Professional Documents
Culture Documents
separated into two categories: schemes which extend the robustness to (i) spin or (ii) motional
decoherence. We choose to focus on extensions of the σx ⊗ σx Mølmer-Sørensen interaction using
microwaves and a static magnetic field gradient. Nevertheless, some of the techniques discussed
here can be relevant to other trapped ion architectures or physical qubit implementations. Finally,
we experimentally realize a proof-of-concept interaction with simultaneous robustness to spin and
motional decoherence by combining several quantum control methods presented in this manuscript.
to maximize the continuous runtime of the quan- operation, which affects the subsequent phase space tra-
tum processor [33]. Therefore, the duration of the jectory. This led to the development of DD protocols
calibration sequences have a direct impact on the in which refocussing pulses are only applied when the
processor’s duty cycle. This introduces a trade-off spin and motion are disentangled [44–47]. In practice, a
between a quantum control method’s complexity k-loop√MS gate is implemented by choosing a detuning
and the processor’s available runtime. We quan- δ = 2 kΩ0 , with the Lamb-Dicke parameter and Ω0
tify this trade-off by representing the calibrations the gate fields’ Rabi frequency, and π-pulses are added
of each scheme as a Directed Acyclic Graph (DAG) at the completion of each loop.
[34], from which the number of nodes and depen- The PDD protocols can be made compatible with clas-
dencies provide a proxy for the calibration complex- sical digital circuitry by building the pulse timings as
ity and duration. multiples of a clock cycle [48–50]. The timings are en-
coded by Walsh functions, which are a series of square
Experimental overhead: pulses with values {−1, +1}, where π-pulses are applied
The final figure of merit is the experimental over- during zero-crossings. These are complementary to a k-
head of a particular quantum control scheme. This loop MS gate since the duration of a single loop can be
includes the required physical resources, the gen- viewed as a single clock cycle. The maximum number
eral complexity of the scheme both conceptually of π pulses is then Nπ =√k − 1. The gate speed √ suffers
and physically, and the stringency of hardware re- by a reduction factor of k, such that τ = kτ0 . It is
quirements. Since it is difficult to find a common therefore difficult to implement a large number of pulses
metric for the experimental overhead, this figure of due to the poor gate time scaling. Note that the applica-
merit should serve as a summary of future chal- tion of a π-pulse when spin and motion are disentangled
lenges that may arise when experimentally imple- still alters the resulting phase space trajectory, usually
menting a particular scheme. causing a symmetry of the path about the origin. This
was found to also increase motional robustness [45], and
is further discussed in section III.
A. Pulsed Dynamical Decoupling There exist PDD sequences with more complex timings
that are difficult to encode with Walsh functions (e.g.
The first demonstrations of quantum control within Uhrig [39]). Nevertheless, combining these sequences
the nuclear magnetic resonance community extended the with the MS interaction is in principle possible. The du-
spin’s coherence time via the application of π pulses ration of a single loop can be altered by varying the MS
[35, 36]. The seminal Hahn spin echo consists of a sin- fields’ detuning and Rabi frequency. One could there-
gle π-pulse at half the duration τ /2, which refocusses fore imagine a pulse sequence that implements a k-loop
qubit frequency fluctuations oscillating at frequencies MS gate with non-constant durations which match the
ω < 2/(τ ) [37]. It was later shown that interleaving many timings of the PDD pulses. The total entangling phase
π-pulses during the evolution may extend the coherence picked up by all loops should then be identical to that of
time further [21, 38, 39]. It was also found that per- a primitive Mølmer-Sørensen gate.
forming pulses along alternating orthogonal bases could It was later shown that π-pulses can be applied at any
efficiently decouple the qubit from noise in all three axes, point in the MS evolution, even if the spin and motion re-
and similar techniques can mitigate the effects of imper- main entangled [51]. A π-pulse effectively flips the sign of
fections in the decoupling pulses themselves [40–43]. All the operators in the Mølmer-Sørensen unitary. In phase
together, this vast library of pulse sequences form what space, this is equivalent to reversing the direction of the
we refer to as Pulsed Dynamical Decoupling (PDD). trajectory. The timings of the pulses can therefore be en-
PDD schemes may be used to extend the coherence gineered such that fast reversals of the phase space tra-
of spins during an entangling interaction. For exam- jectory disentangle the spin and motion at the gate time.
ple, π-pulses interleaved during a σz ⊗ σz entangling gate The resulting trajectory traces out what is commonly re-
have been shown to suppress both static and time-varying ferred to as a flower, where a π-pulse is added at every
qubit frequency noise [3, 21, 22]. Protecting a Mølmer- intersection of two petals (see Ref. [51] for a detailed ex-
Sørensen evolution in a similar manner is, however, more planation). The duration between two pulses required to
difficult, as the dynamical decoupling pulses do not nec- implement a periodic PDD sequence is π(2 + Nπ )/Nπ δ,
essarily commute with the gate fields. The nature of the where Nπ are the number of π-pulses. For a given MS
bichromatic interaction causes the spin and motion to detuning δ, the gate time is chosen such that the flower
remain entangled throughout the evolution, further com- trajectory encloses an area equal to the maximal entan-
plicating the timings of the π pulses. In what follows, gling phase π/2. In the limit of many π-pulses, the total
we outline a variety of gate schemes that combine PDD required gate time is τ = π2 τ0 [51]. The advantage of this
with a σx ⊗ σx MS interaction. scheme therefore comes from the gate time scaling which
Earlier works assumed that adding π-pulses while the is independent of the number of pulses and allows for fast
spin and motion are still entangled will inevitably dam- PDD gates (assuming instantaneous π-pulses; see section
age the fidelity since they do not commute with the gate II A 2 for a discussion on non-instantaneous pulses). Note
4
1−F
dynamical decoupling sequences with multi-axis pulses,
10-3 5
such as the XY-4 or XY-8 schemes [52, 53]. 12
10-4
1. Fidelity
10-1 b)
As outlined in equations 2 and 3, the infidelity under δΩ(t)/Ω [10 −3 ]
10-2
PDD is estimated from the spin’s loss of coherence. The 2.5
1−F
1
filter function of an arbitrary sequence is [29] 10-3 0.5
10-4
FP DD (ω, τ ) =
Nπ
1 20 40 60 80 100
|1 + (−1) Nπ +1 iωτ
e +2
X
(−1)j eiδj ωτ cos(ωτπ /2)|2 , (4)
Nπ
j=1
Figure 1: Numerical simulations of the Mølmer-Sørensen
where Nπ is the number of π-pulses, δj are their nor- entangling gate protected by pulsed dynamical decou-
malized timings and τπ is the duration of a single pulse. pling and subject to spin dephasing noise. The cir-
In general, the PDD filter function corresponds to a high- cles represent the simulated infidelities averaged over 200
pass filter, whose low-frequency roll-off largely varies be- noise trajectories, in which qubit frequency fluctuations
tween specific sequences, [30]. An efficient PDD sequence of equation 1 are modelled as an Orstein-Uhlenbeck pro-
is one that results in a steeper roll-off. Furthermore, the cess [11, 54, 55]. Solid lines are predictions from the ana-
corner frequency of the high-pass filter tends to increase lytical infidelity models of equations 5 and 8. (a) Infideli-
with the number of pulses. A large number Nπ is there- ties due to qubit frequency fluctuations as a function of
fore desirable such that the stop-band region of the filter the number of noise-free pulses Nπ . The coherence times
function suppresses most of the noise. The resulting in- T2 are those of the magnetic sensitive transition, and
fidelity from dephasing is, to first order, are chosen such that T2 = 2 ms is achievable on readily
available systems, while T2 = 12 ms can be obtained via
1
Z +∞
FP DD (ω, τ ) additional noise shielding [56]. (b) Infidelities from pulse
I= Sz (ω) dω. (5) imperfections, where amplitude noise is injected into the
4π −∞ ω2
refocussing pulses and the MS evolution is made noise-
We identify another source of infidelity arising from free. Relative amplitude noise levels were chosen from
imperfections in the dynamical decoupling pulses them- [57].
selves. This error is evaluated by following a similar rea-
soning as in Ref. [52]. Because the pulse durations are
generally fast with respect to the qubit frequency fluctu-
This final result can be used to evaluate the infidelity
ations, i.e. τπ T2 , the time-dependent noise βz (t) can
from pulse imperfections,
be modelled as a constant offset during rotations. The
same argument is made for amplitude fluctuations. In
this way, all sources of noise are modelled as constant
shifts and the rotations become instantaneous unitaries T r(U U0 )
I= (8)
with a static error in either the azimuthal Bloch sphere T r(U )T r(U0 )
angle or the rotation (polar) angle. The non-ideal π-pulse
operator is where U0 is the ideal error-free operator, obtained by
setting k = 0, nx/y = nz = 0 and ny,x = 1.
#» ,
Uk = exp − i(π + k )(S · n)) (6) An estimate of the total infidelity can be found from
equations 5 and 8. The accuracy of these results is ver-
where k models over- and under-rotations, S are the ified by numerically simulating the Mølmer-Sørensen in-
Pauli matrices and n#» = (n , n , n ) is the rotation axis.
x y z teraction under the influence of dephasing noise. The
Only considering the dynamical decoupling pulses, the results, reported in figure 1, show good agreement be-
total rotation of N pulses becomes tween predicted and simulated infidelities. A trade-off
also becomes apparent: large numbers of pulses Nπ are
N
Y desired to more efficiently decouple the qubit from low
U= Uki . (7) frequency noise, however this also leads to a an increased
i accumulation of error due to pulse imperfections.
5
2. Gate duration 10 0
3 10 −2
10 −1
The shortest entangling duration that makes use of 2
PDD is achieved by the fast gate scheme of Ref. [51],
which applies π-pulses while spin and motion remain en- 1 10 −3 10 −2
10 −4
δω/δ0
1−F
tangled. The gate duration must be slightly prolonged 0 10 −3
given that the enclosed area in phase space is reduced.
1 10 −4
We recall that, in the limit of many π-pulses, the prim-
itive gate duration τ0 increases by a factor of π2 . Along 2 10 −5
with the finite π-pulse durations, the total gate duration 3
becomes 10 −6
0 20 40 60 80 100
π
Nπ
τP DD = τ0 + Nπ τπ . (9)
2
Figure 2: Robustness of the Mølmer-Sørensen entangling
We note that the duration of the fast PDD scheme gate protected by pulsed dynamical decoupling to static
can, in principle be shortened by employing modulation qubit frequency shifts. The Bell state fidelities are nu-
techniques which alter the phase space trajectory (c.f. merically simulated for a range of π-pulse numbers Nπ
Section III B 1). For example, instantaneous phase shifts and normalised shifts δω/δ0 . The dashed, dotted and
in the bichromatic fields alters the MS motional phase, dash-dotted lines are contours corresponding to the infi-
which subsequently changes the direction of the phase delities 10−4 , 10−3 and 10−2 .
space trajectory. By appending such a modulation to
every π-pulse, the phase space trajectory could be engi-
neered to maximize the enclosed area. If the enclosed The model of equation 11 estimates the tolerable qubit
area corresponds to that of a primitive sequence, i.e. a frequency shift that still allows infidelities below a certain
circle, the gate duration becomes threshold. For example, one can achieve infidelities below
10−3 despite a shift of δω = 0.108δ0 if Nπ = 10 pulses
are chosen. Alternatively, a larger shift of δω = 1.008δ0
τP DD = τ0 + Nπ τπ . (10) is possible if one chooses Nπ = 100 pulses. For a gate
This result places an additional trade-off on the num- duration of τ0 = 1 ms, this would correspond to δω/2π ≈
ber of pulses Nπ . While many π-pulses are desired to 1 kHz.
improve the filtering properties (c.f. equation 4), this di-
rectly increases the gate duration, making the interaction
4. Calibration requirements
more sensitive to other sources of noise, such as motional
decoherence.
A calibration involves the determination of a parame-
ter by fitting the results of an experiment. The results
3. Robustness to static shifts of a calibration experiment may be skewed by another
parameter that is mis-calibrated. This introduces the
Robustness of the PDD scheme to static qubit tran- notion of dependency, i.e. some calibrations should pre-
sition frequency shifts is investigated by means of nu- cede others. A calibration sequence is a collection of cal-
merical simulations. The noise Hamiltonian (equation 1) ibration experiments that should be executed in a spe-
generalized to multiple ions is integrated with the stan- cific order to take into account any possible dependen-
dard MS Hamiltonian after replacing the time-dependent cies of a gate scheme. In this way, a calibration sequence
noise by a static shift δω. A periodic PDD sequence is can be modelled as a Directed Acyclic Graph (DAG),
integrated by applying equally spaced π-pulses that are wherein vertices (or nodes) designate a parameter and
error-free and instantaneous. Figure 2 reports the Bell arcs (edges) represent a dependency [34]. We also make
state fidelity for a range of pulse numbers Nπ and normal- the distinction between weak and strong dependencies.
ized detuning errors δω/δ0 , where δ0 is the MS detuning. In the case of a weak (strong) dependency, the child pa-
An empirical model of the robustness is constructed by rameter’s calibration is (in)valid despite a small misset
fitting the contours of figure 2 to a linear function, and in the parent vertex.
one finds The PDD scheme necessitates three fields per ion: a
pair of sidebands for the bichromatic interaction and a
single carrier for dynamical decoupling. Each field is
−2 −2
(2.8 + 3.2Nπ ) × 10 , I ≤ 10 ,
parametrized by a transition frequency ω, a Rabi fre-
δω/δ0 −2
≤ (0.8 + Nπ ) × 10 , −3
I ≤ 10 , (11) quency Ω and a phase φ. Calibrations of the field’s phases
(0.3 + 0.3N ) × 10−2 , I ≤ 10−4 .
are not considered here, as typical microwave sources ex-
π hibit low phase drift and a high resolution. The resulting
6
a)
ν
F=1
(i)
Ω r,b
δ0
Fidelity
ω0(i) δω0
F=0
b) Continuous Dynamical Decoupling
Ωc(i)
Ωr,b
Weak Strong τ
dependency Ωc
dependency
τ
Figure 3: Directed acyclic graph describing the calibra-
tion requirements of both the pulsed and continuous dy- c) Pulsed Dynamical Decoupling
namical decoupling schemes. The nodes represent pa- Ωr,b
rameters that need to be calibrated, and vertices repre-
sent dependencies. Yellow (red) nodes consist of transi- τ
tion frequencies (Rabi rates). Thin (thick) arrows rep- Ωc
resent weak (strong) dependencies. Nodes within paren- τ
theses indicate that this sub-graph should be repeated
for every ith ion. Figure 4: Quantum control scheme for PDD and CDD.
Sideband (carrier) fields are in blue (red). (a) Energy
levels of a typical hyperfine ground state. Ions have dif-
DAG for the PDD scheme is illustrated in figure 3. The ferent resonance frequencies due to the static magnetic
graph contains calibrations for the qubit transition fre- field gradient. PDD and CDD fields must address a mag-
quency ω0 , the bichromatic and carrier Rabi frequencies netic sensitive transition between F=0 and F=1 (either
Ωr,b and Ωc , a possible stark shift δω0 , the MS detuning of the opaque or transparent fields are possible). (b-c)
δ0 and finally the secular frequency ν. In total, we iden- Pulse sequences for both schemes. In CDD (b), both
tify 12 nodes, with 12 strong and 10 weak dependencies. sidebands and carrier fields are continuously applied. In
PDD (c), the carrier field is interleaved throughout the
evolution.
5. Experimental overhead
1−F
5
the carrier drive, which now only provides additional dy- 12
namical decoupling [15]. 10-4
In order to elucidate the beneficial effects of the drive, 10-5
we consider the usual MS Hamiltonian along with a field
resonant with each qubit carrier transition and a noise 10-1
term, b)
10-2 δΩ(t)/Ω [10 −3 ]
2.5
10-3 1
1−F
0.5
H = HM S + Hdrive + Hnoise (12) 10-4
~Ωc X (i)
Hdrive = σ (13) 10-5
2 i x
0 20 40 60 80
~βz (t) X (i) N = Ω c /δ0
Hnoise = σz . (14)
2 i
Figure 5: Numerical simulations of the Mølmer-Sørensen
Provided that the basis of the drive is identical to entangling gate protected by continuous dynamical de-
that of the MS interaction, Hdrive and HM S commute coupling and subject to spin dephasing noise. (a-b) as
with one another. Therefore, in an interaction basis captioned in figure 1, namely, numerical simulations used
with respect to the drive, i.e. H̃ = eitHdrive /~ (H − the exact Hamiltonian of equation 12, while analytical
Hdrive )e−itHdrive /~ , HM S is unaffected and equation 12 fidelities were obtained from equation 17 and 18. (a) In-
becomes fidelities due to qubit frequency fluctuations for varying
carrier Rabi frequencies, parametrized by the number of
2π pulse performed at the gate duration N = Ωc /δ0 . The
H̃ = HM S + H̃noise , (15)
dynamical decoupling field is noise-free. (b) Infidelities
~βz (t) X h i
from amplitude noise in the carrier dynamical decoupling
H̃noise = cos(Ωc t)σz(i) + sin(Ωc t)σy(i) . (16)
2 i
fields, modelled by the replacement Ωc → βx (t)Ωc , where
βx (t) is an Ornstein-Uhlenbeck process [11, 54, 55]. Here,
The addition of a resonant drive therefore continuously dephasing noise is ignored by setting βz (t) = 0.
rotates the qubit frequency noise βz (t) around the z and
y axes. The terms of H̃noise are neglected under a ro-
tating wave approximation in the limit βz (t) Ωc . The
drives decouple the qubits from frequency fluctuations,
Sz (Ωc )τ
and only noise at a frequency nearing Ωc will decohere I= . (17)
the spins. It is generally beneficial to introduce a large 4
carrier Rabi frequency to better suppress qubit frequency An additional infidelity term arises when consider-
fluctuations. We note that the interaction picture of the ing noise in the dynamical decoupling carrier field it-
carrier drive must coincide with that of the MS evolu- self. Amplitude fluctuations introduce Rabi frequency
tion at the completion of the entangling gate. To this noise, which is modelled as Ωc βx (t)σx , where βx (t) is
end, the carrier should perform an integer number of ro- the fractional Rabi frequency noise with PSD Sx (ω) =
tations during the evolution, which places a constraint R +∞
hβx (0)βx (τ )ie−iωτ dτ . The noise amplitude there-
on its Rabi freuqency, Ωc = 2πk/τ0 where k ∈ Z+ . −∞
fore increases with Ωc , which limits the efficacy of the
dynamical decoupling and introduces a trade-off. The
infidelity from this additional noise is, from equation 2
1. Fidelity
+∞
Ω2
Z
The infidelity is estimated from the decay of the spin’s Fx (ω, τ )
coherence. From equations 2 and 3, we aim to find a I= c dωSx (ω) . (18)
2π −∞ ω2
filter function for the continuous drive. In an interac-
tion picture with respect to Hnoise , dephasing noise is With no further dynamical decoupling, the filter func-
suppressed by the energy gap that is opened in the new tion Fx (ω, t) coincides with that of a free induction decay
eigenstates. One then expects a sinc-like filter function, sequence, i.e. Fx (ω, t) = 2 sin2 ( ωt
2 ). A trade-off becomes
which consists of a narrow-bandwidth band-pass filter apparent from equations 17 and 18. A large carrier Rabi
(see appendix D). This is approximated by a Dirac delta frequency is desired to filter higher frequencies of noise
function [62] and the infidelity becomes, to first order, which, for a typical correlated noise spectrum, contain
8
1−F
cies. The infidelities of equations 17 and 18 are verified by 0.5
numerically integrating the full Hamiltonian of equation 10-4
12. The analytical model and simulation results show
good agreement, and further illustrate the trade-off in 10-5
the choice of the carrier Rabi frequency. 0 10 20 30
There exist various extensions to the CDD scheme that Npf
make the filter function of equation 18 more efficient
at decoupling the spins from amplitude noise. A first
Figure 6: Infidelities of the Mølmer-Sørensen entangling
scheme consists in repeatedly adding dynamical decou-
gate with continuous dynamical decoupling due to am-
pling drives with decreasing powers [61]. The addition of
plitude noise in the carrier field itself. A number of phase
a second drive effectively opens a new energy gap which
flips Npf are introduced in the dynamical decoupling field
decouples the first drive’s amplitude noise. In princi-
to refocus noise by means of rotary echoes. Circles are
ple, any number of concatenated drives can be imple-
the result of numerical simulations (c.f. figure 5) while
mented, each mitigating the noise introduced by the pre-
solid lines are analytical predictions from equation 18.
vious one. A different approach uses continuous phase
modulation on the principle dynamical decoupling car-
rier field [63, 64]. Similarly to the concatenated drives,
this also opens an additional energy gap which decouples 19 can be suppressed by introducing a time-dependent
the qubit from amplitude noise. Alternatively, one can phase modulation on all of the fields. The scheme can
apply refocussing pulses within the dressed eigenbasis to also be combined with rotary echoes to mitigate imper-
mitigate noise in the drive itself [65]. fections in the carrier drive itself.
While the previous methods of decoupling noise from
the drive itself are efficient, they each involve additional
2. Gate duration
fields or modulations which increase the complexity of the
scheme. An experimentally simpler solution consists of a
rotary echo sequence [66, 67]. Rotary echos are analogous Since all participating fields are always on, the total
to pulsed dynamical decoupling, however refocussing oc- available power of the microwave synthesis chain must
curs in a frame rotating with the continuous drive. In- be shared between several tones: the four MS sidebands
stead of introducing π-pulses, phase flips in the drive are and the two dynamical decoupling carriers. This intro-
applied throughout the evolution. The filter function of duces a trade-off in the choice of the carrier Rabi fre-
equation 18 is then found from the timings of the phase quency, wherein larger powers are desired to more effi-
flips, and one can use those derived in the context of ciently decouple the qubit from noise (c.f. equation 17)
PDD. The resulting infidelity model is numerically veri- but also take away power from the gate fields, thereby
fied in figure 6. We stipulate that mitigating amplitude reducing the gate speed. Assuming a total power budget
noise by means of rotary echos is more beneficial than Ωmax = 4ΩM S + 2Ωc , the gate duration suffers from a
other methods that were mentioned, since phase flips do reduction of
not affect the power budget and are more straightforward
to implement experimentally. Ωmax
A final infidelity source arises from the off-resonant τcdd = τ0 (20)
coupling of the carrier field to the motional states. An Ωmax − 2Ωc
order of magnitude estimate of the error is obtained from This result can directly be plugged into the previously
the excitation probability of the carrier field in the bichro- derived infidelities of equations 17, 18 and 19 to find an
matic interaction picture (see appendix E), optimal carrier Rabi frequency Ωc . Note that this trade-
off for CDD is very similar to that of PDD, in that in-
ν 4 −1 creasing the dynamical decoupling quality by means of
I = (1 + ) , (19) increasing the number of π-pulses or the Rabi frequency
Ω2c Ω20
leads to an increased gate duration.
where ν is the motional frequency and Ω0 is the sideband
Rabi frequency. This interaction is suppressed by ensur-
ing that Ωc ν. While large carrier Rabi frequencies 3. Robustness to static shifts
are desired to reduce the spin dephasing infidelity term
of equation 17, errors due to off-resonant coupling scale as In an identical manner to PDD (c.f. section II A 3), the
∝ Ω2c and may quickly deteriorate the fidelity. Neverthe- robustness to static qubit frequency shifts is investigated
less, Ref. [57] shows that the infidelity term of equation by numerically simulating the Hamiltonian of equation
9
1−F
0 10 −3
1 10 −4 C. Multi-level Continuous Dynamical Decoupling
2 10 −5 Both pulsed and continuous dynamical decoupling
3 −6 schemes made use of a simple two-level system and en-
10
0 50 100 150 coded the qubit with a magnetically sensitive transition.
N = Ω c /δ0 Furthermore, both schemes suffer from susceptibility to
noise in the dynamical decoupling fields themselves. One
can instead make use of the multi-level structure that
Figure 7: Robustness of the Mølmer-Sørensen entangling naturally appears in hyperfine ground states and encode
gate protected by continuous dynamical decoupling to qubits in a decoherence-free subspace. We refer to this
static qubit frequency shifts. The Bell state fidelities are approach as multi-level continuous dynamical decoupling
numerically simulated for a range of carrier Rabi frequen- (MLCDD) [23–25, 68–73]. Similarly to CDD, a pair of
cies Ωc and normalised shifts δω/δ0 , where δ0 is the MS carrier fields are continuously applied to drive two mag-
detuning. The carrier Rabi frequency is parametrized by netically sensitive mf = ±1 states which we label |±1i.
N = Ωc /δ0 , such that N represents the number of 2π The dynamical decoupling Hamiltonian is, for σx drives,
carrier oscillations completed at the gate duration. The
dashed, dotted and dash-dotted lines are contours corre-
sponding to the infidelities 10−4 , 10−3 and 10−2 . X ~Ωc 1
Hdd = √ |0i(j) h−1|(j) + h+1|(j) + H.C. .
j
2 2
(22)
12 after replacing δβz (t) → δω (see figure 7). The carrier In the eigenbasis of these carriers, one finds the eigen-
Rabi frequency is parametrized by N = Ωc /δ0 , which state |Di = √12 (|−1i + |+1i) whose eigenenergy is degen-
represents the number of carrier oscillations performed erate with the |00 i state. For this reason, the transition
during the interaction. An empirical model is fitted to |00 i → |Di is insensitive to noise in the dynamical de-
the resulting infidelities, coupling drives themselves, and makes for an excellent
qubit. To better understand the dynamical decoupling
√ −2
interaction, one can consider noise of the form
0.22√N , I ≤ 10 ,
δω/δ0 ≤ 0.13 N , I ≤ 10−3 , (21)
0.07√N , I ≤ 10−4 .
~βz (t)
Hnoise = (|+1i h+1| − |−1i h−1|), (23)
2
This model allows one to estimate the largest tolerable where we only include the first order Zeeman
qubit frequency shift that still allows infidelities below a shifts of the magnetically sensitive transitions, as
specific threshold. For example, for a gate duration of this generally dominates the infidelity. Transforming
τ0 = 1 ms, one could achieve infidelities below 10−3 for Hnoise in an interaction picture with respect to Hdd ,
shifts of up to δω/2π = 1.3 kHz with a carrier Rabi fre- e−itHdd /~ Hnoise eitHdd /~ ,
quency Ωc /2π = 100 kHz, as opposed to δω/2π ≈ 1 kHz
with Nπ = 100 pulses using PDD (c.f. section II A 3). √
δβz (t)
H̃noise = √ S+ eitΩc / 2 + H.C. . (24)
2 2
4. Calibration requirements Here, the multi-level ladder operators S+ = |Di hd| +
|ui hD| and S− = |di hD| + |Di hu| describe transitions
Similarly to PDD, the CDD scheme requires three between the eigenstates. The effects of the dynamical
fields per ion: two sideband fields driving the MS inter- decoupling drive are elucidated by equation 24, which
action and one carrier that is continuously applied. Due shows that only
√ noise nearly resonant with the eigenstate
to the static magnetic field gradient imparting different splitting Ωc / 2 may drive population out of the logical
transition frequencies to each qubit, a total of six fields qubit state |Di and into spectator states |ui and |di.
are required. The calibration DAG is identical to that of Consequently, dephasing of the magnetic sensitive states
the PDD scheme (c.f. figure 3); there are 12 nodes, with under the action of continuous dynamical decoupling re-
10 strong dependencies and 8 weak dependencies. sults in an equal distribution of population among the
10
F=1
analogous to a leakage error mechanism. Finally, we note
that transitions within
√ the eigenstates are slightly weaker
by a factor of 1/ 2.
Obtaining an MS type interaction within the dressed
F=0
states involves applying RF fields on either of the |00 i →
|±1i transitions, and introducing a detuning equal to the
b) Multi-Level Continuous Dynamical Decoupling
motional frequency. In this way, sideband transitions are
driven on the |00 i → |Di transition. Note that, due to Ωr,b
the second order Zeeman shift, both |00 i → |±1i tran- τ
sitions are resolvable and can lead to transitions with Ωc
the |Di state [69]. The full Hamiltonian describing the τ
bichromatic interaction is
Figure 8: Required fields (a) and pulse sequence (b) for a
X ~Ωrf (j) 0 (j) −iδrf t j (a† −a) MS entangling gate that uses the multi-level continuous
Hrf = |+1i h0 | e e dynamical decoupling scheme. The RF gate fields (blue)
j
2
are near resonance with a magnetic sensitive transition
(j) (j) i(δrf −∆ω± )t −j (a† −a)
h00 |
+ |−1i e e + H.C. , (25) within the F = 1 manifold. The microwave dynami-
cal decoupling fields (red) bridge both magnetic sensitive
where δrf is the detuning from the carrier transition, j transitions between F = 0 and F = 1.
is the Lamb-Dicke parameter and ∆ω± is the frequency
splitting between the |00 i → |+1i transitions that arises
due to the second order Zeeman shift. Moving Hrf into bandwidths (c.f. appendix D).√Conversely, the passband
an interaction picture with respect to Hdd of equation filter’s center frequency is Ωc / 2 instead of Ωc due to the
22, and dropping fast oscillating terms under a rotating different eigenenergies of the dressed states. With these
wave approximation
√ in the limits Ωrf /2 ∆ω± and modifications, the infidelity due to dephasing becomes,
δ0 Ωc / 2 ν (where δ0 = δrf − ν), to first order,
Ω0 (1) √
(σ̃x − σ̃x(2) ) aeiδ0 t − a† e−iδ0 t .
H̃rf = − (26) Sz (Ωc / 2)τ
2 I= . (28)
12
This final Hamiltonian corresponds to the usual
(i) The interaction is protected from noise in the con-
Mølmer-Sørensen interaction, where σ̃x = |00 i hD|√− tinuous dynamical decoupling fields themselves, since
|Di h00 | is the modified Pauli operator, and Ω0 = Ωrf / 2 the transition frequency of the qubit {|00 i , |Di} is in-
is the MS Rabi frequency. dependent of the amplitude Ωc . Therefore, the MLCDD
scheme does not have infidelity terms originating from
the drive’s noise. Nevertheless, the multiple levels that
1. Fidelity
are used in this scheme lead to several unwanted transi-
tions and parasitic couplings.
The decay function of the spin’s coherence under ML- A dominant infidelity term of order Ω2c /ν 2 appears
CDD is slightly different than for PDD and CDD (c.f. from the off-resonant coupling of the carrier fields to the
equation 2). As outlined in the noise mechanism of equa- motional sidebands [71]. This term, however, oscillates
tion 24, dephasing causes population leakage into spec- at a frequency ν, and can therefore be eliminated by care-
tator eigenstates. In the limit of large durations, the fully choosing the gate time. This nevertheless requires
state decoheres into an equal distribution of |Di, |ui and a timing resolution that is much smaller than 1/ν, and
|di. Taking this into account, the modified infidelity from the interaction is more susceptible to drifts of the mo-
equation 2 becomes tional mode frequency. Another similar error term of
order Ω20 /ν 2 originates from the off-resonant coupling of
1 the MS fields to the carrier transitions. Nevertheless, am-
I= (1 − e−χ(t) ). (27) plitude pulse shaping can be used to adiabatically drive
3
transitions and reduce the timing sensitivity at the gate
The filter function of the MLCDD scheme is almost duration [25].
identical to that of CDD, i.e. the transfer function corre- Other higher order terms arise from the coupling of the
sponds to a sinc-like passband filter that can be approx- sideband fields to the carrier transitions, leading to pop-
imated by a Dirac delta function in the limit of small ulation leakage and unwanted energy shifts [23]. These
11
10 −2 10 −3
10-1 10 0
δΩ(t) = 0 3 −4
δΩ(t) 0 10
10-2 T2 [ms] 2 10 −1
1−F
1
2 1 10 −2
10-3 4
δω/δ0
1−F
0 10 −3
10-4
1 10 −4
1 15 30 45
Ω c /2π [kHz] 2 10 −5
3
10 −6
Figure 9: Numerical simulations of the Mølmer-Sørensen 0 5 10 15 20 25 30
entangling gate protected by multi-level continuous dy- N
namical decoupling and subject to spin dephasing noise.
See caption of figure 1 for numerical simulation details, Figure 10: Robustness of the Mølmer-Sørensen entan-
namely, the Hamiltonians of equations 22, 23 and 25 were gling gate protected by multi-level continuous dynamical
used. Analytical fidelities were obtained from equation decoupling to static qubit frequency shifts. The Bell state
28. Numerical simulations modelled both an ideal (cir- fidelities are numerically simulated for a range of carrier
cles) and a noisy (crosses) dynamical decoupling carrier. Rabi frequencies Ωc and normalised shifts δω/δ0 , where
δ0 is the MS detuning. The carrier Rabi frequency is
parametrized by N = Ωc /δ0 , such that N represents the
terms can be made vanishingly small by driving the car- number of 2π carrier oscillations completed at the gate
rier transitions between the two qubits at different Rabi duration. The dashed, dotted and dash-dotted lines are
frequencies. contours corresponding to the infidelities 10−4 , 10−3 and
Finally, off-resonant coupling of the carriers to the mo- 10−2 .
tional sidebands as well as coupling of the sideband fields
to the spectator’s motional states generate an MS type
interaction of strength Ω2c /(2ν 2 − Ω2c ). This interaction imperfect frequency in the control fields set by the ex-
can nevertheless be compensated for by appropriately ad- perimentalist. In the MLCDD scheme, however, these
justing the gate’s duration. two error sources do not coincide with the same robust-
ness. On the one hand, a change in the magnetic field
will, to first order shift the magnetic sensitive transitions
2. Gate duration by an equal and opposite amount (c.f. equation 23). On
the other hand, a misset in the frequency of the control
The gate speed of the MLCDD scheme is different from field results in a detuning error with the |00 i → |Di tran-
the PDD and CDD methods in that the gate fields (RF) sition. The robustness to either, as will be shown, is very
and dynamical decoupling fields (microwave) originate different.
from separate synthesis chains. Their power budgets are We first consider the robustness to a change in mag-
therefore independent, and there is no longer a trade-off netic field by numerically simulating the Hamiltonians
between increasing the dynamical decoupling quality and of equations 22, 23 and 25 (after dropping terms rotat-
reducing the interaction strength. The gate duration is ing with ∆ω±1 and ν) and performing the replacement
then βz (t) → δω. The results are reported in figure 10 for a
range of carrier Rabi frequencies, where N = Ωc /δ0 . An
empirical model is then constructed by fitting the linear
τM LCDD = τ0 , (29) portion of the contours (N 1),
where τ0 is the duration of a primitive MS entangling I ≤ 10−2 ,
−0.33 + 0.30N,
gate after replacing the usual Rabi frequency with Ω0 →
δω/δ0 ≤ −0.40 + 0.17N, I ≤ 10−3 , (30)
Ωrf .
I ≤ 10−4 .
−0.53 + 0.10N,
ν
and measure the stark shift δω0 . The final nodes in the
DAG are the secular frequency ν and the MS detuning
ω+-1
(i)
δ0 . In total, the DAG comprises of 22 parameters, with
Ωr,b(i)
Ω+-1
(i) 20 strong dependencies and 14 weak dependencies.
δ0
ωD(i)
Fidelity
5. Experimental overhead
ωclk
(i)
δω0
Ωclk
(i) The MLCDD scheme requires 2 MS sideband fields and
2 dynamical decoupling drives per ion for a total of 8
fields. The carrier drives should be resonant with the
Weak Strong
hyperfine |F = 0, mf = 0i → |F = 1, mf = ±1i transi-
dependency dependency tions, requiring near equal amounts of σ + and σ − po-
larization. The amplitudes of the carrier fields within
an ion should be identical in order to maximize the co-
Figure 11: Directed acyclic graph describing the calibra- herence of the eigenstate |Di [74]. The sideband fields
tion requirements of the multi-level continuous dynami- should be detuned from either of the |F = 1, mf = 0i →
cal decoupling scheme. Labels and caption identical to |F = 1, mf = ±1i transitions within the |F = 1i triplet,
figure 3. requiring a single component of σ + or σ − polarization.
An additional microwave field resonant with the
|F = 0, mf = 0i → |F = 1, mf = 0i clock transition is
|00 i h00 |). Note that a shift of this nature could also occur required to map population in and out of the computa-
from stark shifts due to the sideband fields [25, 73]. With tional subspace {|00 i , |Di} [24, 72], necessitating a com-
this new definition of Hnoise , we effectively retrieve the ponent of π polarization.
robustness of a primitive Mølmer-Sørensen gate within a
two-level system in the absence of any dynamical decou-
pling, D. Summary
−2 −2
The performances of PDD, CDD and MLCDD are
2.8 × 10 , I ≤ 10 ,
summarized in table I. We first note that the resilience
≤ −2 −3
δω/δ0 0.9 × 10 , I ≤ 10 , (31) of all three schemes to spin dephasing is very similar.
0.3 × 10−2 , I ≤ 10−4 .
Within the filter function framework, the transfer func-
tions that are applied to the noise’s PSD are akin to a
The robustness demonstrated by equation 31 indi- narrow bandwidth passband filter. The filter’s center fre-
cates that the MLCDD scheme is much less forgiving to quency is proportional to the number of π-pulses for PDD
static errors arising from imperfect experimental knowl- and the carrier Rabi frequency for CDD and MLCDD.
edge, e.g. calibrations that aren’t precise enough or unac- Increasing these parameters is desirable to displace the
counted stark shifts. Furthermore, infidelities from these bandpass filter to higher frequencies which, for a typical
errors can not be reduced by increasing the dynamical correlated noise spectrum, contains lower powers of noise.
decoupling power. The filter functions of the pulsed and continuous schemes
differ slightly in that CDD and PDD implement a sinc-
like function that can be approximated by a Dirac delta,
4. Calibration requirements while the PDD implements a high-pass filter, whose low-
frequency roll-off is highly dependent on the timings of
The calibration requirements of the MLCDD scheme the π pulses.
are determined from the corresponding Directed Acyclic The PDD and CDD schemes have an additional source
Graph (DAG) (c.f. section II A 4). Four fields are re- of infidelity that arises from the dynamical decoupling
quired per ion: a pair of sideband fields which drive the fields themselves. In the pulsed scheme, noise and im-
MS interaction and a pair of carrier dynamical decou- perfections in the π-pulses can be treated as static errors
pling fields. The DAG, illustrated in figure 11, contains that accumulate with the number of pulses. For the con-
nodes corresponding to the frequencies and amplitudes tinuous scheme, amplitude noise in the carrier may de-
of each of these fields. Additional nodes are included to cohere the spins and increases with the carrier Rabi fre-
calibrate the |0i → |00 i clock transition (ωclk and Ωclk ), quency. In both cases, this introduces a trade-off with the
since it is required to map in and out of the decoherence number of pulses (PDD) and the carrier Rabi frequency
free subspace [72]. Each ion requires the calibrations of (CDD). Nevertheless, one can implement noise mitiga-
the |0i → |±1i transition frequencies ω±1 and amplitudes tion extensions such as pulses along alternating axes for
Ω±1 , as well as the blue and red sideband Rabi rates Ωr,b . PDD and rotary echoes for CDD. In the case of the ML-
One can then calibrate the qubit transition frequency ωD CDD scheme, errors in the continuous drives do not lead
13
Table I: Summary of the performances of pulsed dynamical decoupling (PDD), continuous dynamical decoupling
(CDD) and multi-level continuous dynamical decoupling (MLCDD). The various metrics are obtained from sections
II A, II B and II C. The symbols and nomenclature are explained throughout the main text. The infidelities are
classified into dephasing errors, representing dephasing noise that is mitigated by the dynamical decoupling scheme,
and intrinsic errors, which consider noise arising from the scheme itself. The MLCDD scheme contains an empirical
robustness model for shifts arising from magnetic field changes, and for shifts due to errors in the control field’s
frequency (c.f. section II C 3). The calibration triplets represent {number of parameters, weak dependencies, strong
dependencies}.
to infidelities by design. Furthermore, leading terms in necessitating many more calibrations for each transition.
the full Hamiltonian suggest that larger carrier Rabi fre- PDD and CDD however make use of an approximate 2-
quencies affect the interaction by introducing static qubit level system, greatly reducing the number of parameters
frequency shifts and modified interaction strengths, both to be calibrated. A similar comment can be made on the
of which can be corrected for by calibrations. We there- experimental implementation and overhead. The added
fore do not include an intrinsic infidelity term, however benefits of MLCDD come at the cost of a greater experi-
precautions should be taken as higher order terms that mental complexity. For example, one requires a synthesis
are analytically and numerically difficult to analyse may chain for both RF and microwaves, all types of polariza-
lead to infidelities in certain parameter regimes. tion and stringent requirements on the amplitudes be-
The gate durations of the PDD and CDD scheme are tween carrier fields.
subject to a similar trade-off, in that increasing the qual-
ity of the dynamical decoupling leads to prolonged gate
durations. Along with the intrinsic infidelities appearing
from noise in the dynamical decoupling fields themselves,
this offers a more practical optimisation constraint. This From the results of table I and from previous discus-
additional trade-off does not appear within the MLCDD sions, it has hopefully become clear to the reader that
scheme, since the gate duration depends only on the MS there is not one quantum control method which solves
field’s power which originates from a different physical every experimental problem, be it high fidelities or fast
signal source to the carrier dynamical decoupling fields. gates. On the contrary, each gate scheme addresses a
The calibration requirements of the PDD and CDD particular subset of a problem, while introducing a new
scheme are much lower than for MLCDD. This can be set of constraints and errors. It is therefore important
attributed to the MLCDD schemes using multiple lev- to characterize the various trade-offs to understand the
els within the hyperfine ground state to encode a qubit, suitability of one scheme over another.
14
2
!
YY XX 1
|αk (τ )|2 (n̄k + ) ,
I = 1 − cos(Φm,n − Ψm,n ) 1 − (33)
m n j
2
k
where n̄k is the average motional occupation. The first for the PSTs: the paths from all participating motional
factor multiplies over all pairs of ions (m, n) and evalu- modes must return to the origin at the gate time τ ,
ates errors between the desired entangling phase Φm,n
and the obtained phase Ψm,n . Errors in Ψm,n arise from X
deviations in the enclosed area of the target mode, or Rend := |αk (τ )|2 = 0. (34)
parasitic phase accumulations of the spectator motional k
modes. The second factor of equation 33 captures infi-
delities from residual spin-motion coupling for all modes In the primitive Mølmer-Sørensen interaction, this is
k and ions j. If αk (τ ) = 0, this second infidelity term re- ensured by carefully choosing the gate’s parameters, i.e.
duces to zero, regardless of the mode’s temperature n̄k . setting δ0 = 2Ω0 and τ = 2π/δ0 .
This further demonstrates the Mølmer-Sørensen gate’s
robustness to the ion chain’s temperature. However, if
the PST does not finish at the origin (αk (τ ) 6= 0), larger 1. Motional dephasing
n̄k amplify the infidelity term. This places a practical
constraint on the maximal temperature, since experi- We now turn our attention towards robustness to mo-
mentally achieved PSTs are never ideal. Furthermore, tional dephasing, and first focus on time-independent
this leads to the first and most fundamental requirement static frequency shifts. This is modelled by adding a
15
small error to the detuning in equation 32, δk → δk + δ. which disturb the PST, preventing it from returning to
The PST’s sensitivity to motional frequency shifts is the origin. We can therefore put forth a qualitative re-
eliminated to first order by setting ∂αk (τ )/∂δ = 0, and quirement and stipulate that the average distance from
one finds [75, 76] the origin must be minimized to mitigate the effects of
motional heating. Quantitatively, the distance of the
Z τ PST at any time t is |αRk (t)|2 , and the time-averaged dis-
τ
iτ αk (τ ) − i dtαk (t) = 0. (35) tance is h|αk (t)|2 i = τ1 0 dt|αk (t)2 |. In Ref. [78], this is
0
placed on a firmer footing and it is found that the best
The α(τ ) of the first term evaluates to zero as it is performance is obtained when h|αk (t)|2 i is minimized.
fulfilled by the requirement Rend of equation 34. The This allows us to write the final requirement for robust-
second term describes the average position of the PST ness to motional heating,
throughout the evolution. The PST should therefore
be designed such that this term reduces to zero. The Z τ
requirement for robustness to static motional frequency 1 2
Rheating := min dt|αk (t)| . (39)
offsets of the k th mode is then τ 0
The filter
P function is summed over all motional modes, 5 1 1
Fδ (ω) = k Fδ,k (ω). For the k th mode, the filter func- I= ˙ k |2 iτ ] − Exp[−4n̄h|α
− Exp[−n̄h|α ˙ k |2 iτ ]. (42)
tion corresponding to a pair of ions m, n is 8 2 8
Taking the first order Taylor expansion of equation 42,
Fδ,k (ω) = the infidelity due to motional heating is approximated by
Z τ
Tk (m) 2 (n) 2
(| | + |k | )| dtΩ(t)ei(δk (t)−ω)t e−iφ(t) t|2 ,
4 k 0 ˙ k |2 iτ.
I = n̄h|α (43)
(38)
From equation 43, it becomes clear that minimizing
where we have defined Tk = 2(n̄k + 12 ). Satisfying the ro- the average distance from the origin h|αk |2 i will minimize
bustness condition of 36 will minimize the integral term the infidelity. For a primitive two-qubit Mølmer-Sørensen
of equation 38 for small ω, thereby increasing the effi- gate with constant parameters, the average displacement
ciency of the filter function. is h|αk |2 i0 = 1/2. The infidelity therefore simplifies to
˙ /2. Alternatively, the heating rate may be re-
I = n̄τ
placed with an effective heating rate such that
2. Motional heating
(1) (2)
A first criterion is the scaling of the gate time. For with Sx = σx + σx , and the PST is
larger ion chains (N 2), the total time of AM and
PM gates scales linearly with the distance between the N
ion pair since non-neighbour interactions are weaker [81].
X
αk (t) = αj,k (t), (48)
The duration of FM gates, however, scales linearly with j=1
the length of the ion chain. The different scalings lead
to vastly different performances for various algorithms where αj,k (t) is the displacement of the j th tone with
within large chains and it is therefore crucial to make Rabi frequency Ωj and detuning δj . By choosing ap-
the appropriate selection [81]. Nevertheless, in what fol- propriate coefficients cj , an N -tone MS gate can imple-
lows we only consider small ion chains that would be ment robustPPSTs. The coefficients are first constrained
used within a QCCD type architecture, such as the one by setting c2j /j = 1, which is required to generate a
considered in Ref. [18]. maximally entangled state. The zero-averaged position
A variety of modulated sideband schemes were also requirement ofP equation 36 (Rdephasing ) is further
P satis-
developed with the aim of mitigating errors from off- fied by setting cj /j = 0. Finally, the quantity c2j /j 2
resonant coupling to other motional states. In laser- should be minimized as per the requirement of equation
based architectures with large strings of ions, the mo- 39 (Rheating ) which leads to gates that are robust to
tional frequencies are spectrally crowded and spectator heating. The reduction factor of the heating rate that
modes strongly interact with the gate field. The mod- was defined in equation 44 now becomes
ulation sequences are therefore designed such that the
total entangling phase from every contributing motional n̄˙ ef f 1 X c2j
mode leads to maximal entanglement [77, 92]. Further- Rheat = = . (49)
more, the sequence ensures that all modes of vibration n̄˙ 2 j j2
are decoupled from the spin states at the completion of
An optimal set of coefficients are derived from the pre-
the gate. These restrictions are fortunately alleviated jb
viously mentioned constraints, resulting in cj = 4 1−jλ
for a QCCD architecture with global microwave fields j −1/2
where b = − 41 ( j (1−jλ)
P P
and static magnetic field gradients. On the one hand, 2) and j 1/(1 − jλ) = 0
the small Lamb-Dicke parameter strongly suppresses cou- [78].
pling to spectator modes of vibration. On the other It is also interesting to note the power requirements
hand, the achievable Rabi frequencies with global radia- of an N -tone gate. Given the close spectral proximity
tion fields are smaller than for laser beams. The resulting of the tones (δ0 ) and the time-scale of the interaction
infidelity is on the order of 10−4 to 10−6 (see appendix (2π/δ0 ), additive and destructive interferences introduce
A). For this reason, the sideband modulation schemes a time-dependence on the amplitude. The total signal is
PN
will only be discussed in the context of improving ro- modelled as fN (t) = j cj ei(ω+jδ0 )t , and the maximum
bustness towards motional decoherence. The additional amplitude of an N -tone gate is max(|fN (t)|)t∈[0,τ ] [94].
constraint of managing parasitic couplings to spectator Constraints on the maximum amplitude of the physical
modes is not an issue for the architecture considered here, fields arise from classical hardware limitations. There-
which grants more freedom in designing efficient PSTs. fore, the primitive Rabi frequency Ω0 of a multi-tone
gate should be adjusted, as both the amplitude and gate
duration are scaled by max(|fN (t)|). For example, a 2-
2. Multi-Tone Mølmer-Sørensen Gate tone gate involves two red (blue) sideband fields with
frequency separation
p δ0 . The amplitude of the total sig-
In the previous section, it was shown that arbitrary nal is |fN (t)| = (5 − 4 cos(tδ0 ))/3, which is maximized
√
PSTs can be engineered by modulating the parameters at half the gate duration where |fN (τ /2)| = 3. The
√
of the sideband fields. Alternatively, one can keep the peak amplitude is therefore 3 times larger compared to
sideband fields constant and add bichromatic tones of a primitive Mølmer-Sørensen gate and the gate √ duration
varying amplitudes, which also contribute to the Mølmer- and Rabi frequency should be adjusted by τ → 3τ and
√
Sørensen interaction [73, 78, 93]. In this way, an N tone Ω0 → Ω0 / 3 to satisfy the power requirements. The
gate involves 2N sideband fields per ion. The detuning gate time scaling of an N -tone gate is
of the j th bichromatic fields are δj = jδ0 and their am-
plitudes are Ωj = cj Ω0 . The real valued coefficients cj
greatly influence the dynamics of the interaction and are Rtime = max(|fN (t)|)t∈[0,τ ] . (50)
chosen to implement robust sequences. The Hamiltonian The coefficients, heating and gate time scalings for up
describing an N -tone Mølmer-Sørensen gate is (setting to N = 5 tones are reported in table II. For increasing
φ = 0) numbers of tones, Rheat decreases and leads to larger re-
ductions of the heating rate. For example, a 5-tone gate
N leads to a tenfold improvement. However, the time scal-
~Ω0 X
H= Sx cj (a† eijδ0 t + ae−ijδ0 t ), (47) ings also progressively worsen, and the tenfold improve-
2 j=1 ment comes at the cost of a gate that is 2.66× slower.
18
Table II: Optimal coefficients and performance of an N -tone Mølmer-Sørensen gate. The factors Rheat and Rtime are
calculated from (49) and (50) respectively. The time-scaled heating reduction factor is R̃heat = Rheat Rtime .
functions.
We conclude that AM and PM are more efficient as
they lead to the smallest gate time scalings. Furthermore,
their experimental complexity is minimal compared to
multi-tone gates. In the following sections, only PM
is considered as it is conceptually simpler to work with
piecewise constant functions.
Primitive Robust
×10 −3 1-F
15
Primitive PST -1
10
Simulation
Theory
Cold
10 -2
10
Infidelity I
-3
10
5
-4
10
Rtime = 1
Hot
1
0.0 0.2 0.4 0.6 0.8 1.0 10
-3
Rheat -4
10
E. Outlook Ω0
Ω 0 (t)
0
The previous sections have identified extensions to the
Mølmer-Sørensen gate which provide robustness to both
φm (t)
φ0 φ1 φ2 φ3 φ4 φ5 φ6 φ7 φN
spin and motional decoherence. While these can be used 0
to independently increase the coherence of either inter- Ωc
nal or external degrees of freedom, they can more im-
Ω c (t)
0
portantly be combined to provide simultaneous robust- −Ω c
ness. For example, phase modulation of the sideband π
φc (t)
fields can be combined with a continuous dynamical de-
coupling carrier field. A proof of concept demonstration 0
of this particular interaction is presented in section IV. 0 τ
Another example of such combinations can be found in Time
Ref. [57], wherein the Mølmer-Sørensen interaction is
subject to continuous dynamical decoupling and addi- Figure 17: Pulse sequence for a Mølmer-Sørensen gate
tional π-pulses that lead to motional robustness. that is simultaneously robust to spin and motional deco-
More generally, the various quantum control schemes herence. Both the sideband fields (Ω0 , φm ) and the car-
previously outlined make up a library of gate extensions rier (Ωc , φc ) are represented. Phases are represented in
that can be combined with one another. The particular red and amplitudes in blue. The motional phase is mod-
choice of which schemes to use is then motivated by the ulated by a piecewise constant function. The sideband
desired performance of an experimental system and its Rabi frequency is kept constant throughout the gate in-
available hardware. For example, if the fidelity of the en- teraction. The carrier’s phase alternates between 0 and
tangling operation should be maximized, one can build π, causing a change in the sign of the carrier Rabi fre-
an error model from the infidelity terms derived in the quency. The frequency with which phase flips occur are
previous sections and optimize over parameters such as determined from the number of motional phase modula-
the gate duration or the dynamical decoupling Rabi fre- tion segments.
quency. Furthermore, the gate scheme library may be
constrained by factors such as the available polarisation
of the microwave fields or their achievable bandwidth. of the red and blue sidebands, however, are subject to a
discrete phase modulation sequence. The amplitude of
the dynamical decoupling carrier field is constant, how-
IV. EXPERIMENTAL DEMONSTRATION ever phase flips are introduced after every sideband phase
change to implement rotary echoes.
Sections II and III outlined a number of quantum con-
trol schemes that extend the robustness of the primi-
tive Mølmer-Sørensen interaction to either spin or mo- A. Robustness to spin decoherence
tional decoherence. In principle, multiple schemes can
be combined to enable a simultaneous robustness to both The increased robustness to spin decoherence is first
sources of error. For example, one could extend the spin’s characterized. In the absence of any dynamical de-
coherence via PDD, while increasing the robustness to coupling, the coherence time is measured to be T2∗ =
motional heating through frequency modulation of the 358(8) µs via a Free Induction Decay (FID) Ramsey ex-
sideband fields (although extra care should be taken in periment (see figure 18). The coherence time is domi-
keeping track of additional phase offsets introducing by nated by magnetic field noise which couples to the mag-
the π-pulses). netically sensitive transition. We then add a continuous
Here, we experimentally demonstrate the construc- carrier field with a Rabi frequency Ωc /2π = 29.9 kHz,
tion of a robust gate from a library of quantum con- and find that the driven coherence time is on the order
trol schemes by combining continuous dynamical de- of 500 µs. It was found that decoherence during driven
coupling, rotary echoes and sideband phase modula- evolution was dominated by amplitude noise in the car-
tion. The single-ion equivalent of the bichromatic rier field itself. This is further mitigated by introducing
Mølmer-Sørensen interaction is used as a proof-of- rotary echoes, i.e. phase flips which alternate the car-
concept experiment. The qubit is encoded within the rier Rabi frequency from Ωc to −Ωc and refocus ampli-
{|F = 0, mf = 0i , |F = 1, mf = 1i} states of the 2 S1/2 tude noise. The phase flips are applied after every 2π
hyperfine ground state of a 171 Yb+ ion. All fields in- rotation of the carrier, i.e. with a period of 33.44 µs.
volved in the interaction are in the microwave regime, The coherence time under CDD with rotary echoes is fi-
with a frequency nearing 12.64 GHz (see appendix C for nally measured via a Ramsey-type experiment, resulting
further experimental details). The pulse sequence is illus- in T2 = 22.6(3) ms. Continuous dynamical decoupling
trated in figure 17. The sideband fields are applied with therefore extends the spin’s coherence by almost two or-
a square pulse at a fixed Rabi frequency Ω0 . The phases ders of magnitude, despite significant noise in the carrier
22
1.0 a)
1 2
P↑ = (1 − e−2|α(t)| (1+2n̄) ). (54)
2 Rtime = 1.218. The modified bichromatic detuning is
Under the correct Mølmer-Sørensen detuning, the δ 0 /2π = 264 Hz, with the gate duration τ 0 = 3.79 ms.
phase space trajectory α(t) returns to zero at the gate This corresponds to a phase sequence of N = 114 seg-
duration τ , and therefore P↑ = 0. Errors such as param- ments.
eter missets or motional decoherence directly result in a The primitive and robust interactions are compared in
non-zero spin probability. Therefore, the robustness of figure 19. The fields are applied for a duration τ (τ 0 )
the PST can be inferred from measurements of P↑ . for the primitive (robust) sequence and the bichromatic
Robustness to motional decoherence is demonstrated detuning is varied from −2δ0 to 2δ0 (−2δ00 to 2δ00 ). The
by comparing the probabilities resulting from a primitive correct interaction of the primitive (robust) sequence is
bichromatic interaction to one with sideband PM. The obtained for δ = ±δ0 (δ = δ00 ). Around these detun-
sideband fields are set to Ω0 /2π = 30 kHz. The bichro- ings, the response of the measured probability differs be-
matic detuning of the primitive Mølmer-Sørensen inter- tween both sequences. The primitive interaction exhibits
action is δ0 /2π = 321 Hz, resulting in the gate duration a narrow region, while the robust interaction results in
τ = 3.12 ms. Note that at time τ , the interaction picture a flatter region. The latter is therefore more tolerant to
of the bichromatic interaction and the carrier fields must missets in the bichromatic detuning, and by extension is
coincide. In other words, the carrier dynamical decou- more robust to motional decoherence mechanisms. This
pling field should have completed an integer number of is further verified by noting that its PST satisfies all the
2π rotations. This is satisfied by setting Ωc = N δ0 , and requirements for robustness (c.f. figure 19): (i) the path
the carrier Rabi rate was chosen such that N = 93. This is symmetric about an axis, (ii) the average position is
further constrains the efficiency of the rotary echoes, as zero and (iii) the average distance to the centre is smaller.
one can now only implement N = 93 phase flips on the Finally, it is interesting to note that the experimentally
carrier. measured probability P↑ of the robust sequence is smaller
A robust PST is obtained via numerical optimiza- than for the primitive sequence, as this is a reflection of
tions [82, 92] and obeys all robustness conditions out- the interaction’s increased resilience.
lined in section III A. The resulting gate time scaling is An important advantage in using microwave radia-
23
ω = 1/τ
10-5
FID
Fz (ω, t)/ω 2
SE
10-10 PDD, n=9
CPMG, n = 10
10-15 1
10 102 103 104
ω/2π[Hz]
Figure 23: Filter function corresponding to qubit fre-
Figure 21: Filter functions corresponding to various dy- quency noise for the dressed states, i.e. multi-level con-
namical decoupling sequences: Free Induction Decay tinuous dynamical decoupling. The system is modelled
(FID), Spin Echo (SE), Periodic Dynamical Decoupling with the Hamiltonian of equation 22 of the main text,
(PDD) and Carr-Purcell-Meiboom-Gill (CPMG). The to- where Ωc /2π = 30 kHz, and noise is included via equa-
tal duration is set to τ = 1 ms. The PDD and CPMG tion 23. The inset shows that the filter function is well
sequences are implemented with 9 and 10 π-pulses re- approximated by a sinc function
√ that is centred around
spectively. Ωc / 2.
H = H0 + HM S + Hc ,
X ~ω (j)
H0 = 0
σz(j) + ~νa† a,
j
2
which corresponds to equation 19 of the main text un- der the assumption that δ ≈ ν.
view A 75, 10.1103/physreva.75.022329 (2007). tion for configurable, parallel, error-robust entangling
[61] J.-M. Cai, B. Naydenov, R. Pfeiffer, L. P. McGuinness, gates in large ion registers, Advanced Quantum Tech-
K. D. Jahnke, F. Jelezko, M. B. Plenio, and A. Retzker, nologies 3, 10.1002/qute.202000044 (2020).
Robust dynamical decoupling with concatenated contin- [76] P. H. Leung, K. A. Landsman, C. Figgatt, N. M. Linke,
uous driving, New Journal of Physics 14, 10.1088/1367- C. Monroe, and K. Brown, Robust 2-qubit gates in
2630/14/11/113023 (2012). a linear ion crystal using a frequency-modulated driv-
[62] F. Yan, S. Gustavsson, J. Bylander, X. Jin, F. Yoshi- ing force, Physical Review Letters 120, 10.1103/Phys-
hara, D. G. Cory, Y. Nakamura, T. P. Orlando, and RevLett.120.020501 (2018).
W. D. Oliver, Rotating-frame relaxation as a noise [77] A. R. Milne, C. L. Edmunds, C. Hempel, F. Roy,
spectrum analyser of a superconducting qubit un- S. Mavadia, and M. J. Biercuk, Phase-modulated entan-
dergoing driven evolution, Nature Communications 4, gling gates robust to static and time-varying errors, Phys-
10.1038/ncomms3337 (2013). ical Review A 13, 10.1103/PhysRevApplied.13.024022
[63] I. Cohen, N. Aharon, and A. Retzker, Continuous dy- (2020).
namical decoupling utilizing time-dependent detuning, [78] F. Haddadfarshi and F. Mintert, High fidelity quan-
Fortschritte der Physik 65, 1600071 (2016). tum gates of trapped ions in the presence of mo-
[64] D. Farfurnik, N. Aharon, I. Cohen, Y. Hovav, A. Ret- tional heating, New Journal of Physics 18, 10.1088/1367-
zker, and N. Bar-Gill, Experimental realization of 2630/18/12/123007 (2016).
time-dependent phase-modulated continuous dynami- [79] K. Wright, K. M. Beck, S. Debnath, J. M. Amini,
cal decoupling, Physical Review A 96, 10.1103/phys- Y. Nam, N. Grzesiak, J.-S. Chen, N. C. Pisenti,
reva.96.013850 (2017). M. Chmielewski, C. Collins, K. M. Hudek, J. Mizrahi,
[65] G. T. Genov, N. Aharon, F. Jelezko, and A. Ret- J. D. Wong-Campos, S. Allen, J. Apisdorf, P. Solomon,
zker, Mixed dynamical decoupling, Quantum Science and M. Williams, A. M. Ducore, A. Blinov, S. M. Kreike-
Technology 4, 035010 (2019). meier, V. Chaplin, M. Keesan, C. Monroe, and J. Kim,
[66] I. Solomon, Rotary spin echoes, Physical Review Letters Benchmarking an 11-qubit quantum computer, Nature
2, 301 (1959). communications 10, 10.1038/s41467-019-13534-2 (2019).
[67] S. Gustavsson, J. Bylander, F. Yan, P. Forn-Dı́az, [80] R. Blümel, N. Grzesiak, N. Pisenti, K. Wright, and
V. Bolkhovsky, D. Braje, G. Fitch, K. Harrabi, Y. Nam, Power-optimal, stabilized entangling gate be-
D. Lennon, J. Miloshi, P. Murphy, R. Slattery, S. Spec- tween trapped-ion qubits, npj Quantum Information 7,
tor, B. Turek, T. Weir, P. B. Welander, F. Yoshihara, 10.1038/s41534-021-00489-w (2021).
D. G. Cory, Y. Nakamura, T. P. Orlando, and W. D. [81] P. Murali, D. M. Debroy, K. R. Brown, and
Oliver, Driven dynamics and rotary echo of a qubit tun- M. Martonosi, Architecting noisy intermediate-scale
ably coupled to a harmonic oscillator, Physical Review trapped ion quantum computers, in 2020 ACM/IEEE
Letters 108, 10.1103/physrevlett.108.170503 (2012). 47th Annual International Symposium on Computer Ar-
[68] N. Timoney, I. Baumgart, M. Johanning, A. F. Varón, chitecture (ISCA) (2020) pp. 529–542.
M. B. Plenio, A. Retzker, and C. Wunderlich, Quantum [82] H. Ball, M. J. Biercuk, A. Carvalho, J. Chen, M. B. Hush,
gates and memory using microwave-dressed states, Na- L. A. de Castro, L. Li, P. J. Liebermann, H. Slatyer,
ture 476, 185 (2011). C. Edmunds, V. Frey, C. Hempel, and A. Milne, Software
[69] S. C. Webster, S. Weidt, K. Lake, J. J. McLoughlin, and tools for quantum control: Improving quantum computer
W. K. Hensinger, Simple manipulation of a microwave performance through noise and error surpression, Quan-
dressed-state ion qubit, Physical Review Letters 111, tum Science and Technology 6 (2021).
10.1103/physrevlett.111.140501 (2013). [83] T. Choi, S. Debnath, T. A. Manning, C. Figgatt, Z.-
[70] N. Aharon, M. Drewsen, and A. Retzker, General scheme X. Gong, L.-M. Duan, and C. Monroe, Optimal quan-
for the construction of a protected qubit subspace, Phys- tum control of multimode couplings between trapped ion
ical Review Letters 111, 10.1103/physrevlett.111.230507 qubits for scalable entanglement, Physical Review Let-
(2013). ters 112, 10.1103/PhysRevLett.112.190502 (2014).
[71] G. Mikelsons, I. Cohen, A. Retzker, and M. B. Ple- [84] G. Zarantonello, H. Hahn, J. Morgner, M. Schulte,
nio, Universal set of gates for microwave dressed-state A. Bautista-Salvador, R. F. Werner, K. Hammerer, and
quantum computing, New Journal of Physics 17, 053032 C. Ospelkaus, Robust and resource-efficient microwave
(2015). near-field entangling 9be+ gate, Physical Review Letters
[72] J. Randall, A. M. Lawrence, S. C. Webster, S. Weidt, 123, 10.1103/PhysRevLett.123.260503 (2019).
N. V. Vitanov, and W. K. Hensinger, Generation of [85] N. Grzesiak, R. Blumel, K. Wright, K. M. Beck, N. C.
high-fidelity quantum control methods for multilevel sys- Pisenti, M. Li, V. Chaplin, J. M. Amini, S. Debnath, J.-
tems, Physical Review A 98, 10.1103/physreva.98.043414 S. Chen, and Y. Nam, Efficient arbitrary simultaneously
(2018). entangling gates on a trapped-ion quantum computer,
[73] A. Webb, S. Webster, S. Collingbourne, D. Bretaud, Nature communications 11, 10.1038/s41467-020-16790-9
A. Lawrence, S. Weidt, F. Mintert, and W. Hensinger, (2020).
Resilient entangling gates for trapped ions, Physi- [86] P. H. Leung and K. R. Brown, Entangling an arbitrary
cal Review Letters 121, 10.1103/physrevlett.121.180501 pair of qubits in a long ion crystal, Physical Review A
(2018). 98, 10.1103/PhysRevA.98.032318 (2018).
[74] J. Randall, High-fidelity entanglement of trapped ions us- [87] Y. Wang, S. Crain, C. Fang, B. Z. gand S. Huang,
ing long-wavelength radiation, Ph.D. thesis, Imperial Col- Q. Liang, P. H. Leung, K. R. Brown, and J. Kim, High-
lege of London (2016). fidelity two-qubit gates using a microelectromechanical-
[75] C. D. B. Bentley, H. Ball, M. J. Biercuk, A. R. R. Car- system-based beam steering for individual qubit ad-
valho, M. R. hush, and H. J. Slatyer, Numeric optimiza- dressing, Physical Review Letters 125, 10.1103/Phys-
29