Suppl Materials
Suppl Materials
Suppl Materials
p
Let a charged particle have one dimensional wave function ψ(x) ≡ p(x)eiϕ(x) . Consider
a fluctuating electromagnetic field potential V (x, t) ofP
the electromagnetic field which is ap-
proximated by the N -cell stepwise function V (x, t) = N n=1 In (x)V (xn , t), where In (x) is an
1
indicator function of the cell with the index n. Let us assume that during the short time in-
terval a relatively strong non-homogenous fluctuation has emerged and the Pwave packet ψ(x)
´ the coordinate dependent phase shift ψ(x) → ψ̃(x) = ψ(x) exp(i n In (x)φn ), where
acquires
φn = dt eV (xn , t)/~. Consider then the specific fluctuation with φn (x) = −2ϕ(xn ) which
drives the original wave packet ψ(x) into its approximate complex conjugated form ψ̃ ∗ (x). The
accuracy of such a conjugation procedure is defined through the overlap of the exact conjugated
state ψ ∗ (x) with the approximate conjugated state ψ̃ ∗ (x), S = hψ ∗ (x)|ψ̃ ∗ (x)i,
X∞ ˆ
S= dx In (x)p(x) e2i(ϕ(x)−ϕ(xn )) . (1)
n=1
Then the probability of the correct reversion is given by |S|2 . Assuming that the particle density
p(x) changes slowly on the scale of large fluctuations of the particle phase ϕ(x) one arrives at
N
1X 2
|S|2 ≈ 1 − p(xn )δxn ϕ0 (xn )δxn ,
(2)
3 n=1
for the sufficiently small δxn of the cells defined through the condition g(xn ) ≡ ϕ0 (xn )δxn 1.
Then the error probability of the incorrect conjugation of the wavepacket is given by |S|2 =
1 − and in the continuous limit one has
ˆ
1
= dx p(x) g 2 (x). (3)
3
Let us now find the number of cells N needed to approximate the electromagnetic field com-
plex conjugation procedure with a given error probability level . From the definition g(x) =
ϕ0 (x)δx one has ˆ
|ϕ0 (x)|
N = dx . (4)
g(x)
Minimizing the functional N [g(x)] under the constraint Eq. (3) one finds
λ3 (ψ) 1/2 ˆ
1/3
N= , λ(ψ) ≡ dx |ψ(x)|2 [φ0 (x)]2 . (5)
3
Generalization of the above result to a d-dimensional case is straightforward
λ(ψ) d/2 ˆ
, λ(ψ) = dd~x |ψ(~x)|2 ∇ ~ 2 ϕ(~x) d/(d+2) .
Nd = λ(ψ) (6)
3
Applying these results ´ to the wave packet given by the Eq. (1) of the main text, one obtains
2/3
λ(Ψ) ∼ (~τ /m) dx(f (x)x2 )1/3 . We assume that initially at τ = 0 the wave packet has
2
the size L0 , so that f (k) ∼ L0 for |k| ≤ 1/L0 and, therefore, λ(Ψ) ∼ (Lτ /L0 )2/3 where
2
Lτ = ~τ /mL0 is the size of the wave packet after the free evolution during the time τ . There-
fore, the number of the elementary cells needed to arrange the electromagnetic potential fluc-
tuation which reverses the dynamics of a one dimensional wave packet is linear in τ since
N ∼ −1/2 Lτ /L0 , see also (S1). For a d-dimensional wave packet the number of cells grows
polynomially with τ as N ∼ −d/2 (Lτ /L0 )d .
2
SI 2. Reversal of the qubit register dynamics
P −1
Let the forward time dynamics of the n-qubit register state |ψ(t)i = N i=0 ψi (t)|ii be governed
by the Hamiltonian Ĥ, i~∂t |ψ(t)i = Ĥ|ψ(t)i. The time-reversal symmetry of the Schrödinger
equation implies that if there is a forward time solution |ψ(t)i then the backward time solution
|ψ̃(t)i
− i~|ψ̃(t)i = Ĥ|ψ̃(t)i (7)
also exists and is uniquely defined through the forward time solution via the time-reversal oper-
ation R̂ such that |ψ̃(t)i = R̂|ψ(t)i. The time-reversal operation R̂ is an anti-unitary operation:
hR̂ψ1 |R̂ψ2 i = hψ1 |ψ2 i∗ and can be presented as a product R̂ = ÛR K̂ of some unitary operator
ÛR and the complex conjugation operation K̂ which we define with respect to the computational
basis |ii of the qubit register as
X X
K̂ ψi |ii = ψi∗ |ii . (8)
i i
The relation (10) defines the unitary ÛR . Indeed, the hermitian operator Ĥ can be represented
in a form Ĥ = ÛH† Ê ÛH , where Ê is a real diagonal operator and ÛH is unitary. Then it follows
from the Eq. (10)
ÛR = ÛH† ÛH∗ . (11)
The forward time evolution operator Û (τ ) = exp(−iĤτ /~) applied to the time reversed state
|ψ̃(τ )i drives it into the new state
Û (τ )|ψ̃(τ )i = R̂|ψ(0)i. (12)
Indeed,
i i
Û (τ )|ψ̃(τ )i ≡ exp − Ĥτ ÛR K̂ exp − Ĥτ |ψ(0)i (13)
~ ~
i i
∗ †
= exp − Ĥτ exp + ÛR Ĥ ÛR τ R̂|ψ(0)i.
~ ~
Making use of the explicit form of the ÛR operator, see Eq. (11), one has ÛR Ĥ t ÛR† = Ĥ that
proves Eq. (12). Therefore, in order to restore the original state |ψ(0)i from the time-evolved
state |ψ(τ )i one has to apply the following sequence of operations
|ψ(0)i = R̂−1 Û (τ )R̂ |ψ(τ )i . (14)
3
SI 3. Optimal phase shifts arrangement
Here we outline an optimal arrangement of the state selective Q nphase shift operations Φ̂i (ϕ) =
|iihi|eiϕ entering the complex conjugation operation Ûψ = 2i=0−1 Φ̂i (−2ϕi ) for the qubit state
P2n −1 iϕi
|ψi = i=0 |ψi |e |ii. Let us consider 2k−2 operations Φ̂k with index k having the same
values of two highest bits b0 = b1 = 1: k(k 0 ) = 2n−1 + 2n−2 + k 0 , k 0 = 0, . . . , 2n−2 − 1.
Q n−2
Then in the product 2k0 =0 −1 Φ̂k (−2ϕk ) one needs to check the values of the bits b0 and b1 only
once, and this reduces the number of Toffoli gates. This recipe can be recursively repeated for
the next lower bits b2 , b3 and so on, see Fig. 1B. Then the resulting quantum circuit comprises
the sequence of nested blocks or subroutines A11b2 ...bn−1 ⊃ A111b3 ...bn−1 ⊃ · · · ⊃ A1...1bn−1
where each subroutine A1...1bm ...bn−1 performs the controlled phase shift on all components |ki
with first m highest bits equal to 1. As follows from the Fig. 1B, the subroutine A11...bm ...bn−1
involves two subroutines of the next lower level A11...1bm+1 ...bn−1 and A11...0bm+1 ...bn−1 , and two
additional Toffoli gates that are needed tocheck the value of the bit bm+1 . Therefore, the
number of Toffoli gates NΛ2 A11...1bm...bn−1 neededfor the implementation
of the subroutine
A11...1bm ...bn−1 obeys the relation NΛ 2 A11...1bm ...bn−1 = 2 + 2NΛ2 A11...1bm+1 ...bn−1 with the
boundary condition NΛ2 A11...1bn−1 = 2, that gives NΛ2 A11b2 ...bn−1 = 2n − 2. The full n-
(2)
qubit complex conjugation procedure Ûψ involves four different qubit subroutines A00b2 ...bn−1 ,
(2) (2)
A01b2 ...bn−1 and so on. This, finally, yields NΛ2 Ûψ = 4(2n − 2) and hence N⊕ Ûψ =
24(2n − 2) ∼ 24N .
Here we describe the time-reversal procedure of a qubit register based on the arithmetic repre-
sentation of a n-qubit Boolean function,
1, b0 = b1 = · · · = bn−1 = 1
bn−1 ∧ bn−2 ∧ · · · ∧ b0 = . (15)
0, overwise
We find the minimal number of CNOT gates needed for the implementation of this procedure.
Let us start with the two-qubit situation where one wishes to reverse the general two-qubit state
|ψ2 i = eiϕ00 |00i + eiϕ01 |01i + eiϕ10 |10i + eiϕ11 |11i. This requires to implement the complex
conjugation procedure, which for a given state can be realized by the two-qubit unitary operation
X
K̂2 = e−2iϕb1 b0 |b1 b0 ihb1 b0 | ≡ e−2iF̂ (b1 ,b0 ) , (16)
b0 ,b1 =0,1
F̂ (b1 , b0 ) = ϕ00 b̄1 ∧ b̄0 + ϕ10 b1 ∧ b̄0 + ϕ01 b̄1 ∧ b0 + ϕ11 b1 ∧ b0 , (17)
4
and b̄i denotes the logical negation of the bit bi , b̄i = NOT(bi ). Making use of the arithmetic
representation of b1 ∧ b0 , see Eq. (4) of the main text, one finds
ϕ01 +ϕ11 ϕ10 +ϕ00 ϕ10 +ϕ11 ϕ01 +ϕ10
F̂ (b1 , b0 ) = b0 + b̄0 + b1 + b̄1
2 2 2 2
ϕ00 + ϕ11 ϕ10 + ϕ01
− b1 ⊕ b0 − b1 ⊕ b0 , (18)
2 2
where b1 ⊕ b0 denotes a bit summation by modulo 2,
0, b0 = b1
b1 ⊕ b0 = . (19)
1, b0 6= b1
The first four terms in the Eq. (18) correspond to the one-qubit state dependent phase shifts
and can be realized only via the single-qubit gates
1 0 0 1
T̂ (α) = , X̂ = , (20)
0 eiα 1 0
available on the public IBM quantum computer. The last two-qubit terms in Eq. (18) will
require
two-qubit CNOT gates. The overall quantum circuit which realizes the unitary operation
exp −2iF̂ (b1 , b0 ) is described by the following sequence of unitary operations
h i
exp −2iF̂ (b1 , b0 ) = CNOT0,1 · TXTX1 ϕ00 +ϕ11 , ϕ10 +ϕ01 ⊗ 10 · CNOT0,1 (21)
h i
· TXTX1 −ϕ10 −ϕ11 , −ϕ00 −ϕ01 ⊗ TXTX0 −ϕ01 −ϕ11 , −ϕ10 −ϕ00 ,
where TXTXi (ϕ, ϕ̄) ≡ T̂i (ϕ)X̂i T̂i (ϕ̄)X̂i is a single-qubit unitary operation which adds spec-
ified phase shifts to the state components of the ith qubit: TXTXi (ϕ, ϕ̄) a|0i i + b|1i i =
aeiϕ̄ |0i i + beiϕ |1i i. The corresponding quantum circuit is shown in the Fig. 1D and involves
only two CNOT0,1 gates, where |b0 i qubit serves as control bit and |b1 i as a target.
The above two-qubit complex conjugation procedure can be further extended onto a general
n-qubit state. As follows from the Eq. (4) of the maintext, the quantum circuit performing
complex conjugation of a given n-qubit state requires n2 two-qubit operations,
T̂i2 (ϕ), bi1 ⊕ bi2 = 1
CTXTXi1 i2 (ϕ, ϕ̄) ≡ , 1 ≤ i1 < i2 ≤ n, (22)
T̂i2 (ϕ̄), bi1 ⊕ bi2 = 0
n
3
three-qubit operations,
T̂i3 (ϕ), bi1 ⊕ bi2 ⊕ bi3 = 1
CTXTXi1 i2 i3 (ϕ, ϕ̄) ≡ , 1 ≤ i1 < i2 ≤ i3 ≤ n, (23)
T̂i3 (ϕ̄), bi1 ⊕ bi2 ⊕ bi3 = 0
and so on. The general n-qubit operation CTXTXi1 ...in (ϕ, ϕ̄), 1 ≤ i1 < i2 < · · · < in is
implemented with the help of 2(n − 1) CNOT gates as shown on the Fig. S1. Therefore, one
5
11
00
0
1
111111111111111111
000000000000000000 11
00
0
1
0
1
111
000
00
11 0
1
000
111
00
11
0
1
000
111
00
11 00
11 00
11
111111111111111111
000000000000000000
00
11 0
1 0
1 0
1
000
111
00
11
00
11
000
111
00
11
00
1100
110
1 00
11000
111
0
100
11
00
11
000 1
111 0 000
111
0
1
00
110
1
00
11 0
1
000
111
0
1
00
11
11
00
00
11 11
00 11
00
0
1 000
111
1
0
111111111111111111
000000000000000000
0
1 0
1
00
110
1
000
111 0
1
000
111000
111
00
11
0
1
000
111
00
11
1111111
0000000 00
11
0
1
000
111
00
11
1111111
0000000
00
11
000
111
00
11 TXTX 00
11
000
111
00
11
00
11
000
111 00
11
000
111
Figure S 1: The quantum circuit which implements the four-qubit quantum gate CTXTX0123 .
One can check that for any computational basis state |b3 b2 b1 b0 i the state of the elder bit b3 is
given by b0 ⊕ b1 ⊕ b2 ⊕ b3 right after the first ladder CNOT gates. The remaining symmetric half
of CNOT gates is required in order to restore the original quantum state of the qubit register.
6
Making an optimization procedure one can group these operations as
CTXTX01 ·CTXTX012 ·CTXTX0123 · CTXTX02 ·CTXTX023 ·CTXTX013 ·CTXTX03 , (25)
and hence
N⊕ K̂4 ] = N⊕ CTXTX0123 + N⊕ CTXTX023 + N⊕ CTXTX013 (26)
+N⊕ CTXTX03 + N⊕ K̂3 ,
where N⊕ CTXTXs is the number of CNOT gates needed for the operation CTXTXs . One
can note, that only generalized operations CTXTXs with the inputs strings s = i1 . . . ik where
first and last indices are equal to 0 and 3, respectively are counted for the total number of the
CNOT gates. Therefore, for a general case, the following relation holds
X
N⊕ K̂n = N⊕ K̂n−1 + N⊕ CTXTX1n + N⊕ CTXTX1k1 n (27)
1<k1 <n
X
+ N⊕ CTXTX1k1 k2 n + · · · + N⊕ CTXTX1...n
1<k1 <k2 <n
n−2
X n−2
= N⊕ K̂n−1 + n2n−2 ,
= N⊕ K̂n−1 + 2(k + 1) (28)
k=0
k
and, therefore,
N⊕ K̂n = (n − 1)2n−1 ,
n > 1. (29)
Here we discuss a spinless particle which scatters on a two-level impurity (TLI). The free dy-
namics of the TLI is governed by a Hamiltonian
Ĥi = ~ω cos(α) σ̂z + sin(α) σ̂x . (30)
The scattering process is described by the 2 × 2 scattering matrix Ŝi , i = 0, 1 whose form de-
pends on the impurity state.
PThe quantum state of the particle-impurity system can be described
as the two-bit state |ψi = b0 ,b1 =0,1 Ab1 b0 |b1 i ⊗ |b0 i where the first qubit describes the TLI and
the second one describes the propagation direction of an incoming/scattered particle. Let the
system start in the state |ψ(0)i = |0i ⊗ |Li with the particle coming from the left. Let after
the time τ > 0 the particle be scattered on the TLI. The resulting
state |ψ(τ )i is generated by
the sequence of unitary operations |ψ(τ )i = Ŝψ · Ûi (τ ) ⊗ 1 |ψ(0)i, where the unitary operator
Ûi (τ ) ≡ exp −iĤi τ /~ describes the free evolution of TLI and
7
describes the state dependent scattering process
of the incoming particle. The unitary operator
Ûi (τ ) = exp −iωτ (σ̂x cos α + σ̂z sin α) is symmetric. In the absence of the magnetic field,
the scattering operator Ŝψ is symmetric as well. Let the state freely evolve after the scattering at
the t = τ during the same time period τ . Then the resulting state |ψ(2τ )i = [Ûi (τ ) ⊗ 1]|ψ(τ )i
can be generated from the initial state |ψ(0)i by the symmetric 2-qubit unitary operator
Û2bit = Ûi (τ ) ⊗ 1 · Ŝψ · Ûi (τ ) ⊗ 1 . (32)
Therefore, as we have already discussed in SI 2, the time reversal procedure of the 2-qubit
state |ψ(2τ )i requires only the unitary implementation of the complex conjugation operation
|ψ(2τ )i → |ψ ∗ (2τ )i.
Our goal is to implement the unitary operation Û2bit via the set of quantum gates available
on the IBM public quantum computer. The only available two-qubit gate is the CNOTbc ,bt gate,
where bc the qubit serves as a control and bt qubit serves a target. Among the standard 1-qubit
gates we will need two available generalized 1-qubit gates: the relative phase shift gate T̂ (α),
introduced in the SI 4 and the full 1-qubit unitary rotation
where
cos 2θ − sin 2θ
R̂(θ) = . (34)
sin 2θ cos 2θ
Any 2 × 2 unitary matrix Û can be represented in the form (33) up to some phase factor:
Û = eiδ Û3 (θ, α, β). In particular, any symmetric 2 × 2 unitary matrix Û = Û t has the form
eiδ Û3 (θ, α, α+π). Therefore, a given set of matrices Ûi (τ ), Ŝ0 and Ŝ1 entering into the definition
of the model can be presented as
The phase exponent eiφ gives only a trivial common phase factor for the system state and will
be omitted in what follows. Without any loss of generality we assume δ0 = 0 as well.
Next, let us construct the 2-qubit operation Ŝψ using as less CNOT gates as possible. It turns
out that Ŝψ can be constructed with the help of only two CNOT gates. Indeed,
Ŝψ = |1ih1| ⊗ Ŝ1 Ŝ0† + |0ih0| ⊗ 1 · 1 ⊗ Ŝ0 ≡ Λb1 ,b0 (Ŝ1 Ŝ0† ) · 1 ⊗ Ŝ0 ,
(37)
8
The unitary matrix Ŵ = Ŝ1 Ŝ0† ≡ eiδ Û3 (θ, α, β) can be represented as,
α+β θ θ α+β β−α
Ŵ = ei(δ+ 2
)
T̂ (α)R̂( ) σ̂x R̂(− )T̂ (− ) σ̂x T̂ ( ). (39)
2 2 2 2
The advantage of the latter representation is that if one replaces in the Eq. (39) two Pauli matri-
ces σ̂x by the identity operator, one gets a phase shift exp[i(δ + α+β
2
)] only. Therefore,
α+β θ
Λb1 ,b0 (Ŵ ) = T̂ (δ + ) ⊗ Û3 ( , α, 0) (40)
2 2
θ α+β β −α
·CNOTb1 ,b0 · 1 ⊗ Û3 (− , 0, − ) · CNOTb1 ,b0 · 1 ⊗ T̂ ( ) ,
2 2 2
and the whole evolution operator, see Eq. (32) can be presented as,
Û2bit = Û3 (ξ, η, η + π) ⊗ 1 · Λb1 ,b0 (Sˆ1 Ŝ0† ) · Û3 (ξ, η, η + π) ⊗ Û3 (θ0 , ϕ0 , ϕ0 + π) .
(41)
where the first (eldest) bit describes the state of the TLI and the second and third qubits describe
the scattering state of the first and second particles correspondingly. The quantum circuit which
implements the evolution operator Û3bit is shown in the Fig. 1E of the main text.
SI 6. Time-reversal experiment
In the simulation experiment we choose fixed scattering matrices of the two-level impurity
(TLI),
" √ # " √ #
1 3 3 1 iπ/3
e
Ŝ0 = √23 2 1 , Ŝ1 = 1 iπ/3 2 2
√
3 2πi/3
, (43)
2
− 2 2
e − 2
e
for the |0i and |1i impurity states correspondingly. Then the state dependent scattering operator
Ŝψ , see Eqs. (37) and (40), is given by the following sequence of quantum gates,
Ŝψ ≈ 1 ⊗ Û3 (0.723, −1.27) · CNOT1,2 · T̂ (1.047) ⊗ Û3 (−0.723, 0, −0.523) (44)
2π
·CNOT1,2 1 ⊗ T̂ (1.761) · 1 ⊗ Û3 , 0, π ,
3
9
where the first (control) qubit describes a state of TLI and the second (target) qubit describes a
scattering state of the particle, Û3 (α, ϕ, λ) and T̂ (ϕ) are generalized one-qubit gates available
on the IBM quantum computer.
The free evolution operator Ûi (τ ) = exp(−iĤi τ /~) with Ĥi = ~ω(cos α σ̂z + sin ασ̂x ) is
parameterized by two parameters ωτ and α. The unitary operator Ûi (τ ) is symmetric and for a
fixed values of ωτ and α can be presented in the form,
Ûi (τ ) = eiδ Û3 (ξ, η, η + π), ξ = ξ(ωτ, α), η = η(ωτ, α), (45)
where eiδ some phase factor which changes only an overall phase of the qubit register; ξ and η
are parameters which uniquely defined by ωτ and α. In the following we choose ωτ = π/6 and
vary the parameter α among four values π/6, π/4, π/3 and π/2 with the corresponding gate
parameters,
π π π π
ξ , ≈ 0.505, η , ≈ −1.107, (46)
6 6 6 6
π π π π
ξ , ≈ 0.723, η , ≈ −1.183,
6 4 6 4
π π π π
ξ , ≈ 0.896, η , ≈ −1.290,
6 3 6 3
π π π π
ξ , ≈ 1.047, η , ≈ −π/2.
6 2 6 2
The occurrence rates of the computational basis states for 2-qubit and 3-qubit experiments
are shown in Tables 1 and 2 for the different input parameters of the model. The 2-qubit ex-
periment used q1 and q2 qubit lines of the ‘ibmqx4’ five qubit quantum processor. The 3-qubit
experiment used in addition a q0 qubit line. In both experiments the q2 qubit line has modeled
a state of TLI. The calibration state of the quantum computer was the same for all experiments.
The qubit’s relaxation times T1 , coherence times T2 , readout errors r and one-qubit gate errors
1 for each qubit line are shown in the Table. 3. The errors of the CNOT gates CNOTq2,q0 ,
CNOTq2,q1 and CNOTq1,q0 used in the experiments are g20 = 1.91%, g21 = 2.68% and
g10 = 1.70% respectively. These processor’s state parameters allows us to estimate a theoret-
ical value of a time-reversal fidelity F = |h0 . . . 0|ψ̃0 i|2 , where |ψ̃0 i is a final state of the qubit
register. For the used gate arrangement one has,
theor
F2bit = (1 − g21 )6 (1 − r1 )(1 − r2 ) ≈ 79.6% (47)
theor 6 6 4
F3bit = (1 − g21 ) (1 − g20 ) (1 − g10 ) (1 − r0 )(1 − r1 )(1 − r2 ) ≈ 63.4%, (48)
while the experimentally observed values of the time-reversal fidelity are shown in Tables. 1
and 2.
10
Table 1:
ωτ α |00i |10i |01i |11i F
π/6 π/6 6949 437 562 244 84.8 ± 0.4%
π/6 π/4 6916 440 576 260 84.4 ± 0.4%
π/6 π/3 6983 370 560 279 85.2 ± 0.4%
π/6 π/2 6950 338 551 353 84.8 ± 0.4%
Table 2:
ωτ α |000i |001i |010i |011i |100i |101i |110i |111i F
π/6 π/6 3909 1380 1069 487 482 309 332 224 47.7 ± 0.5%
π/6 π/4 3934 1157 981 380 618 360 407 355 48.0 ± 0.5%
π/6 π/3 3957 832 884 327 859 359 531 443 48.3 ± 0.5%
π/6 π/2 3879 355 1050 425 964 418 630 471 47.3 ± 0.5%
Supplementary references
S1 Lesovik, G. B. On the law of increasing entropy and the cause of the dynamics irreversibil-
ity of quantum systems. JETP Lett. 98, 184–189 (2013).
11
Table 3:
qn T1 (µs) T2 (µs) r (%) 1 (%)
q0 52.4 47.3 4.2 0.077
q1 58.0 40.6 3.6 0.103
q2 46.9 47.4 2.8 0.137
12