DÞ
shown a reduction of high intrinsic or artificially introduced errors in pffiffiffi
logical qubits (6–14), fault-tolerant encoding of a logical qubit has never and with j±i ¼ ðj0i±j1〉Þ= 2 , we can write them down along X as
been demonstrated. We note that there are subtle issues with respect to follows
the definition of fault tolerance that are beyond the scope of this paper.
Here, we implement a four-qubit error detection code with two stabili- pffiffiffi
zers (see Fig. 1). This leaves two possible encoded qubits, La and Lb, for which jþþiL ¼ ðj þ þ þ þi þ j    〉Þ= 2 ð2AÞ
errors can be detected: a [[4, 2, 2]] code (15, 16). The preparation and error
detection procedures considered here are fault-tolerant on only a single pffiffiffi
encoded qubit. From a fault tolerance perspective, this is a [[4, 1, 2]] sub- jþiL ¼ ðj þ  þ i þ j  þ  þ〉Þ= 2 ð2BÞ
system code where the logical qubit La is protected and the gauge qubit Lb
is not. As such, the code was used in experiments with photonic qubits pffiffiffi
(17, 18). By instead considering errors on both encoded qubits, we jþiL ¼ ðj þ þ  i þ j   þ þ〉Þ= 2 ð2CÞ
highlight the importance of fault tolerance for reducing intrinsic errors
and managing error propagation. The non–fault-tolerant procedures pffiffiffi
that generate Lb still succeed in reducing added errors. jiL ¼ ðj þ   þi þ j  þ þ 〉Þ= 2 ð2DÞ
The code implements La and Lb on only four physical qubits and
hence violates the quantum Hamming bound (5), which means that The encoding of different initial states is shown in Fig. 2 (A to D).
detected errors cannot be uniquely identified and corrected. We The fault tolerance arises because the circuits for encoding and syn-
must therefore rely on postselection to find and discard cases where drome extraction are carefully constructed such that a single physical
an error occurred. The code does have the advantage of requiring qubit error occurring anywhere cannot lead to an undetectable error
only five physical qubits for the fault-tolerant encoding of La: four on logical qubit La. It comes at the cost of the logical gauge qubit Lb,
data qubits and one ancilla qubit. for which there is such an undetectable error channel. An example of
The logical codewords |LaLb〉L in the computational or Z-basis are this is shown in Fig. 2G.
With the Pauli operators X, Y, and Z and the identity I, the logical
pffiffiffi operators are
j00iL ¼ ðj0000i þ j1111〉Þ= 2 ð1AÞ
1
Joint Quantum Institute and Joint Center for Quantum Information and Computer Za ¼ Z⊗Z⊗I⊗I ð3AÞ
Science, University of Maryland Department of Physics and National Institute of
Standards and Technology, College Park, MD 20742, USA. 2Schools of Chemistry and
Biochemistry, Computational Science and Engineering, and Physics, Georgia Institute Zb ¼ Z⊗I⊗Z⊗I ð3BÞ
of Technology, Atlanta, GA 30332, USA. 3IonQ Inc., College Park, MD 20742, USA.
*Corresponding author. Email: [email protected]
†Present address: Department of Physics, College of Science, Swansea University, Xa ¼ X⊗I⊗X⊗I ð3CÞ
Singleton Park, Swansea SA2 8PP, UK.
‡Present address: Department of Physics, University of California, Berkeley, Berkeley,
CA 94720, USA. Xb ¼ X⊗X⊗I⊗I ð3DÞ

Linke et al., Sci. Adv. 2017; 3 : e1701074 20 October 2017 1 of 6

SCIENCE ADVANCES | RESEARCH ARTICLE

A B
1 2 Za 1 2 Xb 1 2 1 2

3 4 3 4 3 4 3 4
Xa Zb Sz Sx
Fig. 1. Graphical representation of the logical operators and stabilizers de-
fining the [[4, 2, 2]] code on physical qubits 1 to 4. The structure of the logical C D
operators X and Z for the two encoded qubits La and Lb, and for the two stabi-
lizers Sx and Sz, is defined in Eqs. 3 and 4.

With these operators and the circuits given in Fig. 2 (A to D), any E F
state |LaLb〉L can be generated maintaining the fault tolerance of La.
The [[4, 2, 2]] code has the additional advantage that, in contrast to
other codes (19), fault-tolerant syndrome extraction for the logical qubit
La can be achieved using a bare ancilla, that is, an ancilla qubit that is
not itself a logical qubit. The stabilizers to extract logical phase-flip (Z)
and bit-flip (X) errors are Sx and Sz, respectively

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

G
Sx ¼ X⊗X⊗X⊗X ð4AÞ

Sz ¼ Z⊗Z⊗Z⊗Z ð4BÞ

As in a Bacon-Shor code block (20, 21), the code space together with Fig. 2. Circuit diagrams. (A to D) Circuits for the encoding of four different logical
the logical operators and stabilizers form a subsystem that allows local states constructed such that logical qubit La is prepared fault-tolerantly. Any logical
syndrome extraction similar to that of Napp and Preskill (22), as de- state can be achieved by applying single logical qubit operators to states en-
picted in Fig. 1. The difference is that the stabilizers have weight 4 coded as shown here. (E and F) Circuits for the two stabilizers Sx and Sz, which
because we simultaneously extract information about the gauge qubit project Z- and X-type errors, respectively, onto an ancilla qubit a. Note that a
controlled Z-gate is realized by an inverted CNOT with the ancilla in the Z-basis
Lb. Applying these stabilizers conditional on the state of an ancilla qubit
as the target. (G) Example of fault-tolerant construction of circuits for logical qubit
extracts the parity of the data qubits along X or Z (see Fig. 2, E and F). La: The encoding circuit for |00〉L has a single nondetectable error channel. A bit-flip
Measuring the ancilla yields either |0〉, indicating no error, or |1〉, error E occurring as shown can change the state to |01〉L, which is an error on the
meaning an error has occurred and the run is to be discarded. With only logical gauge qubit Lb. Logical qubit La is prepared fault-tolerantly. This property
one ancilla qubit available, we measure the two stabilizers in separate holds for all circuits (A to F).
experiments. Because we prepare eigenstates of logical Pauli operators,
only logical Pauli operations that change the ideal state result in errors.
Both stabilizer measurements serve to determine the overall yield, that
is, the fraction of runs for which no error was indicated. In addition to
the error checks provided by stabilizer measurements, only even-parity probability. The data yield 91.1(4)% even-parity outcomes from the
outcomes are accepted when the data qubits are measured at the end of four data qubits. Breaking these results down by logical state gives
the circuit. Note that similar weight 4 stabilizers have recently been im- 98.0(2)% population in the target state |00〉L. The error falls almost
plemented in superconducting qubits (23). entirely on |01〉L, which corresponds to a 1.7(2)% error on the non–fault-
We implement the [[4, 2, 2]] code on a fully connected quantum tolerantly prepared gauge qubit Lb. The 0.11(6)% error exclusively on
computer comprising a chain of five single 171Yb+ ions confined in La is an order of magnitude lower than this, and at a similar level to
a Paul trap (see Materials and Methods). The state-detection fidelity the 0.18(7)% logical two-qubit error resulting in |11〉L. For the logical
for a single qubit is 99.7(1)% for state |0〉 and 99.1(1)% for state |1〉. A state preparation step, both of these small erroneous state populations
general five-qubit state is detected with 95.7(1)% fidelity. Single- and are dominated by physical readout errors.
two-qubit gate fidelities are typically 99.1(5) and 97(1)%, respectively. With |00〉L thus prepared, we apply in turn the two stabilizers Sz
Typical gate times are 20 and 250 ms for single- and two-qubit gates, and Sx, shown in Fig. 2 (E and F), for nondemolition syndrome extrac-
respectively. The computational gates H and CNOT are generated by tion. The results are shown in Fig. 3B. Populations in the odd-numbered
combining several physical-level single- and two-qubit gates in a states reflect events where an error is detected by a stabilizer. The results
modular fashion (24). of the logical states are similar, with |00〉L populations of 97.8(2) and
97.1(3)%, respectively, and the errors occur predominantly in the
non–fault-tolerantly prepared gauge qubit Lb. The errors on La are
RESULTS 0.2(1)%, similar to the error floor given by the |11〉L population, which is
We start by preparing state |00〉L using the circuit shown in Fig. 2A. slightly higher than after mere state preparation due to the additional
The results of measuring this state directly after preparation are gates introduced by the stabilizers. Sx introduces X-type errors in the
shown in Fig. 3A. The target states 0 (|00000〉) and 30 (|11110〉) show system, which can be seen from a higher Lb error. Sz introduces Z-type
that we succeed in preparing and measuring this state with ≃90% errors, which do not affect |00〉L. The opposite is true when applying

Linke et al., Sci. Adv. 2017; 3 : e1701074 20 October 2017 2 of 6

SCIENCE ADVANCES | RESEARCH ARTICLE

A B
50 50
Prep, S
z
40 40 Prep, S
x
Probability / %

Probability / %
100 98.0 100 97.8 100 97.1
30 30

Prob. / %

Prob. / %

Prob. / %
10 10 10
1.7 1.7 2.4
1 1 1
20 20
0.1 0.2 0.1 0.3 0.1 0.3

0 0 0.2 0 0.2
0.1 1 1 1
10 1 0 10 1 0 1 0
La (FT) Lb (NFT) La (FT) Lb (NFT) La (FT) Lb (NFT)

0 0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
Detected state Detected state
Fig. 3. Results achieved with logical state |00〉L. (A) Results from the preparation of state |00〉L. The abscissa represents the five-qubit states in decimal. We succeed in
preparing (prep) and measuring the state with ~90% probability. The inset shows the result after postselection on the state being in the logical basis, that is, even
parity. It is broken down by the logical state of the fault-tolerantly (FT) prepared qubit La and the non–fault-tolerantly (NFT) prepared qubit Lb. (B) Results of the
stabilizer measurements after preparation of |00〉L. The yields are 77.8(6) and 65.2(7)% for Sz and Sx, respectively. The insets show that the error probability on the
fault-tolerantly prepared and stabilized logical qubit La is an order of magnitude below the non–fault-tolerantly prepared and stabilized qubit Lb.

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

Table 1. Probability distributions (in percentage) of measured logical states |LaLb〉 for various prepared logical states in each row, with and without
stabilizers Sx or Sz applied. The measurement basis is shown in the last column. The logical states are |00 > L … |11 > L, measured in the Z-basis, and | + + 〉L …
| − − 〉L, measured in the X-basis. The very low error probability on the first logical qubit La compared to Lb shows the action of its fault-tolerant construction. We
run every circuit 5000 to 6000 times. The results without stabilizer show the number of rejected runs from the parity check on the data qubits (typically ~8%)
whereas the additional discard on the results with stabilizer (typically ~20%) is due to the ancilla result. The physical errors for state preparation and mea-
surement are 0.3(1)% for states |0〉 and | + 〉, and 1.2(1)% for states |1〉 and | − 〉.
Yield Measured logical state |LaLb〉 Measurement basis
|00〉 |01〉 |10〉 |11〉 Z
|++〉 |+−〉 |−+〉 |−−〉 X

|00〉L 91.1(4) 98.0(2) 1.7(2) 0.11(6) 0.18(7) Z

|00〉LSz 77.8(6) 97.8(2) 1.7(2) 0.18(8) 0.3(1) Z

|00〉LSx 65.2(7) 97.1(3) 2.4(3) 0.2(1) 0.3(1) Z

| + + 〉L 91.1(4) 95.7(3) 3.9(3) 0.24(8) 0.22(8) X

| + + 〉LSz 68.2(7) 93.0(5) 4.2(4) 1.3(2) 1.5(2) X

| + + 〉LSx 72.1(6) 94.3(4) 4.5(4) 0.5(1) 0.7(2) X

| − 1〉L 90.1(4) 0.22(8) 50.5(8) 0.09(6) 49.2(8) Z

| − 1〉L 87.0(5) 0.3(1) 0.3(1) 50.4(8) 48.9(8) X

| − 1〉LSz 79.9(6) 0.15(7) 50.0(8) 0.10(6) 49.8(8) Z

| − 1〉LSz 75.5(6) 0.4(1) 0.3(1) 50.1(8) 49.2(8) X

| − 1〉LSx 72.1(6) 0.6(1) 50.2(8) 0.5(1) 48.7(8) Z

| − 1〉LSx 76.2(5) 0.4(1) 0.4(1) 50.0(7) 49.2(7) X

|0 + 〉L 93.2(3) 47.4(5) 52.5(5) 0.06(3) 0.05(3) Z

|0 + 〉L 92.4(4) 50.0(8) 0.04(4) 49.8(8) 0.09(5) X

|0 + 〉LSz 81.6(6) 48.3(8) 51.4(8) 0.17(8) 0.17(8) Z

|0 + 〉LSz 68.5(7) 47.1(9) 2.4(3) 47.4(9) 3.1(3) X

|0 + 〉LSx 72.0(6) 48.3(8) 51.5(8) 0.16(7) 0.12(7) Z

|0 + 〉LSx 70.9(7) 49.4(9) 0.4(1) 49.7(9) 0.5(1) X

|11〉LSz 73.3(6) 0.4(1) 0.3(1) 2.8(3) 96.5(3) Z

Linke et al., Sci. Adv. 2017; 3 : e1701074 20 October 2017 3 of 6

SCIENCE ADVANCES | RESEARCH ARTICLE

the stabilizers to |++〉 instead. Table 1 (see Materials and Methods) of a physical qubit is measured with 98.8(2)% fidelity, nearly a factor
summarizes results for different logical states prepared with the circuits of 2 worse than La. The infidelity in this case is dominated by the single-
shown in Fig. 2 (A to D). The circuit elements that dominate the intrin- qubit detection error of 0.9(1)% for |1〉, which the code successfully
sic errors in our system are the two-qubit gates. Note that after a circuit suppresses in La.
with seven CNOT gates, each of which introduces 3 to 4% infidelity, we To further investigate the robustness of the code, we add two kinds
obtain the correct answer |00〉L with 97 to 98% probability. The gauge of error to the system. First, we deliberately introduce single- and two-
qubit Lb circuit failures occur at approximately the error rate of one qubit Pauli errors and study how errors on La and Lb scale with increasing
two-qubit gate, whereas La errors are suppressed substantially below physical qubit errors. Instead of trying to reproduce a stochastic error
that level to <1%. These results show the power of fault-tolerant channel, which can be tedious for low error rates (25), we sample the
preparation and stabilizer measurement. The circuits succeed in dis- various error configurations and then multiply them by their respective
carding nearly all errors, but we pay a price because the yield is in the statistical importance to obtain a logical error probability (see Materials
65 to 75% range. We must expect to discard around half of the runs and Methods). We further compare our experimental results to an exact
when measuring both stabilizers. The yield is higher for preparation simulation with optimized error parameters (see Materials and Methods).
of logical states without syndrome measurements because there are The results are shown in Fig. 4A. The clear separation between the two
fewer gates to introduce error and there is only a single selection step. logical qubits is persistent until they converge above 20% introduced
The [[4, 2, 2]] code allows transversal operations, that is, single-qubit error and approach the curve for the theoretical case without intrinsic
logical gates that are generated by applying single-qubit physical gates. errors. In this example, a physical qubit prepared in state |0〉 is outper-
To show an example of this, we prepare |00〉L followed by the logical formed by logical qubit La over the entire range, although our measure-

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

XaXb operation consisting of X-gates on physical qubits 2 and 3. This ment uncertainty limits the significance of this to above p ~ 0.07%. Lb
gives |11〉L, on which we apply the Sx stabilizer followed by readout. outperforms the physical qubit above 4% added error (solid black line in
The yield is 77.3(6)%, and the logical state populations are as follows: Fig. 4A). For state preparations |−〉 and |1〉, La also outperforms the
|00〉L, 0.4(1)%; |01〉L, 0.3(1)%; |10〉L, 2.8(3)%; and |11〉L, 96.5(3)%. physical error based on circuits of preparation and measurement,
Apart from surpassing Lb as before, the La error of 0.7% also outper- whereas for |+〉, the errors are consistent within statistical uncertainty
forms the physical qubit. We find that after an X-gate, the correct state (see Table 1).

A Physical qubit B Prep, Sz C Prep, Sz − Lb

Logical qubit L b experiment Prep, Sx − La Prep, Sx − Lb
Logical qubit L b simulation Prep, Sz − La
90
Logical qubit L a experiment Prep, Sx − La
Logical qubit L a simulation 85
10
No intrinsic error
80

10 75
Logical error probability / %
Logical error probability / %

Yield / %

70

65 1
NFT
al
ic

1 60
ys
Ph

55

50
FT

0.1 45 0.1
0.1 1 10 100 −0.05 0 0.05 0.1 −0.05 0 0.05 0.1
Inserted error probability / % Miscalibration α Miscalibration α
Fig. 4. Performance of the code under different kinds of artificial errors. (A) Logical error probability under artificially introduced stochastic Pauli errors. Uncertainties
shown in gray with dashed outlines. We prepare state |00〉L, introduce a specific error, and apply Sz before readout. The parameter values for the curves (see Materials and
Methods) corresponding to the two logical qubits are determined either experimentally (solid lines) or from simulation (dashed lines). The black curve shows the limiting
theoretical case without intrinsic errors (see Materials and Methods). At low added error rates, the intrinsic errors dominate, and the fault-tolerantly constructed qubit La starts
about an order of magnitude below the non–fault-tolerantly constructed qubit Lb. With increasing inserted error probability, the added Pauli errors become dominant, and the
La/b curves converge and approach the theory curve without intrinsic error. The solid black line shows the error rate for a single physical qubit. La results in a lower error across
the entire range relative to the physical qubit, although our measurement uncertainty means that this is no longer significant below p ~ 0.07%. The Lb error is lower than the
physical qubit for added errors >4%. (B and C) Preparing |00〉L p and measuring Sx/z with purposefully miscalibrated two-qubit gates, known as XX-gates. A miscalibration of a
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
means that the Bell state produced by the gate is imbalanced: 0:5  aj00i þ i 0:5 þ aj11i. The yields diminishing with miscalibration for the stabilizer measurements are
shown in (B), whereas the errors on the logical qubits presented in (C) remain similar, with La errors about an order of magnitude lower than Lb errors.

Linke et al., Sci. Adv. 2017; 3 : e1701074 20 October 2017 4 of 6

SCIENCE ADVANCES | RESEARCH ARTICLE

Second, we run the |00〉L data with purposefully miscalibrated two- appear in a particular error configuration is given by its weight w,
qubit gates. The results for the logical errors are shown in Fig. 4C. The and its probability of occurrence or statistical importance is po =
error gap of nearly an order of magnitude between La and Lb persists (p/3)w(1 − p)4−w. The probability of a logical error is given by
over a wide range of calibration errors, which are absorbed into a
reduced yield as shown in Fig. 4B. This proves that the code succeeds
in protecting qubit La against intrinsic systematic errors. pL ¼
∑e poðeÞ⋅pa ðeÞ⋅pf ðeÞ ð5Þ
∑e po ðeÞ⋅pa ðeÞ
DISCUSSION The sum runs over all error configurations. pa is the yield, that is, the
Note that the [[4, 2, 2]] code is relevant beyond its limited immediate probability that a run is accepted, and pf is the probability of failure after
application as an error detection code. It forms the base encoding layer postselection, that is, the probability that an accepted run suffers a log-
of the high-threshold Knill C4/C6 code (26) and of a recent proposal for ical error. The dividend is the number of accepted runs with a logical
a topological code (27), and it is equivalent to one face of the distance error, whereas the divisor is the number of accepted runs (both divided
3 color code (28) or the Steane code (2). The code is robust to the high by the total number of runs). The parameters pa and pf were found from
levels of intrinsic errors present in current realizations of quantum com- either experiment or simulation (see Fig. 4A). Out of the total number of
puters. We find no evidence of unexpected two-qubit correlated errors, error configurations, nðwÞ ¼ w3 ð 4 Þ. We covered the error configura-
w
which are always assumed absent when constructing error correction tions of weight 0 and 1 exhaustively. For the w = 2 subset, we only
procedures. Therefore, our results both serve as a demonstration that sampled 27 representative configurations out of the total 54 and doubled

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

this underlying model is correct and pave the way toward error-corrected their weight. The weight 3 and weight 4 subsets were not sampled,
quantum computations on a larger scale. and their logical error rates were set to zero, because their statistical
importance is significant only at very high added error rates. In the
limit of no intrinsic errors, that is, perfect gates, preparation, and mea-
MATERIALS AND METHODS surement, both logical qubits had the same error rate under this model
Experimental system (dash-dotted line in Fig. 4A). We found this error rate from Eq. 5 by
The experiment was performed on a quantum computer consisting of counting accepted error configurations with w ≤ 2 (denominator)
a chain of five single 171Yb+ ions confined in a Paul trap and laser-cooled and checking which of those caused an error (numerator).
near the motional ground state. Each ion provided one physical qubit
in the form of a pair of states in the hyperfine-split 2S1/2 ground level
16ð1  pÞ2 ðp=3Þ2
with an energy difference of 12.642821 GHz, which is magnetic field– p*L ¼ ð6Þ
independent to first order. This so-called “atomic clock” qubit has a ð1  pÞ4 þ 4ð1  pÞ3 p=3 þ 30ð1  pÞ2 ðp=3Þ2
typical coherence time of 0.5 s, which could be straightforwardly extended
by suppressing magnetic field noise. All qubits were collectively initia- The dashed curves in Fig. 4A were obtained by performing a full
lized by optical pumping and measured via state-dependent fluores- density matrix simulation of the five-qubit circuits. We used a simplified
cence detection (29). Each ion was mapped to a distinct channel of a error model to emulate experimental errors. The model had three
photomultiplier tube array. Its state could be detected with 99.4(1)% independent parameters corresponding to errors associated with over-
average fidelity, although a five-qubit state was read out with 95.7(1)% or underrotations after (i) single-qubit and (ii) two-qubit gates and (iii)
average fidelity, limited by channel-to-channel cross-talk. Qubit ma- phase errors caused by Stark shifts. An experimentally found state-
nipulation was achieved by applying two Raman beams from a single transfer matrix was used to take state preparation and detection errors,
355-nm mode-locked laser, which formed beat notes near the qubit including cross-talk, into account. We then optimized the model over
frequency. The first Raman beam was a global beam applied to the entire the parameter space to minimize the difference between the final state
chain, whereas the second was split into individual addressing beams, populations of the experimental and simulated circuits. The resulting
each of which could be switched independently to target any single qubit values for the error rates were 0.50, 1.0, and 1.4%, respectively.
(24). Single qubit gates were generated by driving resonant Rabi rota- The physical error curve in Fig. 4A is the straight line pp = r +
tions of defined phase, amplitude, and duration. Two-qubit gates (so- (2/3 Fx)p, where r = 0.003 is the readout error for a physical qubit in
called XX-gates) were realized by illuminating two ions with beat-note state |0〉. The slope is 2/3 Fx because one in three added errors is a
frequencies near the motional sidebands and creating an effective spin- Z-type error, which does not affect |0〉, and Fx = 0.997 is the success
spin (Ising) interaction via transient entanglement between the state of probability of a physical spin flip operation.
two ions and all modes of motion (30–32). To ensure that the motion
was left disentangled from the qubit states at the end of the interaction,
we used a pulse shaping scheme by modulating the amplitude of the REFERENCES AND NOTES
global beam (33, 34). 1. A. R. Calderbank, P. W. Shor, Good quantum error-correcting codes exist. Phys. Rev. A 54,
1098–1105 (1996).
2. A. M. Steane, Multiple-particle interference and quantum error correction. Proc. R. Soc. A
Artificial stochastic errors 452, 2551–2577 (1996).
To analyze how the code copes with artificially introduced stochastic 3. J. Preskill, Reliable quantum computers. Proc. R. Soc. A 454, 385 (1998).
errors, we prepared logical state |00〉L and added a specific Pauli error, 4. M. A. Nielsen, I. L. Chuang, Quantum Computation and Quantum Information: 10th
for example, I ⊗ X ⊗ Y ⊗ I. We then applied the Sz stabilizer and Anniversary Edition (Cambridge Univ. Press, ed. 10, 2011).
5. D. Gottesman, An introduction to quantum error correction and fault-tolerant
measured the state. We repeated this for different error configurations quantum computation, in Quantum Information Science and Its Contributions to
e. The error probability p on a physical qubit corresponds to an X, Y, or Mathematics, S. J. Lomonaco Jr., Ed. (American Mathematical Society, 2009),
Z error, each occurring with probability p/3. The number of errors that vol. 68, pp. 13–59.

Linke et al., Sci. Adv. 2017; 3 : e1701074 20 October 2017 5 of 6

SCIENCE ADVANCES | RESEARCH ARTICLE

6. J. Chiaverini, D. Leibfried, T. Schaetz, M. D. Barrett, R. B. Blakestad, J. Britton, W. M. Itano, 24. S. Debnath, N. M. Linke, C. Figgatt, K. A. Landsman, K. Wright, C. Monroe, Demonstration
J. D. Jost, E. Knill, C. Langer, R. Ozeri, D. J. Wineland, Realization of quantum error of a small programmable quantum computer with atomic qubits. Nature 536, 63–66
correction. Nature 432, 602–605 (2004). (2016).
7. P. Schindler, J. T. Barreiro, T. Monz, V. Nebendahl, D. Nigg, M. Chwalla, M. Hennrich, R. Blatt 25. S. Bravyi, A. Vargo, Simulation of rare events in quantum error correction. Phys. Rev. A 88,
, Experimental repetitive quantum error correction. Science 332, 1059–1061 (2011). 062308 (2013).
8. B. P. Lanyon, P. Jurcevic, M. Zwerger, C. Hempel, E. A. Martinez, W. Dür, H. J. Briegel, 26. E. Knill, Quantum computing with realistically noisy devices. Nature 434, 39–44 (2005).
R. Blatt, C. F. Roos, Measurement-based quantum computation with trapped ions. 27. B. Criger, B. Terhal, Noise thresholds for the [[4, 2, 2]]-concatenated toric code. QIC 16,
Phys. Rev. Lett. 111, 210501 (2013). 1261–1281 (2016).
9. G. Waldherr, Y. Wang, S. Zaiser, M. Jamali, T. Schulte-Herbrüggen, H. Abe, T. Ohshima, 28. H. Bombin, M. A. Martin-Delgado, Topological quantum distillation. Phys. Rev. Lett. 97,
J. Isoya, J. F. Du, P. Neumann, J. Wrachtrup, Quantum error correction in a solid-state 180501 (2006).
hybrid spin register. Nature 506, 204–207 (2014). 29. S. Olmschenk, K. C. Younge, D. L. Moehring, D. N. Matsukevich, P. Maunz, C. Monroe,
10. D. Nigg, M. Müller, E. A. Martinez, P. Schindler, M. Hennrich, T. Monz, Manipulation and detection of a trapped Yb+ hyperfine qubit. Phys. Rev. A 76, 052314
M. A. Martin-Delgado, R. Blatt, Quantum computations on a topologically encoded qubit. (2007).
Science 345, 302–305 (2014). 30. K. Mølmer, A. Sørensen, Multiparticle entanglement of hot trapped ions. Phys. Rev. Lett.
11. J. Kelly, R. Barends, A. G. Fowler, A. Megrant, E. Jeffrey, T. C. White, D. Sank, J. Y. Mutus, 82, 1835 (1999).
B. Campbell, Y. Chen, Z. Chen, B. Chiaro, A. Dunsworth, I.-C. Hoi, C. Neill, P. J. J. O’Malley, 31. E. Solano, R. L. de Matos Filho, N. Zagury, Deterministic Bell states and measurement of
w?>C. Quintana, P. Roushan, A. Vainsencher, J. Wenner, A. N. Cleland, J. M. Martinis, the motional state of two trapped ions. Phys. Rev. A 59, R2539 (1999).
State preservation by repetitive error detection in a superconducting quantum circuit. 32. G. J. Milburn, S. Schneider, D. F. V. James, Ion trap quantum computing with warm ions.
Nature 519, 66–69 (2015). Fortschritte der Physik 48, 801–810 (2000).
12. A. D. Córcoles, E. Magesan, S. J. Srinivasan, Andrew W. Cross, M. Steffen, Jay M. Gambetta, 33. S.-L. Zhu, C. Monroe, L.-M. Duan, Arbitrary-speed quantum gates within large ion crystals
Jerry M. Chow, Demonstration of a quantum error detection code using a square lattice through minimum control of laser beams. Europhys. Lett. 73, 485 (2006).
of four superconducting qubits. Nat. Commun. 6, 6979 (2015). 34. T. Choi, S. Debnath, T. A. Manning, C. Figgatt, Z.-X. Gong, L.-M. Duan, C. Monroe, Optimal

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

13. J. Cramer, N. Kalb, M. A. Rol, B. Hensen, M. S. Blok, M. Markham, D. J. Twitchen, R. Hanson, quantum control of multimode couplings between trapped ion qubits for scalable
T. H. Taminiau, Repeated quantum error correction on a continuously encoded qubit by entanglement. Phys. Rev. Lett. 112, 190502 (2014).
real-time feedback. Nat. Commun. 7, 11526 (2016).
14. N. Ofek, A. Petrenko, R. Heeres, P. Reinhold, Z. Leghtas, B. Vlastakis, Y. Liu, L. Frunzio, Acknowledgments: We thank J. Kim and D. Maslov for helpful discussions and D. Gottesman
S. M. Girvin, L. Jiang, M. Mirrahimi, M. H. Devoret, R. J. Schoelkopf, Extending the lifetime of a for useful comments. Funding: This work was supported by the Army Research Office with
quantum bit with error correction in superconducting circuits. Nature 536, 441–445 (2016). funds from the Intelligence Advanced Research Projects Activity LogiQ program (grant
15. M. Grassl, T. Beth, T. Pellizzari, Codes for the quantum erasure channel. Phys. Rev. 3130638), the Air Force Office of Scientific Research Multidisciplinary University Research
A 56, 33–38 (1997). Initiative program on Quantum Measurement and Verification (grant 5710003628), and the
16. D. Gottesman, Quantum fault tolerance in small experiments. arXiv:1610.03507 (2016). NSF Physics Frontier Center at the Joint Quantum Institute (grant PHY0822671). Author
17. C.-Y. Lu, W.-B. Gao, J. Zhang, X.-Q. Zhou, T. Yang, J.-W. Pan, Experimental quantum coding contributions: N.M.L., M.G., K.A.L., C.F., S.D., K.R.B., and C.M. all contributed to the experimental
against qubit loss error. Proc. Natl. Acad. Sci. U.S.A. 105, 11050–11054 (2008). design, construction, data collection, and analysis of this experiment. All authors contributed
18. B. A. Bell, D. A. Herrera-Martí, M. S. Tame, D. Markham, W. J. Wadsworth, J. G. Rarity, to this manuscript. Competing interests: C.M. is a founding scientist of IonQ Inc. All other
Experimental demonstration of a graph state quantum error-correction code. Nat. authors declare that they have no competing interests. Data and materials availability: All
Commun. 5, 3658 (2014). data needed to evaluate the conclusions in the paper are present in the paper. Additional data
19. A. W. Cross, D. P. DiVincenzo, B. M. Terhal, A comparative code study for quantum fault related to this paper may be requested from the authors.
tolerance. Quantum Inf. Comput. 9, 0541–0572 (2009).
20. D. Bacon, Operator quantum error-correcting subsystems for self-correcting quantum
memories. Phys. Rev. A 73, 012340 (2006). Submitted 5 April 2017
21. P. W. Shor, Scheme for reducing decoherence in quantum computer memory. Phys. Rev. Accepted 21 September 2017
A 52, R2493 (1995). Published 20 October 2017
22. J. Napp, J. Preskill, Optimal Bacon-Shor codes. Quantum Inf. Comput. 13, 490–510 (2013). 10.1126/sciadv.1701074
23. M. Takita, A. D. Córcoles, E. Magesan, B. Abdo, M. Brink, A. Cross, J. M. Chow,
J. M. Gambetta, Demonstration of weight-four parity measurements in the surface code Citation: N. M. Linke, M. Gutierrez, K. A. Landsman, C. Figgatt, S. Debnath, K. R. Brown,
architecture. Phys. Rev. Lett. 117, 210505 (2016). C. Monroe, Fault-tolerant quantum error detection. Sci. Adv. 3, e1701074 (2017).

Linke et al., Sci. Adv. 2017; 3 : e1701074 20 October 2017 6 of 6

Fault-tolerant quantum error detection
Norbert M. Linke, Mauricio Gutierrez, Kevin A. Landsman, Caroline Figgatt, Shantanu Debnath, Kenneth R. Brown and
Christopher Monroe

Sci Adv 3 (10), e1701074.

DOI: 10.1126/sciadv.1701074

Downloaded from http://advances.sciencemag.org/ on January 17, 2019

ARTICLE TOOLS http://advances.sciencemag.org/content/3/10/e1701074

REFERENCES This article cites 31 articles, 3 of which you can access for free
http://advances.sciencemag.org/content/3/10/e1701074#BIBL

PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions

Use of this article is subject to the Terms of Service

Science Advances (ISSN 2375-2548) is published by the American Association for the Advancement of Science, 1200 New
York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive licensee American
Association for the Advancement of Science. No claim to original U.S. Government Works. The title Science Advances is a
registered trademark of AAAS.