PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

Letter

Super-robust nonadiabatic geometric quantum control

Bao-Jie Liu ,1 Yuan-Sheng Wang ,1,2 and Man-Hong Yung1,3,4,5,*

1
Department of Physics, Southern University of Science and Technology, Shenzhen 518055, China
2
School of Physical Sciences, University of Science and Technology of China, Hefei 230026, China
3
Shenzhen Institute for Quantum Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, China
4
Guangdong Provincial Key Laboratory of Quantum Science and Engineering, Southern University of Science
and Technology, Shenzhen 518055, China
5
Shenzhen Key Laboratory of Quantum Science and Engineering, Southern University of Science and Technology, Shenzhen 518055, China

(Received 18 March 2021; accepted 7 September 2021; published 16 September 2021)

Nonadiabatic geometric quantum computation (NGQC) and nonadiabatic holonomic quantum computation
(NHQC) have been proposed to reduce the run time of geometric quantum gates. However, in terms of robustness
against experimental control errors, the existing NGQC and NHQC scenarios have no advantage over standard
dynamical gates in most cases. Here, we give the reasons why nonadiabatic geometric gates are sensitive to
the control errors and, further, we propose a scheme of super-robust nonadiabatic geometric quantum control,
in which the super-robust condition can guarantee both high speed and robustness of the geometric gate. To
illustrate the working mechanism of super-robust geometric quantum gates, we give two simple examples of
SR-NGQC and SR-NHQC for two- and three-level quantum systems, respectively. Theoretical and numerical
results with the experimental parameters indicate that our scheme can significantly improve the gate performance
compared to the previous NGQC, NHQC, and standard dynamical schemes. Super-robust geometric quantum
computation can be applied to various physical platforms such as superconducting qubits, quantum dots, and
trapped ions. All of these sufficiently show that our scheme provides a promising way towards robust geometric
quantum computation.

DOI: 10.1103/PhysRevResearch.3.L032066

Introduction. Realizing high-fidelity and fault-tolerant noise-resilient scenario against fluctuations of control pa-
quantum gates is very essential for quantum information pro- rameters [13,14]. However, adiabatic quantum dynamics
cessing since control errors and environment-induced noises implies lengthy gate time and thus long exposure time to
are ubiquitous in operating real quantum devices. Geomet- the environment-induced decoherence. To overcome such
ric quantum computation (GQC) utilizes a unique property a problem, nonadiabatic geometric quantum computation
that the time-dependent quantum state would accumulate (NGQC) [17–23] and nonadiabatic holonomic quantum com-
an Abelian geometric phase [1,2] or non-Abelian holonomy putation (NHQC) [24–36] based on a nonadiabatic Abelian
[3–5] under a cyclic quantum evolution. The geometric phase and non-Abelian geometric phase [2,4], respectively, have
and holonomy depend only on the global properties of the evo- been proposed to reduce the run times of geometric quan-
lution trajectories. Consequently, geometric quantum gates tum gates. Recently, nonadiabatic geometric gates have been
are robust against local disturbances during the evolution experimentally demonstrated in different physical platforms,
[6–11]. More specifically, geometric quantum computation including superconducting qubits [37–45], nuclear magnetic
can be divided into Abelian GQC and holonomic quantum resonance (NMR) [46–48], and nitrogen-vacancy centers in
computation (HQC) depending on whether the geometric diamond [49–54]. However, in terms of robustness against ex-
phase is a real number [1] or a matrix [4] (non-Abelian holon- perimental errors, the existing NGQC and NHQC gates have
omy). no sufficient preponderance over standard dynamical gates
Early applications of GQC are dependent on adiabatic in most cases [19,34–36]. Therefore, it is natural to ask (i)
quantum evolutions to suppress transitions between differ- why the existing nonadiabatic geometric gates lack robustness
ent instantaneous eigenstates of the Hamiltonian [12–16]. against the control errors and (ii) how to maintain both the
Adiabatic GQC has been experimentally verified as a speed and the robustness of geometric gates.
In this Letter, we give clear answers for the above two
important issues using the super-robust condition proposed
*
yung@sustech.edu.cn here. And on that basis, we demonstrate a class of super-robust
nonadiabatic geometric gates in which robustness against the
Published by the American Physical Society under the terms of the control errors is ensured by a super-robust control condition.
Creative Commons Attribution 4.0 International license. Further We implement our schemes in two- and three-level systems,
distribution of this work must maintain attribution to the author(s) respectively, to realize super-robust Abelian (non-Abelian)
and the published article’s title, journal citation, and DOI. nonadiabatic geometric (holonomic) quantum gates, called

2643-1564/2021/3(3)/L032066(7) L032066-1 Published by the American Physical Society

LIU, WANG, AND YUNG PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

SR-NGQC (SR-NHQC). Using the experimental parameters, (a)

the numerical results indicate that our scheme can signifi-
cantly improve the gate performance compared to the existing
NGQC, NHQC, and standard dynamical schemes, which are
in good agreement with the theoretical results. In addition,
these super-robust geometric quantum computations can be
easily applied to various physical platforms. (b)
General framework of nonadiabatic geometric quan-
tum control. Let us start with a nondegenerate quantum
system described by (M+N)-dimensional Hilbert space,
with its evolution governed by the Hamiltonian H (t ).
For any complete set of basis vectors {|ψk (0)}M+N k=1 at
t = 0, the time-evolution operator can be written as
t    (c)
U (t ) = T e−i 0 H (t )dt = m |ψm (t )ψm (0)|, where the time-
t  
dependent state, |ψm (t ) = T e−i 0 H (t )dt |ψm (0), follows the
Schrödinger equation. Here, we choose a different set of time-
dependent auxiliary basis states {|μk (t )}M+N k=1 , which makes
the Hamiltonian H (t ) satisfy the decomposition condition.
The decomposition condition ensures that the operations in FIG. 1. The illustration of our proposed implementation.
the noncomputational basis {|νk (t )}M+N k=M+1 are completely ir- (a) Conceptual explanation for the ideal nonadiabatic geometric
relevant for the geometric operations in the computational quantum gates in a (M + N )-dimensional Hilbert space. The evo-
subspace {|μk (t )}M k=1 , which is given by lution state |μk (t ) acquired a pure nonadiabatic geometric phase
γk under a cyclic evolution in computational subspace to construct
HR (t ) = V + (t )[H (t ) − i∂t ]V (t ) geometric gates. Meanwhile, the state evolution in a noncomputa-
(1) tional basis is completely irrelevant for geometric operations in the
= HC (t ) ⊕ HN (t ), Hilbert space. (b) Without super-robust condition (SRC) protection,
  the fidelity of the nonadiabatic geometric quantum gate with the
where V (t ) ≡ k |μk (t )μk (0)|, HC (t ) ≡ M m,k=1 [μm |H (t )− presence of global control errors is limited because of the unwanted
i∂t |μk ]mk (0) is a Hamiltonian acting on the M-dimensional couplings of the time-dependent auxiliary states in the computational
computational  subspace with mk (0) ≡ |μm (0)μk (0)|, and and noncomputational subspace. (c) With SRC protection, the effects
HN (t ) = M+N m,k=M+1 [μm |H (t ) − i∂t |μk ]mk (0) is Hamilto-
of the above couplings can be greatly suppressed.
nian acting on the noncomputational subspace, as shown in
Fig. 1(a).
which is a nonadiabatic geometric quantum gate in the com-
Now, we explain how to construct a nonadiabatic
putational subspace {|μm (t )}M
m=1 .
geometric gate in the computational subspace. With
Super-robust condition. Now, we consider the effect of
the help of {|μk (t )}M k=1 , |ψm (t ) can be expressed as
M control error on the quantum evolution. The ideal Hamiltonian
|ψm (t ) = l=1 Clm (t )|μl (t ), and the time-evolution H (t ) then becomes
operator in the computational subspace becomes

UC (t ) = M l,m=1 Clm (t )|μl (t )μm (0)|. After a cyclic H  (t ) = H (t ) + βV (t ), (4)
evolution, we obtain the final time-evolution operator
 τ
[24], UC (τ ) = M m,k=1 [T e
i 0 A(t )+K(t )dt
]mk mk (0), where where we assume that β is a small constant, i.e., |β|  1,
T is the time-ordering operator, Alm ≡ iμl (t )|∂t |μm (t ) which corresponds to a quasistatic noise [55,56]. Here, V (t ) is
is the matrix-valued connection one-form, and Klm (t ) ≡ the noise Hamiltonian that destroys the ideal dynamics, which
−μl (t )|H (t )|μm (t ) is the dynamical part. can be regarded as a perturbation [57]. Under this assump-
To meet the decomposition condition given by Eq. (1), one tion, the decomposition condition given by Eq. (1) is broken.
possible set of the auxiliary state |μk (t ) is found to be pro- In other words, the rotating Hamiltonian becomes HR (t ) =
M+N M+N
portional to the time-evolution states |ψk (t ), i.e., |μk (t ) = m=1 Amm (t )mm (0) + β m=k Kmk (t )mk (0). Then, we
ei fk (t ) |ψk (t ). Then, we obtain the nondiagonal parts of A and τ 
obtain the evolution operator as UE (τ ) = T e−i 0 HR (t )dt . In
K, which satisfy the relation general, to analytically solve this equation is difficult due to
the time-ordering operator.
Alm (t ) = −Klm (t ) = e−iδkm (t ) ψl (t )|H (t )|ψm (t ), (2) Here, we use the Magnus expansion [57–59] to pertur-
batively process the evolution operator UE (τ ). Before that,
where δkm (t ) = fk (t ) − fm (t ). Using the spin-echo technique we transform to the interaction picture by defining UI (t ) =
[12] or the pulse-shaping method [24], the accumulated dy- M+N −i  t Amm (t  )dt 
τ e 0 mm (0), and the transformed Hamil-
namical phases can be erased, i.e., 0 Kmm (t )dt = 0, m = m=1 
1, . . . , M. In this way, we will obtain tonian is HIR 
(t ) = βUI+ (t ) M+N m=k Kmk (t )mk (0)UI (t ). The
corresponding evolution operator in the interaction frame is

M τ
given by
UC (τ ) = ei 0 Amm (t )dt
mm (0), (3) τ 
m=1 UIE (τ ) = T e−i 0 HIR (t )dt
. (5)

L032066-2
SUPER-ROBUST NONADIABATIC GEOMETRIC QUANTUM … PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

Applying the Magnus expansion to Eq. (5), we have (a) (b)

∞
UIE (τ ) = e k=1 k (τ )
, (6)
where k denotes the terms of the Magnus expansion, and the
first two terms of the series are given by (see, e.g., [57–59])
 t

1 (t ) = −i HIR (t1 )dt1 ,
0
 t  t1
1  
2 (t ) = dt1 dt2 [HIR (t1 ), HIR (t2 )]. (7)
2 0 0
FIG. 2. The level structure and coupling configuration for the
Expanding Eq. (6) in powers of β and using Eq. (7), we obtain construction of SR-NHQC and SR-NGQC gates. (a) The driven
  pulses with amplitudes 0 and 1 resonantly couple |0 and |1
 β2
U (τ, 0) = U (τ, 0) I − iβD(τ ) − G(τ ) + O(β ) , (8)
3 to |e with the phases φ0 and φ1 , respectively. (b) The driven pulse
2 with amplitudes R resonantly couples |0 and to |1 with the time-
where the element of matrix Dkm (t ) ≡ independent phase φR .
t    
τ
0 ψk (t )|V (t )|ψm (t )dt and G(τ ) ≡ 0 dt[Ḋ(t ), D(t )] +
D2 (τ ). Consequently, the geometric gate in Eq. (3)
under the control error becomes UC (τ, 0) =
M 
M 
m=1 |ψm (τ )ψm (0)| = m=1 NrmU (τ, 0)mm (0), where
 Therefore, we can explain why some previous NGQC and
Nrm = 1/ 1 + β 2 M+N k=1 |Dkm |
2 is the state normalized
NHQC lack robustness to the global control errors.
coefficient. To further evaluate the performance of the In addition, we can further explain the origin of the ro-
quantum operation caused by the control error, the gate bustness of adiabatic GQC [12–14] against the control errors.
fidelity [60,61] is taken by The adiabatic condition [2,62] for a nondegenerate system is
1 |Akm | = | ekE|∂k −E
t H |em 
|  1 for k = m, where {Ek (t )} and |ek (t )
F= | Tr(UC UC† )|
m
denote the instantaneous eigenvalues and eigenvectors of
M
H (t ). In this case, we consider the auxiliary basis states
(9)
β2  
M M+N
|μk (t ) to be identical to the eigenvectors |ek (t ). The adia-
≈1− |Dkm |2 − O(β 4 ). batic condition ensures that the conditions of Eqs. (1) and (11)
2M m=1 k=1
are met. Consequently, we can verify that the fidelity of the
Consequently, we can achieve super-robust gates with the adiabatic GQC has, at least, a fourth-order error dependence
fourth-order error dependence against the control error, as against global control errors, using Eq. (9). In the following,
long as the following condition is satisfied: to illustrate the working mechanism of super-robust geometric
 τ gates, we give two simple examples of SR-NGQC and SR-
Dkm ≡ ψk (t )|V (t )|ψm (t )dt = 0, (10) NHQC for two- and three-level quantum systems, respectively
0 Example 1: SR-NHQC. The main idea of the previous
where k = 1, . . . , M + N and m = 1, 2, . . . , M. NHQC is to generate a nonadiabatic non-Abelian geometric
As a demonstration of Eq. (10), we consider that the control gate in a three-level system, as shown in Fig. 2(a). Un-
error is global with V (t ) = H (t ). In this case, the super-robust der the rotating-wave approximation, the system Hamiltonian

condition can be expressed by is given by H (t ) = 1i=0 21 [ i (t )eiφi |ie| + H.c.]. We de-
 τ fine a bright state, |b ≡ sin( θ2 )eiφ |0 + cos( θ2 )|1, where φ ≡

Dkm = ψk (t )|H (t )|ψm (t )dt = 0. (11) φ0 (t ) − φ1 (t ) and tan(θ /2) ≡ 0 (t )/ 1 (t ). We shall keep θ
0
and φ, and hence |b, to be time independent. The Hamiltonian
In this way, the effects of global control errors can be greatly H (t ) can then be rewritten as H (t ) = 21 ( (t )eiφ1 (t ) |be| +
suppressed. Here, we further illustrate the geometric meaning
 H.c.), where (t ) ≡ 0 (t ) + 1 (t ) is the Rabi frequency
2 2
of Eq. (11): Dmm = 0 with k= m erases the accumulated
τ of H (t ).
dynamical phases, and Dkm = 0 eiδkm Akm dt = 0 with k = m
Here, we choose the time-dependent auxiliary basis states
suppresses both couplings between the time-dependent auxil-
as |μ1 (t ) = |d, |μ2 (t ) = cos α(t2 ) |b − i sin α(t2 ) e−iφ1 (t ) |e,
iary states |μm (t ) and |μn (t ) in the computational subspace
and |μ3 (t ) = −i sin α(t2 ) eiφ1 (t ) |b + cos α(t2 ) |e with α(t ) =
and noncomputational subspace under the control errors, as t
shown in Figs. 1(b) and 1(c). 0 (t  )dt  (see the Supplemental Material [63] for details).
Note that the key difference between the previous NGQC Therefore, the rotating Hamiltonian given by Eq. (1) in the
α
and NHQC schemes and the super-robust condition-based auxiliary basis states is given by HR (t ) = 1−cos 2
φ̇1 |bb| ⊕
cos α−1
scheme proposed here is that the Hamiltonians have differ- 2
φ̇ 1 |ee|. The corresponding computational Hamilto-
ent constraints. In the NGQC case, the Hamiltonian is only nian and noncomputational Hamiltonian are given by HC (t ) =
1−cos α
required to satisfy the constraint Dmn 
= 0. The constraint 2
φ̇1 |bb| + 0|dd| and HN (t ) = cos 2α−1 φ̇1 |ee|, re-

of NHQC is set as Dmn = 0 with m, n = 1, . . . , M. We can spectively. Since the parallel transport conditions of the
clearly see that the constraints of NGQC and NHQC are a previous NHQC [17–23,37–39], i.e., ψm (t )|H (t )|ψn (t ) =
necessary but not sufficient condition of super-robust gates. 0 (n, m = 1, 2), are satisfied, the super-robust condition given

L032066-3
LIU, WANG, AND YUNG PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

TABLE I. The robustness comparison with the NOT gate of the

standard dynamical gate (DG), nonadiabatic geometric gate, holo-
nomic gate, and super-robust geometric quantum gates.

Types System level Gate time Fidelity References

DG 2 π/ 0 1 − β 2 π 2 /8
NGQC 2 2π / 0 1 − β 2 π 2 /8 [17–23,37–39]
SR-NGQC 2 3π / 0 1 − O(β 4 ) This work
NHQC 3 2π / 0 1 − β 2 π 2 /3 [25–36,42–54]
SR-NHQC 3 4π / 0 1 − O(β 4 ) This work

by Eq. (11) can be reduced to

 τ  t
 α̇
D23 = exp −i φ̇1 / cos αdt  dt = 0. (12)
0 2 0

Consequently, we obtain the following unitary transfor-

mation matrix in the basis states: {|μ1 (0),
 τ |μ2 (0)}, i.e.,
UC (τ ) = eiγg |bb| + |dd|, where γg = 0 cos 2α−1 φ̇1 dt =
 
1 φ1 (τ ) α(τ )
2 φ1 (0) α(0)
sin αdαdφ1 , which shows that the geometric
phase γg exactly equals half of the solid angle. Note that the
FIG. 3. Numerical robustness comparison of the NOT gate with
nonadiabatic holonomic gate can be spanned by the logical
various approaches. The NOT gate infidelities 1 − F of DG, NGQC,
basis {|0, |1}, i.e.,
SR-NGQC, NHQC, and SR-NHQC are set as a function of the
γg
U (γg, θ , φ) = ei 2 n·σ , (13) control error of the Rabi frequency, i.e., the relative pulse deviation
β (a) without and (b) with the decoherence. The NOT gate fidelities’
where n = (sin θ cos φ, sin θ sin φ, cos θ ), and σ are the Pauli difference, (d) F = FG − FD and (c) (F = FN − FD ), between the
matrices. Equation (13) describes a rotational operation SR-NGQC (NGQC) and DG are set as the decoherence parameter 
around the n axis by a γg angle, ignoring a global phase factor and the relative pulse deviation β.
γg
e−i 2 . Since both n and γg can take any value, Eq. (13) denotes
a set of universal single-qubit gates in the qubit subspace.
Example 2: SR-NGQC. As another example, one can shown in Fig. 3(a), the SR-NHQC and SR-NGQC are always
also apply our scheme to a two-level system, as shown in more robust than the DG, NHQC, and NGQC gates without
Fig. 2(b). The system Hamiltonian can be written as H1 (t ) = the consideration of decoherence. The numerical results are
R (t ) iφR (t )
e |01| + H.c. The time-dependent auxiliary basis in good agreement with the theoretical results, as shown in
2
states are taken by |ζ1 (t ) = cos αR2(t ) |0 − i sin αR2(t ) e−iφR (t ) |1 Table I.
In fact, the decoherence process is unavoidable. To evaluate
and |ζ2 (t ) = −i sin αR2(t ) eiφR (t ) |0 + cos αR2(t ) |1 with αR (t ) =
t  
the performance of the gates with the consideration of the
0 R (t )dt + αR (0). Under these settings, the condi- influence of decoherence, the Lindblad master equation [64]

 τ of the previous NGQC [17–23,37–39], i.e., Dmm =
tion is used here. To be more eloquent, we set the parameters from
0 Kmm (t )dt = 0, m = 1, 2, is satisfied. Thus, the super- the current experiments [65–67]: the decay and dephasing
robust condition given by Eq. (11) for NGQC becomes rates of the qubit are taken as 1 = 2 =  = 10−4 0 in our
 τ  t numerical simulation. From Fig. 3(b), with the consideration
 α̇R
D12 = exp −i φ̇R / cos αR dt  dt = 0. (14) of both the Rabi control error and the decoherence effect,
0 2 0 we find that the SR-NGQC can significantly improve the
Similar to the SR-based NHQC case, we can also robustness of quantum gates with the relative pulse deviation
obtain the universal single-qubit gates in Eq. (13), |β| > 0.01.
i.e., U [γR , αR (0), φR (0)], with the geometric phase On the other hand, one can find from Fig. 3(b) that the
τ
γR = 0 φ̇R (cos αR − 1)/2dt. DG is better than NGQC and SR-NGQC for the relative
Numerical simulations. To investigate the noise-resilient pulse deviation β ∈ [−0.01, 0.01] since decoherence is the
feature of super-robust geometric gates against the global main factor in this case. In order to balance between the
control error, we take the NOT gate as a typical example to decoherence and the Rabi control errors, we plot the NOT
compare the performances of the SR-NHQC and SR-NGQC gate fidelities’ difference, F = FG − FD (F = FN − FD )
approaches with that of the related standard dynamical gate between the SR-NGQC (NGQC) and DG, with the decoher-
(DG), NGQC, and NHQC approaches. Before that, we choose ence parameter  and the relative pulse deviation β, as shown
time-independent Rabi frequencies of all the gates and keep in Fig. 3(d) [Fig. 3(c)]. We find that the best scheme in this
the same as (t ) = R (t ) = 0 ; the corresponding gate times case depends on the relative importance between the decoher-
are sketched in Table I (see the Supplemental Material [63] ence and the Rabi control error. An open question for further
for details). Here, we assume the relative pulse deviation β of development is whether designing a hybrid SR-NGQC and
the driven pulse to vary in the range of β ∈ [−0.1, 0.1]. As DG scheme is optimal for quantum control. In addition, we

L032066-4
SUPER-ROBUST NONADIABATIC GEOMETRIC QUANTUM … PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

clearly know that the NOT gate of the NGQC has no particular a SR-based geometric scheme with the decoherence-free
advantage compared to DG with both above errors, as shown subspace (DFS) [26–28,68–70] encoding model (surface
in Fig. 3(c). codes [71–75]) to further suppress the dephasing noises
Conclusion and outlook. In conclusion, we have explained (local errors). In addition, it would be interesting to fur-
why the existing nonadiabatic geometric gates are so sensi- ther optimize the magnetic field sensitivity and maximum
tive to the control errors, but also proposed the scheme of field range of geometric-phase magnetometry [76] via our
super-robust nonadiabatic geometric gates. Specifically, we SR-NGQC scheme.
have taken two examples in two- and three-level systems for Acknowledgments. The authors thank Professor S.-L. Su
realizing super-robust condition-based Abelian (non-Abelian) for helpful discussions and improving this Letter. This
nonadiabatic geometric (holonomic) quantum gates, respec- work is supported by the Natural Science Foundation of
tively. The theoretical and numerical results indicate that Guangdong Province (Grant No. 2017B030308003), the
our scheme can significantly improve the gate performance Key R & D Program of Guangdong Province (Grant No.
compared to the existing geometric and standard dynamical 2018B030326001), the Science, Technology and Innova-
schemes with the experimental parameters. Moreover, our tion Commission of Shenzhen Municipality (Grants No.
scheme can also be extended to construct two-qubit geo- JCYJ20170412152620376, No. JCYJ20170817105046702,
metric gates [37,45,49,50] to realize a universal SR-based and No. KYTDPT20181011104202253), the National Nat-
geometric gate set. In addition, this extensible approach of ural Science Foundation of China (Grants No. 11875160
SR-based geometric gates can be applied to various phys- and No. U1801661), the Economy, Trade and Informa-
ical platforms such as superconducting circuits, quantum tion Commission of Shenzhen Municipality (Grant No.
dots, and trapped ions. For future work, it would be at- 201901161512), and Guangdong Provincial Key Laboratory
tractive for fault-tolerant quantum computation to combine (Grant No. 2019B121203002).

[1] M. V. Berry, Quantal phase factors accompanying adiabatic Thewalt, and J. J. L. Morton, Geometric phase gates with
changes, Proc. R. Soc. London A 392, 45 (1984). adiabatic control in electron spin resonance, Phys. Rev. A 87,
[2] Y. Aharonov and J. Anandan, Phase Change During a Cyclic 032326 (2013).
Quantum Evolution, Phys. Rev. Lett. 58, 1593 (1987). [14] Y.-Y. Huang, Y.-K. Wu, F. Wang, P.-Y. Hou, W.-B. Wang, W.-
[3] P. Zanardi and M. Rasetti, Holonomic quantum computation, G. Zhang, W.-Q. Lian, Y.-Q. Liu, H.-Y. Wang, H.-Y. Zhang
Phys. Lett. A 264, 94 (1999). et al., Experimental Realization of Robust Geometric Quantum
[4] J. Anandan, Non-adiabatic non-Abelian geometric phase, Phys. Gates with Solid-State Spins, Phys. Rev. Lett. 122, 010503
Lett. A 133, 171 (1988). (2019).
[5] E. Sjöqvist, Trend: A new phase in quantum computation, [15] L. M. Duan, J. I. Cirac, and P. Zoller, Geometric manipulation
Physics 1, 35 (2008). of trapped ions for quantum computation, Science 292, 1695
[6] S.-L. Zhu and P. Zanardi, Geometric quantum gates that are (2001).
robust against stochastic control errors, Phys. Rev. A 72, [16] L.-A. Wu, P. Zanardi, and D. A. Lidar, Holonomic Quantum
020301(R) (2005). Computation in Decoherence-Free Subspaces, Phys. Rev. Lett.
[7] S. Berger, M. Pechal, A. A. Abdumalikov, Jr., C. Eichler, L. 95, 130501 (2005).
Steffen, A. Fedorov, A. Wallraff, and S. Filipp, Exploring the [17] X. B. Wang and M. Keiji, Nonadiabatic Conditional Geo-
effect of noise on the Berry phase, Phys. Rev. A 87, 060303(R) metric Phase Shift with NMR, Phys. Rev. Lett. 87, 097901
(2013). (2001).
[8] G. D. Chiara and G. M. Palma, Berry Phase for a Spin-1/2 [18] S.-L. Zhu and Z. D. Wang, Implementation of Universal Quan-
Particle in a Classical Fluctuating Field, Phys. Rev. Lett. 91, tum Gates Based on Nonadiabatic Geometric Phases, Phys.
090404 (2003). Rev. Lett. 89, 097902 (2002).
[9] P. J. Leek, J. M. Fink, A. Blais, R. Bianchetti, M. Göppl, J. M. [19] J. T. Thomas, M. Lababidi, and M. Tian, Robustness of single-
Gambetta, D. I. Schuster, L. Frunzio, R. J. Schoelkopf, and A. qubit geometric gate against systematic error, Phys. Rev. A 84,
Wallraff, Observation of Berry’s phase in a solid state qubit, 042335 (2011).
Science 318, 1889 (2007). [20] P. Z. Zhao, X.-D. Cui, G. F. Xu, E. Sjöqvist, and D. M. Tong,
[10] S. Filipp, J. Klepp, Y. Hasegawa, C. Plonka-Spehr, U. Schmidt, Rydberg-atom-based scheme of non-adiabatic geometric quan-
P. Geltenbort, and H. Rauch, Experimental Demonstration of tum computation, Phys. Rev. A 96, 052316 (2017).
the Stability of Berry’s Phase for a Spin-1/2 Particle, Phys. Rev. [21] K. Z. Li, P. Z. Zhao, and D. M. Tong, Approach to realizing
Lett. 102, 030404 (2009). non-adiabatic geometric gates with prescribed evolution paths,
[11] F. Wilczek and A. Zee, Appearance of Gauge Structure in Phys. Rev. Research 2, 023295 (2020).
Simple Dynamical Systems, Phys. Rev. Lett. 52, 2111 (1984). [22] T. Chen and Z.-Y. Xue, Non-Adiabatic Geometric Quantum
[12] J. A. Jones, V. Vedral, A. Ekert, and G. Castagnoli, Geomet- Computation with Parametrically Tunable Coupling, Phys. Rev.
ric quantum computation using nuclear magnetic resonance, Appl. 10, 054051 (2018).
Nature (London) 403, 869 (2000). [23] C. Zhang, T. Chen, S. Li and Z.-Y. Xue, High-fidelity geometric
[13] H. Wu, E. M. Gauger, R. E. George, M. Möttönen, H. Riemann, gate for silicon-based spin qubits, Phys. Rev. A 101, 052302
N. V. Abrosimov, P. Becker, H.-J. Pohl, K. M. Itoh, M. L. W. (2020).

L032066-5
LIU, WANG, AND YUNG PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

[24] B.-J. Liu, X.-K. Song, Z.-Y. Xue, X. Wang, and M.-H. Yung, X. Tan, D. Lan, M.-H. Yung, and Y. Yu, Experimental
Plug-and-Play Approach to Nonadiabatic Geometric Quantum realization of universal time-optimal non-abelian geometric
Gates, Phys. Rev. Lett. 123, 100501 (2019). gates, arXiv:2004.10364.
[25] E. Sjöqvist, D. M. Tong, L. M. Andersson, B. Hessmo, M. [42] A. A. Abdumalikov, J. M. Fink, K. Juliusson, M. Pechal, S.
Johansson, and K. Singh, Non-adiabatic holonomic quantum Berger, A. Wallraff, and S. Filipp, Experimental realization of
computation, New J. Phys. 14, 103035 (2012). non-Abelian non-adiabatic geometric gates, Nature (London)
[26] G. F. Xu, J. Zhang, D. M. Tong, E. Sjöqvist, and L. C. 496, 482 (2013).
Kwek, Nonadiabatic Holonomic Quantum Computation in [43] Y. Xu, W. Cai, Y. Ma, X. Mu, L. Hu, T. Chen, H. Wang,
Decoherence-free Subspaces, Phys. Rev. Lett. 109, 170501 Y. P. Song, Z.-Y. Xue, Z.-Q. Yin, and L. Sun, Single-Loop
(2012). Realization of Arbitrary Nonadiabatic Holonomic Single Qubit
[27] Z.-Y. Xue, J. Zhou, and Z. D. Wang, Universal holonomic quan- Quantum Gates in a Superconducting Circuit, Phys. Rev. Lett.
tum gates in decoherence-free subspace on superconducting 121, 110501 (2018).
circuits, Phys. Rev. A 92, 022320 (2015). [44] S. Danilin, A. Vepsäläinen, and G. S. Paraoanu, Experimental
[28] Z.-Y. Xue, F.-L. Gu, Z.-P. Hong, Z.-H. Yang, D.-W. Zhang, Y. state control by fast non-Abelian holonomic gates with a super-
Hu, and J. Q. You, Non-Adiabatic Holonomic Quantum Com- conducting qutrit, Phys. Scr. 93, 055101 (2018).
putation with Dressed-State Qubits, Phys. Rev. Appl. 7, 054022 [45] D. J. Egger, M. Ganzhorn, G. Salis, A. Fuhrer, P. Muller, P. K.
(2017). Barkoutsos, N. Moll, I. Tavernelli, and S. Filipp, Entanglement
[29] J. Zhou, B. J. Liu, Z. P. Hong, and Z. Y. Xue, Fast holonomic Generation in Superconducting Qubits Using Holonomic Oper-
quantum computation based on solid-state spins with all-optical ations, Phys. Rev. Appl. 11, 014017 (2019).
control, Sci. China-Phys. Mech. Astron. 61, 010312 (2018). [46] G. Feng, G. Xu, and G. Long, Experimental Realization of Non-
[30] Z.-P. Hong, B.-J. Liu, J.-Q. Cai, X.-D. Zhang, Y. Hu, Z. D. Adiabatic Holonomic Quantum Computation, Phys. Rev. Lett.
Wang, and Z.-Y. Xue, Implementing universal non-adiabatic 110, 190501 (2013).
holonomic quantum gates with transmons, Phys. Rev. A 97, [47] H. Li, L. Yang, and G. Long, Experimental realization of single-
022332 (2018). shot non-adiabatic holonomic gates in nuclear spins, Sci. China:
[31] V. Azimi Mousolou, Electric non-adiabatic geometric entan- Phys., Mech. Astron. 60, 080311 (2017).
gling gates on spin qubits, Phys. Rev. A 96, 012307 (2017). [48] Z. Zhu, T. Chen, X. Yang, J. Bian, Z.-Y. Xue, and X. Peng,
[32] P. Z. Zhao, K. Z. Li, G. F. Xu, and D. M. Tong, General Single-Loop and Composite-Loop Realization of non-adiabatic
approach for constructing Hamiltonians for non-adiabatic holo- Holonomic Quantum Gates in a Decoherence-Free Subspace,
nomic quantum computation, Phys. Rev. A 101, 062306 (2020). Phys. Rev. Appl. 12, 024024 (2019).
[33] M. Johansson, E. Sjöqvist, L. M. Andersson, M. Ericsson, [49] C. Zu, W.-B. Wang, L. He, W.-G. Zhang, C.-Y. Dai, F. Wang,
B. Hessmo, K. Singh, and D. M. Tong, Robustness of non- and L.-M. Duan, Experimental realization of universal geomet-
adiabatic holonomic gates, Phys. Rev. A 86, 062322 (2012). ric quantum gates with solid-state spins, Nature (London) 514,
[34] S. B. Zheng, C. P. Yang, and F. Nori, Comparison of the sen- 72 (2014).
sitivity to systematic errors between nonadiabatic non-Abelian [50] K. Nagata, K. Kuramitani, Y. Sekiguchi, and H. Kosaka, Uni-
geometric gates and their dynamical counterparts, Phys. Rev. A versal holonomic quantum gates over geometric spin qubits
93, 032313 (2016). with polarized microwaves, Nat. Commun. 9, 3227 (2018).
[35] N. Ramberg and E. Sjöqvist, Environment-Assisted Holonomic [51] S. Arroyo-Camejo, A. Lazariev, S. W. Hell, and G.
Quantum Maps, Phys. Rev. Lett. 122, 140501 (2019). Balasubramanian, Room temperature high-fidelity holonomic
[36] J. Jing, C.-H. Lam, and L.-A. Wu, Non-Abelian holonomic single-qubit gate on a solid-state spin, Nat. Commun. 5, 4870
transformation in the presence of classical noise, Phys. Rev. A (2014).
95, 012334 (2017). [52] Y. Sekiguchi, N. Niikura, R. Kuroiwa, H. Kano, and H. Kosaka,
[37] Y. Xu, Z. Hua, Tao Chen, X. Pan, X. Li, J. Han, W. Cai, Y. Optical holonomic single quantum gates with a geometric spin
Ma, H. Wang, Y. P. Song, Z.-Y. Xue, and L. Sun, Experimental under a zero field, Nat. Photon. 11, 309 (2017).
Implementation of Universal Non-Adiabatic Geometric Quan- [53] B. B. Zhou, P. C. Jerger, V. O. Shkolnikov, F. Joseph Heremans,
tum Gates in a Superconducting Circuit, Phys. Rev. Lett. 124, G. Burkard, and D. D. Awschalom, Holonomic Quantum Con-
230503 (2020). trol by Coherent Optical Excitation in Diamond, Phys. Rev.
[38] P. Z. Zhao, Z. Dong, Z. Zhang, G. Guo, D. M. Tong, and Y. Yin, Lett. 119, 140503 (2017).
Experimental realization of non-adiabatic geometric gates with [54] N. Ishida, T. Nakamura, T. Tanaka, S. Mishima, H. Kano, R.
a superconducting xmon qubit, Sci. China-Phys. Mech. Astron. Kuroiwa, Y. Sekiguchi, and H. Kosaka, Universal holonomic
64, 250362 (2021). single quantum gates over a geometric spin with phase-
[39] C. Song, S.-B. Zheng, P. Zhang, K. Xu, L. Zhang, Q. Guo, modulated polarized light, Opt. Lett. 43, 2380 (2018).
W. Liu, D. Xu, H. Deng, K. Huang et al., Continuous-variable [55] X. Rong, J. Geng, F. Shi, Y. Liu, K. Xu, W. Ma, F. Kong, Z.
geometric phase and its manipulation for quantum computation Jiang, Y. Wu, and J. Du, Experimental fault-tolerant universal
in a superconducting circuit, Nat. Commun. 8, 1061 (2017). quantum gates with solid-state spins under ambient conditions,
[40] T. Yan, B.-J. Liu, K. Xu, C. Song, S. Liu, Z. Zhang, H. Deng, Nat. Commun. 6, 8748 (2015).
Z. Yan, H. Rong, M.-H. Yung, Y. Chen, and D. Yu, Experi- [56] X. Wang, L. S. Bishop, J. P. Kestner, E. Barnes, K. Sun, and
mental Realization of Non-Adiabatic Shortcut to Non-Abelian S. D. Sarma, Composite pulses for robust universal control of
Geometric Gates, Phys. Rev. Lett. 122, 080501 (2019). singlet-triplet qubits, Nat. Commun. 3, 997 (2012).
[41] Z. Han, Y. Dong, B. Liu, X. Yang, S. Song, L. Qiu, D. [57] H. Ribeiro, A. Baksic, and A. A. Clerk, Systematic Magnus-
Li, J. Chu, W. Zheng, J. Xu, T. Huang, Z. Wang, X. Yu,

L032066-6
SUPER-ROBUST NONADIABATIC GEOMETRIC QUANTUM … PHYSICAL REVIEW RESEARCH 3, L032066 (2021)

Based Approach for Suppressing Leakage and Non-Adiabatic Gustavsson, and W. D. Oliver, A quantum engineer’s guide
Errors in Quantum Dynamics, Phys. Rev. X 7, 011021 (2017). to superconducting qubits, Appl. Phys. Rev. 6, 021318
[58] W. Magnus, On the exponential solution of differential equa- (2019)
tions for a linear operator, Commun. Pure Appl. Math. 7, 649 [68] X. K. Song, H. Zhang, Q. Ai, J. Qiu, and F. G. Deng, Shortcuts
(1954). to adiabatic holonomic quantum computation in decoherence-
[59] S. Blanes, F. Casas, J. A. Oteo, and J. Ros, The magnus expan- free subspace with transitionless quantum driving algorithm,
sion and some of its applications, Phys. Rep. 470, 151 (2009). New J. Phys. 18, 023001 (2016).
[60] A. Souza, G. A. Alvarez, and D. Suter, Robust dynamical de- [69] Z.-T. Liang, Y.-X. Du, W. Huang, Z.-Y. Xue, and H. Yan, non-
coupling, Philos. Trans. R. Soc. A 370, 4748 (2012). adiabatic holonomic quantum computation in decoherence-free
[61] G. T. Genov, D. Schraft, N. V. Vitanov, and T. Halfmann, subspaces with trapped ions, Phys. Rev. A 89, 062312 (2014).
Arbitrarily Accurate Pulse Sequences for Robust Dynamical [70] X.-L. Feng, C. Wu, H. Sun, and C. H. Oh, Geometric Entangling
Decoupling, Phys. Rev. Lett. 118, 133202 (2017). Gates in Decoherence-Free Subspaces with Minimal Require-
[62] M. Kolodrubetz, D. Sels, P. Mehta, and A. Polkovnikov, Ge- ments, Phys. Rev. Lett. 103, 200501 (2009).
ometry and non-adiabatic response in quantum and classical [71] S. B. Bravyi and A. Y. Kitaev, Quantum codes on a lattice with
systems, Phys. Rep. 697, 1 (2017). boundary, arXiv:quant-ph/9811052.
[63] See Supplemental Material at http://link.aps.org/supplemental/ [72] E. Dennis, A. Y. Kitaev, A. Landahl, and J. Preskill, Topological
10.1103/PhysRevResearch.3.L032066 for a discussion of the quantum memory, J. Math. Phys. 43, 4452 (2002).
choices of time-dependent auxiliary states and the parameters [73] A. G. Fowler, M. Mariantoni, J. M. Martinis, and A. N. Cleland,
of a NOT gate for SR-NHQC and SR-NGQC, which includes Surface codes: Towards practical large-scale quantum computa-
Refs. [20,24–26,43]. tion, Phys. Rev. A 86, 032324 (2012).
[64] G. Lindblad, On the generators of quantum dynamical semi- [74] J. Zhang, S. J. Devitt, J. Q. You, and F. Nori, Holonomic surface
groups, Commun. Math. Phys. 48, 119 (1976). codes for fault-tolerant quantum computation, Phys. Rev. A 97,
[65] I. Buluta, S. Ashhab, and F. Nori, Natural and artificial atoms 022335 (2018).
for quantum computation, Rep. Prog. Phys. 74, 104401 (2011). [75] C. Wu, Y. Wang, X.-L. Feng, and J.-L. Chen, Holonomic
[66] Z. Chen, J. Kelly, C. Quintana, R. Barends, B. Campbell, Quantum Computation in Surface Codes, Phys. Rev. Appl. 13,
Y. Chen, B. Chiaro, A. Dunsworth, A. G. Fowler, E. Lucero 014055 (2020).
et al., Measuring and Suppressing Quantum State Leakage in a [76] K. Arai, J. Lee, C. Belthangady, D. R. Glenn, H. Zhang, and
Superconducting Qubit, Phys. Rev. Lett. 116, 020501 (2016). R. L. Walsworth, Geometric phase magnetometry using a solid-
[67] P. Krantz, M. Kjaergaard, F. Yan, T. P. Orlando, S. state spin, Nat. Commun. 9, 4996 (2018).

L032066-7