Arxiv:2104.07581V1 [Cond-Mat.Mes-Hall] 15 Apr 2021

Interplay of charge noise and coupling to phonons in adiabatic electron transfer between quantum dots

Jan A. Krzywda∗ andLukasz Cywi´nski Institute of Physics, Polish Academy of Sciences, al. Lotnik´ow32/46, PL 02-668 Warsaw, Poland

Long-distance transfer of quantum information in architectures based on quantum dot spin qubits will be necessary for their scalability. One way of achieving it is to simply move the electron between two quantum registers. Precise control over the electron shuttling through a chain of tunnel-coupled quantum dots is possible when interdot energy detunings are changed adiabatically. Deterministic character of shuttling is however endangered by coupling of the transferred electron to thermal reservoirs: sources of fluctuations of electric fields, and lattice vibrations. We theoretically analyse how the electron transfer between two quantum dots is affected by electron-phonon scattering, and interaction with sources of 1/f and Johnson charge noise in both detuning and tunnel coupling. The electron-phonon scattering turns out to be irrelevant in Si quantum dots, while a competition between the effects of charge noise and Landau-Zener effect leads to an existence of optimal detuning sweep rate, at which probability of leaving the electron behind is minimal. In GaAs quantum dots, on the other hand, coupling to phonons is strong enough to make the phonon-assisted processes of interdot transfer dominate over influence of charge noise. The probability of leaving the electron behind depends then monotonically on detuning sweep rate, and values much smaller than in silicon can be obtained for slow sweeps. However, after taking into account limitations on transfer time imposed by need for preservation of electron’s spin coherence, minimal probabilities of leaving the electron behind in both GaAs- and Si-based double quantum dots turn out to be of the same order of 3 magnitude. Bringing them down below 10− requires temperatures 100 mK and tunnel couplings above 20 µeV. ≤

I. INTRODUCTION Hilbert space spanned by |L(R)i states, corresponding to electron localized in a local ground states of energy EL(R) In quantum computing architectures based on in left (right) dot: voltage-controlled quantum dots (QDs), developed in tc GaAs/AlGaAs [1–3], Si/SiGe [4, 5], and silicon MOS [6– Hˆ = σˆz + σˆx , (1) 9] structures, scalability will be possible only if quan- 2 2 tum information is transferred between few-qubit regis- where = EL − ER is the so-called interdot detuning ters, separated by distances much larger than the typi- of energy, tc is the tunnel coupling between the QDs, cal QD size. This is caused by short-distance character σˆz = |LihL| − |RihR|, andσ ˆx = |LihR| + |RihL|. For of exchange interaction needed for two-qubit gates, and −tc the lowest energy state is localized in the L dot, spatial extent of wiring needed for controlled application and this is the state that we take as an initial one in all of voltages to the gates defining the dots, which together the considerations below. For tc the lowest-energy put limits on density of a qubit array [10]. Coupling of state is localized in the R dot, and one of course expects electron spins to microwave photons is a possible mean of that for very slow change of from negative to positive coherent coupling of spin qubits in GaAs [11] and silicon values, the evolution will be adiabatic and the system will [12–15]. A conceptually simpler alternative, which has end up in this state. For a linear sweep, ∝ vτ, where been recently pursued in experiments [16–28], is to sim- v is the rate of change of detuning, and constant tc, we ply transfer an electron spin qubit over a large (at least are dealing with classical Landau-Zener model [29], for a few micrometer) distance. which the probability of having the electron in an exited We focus here on electron transfer along a chain of state for (τ) → ∞ (i.e. leaving the electron behind in tunnel-coupled QDs [22–28]. The shuttling is then caused the L dot) is given by by controlled tilting of energy levels of neighboring QDs that makes an electron move from one dot to the other. 2 arXiv:2104.07581v1 [cond-mat.mes-hall] 15 Apr 2021 QLZ = exp −πtc /2v , (2) The basic step in such a process is single electron transfer between two tunnel-coupled QDs. In a simplified situa- so that a near-perfect adiabatic transfer occurs when tion, in which we neglect spin and valley (in case of Si) 2 tc /v 1, i.e. when the sweep rate is low. degrees of freedom of the electron, the basic physics is Changing the interdot detunings slowly is thus an ob- captured by the Hamiltonian acting in a two-dimensional vious way to perform an on-demand deterministic transfer of an electron spin qubit. Of course, the total shuttling time should be much shorter than the spin coherence time of a moving electron, and according to Eq. (2) [email protected] ∗ this requirement will put a lower bound on values of tc 2 characterizing the chain of QDs. However, another is- [67, 68]), which we solve in a way analogous to the one sue needs to be addressed before we can claim to have a described in [53]. realistic estimate of sweep rate v giving the smallest pos- During the detuning sweep, the energy gap between sible probability of error Q in transfer between a pair of eigenstates of instantaneous Hamiltonian varies between dots. Electrons are affected by charge noise unavoidable tc ∼ 10 µeV and largest value of ∼ 1 meV. With tem- in semiconductor nanostructures, and coupled to lattice peratures in experiments typically around 100 mK, cor- vibrations. As we show in this paper, interactions with responding to thermal energy of ≈ 10 µeV, we should sources of electric field noise and phonons in realistic Si- expect a nontrivial role of temperature dependence of and GaAs-based structures are dominating the physics rates of energy absorption and emission by the reservoirs. of charge transfer in a wide range of sweep rates, with Note that in our previous work [39] we have focused on nonadiabatic effects described by Landau-Zener theory influence of classical (i.e. high-temperature) 1/f charge being relevant only for very fast sweeps. noise on electron transfer. Here we address the situa- Since our focus here is on open system character of tion of lower temperatures/larger tunnel couplings, tak- an electron tunneling between two quantum dots, we use ing into account the quantum limit [69] of both 1/f noise the above-described simplest possible two-level model of from two-level fluctuators present in the nanostructure, the closed system. Taking into account the spin degree and Johnson noise from reservoirs of free electrons, while of freedom and spin-orbit cupling that affects its dynam- furthermore considering the coupling of the moving elec- ics during the electron motion in GaAs [30] (and to a tron to phonons. Coupling to all these thermal reservoirs smaller extent in silicon [30, 31]), and then a valley de- gives transition rates, Γ+/−(Ω) for transfer of energy Ω gree of freedom in Si [32–34], leads to 4- or 8-level mod- from/to the environment, that nontrivially depend on els with multiple anticrossings of states [30, 31, 35–41]. Ω. The detailed balance between them, which reads −βΩ The two-level model used here exhibits a simpler behav- Γ+/Γ− =e , has the following general consequence for ior in closed system case, and using it will typically lead the dynamics of the system. With the system initially to an underestimation of unwanted effects due to not- in ground state, transitions into an excited are exponen- slow-enough sweeps (for a closed system), and coupling tially suppressed for large negative detunings, and they to environment (for an open system). The results given start to become increasingly efficient as we approach the in this paper consequently correspond to the best-case anticrossing of levels, at which the gap is minimal and scenario for given tc and assumed magnitudes of charge equal to tc. This effect of enhancement of excitation rate noise and temperature. at the anticrossing is additionally strengthened in the The physics of Landau-Zener effect in presence of cou- considered system by the fact that an electron delocal- pling to environment has obviously been a subject of mul- ized between the two dots is more susceptible to both tiple works. Dissipative adiabatic evolution affected by charge noise and interaction with phonons (as the transi- coupling to bosonic baths having Ohmic spectrum was tions between states localized in each dot that govern the most often considered [42–47]. It is known [48, 49] that dynamics in far-detuned regimes are suppressed by small coupling to zero-temperature bath suppresses the final overlap of wavefunctions). The finite occupation of the occupation of the higher-energy state (“the electron be- “wrong” dot generated during passing through || . tc ing left behind in the initial dot” in the physical scenario region can then be diminished (“healed” in the termi- of interest here), while at finite temperature this occupa- nology used below) by processes of energy emission into tion can be enhanced [43, 50–53]. Such effects of coupling the reservoirs that dominate over processes of energy ab- to low-temperature reservoirs were discussed in many sorption by them when kBT . Arriving at the final re- physical contexts [42, 54, 55]. Stochastic modifications sult of interplay between environment-induced excitation of LZ parameters were also considered [56–58], includ- near the anticrossing, and the subsequent energy relax- ing fast classical fluctuations [59] and noise characterized ation (the environment-assisted dissipative tunneling into by non-trivial spectral density [39, 60–63], including 1/f the “correct” final state), requires consideration of real- type noise, the tail of which also resulted in incoherent istic coupling to all the reservoirs at temperatures and transitions between the states [39, 64, 65]. In this paper sweep rates relevant for experiments in quantum dots. we focus on quantum dots based on silicon and GaAs, Such a careful consideration is the goal of this paper. and employ realistic models of charge noise (having both Our key qualitative result concerning application to re- Johnson/Ohmic and 1/f type spectra, and coupling to alistic quantum dots, is that in Si-based structures (both both and tc), and phonon interaction with an electron Si/SiGe and SiMOS) the dominant process disturbing the confined in a double quantum dot. We use the Adiabatic adiabatic evolution close to anticrossing of levels is due to Master Equation [42, 53, 66], in which the influence of the charge noise (with coupling to phonons giving transition environment (actually a few distinct reservoirs in the case rates 3 order of magnitude smaller than those estimated discussed here) is modeled with energy-dependent rates for charge noise), and the finite probability of leaving the of transitions between instantaneous eigenstates of the electron behind is subsequently diminished by relaxation slowly changing Hamiltonian of the system. For negligi- processes due to charge noise and phonons that occur at ble probability of coherent Landau-Zener excitation, this large detunings only when the transfer is very slow. On approached reduces to a simple differential rate equation the other hand, in GaAs/AlGaAs structures the piezo- 3 electric coupling to phonons dominates over coupling to For any value of (τ) and tunnel coupling tc(τ), the charge noise over a wide range of detunings, and conse- Hamiltonian Hˆ (τ) has eigenstates quently the processes involving energy exchange between the transferred electron and lattice vibrations dominate |+, θ(τ)i = cos[θ(τ)/2] |Ri + sin[θ(τ)/2] |Li the physics of the problem. The longer the charge trans- |−, θ(τ)i = cos[θ(τ)/2] |Li − sin[θ(τ)/2] |Ri , (3) fer takes, the more time the system spends in far-detuned regime in which the energy gap exceeds thermal energy, where θ(τ) = arccot(−(τ)/|tc(τ)|). The discussion of and the closer it gets to a thermalized state characterized nonadiabatic effects due to time-dependence of and tc, by small occupation of higher-energy level, i.e. of the elec- or effects of interaction with the environment, is most tron being in the wrong dot. Phonons thus help in main- transparent if we transform the state of the system into taining a deterministic character of the charge transfer. an “adiabatic frame” [71]: instead of working with |ψ(τ)i These conclusions are quite robust against modifications which fulfills i∂τ |ψ(τ)i = Hˆ (τ) |ψ(τ)i we work with of parameters of high-frequency properties of Johnson |ψ˜(τ)i ≡ Uˆ[θ(τ)] |ψ(τ)i, where a time-dependent unitary and 1/f type charge noises considered here. operator The article is organized in the following way, in Sec. II we set up the problem for the closed system and dis- Uˆ[θ(τ)] = exp[iσˆyθ(τ)/2] . (4) cuss the adiabatic condition for its dynamics, introduce the Adiabatic Master Equation as an approach to open transforms the R/L states into the instantaneous eigen- system dynamics, and discuss a few physically transpar- states of Hˆ (τ): |+(−), θ(τ)i = Uˆ[θ(τ)] |R(L)i. One can ent (and, as we show later, relevant for the case of elec- see that for a perfectly adiabatic evolution of the sys- tron transfer in silicon- and GaAs-based quantum dots) tem, for which an initial superposition of eigenstates of approximate solutions of this equation. In Section III Hˆ (τi) at given τi evolves into the same superposition of we calculate the detuning-dependent transition rates be- eigenstates of Hˆ (τf ) at the final time τf , the transformed tween instantaneous eigenstates of the two-level Hamilto- |ψ˜(τ)i state is time-independent. Indeed, the evolution in nian. We perform calculations for coupling to phonons, the adiabatic frame is controlled by and finite-temperature environments that cause charge noise of both Johnson and 1/f type in detuning and ∂Uˆ †[θ(τ)] H(τ) = Uˆ[θ(τ)]Hˆ (τ)Uˆ †[θ(τ)] − iUˆ[θ(τ)] , tunnel coupling. We give there a discussion of expected ∂τ amplitude of 1/f noise at GHz frequencies relevant for ! transitions during electron transfer in realistic GaAs- and (5) silicon-based quantum dots. Finally, in Section IV we which for the system discussed here reads use these rates to calculate the dynamics of the electron Ω(τ) θ˙(τ) driven adiabatically through an anticrossing of levels as- Hˆ(τ) = ςˆz − ςˆy, (6) sociated with the two dots, and show a qualitative differ- 2 2 ence between resulting probability of “leaving the elec- whereς ˆz,ς ˆy are Pauli operators in |+, θ(τ)i , |−, θ(τ)i tron behind” between GaAs- and silicon-based quantum basis of instantaneous eigenstates of the time-dependent dots. In the last Section we discuss some of the impli- Hamiltonian Hˆ (τ), θ˙ = dθ(τ)/dτ, and the instantaneous cations of these results for experimental efforts aimed at energy splitting is using chains of quantum dots for coherent shuttling of 2 2 electron spin qubits. Ω(τ) = (τ) + tc (τ) . (7) We assume the electronp is initialized in the ground II. MODEL OF SYSTEM’S DYNAMICS state at large negative detuning (−τ∞) tc, such that the initial state |ψ(−τ∞)i = |−, θ(−τ∞)i ≈ |Li. Due to A. Adiabatic condition for closed system non-negligible coupling between the adiabatic states during the system’s evolution (i.e. a nonzero θ˙(τ) term in Eq. (6)), a non-zero occupation of excited state |+, θ(τ)i We consider two energy levels that in the double quan- can be generated. When the detuning sweep terminates tum dot case correspond to the lowest-energy orbital at large (τ ) t , the occupation of excited state de- | i | i ∞ c states localized in each of the two dots, L and R . In fines the transfer error, i.e. the probability of the electron case of silicon QDs we assume that the valley splitting being left behind in the L dot: is large enough for us to consider a single anticrossing 2 2 of two lowest-energy valley-orbital levels. We also ne- Q = | hψ(τ∞)|+, θ(τ∞)i | ≈ | hψ(τ∞)|Li | . (8) glect the spin degree of freedom - interplay between the nonadiabatic effects in charge transfer and dynamics of The calculation of Q for an electron coupled to environ- the spin of the transferred electron will be discussed else- ments relevant for semiconductor-based gated quantum where [70]. We therefore work with the model defined by dots is the main goal of this paper. Hamiltonian from Eq. (1), in which we now assume that For constant tunnel coupling tc, and for (τ) = vτ and tc depend on time τ. we are dealing with the well-known Landau-Zener model 4

[72], in which Q is given by QLZ from Eq. (2). We concen- where trate here on the adiabatic regime, defined by QLZ 1, τ∞ 2 0 which implies tc /v 1, and means that the ratio of χ(τ, τ∞) = (Γ+ + Γ−) dτ . (14) “transverse” and “longitudinal” terms in the effective Zτ Hamiltonian, Eq. (6), fulfills

θ˙ vtc v C. Approximate solutions = 3 < 2 1 . (9) Ω Ω tc This is the adiabatic condition for the dynamics of closed Let us now discuss a few physically motivated approx- and noise-free system. When it is fulﬁlled during detun- imate solutions for the probability of ending up in the ing sweep, the electron remains at all times in the ground excited state at the end of the sweep Q, i.e. the prob- state |−, θ(τ)i, which means it physically moves from the ability that the electron remains in the initial dot. We start with a simplest perturbative approach to rate equa- state initially localized in the left dot |−, θ(−τ∞)i=|Li, tion (12), assuming Γ±τ∞ 1. In the lowest order one to a ﬁnal state |−, θ(τ∞))i=|Ri, located in the right dot. can write:

τ∞ (1) B. Dynamics of an open system Q = Γ+(τ)dτ, (15) Z−τ∞ We use here Adiabatic Master Equation (AME) ap- As the energy needed for transition from ground to ex- proach [51, 53, 55, 66], in which transitions caused by cited state comes from thermal ﬂuctuations of environ- the environment occur between the instantaneous eigen- ment, the excitation rate Γ is strongly suppressed at low ˆ + states of H(τ), which are given by Eqs. (3). Our focus on temperatures, when kBT tc ≤ Ω(τ). At these tempera- 2 the adiabatic regime (tc v), combined with relatively tures the rate of energy relaxation into the environment, weak coupling to charge noise (with noise RMS σ tc) Γ−(Ω), is temperature-independent, as the thermal occu- and short intrinsic correlation time of phonon bath allow pation factor for environmental states of energy ΩkBT us to use here a Lindbladian form of AME [42, 43], which is zero, and Γ−(Ω) depends then only on density of en- reads: vironmental states and coupling matrix elements. For

∂τ %Â = i[Hˆ(τ), %Â] + Γ+(τ)D[ˆς+]ˆ%A + Γ−(τ)D[ˆς−]ˆ%A, all the environments considered in this paper, these de- (10) pendencies lead to a power-law behavior of the rates, a ˆ ˆ † Γ−(Ω) ∝ Ω with a ∈ [−3, 3] depending on the transition where% Â = Uy(θ)ˆ%LRUy (θ) and %LR is the density matrix of the system before switching the description to the mechanism and range of Ω, see derivations in the next Section. As we assume the environment to be in thermal “adiabatic” frame,ς ˆ+ = |+, θih−, θ|,ς ˆ− = |−, θih+, θ|, and equlibrium, the detailed balance condition, which reads Γ (Ω) = Γ (Ω)e−βΩ, leads to Γ (Ω) ∝ Ωae−βΩ with 1 + − + D[ô]ˆ% =o ˆ%oˆ † − oˆ†o,ˆ %ˆ (11) e−βΩ 1 at low temperatures. 2 The excitation process takes then place in a narrow is the Linbladian associated with operatoro ˆ and time- range of detunings around the avoided crossing, as Γ+(Ω) dependent relaxation/excitation rate Γ±(τ). In this ap- very quickly decreases when || increases. As Ω(τ) ≈ 2 2 proach these rates depend on time though their depen- βtc + βv τ /2tc for tc, we neglect in this regime the dence on the value of instantaneous energy splitting Ω(τ) dependence of Γ+(Ω) and replace it with value for Ω=tc from Eq. (7), i.e. Γ±(τ) = Γ±[Ω(τ)]. Below we will use (equivalently: for τ =0), while we keep it in the thermal both notations, Γ±(τ) and Γ±[Ω(τ)], depending on con- factor. The integrand in (15) can then be approximated −βΩ(τ) text. In particular, if noise-induced excitations dominate as Γ+(τ) ≈ Γ−(0)e , and the integration can be over the Landau-Zener effect due to deterministic time- done over a range of || tc. In this way we obtain the ˆ dependence of H(τ), i.e. Qnoise QLZ, the unitary evo- Single Excitation Approximation Limit (SEAL): lution can be safely neglected and Eq. (10) reduces to a ∞ βv2t2 simple rate equation −βtc − QSEAL = Γ−(0)e e 2tc dτ , ˙ Q(τ) ≈ Γ+(τ) − Q(τ) Γ−(τ) + Γ+(τ) , (12) √ Z−∞

2πkBT tc −βtc where Q(τ) = h+, θ| %Â(τ) |+, θi denotes occupation of = Γ−(0)e , (16) the higher energy state |+, θi at time τ. v Given the initial condition Q(−τ∞) = 0, the solution which assumes that at most a single quantum jump from to Eq. (12) reads ground to excited state takes place in the avoided crossing τ∞ R τ∞ 0 0 0 region. − Γ+(τ )+Γ−(τ )dτ Q = dτ Γ+(τ)e τ = The SEAL approximation does not take into account Z−τ∞ τ∞ possibility of electron transition in the opposite direction, −χ(τ,τ∞) = dτΓ+(τ)e . (13) i.e. from excited to ground state, which would lead to Z−τ∞ partial recovery of ground state occupation - an effect 5 that we will refer to as a “healing” of excitation that occurred close to the anticrossing. This effect is captured by the exp[−χ(τ)] factor in Eq. (13) with χ(τ), given in Eq. (14)), evaluated in the low-temperature limit of Γ− Γ+. The effect of transitions occuring during the part of the sweep when (τ)>tc is captured by a Healed Excitation Approximation Limit (HEAL):

τ∞ QHEAL ≈ QSEAL exp − Γ−(τ) dτ . (17) Z0 The physical picture expected to hold at low T is thus the following. A finite Q is generated due to coupling to a thermal reservoir near the anticrossing, and then processes of emission of energy into this reservoir lead to a diminishing of its final value at the end of the sweep, making the final state of the system closer to the one following from an ideal adiabtic evolution. Such a healing process results in environment-assisted inelastic tunneling into ground state at the end of the driving, see Fig. 1. In Sec. IV we will demonstrate in which regimes of parameters the SEAL/HEAL solutions are applicable for realistic DQD devices. Note that up to this moment we have not specified FIG. 1. A schematic picture of adiabatic transition between any particular form of relaxation/excitation rates Γ±(τ), two quantum dots in presence of charge noise and phonon which makes above approximations suitable also for other bath. In panel (a) we show energy of instantaneous states , θ (red/blue lines) as a function of detuning (lower axis) systems described in terms of the L-Z Hamiltonian (1), in |± i 2 and orbital angle θ = acot( /tc) (upper axis). Despite the adiabatic limit (tc v) and coupled to environment − 2 detuning sweep being adiabatic (tc v), the electron ini- at relatively low-temperature (tc & kBT ). tialized in the left dot (blue circle) can still end up with non-zero occupation of excited state localized in right dot Q (red circle), as a result of coupling to environment, which III. TRANSITION RATES FOR AN leads to incoherent transitions between eigenstates of the in- ADIABATICALLY TRANSFERRED ELECTRON stantaneous Hamiltonian characterized by the rates Γ (Ω). ± At low temperatures, the excitation rate is non-negligible A. General properties only in vicinity of avoided crossing, where the gap is smallest, Ω(0) = tc, while relaxation accounts for recovery of We consider now a transfer of an electron between two ground state occupation (the “healing” of the excitation at larger detuning. As the detuning is changed, the charac- quantum dots that is driven by a detuning sweep slow ter of eigenstates of the instantaneous Hamiltonia, , θ , enough to be adiabatic in the closed system limit. Af- |± i is modified from dot-like character at tc to orbital- | | ter turning on a weak coupling to an environment, the molecular-like at tc, which is accompanied by dominant | | transition rates Γ±(τ) in the Adiabatic Master Equa- role of transverse (interdot) and longitudinal transitions, re- tion (AME) from Eq. (10) are evaluated at given τ as spectively, see Sec. III A. To illustrate the difference between 2 2 if the system described by the instantaneous Hamilto- them, in panel (b) we plot cos θ = +, θ σz +, θ (dashed 2 | h2 | | i | ˆ blue) and sin θ = +, θ σx +, θ (dashed green) factor nian H(τ) from Eq. (1), was subjected to an off-diagonal | h | | i | coupling with an environment for a long enough time for that determine the relative importance of Γt and longitudi- Fermi Golden Rule (FGR) calculation to be applicable. nal Γ relaxation channels, respectively. Insets with green and Thus the general form of electron-environment coupling blue frames schematically representing transition mechanisms ˆ 1 ˆ ˆ dominant in regimes of tc, tc, and tc. V = 2 (Vtσˆx + Vσˆz) in the |L/Ri basis, at given θ(τ) − | | should be expressed in the basis of eigenstates of instan- |± i Vˆ ˆ † ˆ ˆ taneous Hamiltonian, , θ , using (θ) = Uy (θ)V Uy(θ), ˆ which leads to the quantum spectral density for the operator Vx(t) = iHÊ t −iHÊ t e Vˆxe , given by [69, 73] ˆ 1 ˆ ˆ V(θ) = 2 (Vx ςˆx + Vz ςˆz), (18) ∞ Q ˆ ˆ iωt where Vˆx = (Vˆt cos θ + Vˆ sin θ) and Vˆz = (Vˆ cos θ − SV (ω) = hVx(t)Vx(0)ie dt (19) −∞ Vˆt sin θ). This means that at every τ we do the FGR Z Vˆ calculation for xςˆx/2 coupling, whereς ˆx acts in basis of where h...i = TrE(ˆρE ...) is the averaging over the envi- ˆ eigenstates of the instantaneous H(τ). With the environmental density matrixρ Ê. The rate of excitation of ronmental Hamiltonian given by HÊ, we calculate then the system, i.e. a transition that involves taking energy 6

Ω(τ) from the environment, is then given by [69, 73] dots [75]. As a direct consequence of Eq. (23) the exact formula for leaving electron in the initial dot reads: 1 Γ (τ) = SQ[−Ω(τ)] , (20) + V τ∞ 4 (m) (m0) Q = dτ Γ+ (τ) exp − χ (τ, τ∞) , while the relaxation rate is Z−τ∞ m m0 ! X X (24) 1 Q Γ−(τ) = SV [Ω(τ)] . (21) 4 where indices m, m0 stands for phonon, 1/f or John- (m) b (m) 0 For an environment in thermal equilibrium which we con- son’s mechanisms, while χ (a, b) = a Γ+ (τ ) + ˆ −βHE (m) 0 0 sider here, we haveρ Ê ∝e and the detailed balance Γ− (τ )dτ . Q Q βΩ R condition, SV (Ω) = SV (−Ω)e , and thus Γ−[Ω(τ)] = Let us discuss now the quantum noise spectra relevant βΩ(τ) Γ+[Ω(τ)]e , is fulfilled. for the two types of reservoirs being the sources of charge As the longitudinal Vˆ and transverse Vˆt couplings in noise, and the lattice vibrations. dots basis are often of different physical origin, we assume hVˆVˆti = 0, so that the transition rate can be written as B. Charge noise Γ±(τ) = Γ,±(τ) + Γt,±(τ), where we introduced

1 2 Q The way in which sources of charge noise couple to (Longitudinal) : Γ,±(τ) = sin θ S [∓Ω(τ)], 4 the electron in a DQD is most easily visible if we con- 1 2 Q sider the high-temperature (or low energy transfer) limit (Transverse) : Γt,±(τ) = cos θ S [∓Ω(τ)] (22) 4 t of βΩ 1. The quantum spectral density becomes Q Q then symmetric in frequency, SV (Ω) = SV (−Ω) (so that contributions, defined using spectral densities of Vˆ and Γ+ =Γ−), and it can be identified with a classical power ˆ Q ˆ ˆ iωt Q Vt operators, S (ω) = hV(t)V(0)ie dt and St (ω) = spectral density of a classical stochastic process describ- iωt hVt(t)Vˆt(0)ie dt. The Vˆσˆz coupling that is longitudi- ing the fluctuations of the electric fields caused by the dy- R nal in the |L/Ri basis (the dot basis) appears due to fluc- namics of the reservoir. These processes manifest them- Rtuations of detuning or phonons coupling to the operator selves as time-dependent corrections to parameters of σˆz. It is most efficient at causing transitions between Hˆ(τ): δ(τ) and δt(τ) for detuning and tunnel coupling |±, θi states when the latter have a molecular-orbital noise, respectively. As long as the amplitude of the noise 2 1 character, i.e. θ ≈π/2, sin θ ≈1, |±, π/2i = 2 (|Li±|Ri), is small (δt, δ tc), the modification of the instanta- 2 2 and ||tc. On the other hand, the transverse coupling neous splitting Ω(τ) = (vτ + δ) + (tc + δt) is negli- Vˆtσˆx is due to fluctuations of tunnel coupling or phonons gible. However, time variation of δ and δt activates cou- p coupling to the operatorσ ˆx. It leads to transitions of in- pling between the eigenstates of the instantaneous Hamil- terdot character between the states |Li ↔ |Ri that cor- tonian from Eq. (6)), as in the lowest order in δ and δt respond to |±, θi states at θ 1 and θ ≈ π (i.e. −tc we have and t , respectively), see Fig. 7b. Below we will c ∂ vτ + δ sin θδ˙ + cos θδt˙ see that for all the considered mechanisms, the trans- θ˙ = acot − ≈ , (25) ∂τ t + δt Ω verse processes are weaker than the longitudinal ones, c 0 i.e. SQ(Ω) SQ(Ω), so the latter could become rele- t where Ω = v2τ 2 + t2, and the last approximation re- vant only in a very far-detuned regime. 0 c lies on t2 v assumption to neglect contributions not For an electron in a double quantum dot, the relevant c p ˙ 2 mechanisms of transitions between the eigenstates are larger than the noiseless coupling θ0 = vtc/Ω0 Ω, see due to coupling of electron charge to two reservoirs: lat- Eq. (9). As we neglect correlations between δ and δt, tice vibrations (phonons) and sources of fluctuations of we treat the transitions induced by these two noises in- electric fields – free electrons in metallic electrodes and dependently. Taking then into account that the classical cl h i iωt h i ungated regions of semiconductor quantum well being spectrum Sx˙ (ω)= x˙(t)x ˙(0) e dt (where ... denotes now averaging over realizations of noise) is related to the the sources of Johnson noise, and bound charges switch- R cl 2 cl classical spectrum of x(t) noise by Sx˙ (ω)=ω Sx (ω), and ing between a discrete number of states being the sources 2 2 2 2 2 2 of 1/f type noise [74]. Due to their distinct physical ori- that sin θ =tc /Ω0 and cos θ =(vτ) /ω0, we have gin, we neglect correlations between different transition 2 1 tc cl mechanisms and write the relaxation rate as: Γ±,(τ) = 2 S [∓Ω0(τ)] , (26) 4 Ω0(τ) (ph) (1/f) (Joh) 2 Γ− = Γ− + Γ− + Γ− . (23) 1 (vτ) cl Γ±,t(τ) = S [∓Ω0(τ)] . (27) 4 Ω2(τ) t In the above, we separated charge noise contribution into 0 (1/f) Γ due to tail of 1/f-like noise from two-level fluctua- In these equations the ∓Ω0 arguments can be, of course, (Joh) tors [65] in the quantum well interface and Γ due to replaced by |Ω0|, as the classical spectra are symmet- Johnson’s noise caused by wiring in vicinity of quantum ric in frequency. In Appendix A we give an alternative 7

cl derivation of these results (in the spirit of methods used where ω1 = 2π/s and s1(T ) = S (ω1) is a commonly previously in [39, 61]). We also show there that the AME reported classical spectral density at f = 1Hz, which at calculation using these rates agrees very well with direct electron temperature of T = 100mK in typical Si/SiGe 2 2 averaging of evolution due to Hˆ (τ) averaged over realiza- device is given by s1(100mK) ≈ (0.3 − 2) µeV /Hz tions of classical noise with experimentally relevant pa- [81–85]. As s1(T ) ∝ T scaling was observed in ex- rameters (discussed below in this Section). In this way periments on quantum dots [83–85], we assume here T we check the applicability of AME to the system of inter- s1(T )=s1(100mK) 100mK . The negative-frequency quan- est in this paper in the classical noise/high temperature tum spectrum follows from Eq. (29) using the detailed regime. balance condition. It is commonly believed that charge Eqs. (26) and (27) connect the rates as given τ with disorder in Si-MOS should have larger amplitude, for ex- 2 (classical) spectra of appropriate noise at ±Ω0 frequen- ample s1(100mK) ≈ 10µeV /Hz was measured in [86] cies. Extension of AME to regime of lower tempera- at T = 300mK. However, following [87] and references tures/higher Ω0 amounts then to replacing the classi- therein, we assume that the 1/f noise amplitude in Si- cl cal spectra, S (±Ω0), by their quantum counterparts, MOS can be made comparable or even smaller then Q S (±Ω0). Let us now discuss the classical and quan- Si/SiGe [83]. tum regimes for the two charge noise spectra relevant for Let us stress that the character of noise generated by semiconductor quantum dots in GHz range (tc ∼10 µeV) an ensemble of TLFs above ∼ MHz frequency is not uni- of energies. versal, as its amplitude and exponent varies between the First we consider electric fluctuations from electron gas DQDs materials and devices. In particular, recent mea- in metallic gates, the Johnson-Nyquist noise of general surements of charge noise in Si/SiGe [88] and Si-MOS [89] form [69, 76, 77]: showed 1/f and 1/f 0.7 scaling up to 100MHz and 1MHz respectively, which contrasted with few orders of magni- Q,(J) Re{Z} ω tude weaker amplitude of charge noise at MHz frequen- S (ω) = −βω (28) Rq 1 − e cies in some of GaAs singlet-triplet qubits [90, 91]. Addi- tionally, in neither experiment a linear scaling of spectral where Rq is the inverse of conductance quantum Rq = density with temperature was seen at highest frequencies, 2 π/e = 13 kΩ and Z is the impedance of a noise source, and in particular the Si/SiGe case showed only weak de- which we model here as an ideal resistor (R) of the pendence on the temperature, confirmed also elsewhere impedance given by typical for microwaves ZR = R = for SiMOS [87, 92], which stood in contrast to GaAs de- 50Ω. The temperature-dependent part of Eq. (28) re- vice, where S(ω) ∝ T 2 and the spectrum became flat, i.e βω duces to Bose-Einstein distribution n(ω) = 1/(e − 1) α → 0 as T was increased [90]. A recent theoretical study for negative frequencies ω < 0 (absorption) and n(ω) + 1 [65] of qubit relaxation caused by interaction with an en- for ω > 0 (stimulated and spontaneous emission). In the semble of TLFs coupled to thermal bath (which create & GHz frequency range relevant here, Johnson noise from 1/f noise at low frequencies) showed that at high positive a lossy transmission line discussed in [78] for Si/SiGe frequencies (between MHz and GHz, depending on tem- quantum dot, gives at most an order of magnitude larger perature), a crossover first to SQ(ω) ∝ 1/ω2, and then to noise power. a flat or Ohmic spectrum (depending on details of distri- Next, we consider 1/f-type fluctuations of electric field bution of energy splitting of the TLFs) occurs. One can due to two-level fluctuators (TLFs) localized in the insu- thus expect that in measurement of high-frequency quan- lating regions of the nanostructure [74]. We focus first tum noise, it is difficult to distinguish the noise caused on noise in detuning, as there are numerous measure- by TLFs from other sources of electric field fluctuations, ments of spectrum of this noise in DQDs. Due to very as a flat spectrum has been observed already at MHz high spectral weight at low frequencies such a 1/f noise frequency in SiMOS QD spectroscopy [93]. Let us note dominates the dephasing of qubits the energy splitting of that one of the models of distribution of energies of TLFs which depends on electric fields [74, 79]. Here, however, considered in [65] led to SQ(ω > 0)∝T at high frequen- we focus on high (GHz range) positive and negative fre- cies. In light of the above discussion, we use the above Q,(1/f) α quency behavior of S (ω) that is of 1/|ω| character model to estimate the relevance of the tail of 1/f noise in 2 2 at very low frequencies. The behavior of quantum noise silicon-based devices and set s1(0.1K)=1 µeV /Hz. We caused by an ensemble of TLFs at such frequencies de- will use the same spectrum for GaAs, probably overesti- pends on microscopic details of these fluctuators and the mating the noise in this case, but below we will show that distribution of their parameters, see [65] and references for GaAs quantum dots the influence of electron-phonon therein. coupling dominates over that of charge noise having even Here, as in [80], where Si/SiGe charge qubit in a DQD such a large amplitude. was considered, we take α = 1 with noise amplitude For the charge noise in tunnel coupling, we assume that directly extrapolated from the low-frequency regime, it is uncorrelated with the noise in detuning, as the two i.e. for positive-frequency quantum spectrum we have are caused by TLFs from distinct spatial regions. We

Q,(1/f) ω1 parametrize the ratio of rms of ﬂuctuations of the noise S (ω > 0) = s1(T ) , (29) ω in tc and by η = St(ω)/S(ω) ≈ 0.1, with its value p 8

ikr motivated by semiconductor quantum dots experiments where the matrix elements read M(k) = hL| e |Li − ikr ikr ikr [88, 90, 94, 95], and typical values of level arm used to hR| e |Ri, and Mt(k) = hL| e |Ri + hR| e |Li, control the electronic gates during shuttling [25]. We while the temperature-dependent term reduces to Bose- conclude this section by giving the explicit forms of lon- Einstein distribution n(Ω) for Ω < 0 (absorption) and to gitudinal and transverse contributions to relaxation rates n(ω) + 1 for Ω > 0 (emission). The transition rates are due to charge noise given by 2 1 tc Q (ph) 1 2 (ph) Γ−,(τ) = S [Ω0(τ)] , (30) Γ (τ) = sin θ S [∓Ω(τ)] (35) 4 Ω(τ) ±, 4 2 2 (ph) 1 2 (ph) η vτ Q Γ±,t (τ) = cos θ St [∓Ω(τ)] . (36) Γ−,t(τ) = S [Ω0(τ)] , (31) 4 4 Ω(τ) For further calculation we need to specify a model of which are applicable for both 1/f and Johnson noise. hr| L /R i wavefunctions localized in the uncoupled dots. The corresponding excitation rates are obtained via de- 0 0 −βΩ We assume that they are separable and Gaussian: tailed balance condition Γ+(Ω) = Γ−(Ω)e . ∆x 2 2 2 1 (x ∓ 2 ) + y z hr|L0/R0i = exp − − , 3 4 2 1 2 2 C. Electron-phonon interaction (π rxyrz ) 4 ( 2rxy 2rz ) (37) In semiconductors, another mechanism responsible for where the full width at half maximum (FWHM) of elec- transitions between the |±, θi states is associated with tron wavefunction, which deﬁnes dots diameter, is given energy exchange between the electron and lattice vi- by 2rxy in planar, and 2rz in the growth direction of brations. Phonons are assumed to be in thermal equi- structure, while the ∆x gives the distance between the librium, with their free Hamiltonian given by Hph = dots. Next we use Hund-Mulliken approximation [98, 99], ω b† b , where λ and k represents phonon po- to generate a set of orthogonal states in DQDs system k,λ k,λ k,λ k,λ h | i larizations and wavevector respectively. The electron- that fulﬁll L R = 0, which can be done by setting: phononP interaction is given by [96]: |L/Ri = |L0/R0i − g |R0/L0i , (38)

|k| (j) 1 1 −∆x2/4r2 ˆ ˆ† ikr ≈ h | i xy Hel-ph = vk,λ bk,λ + b−k,λ e , where g 2 L0 R0 = 2 e 1. 2%cλV j,k,λ=L,T s As the energy quanta exchanged between the electron X (32) and the lattice are < 1 meV, we take into acocunt only the acoustic phonons with ωk,λ = cλ|k|. The relaxation in which % denotes crystal density, V the crystal volume, rates due to electronphonon interaction are then given and cλ is the speed of λ-polarized phonons. The coupling by: (j) vk,λ stands for piezoelectric (j = p) and deformation po- 2 tential (j = d), evaluated for transverse (λ = T ) and (ph) Ω tc (j) 2 2 Γ (Ω) = dΩk v |F(kλ)| [n(Ω) + 1] longitudinal (λ = L) polarizations of phonons: −, 8π2%c5 kλ,λ λ,j λ X Z 2 (39) (p) χp (d) kz (d) kxykz vk,λ = , vk,L = Ξd+Ξu , vk,T = −Ξu 2 , 2 k k k (ph) Ω 2 (j) 2 2 Γ (Ω) = |hL0|R0i| dΩk v |Ft(kλ)| (33) −,t 8π2%c5 kλ,λ λ,j λ Z where χp is piezoelectric constant, while Ξd,Ξu are di- X latation and shear deformation potentials respectively. In [n(Ω) + 1] , (40) GaAs and Si the coupling to phonons takes a very different form, namely Si lacks the dominant in the GaAs where the integration over solid angle of resonant SiGe wavevector kλ, with length kλ = Ω/cλ was denoted by piezoelectric coupling χp = 0 [97], while the opposite GaAs dΩk = dϑkdϕk sin ϑk, while the form factors read: is true for shear deformation potential since Ξu = 0. We evaluate the matrix elements of interaction from 2 2 2 2 k r + k r 2 2 xy xy z z kx∆x Eq. (32) in the two-dimensional space spanned by |L/Ri |Ft(k)| = exp − 1 − cos 2 2 states (see Appendix B for details), obtain the Vˆ σˆ +Vˆ σˆ ! t x z form of coupling discussed in Sec. III A, and arrive at 2 2 2 2 kxyrxy + kz rz ˆ 2 2 kx∆x quantum spectra associated with longitudinal (V) and |F(k)| = exp − sin ( 2 ). (41) 2 ! transverse (Vˆt) couplings to phonons:

(j) 2 2 2 2 | |2 The common term exp −(kxyrxy + kz rz )/2 is the (ph) k vk,λ 2 δ(Ω(τ) − ωk,λ) S/t [Ω(τ)] = 4π |M/t(k)| −βω , Fourier transform of the electron wavefunction, while %cλV 1 − e k,λ j,k,λ the main diﬀerence between the longitudinal and trans- X (34) verse relaxation is the overlap of bare dots wavefunctions, 9

2 2 2 (ph) −∆x /2rxy 2 |hL0|R0i| = e 1 which makes the transverse enough to π for sin θ term in Eq. (35) suppresses Γ−, relaxation mechanism orders of magnitude weaker, i.e. (ph) (ph) (ph) to the degree that is becomes smaller than Γ−,t . Γ−,t Γ−, , unless detuning is so large that θ is close

Quantity Symbol Values

Tunnel coupling tc 5 60µeV − Eﬀective electron temperature T 50 500mK − Detuning sweep rate v 1 3000µeV/ns − Initial detuning i 500µeV − Final detuning f 500µeV

Time of detuning sweep f i = ∆/v 0.3 1000ns p− − Transverse/longitudinal noise ratio St(ω)/S(ω) 0.1

Resistance of noisy resistor (Johnson noise) ZR 50Ω 2 1/f noise amplitude at T = 0.1K s1(0.1K) 1µeV /Hz Dots separation ∆x 50nm (GaAs), 100nm (SiGe), 150nm (Si-MOS)

Spread of electron wavefunction in XY plane rxy 40nm (GaAs), 20nm (Si/SiGe, Si-MOS)

Width of quantum well 2rz 20nm (GaAs),5nm (Si/SiGe, Si-MOS)

TABLE I. Parameters used in the paper.

As the size of quantum dot in planar direction rxy is of “typical” GaAs, Si/SiGe and Si-MOS double quantum larger then size in the z direction, 2rz rxy, its value dots given in table I, In Fig. 2 we plot zero-temperature can be extracted from splitting between the ground and electron relaxation rate due to scattering with phonons, ∗ (ph) first excited dot state ∆E, i.e. rxy = 1/m ∆E. In Γ− (Ω), as a function of detuning (let us recall that Si at ∆E = 1meV an estimate of rxy ∼ 20nm is con- 2 2 p Ω = tc + ) for three values of tunnel coupling, tc = sistent with reported values of rxy ≈ 15nm [100], 13nm 10, 20, 40µeV. It is clear that the scattering of a single [82] in Si/SiGe and rxy ≈ 21nm [101], 18nm [86] in Si- electronp in a DQD in each of considered nanostructures MOS. The GaAs dots are typically larger (rxy ≈ 55nm is dominated by a different mechanisms. In polar GaAs, [102], 21 nm [103]), mostly due to smaller effective mass, ∗ ∗ the piezoelectric coupling dominates over the deforma- i.e. mSi/mGaAs ≈ 3. The typically reported values of tion potential one, with the fastest relaxation at low de- 2rz ≈ 20nm [102] in GaAs are also larger than those in tuning, where the transitions occur between molecular- Si/SiGe, 2rz ≈ 4nm [104], 6nm [82]. We assume here orbital type states. The relaxation rate, for the energies the extent of electron’s wavefunction in the z direction below c/∆x ≈ 50µeV shows oscillatory behaviour due to in Si-MOS is similar to that in Si/SiGe, and for both we 2 2 |F| ∝ sin (kx∆x/2) term, see Eq. (41). For larger de- take it as 2rz = 5nm. Finally, smaller dots allow for detunings, when the energy transfer Ω ≈ , the relaxation creasing the distance between the sites from typical for rate decreases as its mostly longitudinal character that GaAs ∆x ∼ 150nm [102], ∼ 110 nm [103] to Si/SiGe val- 2 makes it ∝ (tc/) , is combined with phonon spectral ues of ∆x ∼ 100nm [5], to Si-MOS ∆x ∼ 50nm [5]. The density ∝ 3 and piezoelectric coupling |v(piez)|2 ∝ −2, distances between the dots are correlated with reported to produce an overall Γ(piez) ∝ (t )2/ scaling in the far values of t , the largest of which are achieved in SiMOS − c c detuned regime t , until ≈ 500µeV when phonon structures, with examples of t ≈ 450µeV and 50µeV c c bottleneck effects start to become strongly viisble. On for dots separated by ∆x ∼ 40nm [28] and ∼ 100nm the other hand, in Si/SiGe a weaker deformation poten- [101] respectively. However, recently tc ≈ 40µeV was (def) achieved in Si/SiGe across an array of quantum dots with tial scattering gives Γ that first increases with , and rxy ∼ 10nm and ∆x ≈ 70nm [25]. In GaAs, tunnel cou- then becomes suppressed by phonon bottleneck effect at pling of tc ≈ 20 − 40µeV was measured in an array of large detunings. The relaxation time 1/Γ− falls below eight quantum dots with ∆x ≈ 150nm [2] for array of 8 100 ns for ∼ 100 µeV only for the largest considered QDs. Representative parameters for each nanostructure tc = 40µeV. Finally, in Si-MOS the smaller interdot dis- that we will use in subsequent calculations are given in tance makes the transverse relaxation more efficient. Due (def) 3 Tab.I. to its Γt,− ∝ scaling up to phonon bottleneck energy of about 1 meV, it becomes the dominant process We now evaluate numerical values of relaxation rates at larger detunings. Such a transverse relaxation rate from Eqs. (39) and(40) for above-discussed parameters 10

D. Comparison of the transition rates (piez) (piez) (def) (def) Γ Γt Γ Γt 101 (a) GaAs (1/ns) 100 Phonons GaAs

(ph) − Phonons Si/SiGe Γ 1 100 10− Phonons SiMOS

1/f Noise eV 2 µ 10 (1/ns) − 1 Johnson Noise 10− ± = 20 Γ

3 c 10− t

Relaxation Rate 10 2

− mK, 101

(b) Si/SiGe 3 = 100 (1/ns) 10−

0 T 10 Transition rate (ph) −

Γ Γ 1 − 10− 4 10− Γ+

2 10− 0 100 200 300 400 500 Detuning (µeV) 3 10− Relaxation Rate FIG. 3. Relaxation (solid lines) and excitation (dashed lines) 101 rates as a function of detuning at typical tunnel coupling (c) Si-MOS tc = 20µeV and temperature of T = 100mK. Transition (1/ns) 100 rates due to phonons are drawn using green (GaAs), black

(ph) − (Si/SiGe) and blue (Si-MOS) colors, while transition rates Γ 1 10− due to common for all nanostructures charge noise is depicted using red (Johnson) and yellow (1/f) colors. Both excitations 2 10− and relaxations in GaAs are dominated by electron-phonon coupling. In Si the excitations are commonly caused by charge 3 10− noise (either 1/f or Johnson of similar amplitude), while the

Relaxation Rate relaxation at ﬁnite detuning relies on Johnson noise in Si/SiGe 101 102 103 and relatively stronger interdot phonons in Si-MOS, where the Detuning (µeV) dots are closer.

tc = 10µeV tc = 20µeV tc = 40µeV Let us know use the results of the previous Sections FIG. 2. Phonon relaxation rates in (a) GaAs, (b) Si/SiGe, and compare the relative importance of various types (c) Si-MOS double quantum dot devices as a function of of environments on the discussed DQD structures. In dots detuning at fixed tunnel couplings: tc = 10µeV (solid Fig. 3 we plot the relaxation rate Γ−[Ω()] (the excitation −βΩ line), 20µeV (dashed-dotted) and 40µeV (dashed). Contribu- rate Γ+(Ω) = Γ−(Ω)e ) with solid (dashed) lines as tions from different phonon mechanisms are shown with a dis- functions of detuning for all the considered mechanisms (piez) tinct color: piezoelectric longitudinal coupling Γ (green), using the above-discussed representative parameters for (piez) piezoelectric transverse coupling Γt (yellow), deformation GaAs, Si/SiGe, and SiMOS structures, and temperature (def) longitudinal coupling Γ (red) and deformation transverse T = 100mK and tunnel coupling tc = 20µeV. As ex- (def) coupling Γt (blue). Longitudinal phonons couples orbital- pected from discussion in Sec. II C, the excitation rates like states in vincinity of avoided crossing, while tranverse are the largest at the anticrossing, and they become sup- phonons couples dot-like states in detuned regime. Each panel pressed exponentially with increasing above tc. In that represents different device with parameters given in Tab. I. regime the relaxation overwhelmingly dominates over excitation, but the dynamics of the electron will depend on the value of total Γ−: the electron tranfer error will depend on the ratio of timescale of environment-assisted inelastic tunneling between the dots in the far-detuned weakly depends on tunnel coupling (note the presence of regime, 1/Γ−, and the duration of the detuning sweep. −4 single blue lines in Fig. 3, and requires overlap between Note that for tc =20 µeV the requirement of QLZ <10 wavefunctions of L/R dots, which is not large enough means v <200µeV/ns, so the total time of detuning sweep (piez) in the other nanostructures: Γt might be relevant over a meV range is 5ns. This will give a ballpark esti- in GaAs only at highest detunings, see Fig. 3a, and the mate what timescales we should compare 1/Γ− to. transverse process never becomes of similar order of mag- In GaAs the coupling to phonons (the green line in nitude as the longitudinal one in the considered Si/SiGe Fig. 3 dominates the relaxation, with influence of John- structures. son noise possibly becoming dominant at highest con- 11

(J) GaAs Si/SiGe SiMOS longitudinal process, and and Γ−,t ∝ in case of weaker 101 (a) = 100µeV (b) = 400µeV transverse one. For their assumed ratio, the relaxation (J) the rates become equal at J = 10tc, which means Γ− (1/ns) 0 −1 − 10 slowly decreases as up to = 400µeV, and then it Γ starts to slowly increase with ∝ . The relaxation time 1 10− for the assumed amplitude of Johnson noise is ∼ 100 ns in the relevant detuning range. 2 10− Finally, for SiMOS the smaller interdot distance as- (piez) 3

Relaxation rate sumed for this architecture makes Γ ∝ the dom- 10 3 t,− − inant relaxation process at large detunings: as show in 10 20 30 40 50 10 20 30 40 50 Fig. 3 this relaxation channel dominates over the one due Tunnel coupling tc (µeV) Tunnel coupling tc (µeV) to Johnson noise for 200 µeV. The relaxation times (c) = 0 & 101 at large detunings approach ∼ 10 ns, so phonon-assisted interdot tunneling might be an efficient mechanism of 0 10 healing of charge noise-induced excitation that occurred (1/ns) 1 close to the anticrossing in SiMOS. (0) 10− + The other mechanisms only weakly contribute to relax- Γ 2 ation, as longitudinal 1/f noise relaxation rate is strongly 10− (1/f) 2 3 attenuated with increasing detuning, as Γ−, ∝ tc / at 3 10− T > tc, while small overall strength and weak detuning 50mK dependence of longitudinal phonon processes in Si/SiGe, 4 (def) Excitation rate 10− 100mK 2 Γ−. ∝ tc , produces relaxation times above 100 ns only 500mK approaching the order of magnitude of contribution of 5 10− Johnson noise around = 200µeV. 10 20 30 40 50 Let us now discuss the tunnel coupling and tempera- Tunnel coupling tc (µeV) ture dependence of the total Γ+ rate at =0, and of the total Γ− rate at moderate and high detunigs, =100 and FIG. 4. Transition rates relevant for electron charge trans- 400 µeV, respectively. The relaxation rates at moderate fer in GaAs (green), Si/SiGe (red) and Si-MOS (blue): the detuning have a common dependence on tc inherited from excitation rate at avoided crossing Γ+(0) ( = 0) (a), re- the tunneling dependence of the dominant there longitu- laxation rates in detuned regime Γ ( = 100µ)eV (b) and 2 − dinal process, i.e. Γ− ∝ tc . This is not the case at larger Γ ( = 100µ)eV (c) as a function of tunnel coupling tc for − selection of temperatures T = 50, 100, 500mK (solid, dashed- detuning, where transverse processes that are weakly de- dotted, dashed) lines respectively. The excitation rate due pendent on tc can dominate. Similarly for considered to piezoelectric phonons in GaAs in T > 100mK is the only here βtc 1, temperature dependence of relaxation is non-monotonic function of tunnel coupling. Otherwise exci- very weak. We illustrate both statements in Fig. 4a and tation rates decreases for larger tc due to exponential factor 4b where we plot relaxation rates at = 100, 400µeV as a βtc Γ+(0) e− , while relaxation rates increases due to domi- function of tunnel coupling. As it can be seen difference ∝ 2 2 nant role of longitudinal mechanisms Γ (Ω) tc /Ω . Since between Si/SiGe and SiMOS is visible at large detun- 2 2 − ∝ Ω = √tc + increase is stronger at lower . The only discrep- ing where for small tunnel couplings interdot phonons ancy between Si/SiGe and Si-MOS is visible in the relaxation provide order of magnitude faster relaxation rate in the rate at far detuned regime ( = 400µeV) due to presence of latter. interdot phonons in the latter. In Fig. 4c we illustrate the relevant excitation rate Γ+(0), computed at the avoided-crossing at T = (piez) 50, 100, 500mK. In GaAs the only relevant mechanism sidered detunings. As discussed above, Γ ∝ (t )2/ − c is the coupling between orbital states provided by the for most of the considered range of , so for the healing phonons, which has a strong scaling with tunnel coupling of excitation to be significant the time spent at mod- (piez) 3 −βtc erate detunings, up to about 200 µeV (see Fig. 3), has Γ+ (0) ∝ tc e as long as tc c/∆x ∼ 50µeV, wher 3 to be larger than average relaxation time in this range, tc dependence is provided by the piezoelectric coupling (t ) and the resonance term sin2(k ∆x/2) (t2). As a 1/Γ− ∼1 − 10 ns. c x c The situation is more complex in Si nanostructures. result, at smaller temperatures, excitation rate in GaAs For parameters of Si/SiGe DQDs it is the Johnson noise shows a non-monotonic behaviour as a function of tunnel - red line in Fig. 3 - that dominates (more visibly at coupling. In Si, the excitation at tc is caused only by lower )) over the relaxation due to deformation potential charge noise, and hence for amplitude of this noise used coupling to phonons (the black line in Fig. 3). The de- here for both Si/SiGe and SiMOS, it results in the same (1/f) −βtc tuning dependence of this process is rather weak. When rate, in which contributions from Γ+ (0) ∝ e /tc (J) t /k T 1 (in Fig. 3 we have t /k T ≈ 2.3), the John- −βtc c B c B and Γ+ (0) ∝ tce are combined. The latter becomes (J) 2 son noise from 50Ω resistor gives Γ−, ∝ tc / for stronger more relevant at larger tc, for which however the overall 12 charge noise is attenuated due to exponential factor, as tunings. The agreement between the result of evaluation it can be seen in Fig. 4c by a decrease of excitation rate of AME, and the approximation is more visible at lower in Si. T (higher tc), since this agreement is expected to improve as tc/kBT 1. In contrast, in GaAs device, Q (plotted in green color) decreases monotonically as the sweep rate IV. PROBABILITY OF LEAVING THE v gets smaller. This is a consequence of a much stronger ELECTRON BEHIND coupling to environment (specifically piezoelectric coupling to phonons), which on one hand increases probability of transition from ground to excited state in vicinity of Let us use now the above-derived transition rates to the avoided crossing (dashed line), but on the other hand calculate the central quantity of this paper – occupation allows for subsequent relaxation even for not-too-high v. of higher energy state after detuning sweep Q, i.e. the As a result there is no region in which Q ∼ Q ∼ 1/v, probability of leaving the electron in the initial dot. SEAL however as long as Q ≤ 0.1, i.e. the probability of We assume the relevant part of detuning sweep starts SEAL excitation-relaxation-excitation sequence is relatively low and terminates at = ±500µeV, since at ≥ 10t the f c (< Q2 ), the result can be well approximated by tak- dots become uncoupled, i.e. the approximation of con- SEAL ing into account only relaxation processes modifying the stant t breaks down [94, 105], and the detuning sweep c excitation generated at the anticrossing, i.e. Q≈Q . rate used in an experiment can be increased [25]. In Fig. 5 HEAL Finally, when the electron transfer time is long enough we compare the results of a numerical solution of Adia- to allow for a second transition from ground to excited batic Master Equation (AME) from Eq. (10), depicted as state, i.e. when Q ≥ 0.1, the HEAL formula gives squares, against the approximation of single excitation at SEAL only a lower bound for results of the AME, as visible at avoided crossing without relaxation process, Q from SEAL low v when comparing the squares and solid lines. Eq. (15), shown as dashed lines, and the approximation An obvious way to increase the efficiency of charge of an excitation followed by relaxation processes only, transfer, or equivalently decrease Q, would be to bring Q from Eq. (17), shown as solid lines. The dotted HEAL the Q result down, as for Γ < Γ it gives an upper line is the Landau-Zener formula Q from Eq. (2). In SEAL + − LZ bound of excited state occupation induced by environ- the four panels we show results for combinations of tunnel mental fluctuations, i.e. Q < Q in the adiabatic coupling and temperatures: t = 10, 20µeV (columns), SEAL c regime where Q > Q . This can be achieved by low- T = 50, 100mK (rows). With hollow squares we mark LZ ering the temperature or increasing the tunnel coupling. the AME results in the region where Q and Q SEAL HEAL In Fig. 6 we show a rather optimistic result of proba- are no longer an upper and a lower bound on Q, as prob- bility of leaving the electron in the left dot evaluated ability of Landau-Zener transition dominates. We stress for the largest t reported in the array of Si/SiGe quan- that in this region the applicability of AME in secular c tum dots [25], t = 40µeV. As a reference we compare approximation used here is limited [42, 43], however a c it to the other materials considered, and plot results for correction to the L-Z formula computed using different T = 100, 500mK, as larger tunnel couplings should in methods correction is expected to be small for predimi- principle allow for working at higher temperatures [106– nantly longitudinal relaxation Γ Γ [45, 51, 53]. t 108]. We stress that a calculation for T = 50mK (not Let us now concentrate on the region in which Q is shown) gives Q ≤ 10−6 for v < 400µeV/ns. For Si nanos- dominated by effects related to interaction with the ther- tructures the behaviour at higher temperatures is quali- mal environment, where the AME results are plotted as tatively similar to than shown in Fig. 5, with a local min- filled squares. For both Si-based devices in a region of imum of Q = 10−4,10−2 at v = 400µeV/ns, 800µeV/ns moderate v we observe that Q ∝ 1/v. This suggests that for T = 100mK, 500mK respectively. In GaAs the large the value of Q follows from a finite excitation probability value of tc results in strong coupling between transferred in a limited range of detunings (near the anticrossing), electron and the environment, which at higher temper- and the occupation of the excited state grows with in- atures causes flattening of Q as a function of v. This creased time spent in this region. In agreement with this effect can be attributed to reaching thermal equilibrium picture, Q ≈ QSEAL (dashed line), and the electron un- of Qeq( = 0) = Γ+(0)/(Γ+(0)+Γ−(0)) ∼ 0.3 around the dergoes a single transition from ground to excited state avoided crossing, followed by slower relaxation at larger in vicinity of the anticrossing. We note such a transi- . tion from ground to excited state at tc in SI-based A local minimum of Q(v) is thus expected in both Si- devices is caused solely by charge noise. As the sweep based nanostructures considered here. The value of Q rate is decreased, an increasing time spent during the at this minimum can be estimated as the intersection of electron trasnfer in the far-detuned regime, tc, al- Q and Q , i.e. lows for a significant recovery of ground state occupation SEAL LZ by the relaxation mechanism, which is reflected by a de- QSEAL(vopt) = QLZ(vopt), (42) viation from a SEAL approximation and Q ∼ QHEAL (solid lines) for smallest sweep rates. The healing effect the solution to which is expressed in terms of Lambert 2 is stronger for the SiMOS device, due to effective phonon W function [109] as vopt = πtc /2W (a) where W (a) sat- W (a) π 3 relaxation between the dot-like eigenstates at large de- isfies equation W (a)e = a for a = 4 βtc /Γ+(0). p 13

AME QLZ QSEAL QHEAL

tc = 10µeV tc = 20µeV

1 10−

2 mK 10− = 100 3 10− T

(Transfer error) 4 10− Q

1 10−

2 mK 10− = 50

3 T 10− Excited state occuparion

4 10− 100 101 102 100 101 102 Detuning sweep rate v (µeV/ns) GaAs Si/SiGe SiMOS

FIG. 5. Probability of occupying excited state Q, i.e. leaving the electron in the initial dot after detuning sweep, as a function of sweep rate v for ﬁxed tunnel coupling and temperature in two semiconductor DQD devices: GaAs A (Green) and SiGe A for Johnson noise originating from ideal resistor (red) and from transmission line (pink). In the four panels we show combinations of tunnel couplings tc = 10, 20µeV (columns) and temperatures T = 50, 100mK (rows). Squares correspond to numerical solution of Adiabatic Master Equation, where we have used ﬁlled (hollow) squares to denote adiabatic (non-adiabatic) regime. Dashed line corresponds to Single Excitation Approximation Limit QSEAL, see Eq. (16), while the solid line is the Healed Excitation 2 Approximation Limit, QHEAL, see Eq. (17). Dotted black line shows the Landau-Zener result QLZ = exp πtc /2v . Remaining parameters are given in Tab. I in the Appendix. III C. − .

Since typically a 1, the asymptotic form of W (a) = but there are other factors that are limiting v from be- ln(a) − ln(ln(a)) in the low temperature limit βtc > 1 low in GaAs devices (see discussion in the next Section). allows to write vopt ∼ πkBT tc/2. As the value of QSEAL Similarly, in Si quantum dots charge transfer can be in is independent of coupling to environment at larger de- principle improved by going to much lower sweep rates tunings, the sweep rate vopt which minimizes Q depends v < 1µev/ns, however it would make the few-nanosecond on the charge noise amplitude at tc, which is assumed transfer impossible as it has been demonstrated by show- here to be the same in Si/SiGe and Si-MOS. ing the sweep time interval on the right y-axis of Fig. 7a.

In Fig. 7a we show how vopt in Si varies with tc for T = 50, 100, 500 mK. We see that vopt increases as the QLZ curve shifts to higher v (due to an increase of tc), or noise-induced excitations QSEAL become stronger The value of optimal sweep rate and corresponding (here due to an increase of T ). Next, in Fig. 7b we minimum of transfer error Q(vopt) in Si depends on the use vopt to compare the corresponding transfer error amplitude of charge noise at frequency corresponding to in Si Q(vopt) (red) against analogous quantity in GaAs tunnel coupling (which is in the GHz range), where its (green), as a function of tc ∈ 5 − 60µeV. In the Fig- influence dominates over that of phonons. We concen- ure we have put together results of the AME (solid, trate here on the amplitude of 1/f noise, the amplitude dashed, dotted dashed lines) for three different temper- of which can vary by at least an order of magnitude be- atures T = 50, 100, 500 mK. The probability of losing tween different Si DQD devices. In Fig. 8 we plot a min- the electron Q in SiGe appears to be below the value for imal transfer error at typical electron temperature T = GaAs for the tunnel couplings apart from the smallest 100mK as a function of square root of 1/f spectral den- ones of tc ∼ 5 − 10µeV, where the corresponding opti- sity evaluated at 1Hz and at T = 100mK, which we√ have mal sweep rate (vopt ∼ 10 − 50µeV/ns depending on the previously taken as constant s1(100mK) = 1µeV/ Hz temperature) is large enough to make the Landau-Zener (see Sec. III B). We plot the results for range of tc con- p QLZ probability stay above the phonon-induced Q. Of sidered here, and emphasize that noise amplitude can be course, Q in GaAs can be made lower by using v < vopt, directly related to excitation rate at Ω = tc with the 14

(a) 1000 1

QLZ tc = 40µeV T 800 QSEAL ) 50mK 100mK 1 mK 10− AME 600

eV/ns 500mK µ ( 400 = 500

2 opt 10 T − (a) v 200 Sweep time (ns) 10 (Transfer error) (b) 0 Q

GaAs Si/SiGe Si-MOS ) 1 10− opt v

( 1 10− Q

2 mK 10− 2 10− = 100 3

3 T 10− 10− Excited state occuparion 4 10− GaAs 4 10− (b) Si

Excited state occupation 5 10− 100 101 102 103 10 20 30 40 50 60 Detuning sweep rate v (µeV/ns) Tunnel Coupling tc (µeV)

FIG. 6. Probability of leaving the electron behind in the FIG. 7. Optimal transfer in Si. In panel (a) we plot optimal case of high tunnel coupling tc = 40µeV. We compare results sweep rate for Si/SiGe and Si-MOS devices vopt obtained as for T = 500mK (a) and T = 100mK (b), since large tunnel a solution to Eq. (42) for T = 50, 100, 500mK (solid, dashed coupling in general should allow for relatively efficient transfer and dotted-dashed lines). In panel (b) we plot probability of in higher temperature. leaving the electron behind after the detuning sweep with a rate vopt, as a function of tunnel coupling tc and for the same selection of temperatures. We compare results for Si against following formula: phonon dominated transfer with the same sweep rate in GaAs (Green) as a reference. 2 (1/f) 2s1[µeV /Hz] tc Γ (tc) 1/ns ≈ exp − , (43) + t [µeV] k T c B V. DISCUSSION AND SUMMARY where s1 = s1(100mK) for brevity and square brack- ets denoted units in which the quantities should be sub- stituted. The excitation rate obtained using this Equa- We have presented a theory of the dynamics of a sys- tion can be directly used in the SEAL formula, given by tem undergoing a Landau-Zener transition in presence Eq. (16), the result of which was illustrated in Fig. 8 us- of weak transverse and longitudinal couplings to ther- ing dashed lines. As expected, QSEAL agrees well with mal environments: sources of noise of 1/f and Johnson the results of adiabatic Master equation (dots) for rela- type, and a bath of noninteracting bosons, specifically tively small error Q(vopt) 1. Next we analyze transi- acoustic vibrations of a three-dimensional crystal. Our tion between 1/f and Johnson noise dominated excita- focus was on the regime in which the deterministic change tions. The latter can be seen in Fig. 8 as a flattening of parameters of the Hamiltonian is slow enough to ne- of the solid lines, which represents results of adiabatic glect the Landau-Zener coherent excitation, and the ef- master equation with both 1/f and Johnson noise, from fectively nonadiabatic character of the evolution can be R = 50Ω resistor, contributions. By comparing solid and caused only by interaction with the environment. A gen- dashed lines, we conclude that amplitude of 1/f noise at eral theory based on adiabatic Master equation (AME) which it starts to dominate over Johnson noise becomes was then applied to a case of electron transfer between a larger when the tunnel coupling is increased, which can pair of voltage-controlled semiconductor quantum dots, be deduced from scaling of respective excitation rates, i.e. for which we took into account realistic parameters for (1/f) (1/f) (J) Γ+ (tc) ∝ 1/tc and Γ+ (tc) ∝ tc for tc kBT . As electron-phonon interaction and both Johnson and 1/f the optimal sweep rate vopt is too high to allow for any charge noise. We have calculated transition rates be- phonon-mediated suppression of Q in Si DQDs, the dif- tween system’s eigenstates as function of interdot detun- ference between AME and SEAL visible for large noise ing , and used them in AME calculation to obtain the amplitude is attributed to subsequent relaxation (and probability of failure of charge transfer between the two further transitions) caused by 1/f noise of either large dots, Q, as a function of detuning sweep rate v. amplitude (tc ≥ 10µeV) or at relatively high tempera- When v is below the value at which the Landa-Zener ture (tc = 5µeV, for which βtc ∼ 0.5). transition is activated, only a finite temperature of envi- 15

1/f (AME) 1/f (SEAL) Johnson and 1/f (AME) proaching a value exponentially small in ﬁnal , reﬂecting 100 approaching a thermal occupation of the ground state.

) The main qualitative theoretical result of the paper,

opt 1 v 10 which could also apply to systems other than double ( −

Q quantum dots, is thus that a system described by a

2 Landau-Zener model, when coupled to a thermal environ- 10− ment can realize two possible scenarios: one qualitatively similar to the Landau-Zener effect, but with dependence 3 10− tc (µeV) of Q on v renormalized by environment, and another in 5 which dependence of Q on v is nomonotonic, and there 4 is an optimal sweep rate that minimizes Q. The main 10− 10 20 conclusion specific to the considered case of GaAs and Excited state occupation 40 Si-based quantum dots is that for T ≈ 50 mK, in GaAs 5 10− case Q can be made smaller than 10−4 by choosing v 10 1 100 101 − smaller than ≈ 10 (100) µeV/ns for tc = 10 (20) µeV, 1/f noise aplitude s (100mK) (µeV/√Hz) −4 1 while in case of Si having Q=10 requires tc >20 µeV, p and optimal v of a few tens of µeV/ns. Large tunnel cou- FIG. 8. Probability of leaving the electron behind using op- plings and low temperatures are crucial for having small timal vopt sweep rate, as a function of 1/f noise spectral Q. In Si-based DQDs there is a possibility of further density measured at 1 Hz and selection of tunnel couplings suppression of Q by decreasing the level of charge noise tc = 5, 10, 20, 40µeV at T = 100mK. For assumed in Sec. III B at GHz frequencies, corresponding to t ≈10 µeV energy model of high-frequency 1/f noise we compare results of adi- c abatic master equation with (solid line) and without (dots) splitting at the anticrossing. additional contribution from Johnson noise against QSEAL ap- A process of a controlled electron transfer between two proximation with 1/f noise only (dashed lines). For the noise quantum dots is relevant for ongoing attempts at con- t 2 in tunnel couplings we assumed s1 = (0.1) s1. struction of quantum buses based on chains of many tunnel-coupled dots [23, 25, 28, 40, 70]. Let us now discuss the implications of the results of this paper for ronment allows for energy absorption necessary for mod- prospects of coherent shuttling of electron-based spin ification of Q, since otherwise electron would stay in the qubits across N ≈ 100 quantum dots. This number of ground state. This absorption most likely takes place in dots in a 1D chain is motivated by requirement of hav- the vicinity of the anticrossing, where the thermal energy ing ≈ 10 µm distance between few-qubit registers in a needed for transition is the smallest. A specific feature realistic architecture of a quantum computer based on of the system under consideration is that the dominant gate-controlled QDs [10] and typical interdot separation coupling to the environment is most effective at the anti- . 100 nm. crossing, making this effect even stronger. Consequently When the goal of charge shuttling is an on-demand during the process of electron transfer caused by sweeping transfer of qubits, which should be highly coherent, and the detuning, a finite Q is generated at the anticrossing, which are to take part in further coherent manipulations when || ≤ tc. Then, for larger positive detunings the after being moved from one register to another, the de- electron relaxation processes dominate over the excita- terministic character of the shuttling is necessary. Any tion processes, and suppression of Q is expected. randomness in qubit arrival times will complicate the In the considered DQDs there are two possible scenar- application of subsequent coherent operations involving ios. In Si-based dots, coupling to charge noise dominates, that qubit. Furthermore, any stochastic component in and the transition timescale are longer than the typical the duration of the qubit transfer will introduce a ran- transfer times, so that the final Q is very close to the dom contribution to the phase of the qubit. More in- value generated near the anticrossing, which is ∝ 1/v depth discussion of relationship between the indetermin- (proportional to the time spent neart the anticrossing), istic character of electron shuttling and spin qubit de- so it exhibits a dependence on v qualitatively opposite phasing will be given in [70]; here it is enough to realize to the one for Landau-Zener effect dominating at large that large probability of electron arriving at the end of v. Only at lowest v the energy relaxation starts to be N-dot chain at a time other than the desired one, will efficient at lowering Q, with this effect being stronger cause major problems in the context of quantum infor- in SiMOS compared to Si/SiGe structures. The compe- mation processing, and we will treat it here as an error tition between the environment-induced excitation and probability. Assuming that Q 1/N, the probability of Landau-Zener effect leads then to appearance of optimal the electron arriving at the end of the chain not at the de- v, at which Q is minimal. In GaAs, on the other hand a sired time, i.e the probability of qubit transfer-associated strong piezoelectric coupling to phonons dominates, tran- error, is QN ≈ NQ. sition timescales are shorter than the charge transfer time Our results for Si-based quantum dots show that for and consequently many transitions take place, and the tunnel couplings in 5−40µeV range, as recently reported final Q monotonically decreases with decreasing v, ap- in first experimental realizations of electron shuttling 16

−3 over a few dots, achieving QN ≈10 will be possible for ACKNOWLEDGMENTS tc ≥ 30 µeV and at T ≤ 50 mK. Note that a high-fidelity charge transfer between two dots in Si MOS structure was We would like to thank Lars Schreiber and Lieven demonstrated experimentally using tc = 450µeV [28], but Vandersypen for discussions that motivated us to fo- maintaining such a strong tunnel coupling in a 1D array cus on the problem addressed in the paper, and Piotr of N ≈100 quantum dots will be challenging. Szańkowski for multiple comments on earlier versions of In GaAs, on the other hand, QN can be made much this manuscript. This work has been funded by the Na- smaller by decreasing the detuning sweep rate v, so that tional Science Centre (NCN), Poland under QuantERA the time of interdot transfer becomes longer than ∼ 10 programme, Grant No. 2017/25/Z/ST3/03044. This ns. This in fact also holds for Si-based dots, only v has project has received funding from the QuantERA Pro- to be made at least a further order of magnitude smaller. gramme under the acronym Si QuBus. However, for such slow transfers one has to start wor- rying about well-known mechanisms of spin dephasing that affect the coherence of a static electron localized Appendix A: Correction due to dynamics of classical in a QD. In both the considered materials interaction noise ∗ with nuclei leads to dephasing T2 time of the order or 10 ns for GaAs [110] and a few hundreds of ns for natural Here we provide detailed calculations of occupation of Si [111–113] (and up to tens of microsecond for isotopi- excited state in the classical limit, i.e. where the fluctu- cally purified silicon with about 103 ppm of spinful 29Si ations of detuning and tunnel coupling can be modeled [82, 111, 114]). For isotopically purified Si QDs in vicin- by stochastic contribution to Hamiltonian (1), i.e. ity of micromagnets, their spatially inhomogeneous mag- ∗ netic fields together with charge noise lead to T2 ≈ 20 ˆ + δ t + δt ∗ H = σˆz + σˆx. (A1) µs. Let us use now T2 = 10 ns (10 µs) for GaAs and 2 2 Si. In order to avoid significant spin dephasing during the interdot charge transfer, the time of the latter has As pointed out in the main text, in the limit of weak noise ∗ corrections comes from noise dynamics, which in the adi- to be much shorter than T2 . Assuming that the range of detuning sweep corresponding to the transfer is ∼ 1 abatic frame modifies off-diagonal element of Hamilto- meV, the sweep rates have to fulfill v 0.1 µeV/ns for nian (6), written explicitly as Si, and v 100 µeV/ns in GaAs. In Fig. 5 we see that sin θδ˙ + cos θδt˙ it means that in GaAs this lower bound on v severely re- θ˙ ≈ , (A2) stricts the possibility of lowering Q by making the trans- Ω0 fer slower, and in fact a tradeoff between amount of spin 2 2 dephasing and a finite value of Q due to Landau-Zener where Ω0 = + tc , cos θ = −/Ω0, sin θ = tc/Ω0 and δ˙ = ∂τ δ. effect that dominates the behavior of Q(v) for v ≥ 100 p µeV/ns has to be made. In silicon, the lower bound on v is much smaller than vopt, so a local minimum of Q visible in the Figure is attainable - but the viability of 1. Leading order perturbation theory strategy of lowering Q by using v <1 µeV/ns depends on the efficiency of electron relaxation due to charge noise We evaluate the leading order excitation probabil- and electron-phonon relaxation (compare red and blue ity Q(1) due to the noisy term. We use first order lines, corresponding to Si/SiGe and SiMOS in the Fig- time-dependent perturbation theory in adiabatic basis ∗ ˜ ure) and an exact value of T2 . All these observations |ψ(τ)i = a−(τ) |−, θi + a+(τ) |+, θi, and assuming a+ = (1) 2 (2) suggest that from the point of view of coherent transfer λa+ +λ a+ +..., we compute leading order the correc- of a spin qubit, Si-based quantum dot architectures could (1) tion to occupation of excited state as Q(1) = h|a (τ)|2i, have an advantage over GaAs-based ones. + which equals Let us finish by stressing the main message follow- ∞ R τ1 0 0 ing from our calculations for realistic GaAs and Si-based (1) 1 i Ω0(τ )dτ Q = θ˙(τ1)θ˙(τ2) e τ2 dτ1dτ2, (A3) quantum dots: the dynamics of inter-dot electron trans- 4 −∞ fer is very strongly affected by electron’s interaction with Z D E charge noise in Si-based systems and phonons in case of where h...i denotes classical averaging over noise realiza- GaAs-based ones. Effects of energy exchange with these tions. The substitution of Eq. (25) into Eq. (A3) results environments have to be taken into account to correctly in four distinct contributions: describe the basic physics of electron transfer in the cur- (1) (1) (1) (1) (1) rently available devices. More subtle effects appearing in Q = Q + Qtt + Qt + Qt , (A4) closed-system description, associated with spin-orbit and valley-orbit (in case of Si) interactions, could become rel- which correspond to auto- or cross-correlation function of (1) evant if levels of charge noise are significantly suppressed respective noise derivative Qxy ∼ hδx˙(τ1)δy˙(τ2)i, where compared to the currently encountered ones. x, y = or t. For assumed here stationary noises, it 17 is convenient to use the Fourier transform of correla- 1/2, which leads to: tion function, which for the noise derrivative can be −tc 2 2 −1/2 expressed in terms of power spectral density of noise 1 tc tc ∞ −iωτ (1) − (PSD) Sxy(ω) = hδx(τ)δy(0)ie dτ, i.e. Sx˙ y˙ (ω) ≡ Q = dωS(ω) 2 1 2 −∞ 2v −∞ ω ω 2 Z ω Sxy(ω) [76]. This allows us to write the correction as −1/2 R 1 −tc ω2 − t2 t2 Q(1) = dωS(ω) c 1 − c , (A10) ∞ tt 2v ω2 ω2 (1) dω ∗ −∞ Q = Sxy(ω) Fx(ω)F (ω), (A5) Z xy 8π y Z−∞ where the strictly negative value of ω, reflects absorption of energy quanta. Finally we conclude by showing in which we introduced filtering function Fx(ω) using that cross-correlation is negligibly small. We use the ar- Eq. (A3): ∗ gument that Ft = −Ft is strictly imaginary, and as a ∞ τ result we have ω 0 0 Fx(ω) = dτ fx(τ) exp iωτ + i Ω(τ )dτ , ∞ (1) (1) −∞ Ω(τ) 0 ∗ Qt + Qt = (St(ω) − St(ω))F(ω)Ft (ω), (A11) Z Z (A6) Z−∞ with f(τ) = sin θ = tc/Ω and ft(τ) = cos θ = −vτ/Ω. where the integrand is equivalent to imaginary part of cross-spectrum, and hence vanishes for hδ(τ)δti = hδtδi. Non-trivial imaginary part of cross-spectrum 2. Stationary Phase approximation results only from causal relation between δt, δ [115], however even in such special case we argue that ∗ We evaluate the integral (A6) for x = , t, in lead- F(ω)Ft (ω) ∝ cos(ϕ + π/4) sin(ϕ + π/4) which due to ing order of stationary phase approximation, where we zero average is expected to be much smaller than auto- seek for time at which the argument of exponent ϕ(τ) = correlation contributions. As a result corrections to oc- τ 0 cupation of excited state due to weak classical noise can ωτ + 0 Ω(τ )dτ is stationary, i.e. ∂τ ϕ(τ) = 0, from that ω = −Ω(˜τ), which takes place at ±τ˜ = (ω2 − t2)/v2. be written as: R c Additionally since Ω ≥ tc, the ω is strictly negative and ∞ 2 1 S(−Ω) t smaller then −t . The second derivative,p of the phase Q(1) = dΩ c c 2 2 2 2 2 2 2v tc 1 − t /Ω Ω evaluated atτ ˜ reads ∂τ ϕ(τ)|τ=˜τ = ±v 1 − tc /ω . In Z ∞ 2 the leading order, the integral (A6) reads: (1) 1 p tc p Q = dΩ St(−Ω) 1 − , (A12) tt 2v Ω2 ∞ Ztc r tc v 2 2 2 F ≈ exp iϕ(˜τ) + i 1 − t /ω (τ − τ˜) + ω 2 c using which we recovered high frequency limit of [61] Z−∞ n v p o where lower bound of the integrals reflects the minimal + exp iϕ(−τ˜) − i 1 − t2/ω2(τ +τ ˜)2 dτ (A7) 2 c energy needed for the excitation to occur. Due to the n∞ o dominant role of longitudinal component δ we omit here −vτ˜ p v 2 2 2 Ft ≈ exp iϕ(˜τ) + i 1 − t /ω (τ − τ˜) + ω 2 c contributions from frequencies below tc, which are rele- Z−∞ n v p o vant only for transverse δt noise [60, 61]. In particu- − exp iϕ(−τ˜) − i 1 − t2/ω2(τ +τ ˜)2 dτ (A8) lar corrections from quasi-static noise in tunnel coupling 2 c vanishes in assumed here weak noise (δt tc) and adia- n p o 2 iax2 π batic (tc > v) limits [57]. Now we perform Gaussian integration, dxe = ia , using which integrand terms differ by a phase 1/ ± i = R p e∓π/4. Since ϕ(˜τ) = −ϕ(−τ˜) the result can be written 3. Transition rates p as The first order calculation can be interpreted as a prob- 2 −1/4 2t0 cos(ϕ(˜τ) − π/4) 2π tc ability of single transition from ground to excited state F(ω) = 1 − , ω v ω2 during adiabatic transfer, and as such can be written as r (1) 0 0 2 −1/4 an integral of transition rate Q = dτ Γ∞(τ ), see −2ivτ˜ sin(ϕ(˜τ) − π/4) 2π tc Ft(ω) = 1 − Eq. (15). An explicit form of Γ∞ can be deduced from ω v ω2 R r Eq. (A12) as (A9) 2 1 tc cl Γ∞,(τ) = S Ω(τ) First we consider diagonal part (x = y) of Eq. (A5), in 4 Ω(τ) 2 2 2 2 which |F| ∝ cos (ϕ − π/4) and |Ft| ∝ sin (ϕ − π/4). 2 1 tc cl Due to rapidly oscillating nature of both functions, we Γ∞,t(τ) = 1 − S Ω(τ) . (A13) 2 2 4 Ω(τ) t replace them by average values of cos ϕ and sin ϕ equal 18

(∞) magnitude lower Qt . For the white part of Johnson’s 100 1/f δ noise, distinction between different tc is much more vis- Johnson δt ible for noise in detuning (solid lines, filled dots), since Q larger tc significantly increases the time spent in vicinity

1 of avoided crossing ≈ 2tc/v, during which the longitudi- 10− 2 nal transitions Γ∞, ∝ (tc/Ω) are most eﬀective. In the case of noise in tunnel coupling the opposite is true, since larger tc only slightly decreases time spent outside of the 2 2 10− avoided crossing region, while Γ∞,t ∝ (vτ/Ω) .

Excited state occupation tc [µeV] 10 Appendix B: Details of phonon relaxation rate 3 10− 20

101 102 We now turn to evaluation of zero-temperature phonon Sweep rate v (µeV/ns) relaxation rate in more details. First we show how orbital and interdot phonon-related processes emerge when us- FIG. 9. Probability of occupying higher energy state af- ing the |±, θi basis of eigenstates of instantaneous Hamil- ter detuning sweep Q in presence of white (ﬁlled) and 1/f tonian. Next we investigate the elements for Gaussian (hollow dots) classical noise in detuning/tunneling as a func- choice of electron wavefunctions, and discuss the role of tion of sweep rate for tunnel couplings tc = 10µeV (green) harmonic and Hund-Muliken approximation. and 20µeV (red). Points correspond to numerical simula- tion of Schrodinger equation averaged over realization of classical noise process. Lines correspond to analytical expres- ( ) 1. Interdot and orbital processes sion Q ∞ with the rates calculated according to Eq. (A13) for the noise in detunning δ (solid) and tunnel coupling δt (dashed). For illustration we used arbitrary parameters We start with the Phonon spectral density, given by 2 2 ikr 2 for detuning noise S0.1T/T0.1 = (1.5) µeV /Hz (1/f) and Eq. (34), which predicts S(ω) ∝ | h−| e |+i | . We now 2 2 2JkBT = (0.3) µeV /Hz (White part of Johnsons noise). The evaluate the matrix element, by pluging in adiabatic basis tunnel coupling ﬂuctuations are reduced by a factor of 10, i.e. given by Eq. (3), which results in: S(ω) = (102)St(ω). To emulate high temperature limit we set terminal sweep rate to f = 100µeV, which corresponds ikr of thermal energy at T 1.2K . h−, θ| e |+, θi = ≈ cos θ Re hL| eikr |Ri + Im hL| eikr |Ri 1 Finally, we prove that a result obtained by substituting + sin θ(hL| eikr |Li − hR|eikr |Ri), (B1) 2 Γ± = Γ∞ = Γ∞, + Γ∞,t into rate equation Eq. (12), which results in high temperature solution: where tan θ = −/tc, and consecutive terms corresponds

τf to interdot (ˆσx,σ ˆy) and orbitalσ ˆz coupling in dots ba- (∞) 1 0 0 Q = 1 − exp −2 Γ∞(τ )dτ , (A14) sis respectively. In absence of large magnetic field in 2 Zτi z direction, wavefunctions can be assumed real, hence Im hL| eikr |Ri = 0. The exact form of matrix ele- is equivalent to an evolution driven by Hamiltonian ment depends on the assumed form of wavefunctions, i.e. Eq. (1), averaged over realizations of classical fluctua- ψ (r) = hr|L/R i, which will be investigated below. tions of parameters. In Fig. 9 we have separately plot- L/R ted contributions from detuning noise δ (solid line), and tunnel coupling noise δt (dashed line), as a result of 1/f noise (hollow dots) and white noise (filled dots). Inde- 2. Hund-Mulliken approximation pendently of the considered noise type, in the fast sweep 2 rate limit (v tc ) we recover the Landau-Zener solu- Before invoking concrete form of wavefunction of an tion, for which QLZ depends on the relation between v electron localized in a QD, let us comment on the so and tc only, and thus for sufficiently large v the results called Hund-Mulliken approximation, in which one uses group according to tunnel couplings tc = 10µeV (green) orthogonalized orbitals |L/Ri = N (|L0/R0i−g |R0/L0i) and tc = 20µeV (red). In the low sweep rate limit, for built from bare wavefunction of electrons in isolated 1/f noise in detuning (solid line, hollow dots) we obtain quantum dots: |L0/R0i, with N being normalization (∞) (∞) results from [39], for which Q (Q with Γ∞ = Γ∞,) constant. The parameter g is a function of the overlap is independent of tc. The same applies to 1/f noise in l = hL0|R0i 1, with its value given by the orthogonal- tunnel coupling (dashed lines, hollow dots), for which ity condition: a 10-fold decreased noise amplitude (compared to the case of detuning noise) translates into almost 2 orders of hL|Ri = N 2(l − 2g + lg2) = 0 (B2) 19 √ 2 from which g = (1 − 1 − l )/l ∼ l/2 for l 1. Con- rections (ΨL0/R0 (r) = ψL0/R0 (x, y)ψz(z)) and in has a sistently we concentrate on leading order in g or l, ac- Gaussian shape: cording to which and 1 = hL|Li = N 2(1 − 2gl + g2) we N ∼ | i 1 (x ± ∆x/2)2 + y2 take 1. Assuming real wavefunction L0/R0 , we ψ (x, y) = exp − substitute the orthogonalized states into Eq. (B1) from L0/R0 (π2r4 )1/4 2r2 xy xy which we obtain: 1 z2 ψz(z) = exp − , (B4) (πr2)1/4 2r2 h−| eikr |+i ∼ z z

ikr ikr ikr such that for electron wavefunction, FWHMx ≈ 2rxy and ∼ cos θ hL0| e |R0i − g hL0| e |L0i + hR0| e |R0i FWHMz ≈ 2rz. In such case Eq. (B3) reads: 1 ikr ikr 2 2 2 2 + sin θ hL0| e |L0i − hR0| e |R0i , (B3) k r + k r 2 h−| eikr |+i ∼ exp − xy xy z z × ( 4 ) where in the latter term correction linear in the overlap ∆x2 cos θ exp − 1 − cos(kx∆x/2) + g cancels. 4r2 xy − i sin θ sin(kx∆x/2) , (B5) 3. Harmonic approximation where we used that in harmonic approximation g = l/2 = 2 2 1 −∆x /4rxy Finally we substitute concrete form of isolated wave- 2 e . Interdot and orbital relaxation are given functions, and evaluate matrix element h−| eikr |+i. We by real and imaginary part of above matrix element and assume the wavefunction is indepedent in all three di- hence cause relaxation independently.

[1] R. Hanson, L. P. Kouwenhoven, J. R. Petta, S. Tarucha, ing silicon-based quantum computing using cmos tech- and L. M. K. Vandersypen, Spins in few-electron quan- nology: State-of-the-art, challenges and perspectives, tum dots, Rev. Mod. Phys. 79, 1217 (2007). arXiv:2011.11753 (2020). [2] C. Volk, A. M. J. Zwerver, U. Mukhopadhyay, P. T. [9] E. Chanrion, D. J. Niegemann, B. Bertrand, C. Spence, Eendebak, C. J. van Diepen, J. P. Dehollain, T. Hens- B. Jadot, J. Li, P.-A. Mortemousque, L. Hutin, R. Mau- gens, T. Fujita, C. Reichl, W. Wegscheider, and L. M. K. rand, X. Jehl, M. Sanquer, S. De Franceschi, C. B Vandersypen, Loading a quantum-dot based “Qubyte” ¨auerle,F. Balestro, Y.-M. Niquet, M. Vinet, T. Meu- register, npj Quantum Information 5, 29 (2019). nier, and M. Urdampilleta, Charge detection in an array [3] F. Kuemmeth and H. Bluhm, Roadmap for gaas spin of CMOS quantum dots, arXiv:2004.01009 (2020). qubits, arXiv:2011.13907 (2020). [10] L. M. K. Vandersypen, H. Bluhm, J. S. Clarke, A. S. [4] F. Watson, S. G. J. Philips, E. Kawakami, D. R. Ward, Dzurak, R. Ishihara, A. Morello, D. J. Reilly, L. R. P. Scarlino, M. Veldhorst, D. E. Savage, M. G. La- Schreiber, and M. Veldhorst, Interfacing Spin Qubits in gally, M. Friesen, S. N. Coppersmith, M. A. Eriksson, Quantum Dots and Donors—Hot, Dense, and Coherent, and L. M. K. Vandersypen, A programmable two-qubit npj Quantum Information 3, 34 (2017). quantum processor in silicon, Nature 555, 633 (2018). [11] P. Scarlino, D. J. van Woerkom, U. C. Mendes, J. V. [5] W. I. L. Lawrie, H. G. J. Eenink, N. W. Hendrickx, J. M. Koski, A. J. Landig, C. K. Andersen, S. Gasparinetti, Boter, L. Petit, S. V. Amitonov, M. Lodari, B. Paque- C. Reichl, W. Wegscheider, K. Ensslin, T. Ihn, A. Blais, let Wuetz, C. Volk, S. G. J. Philips, G. Droulers, and A. Wallraﬀ, Coherent microwave-photon-mediated N. Kalhor, F. van Riggelen, D. Brousse, A. Sammak, coupling between a semiconductor and a superconduct- L. M. K. Vandersypen, G. Scappucci, and M. Veldhorst, ing qubit, Nature Communications 10, 3011 (2019). Quantum dot arrays in silicon and germanium, Applied [12] X. Mi, J. V. Cady, D. M. Zajac, P. W. Deelman, and Physics Letters 116, 080501 (2020). J. R. Petta, Strong Coupling of a Single Electron in [6] M. Veldhorst, H. G. Eenink, C. H. Yang, and A. S. Silicon to a Microwave Photon, Science 355, 156 (2017). Dzurak, Silicon CMOS Architecture for a Spin-Based [13] M. Benito, X. Mi, J. M. Taylor, J. R. Petta, and Quantum Computer, Nature Communications 8, 1766 G. Burkard, Input-output theory for spin-photon cou- (2017). pling in Si double quantum dots, Phys. Rev. B 96, [7] W. Huang, C. H. Yang, K. W. Chan, T. Tanttu, 235434 (2017). B. Hensen, R. C. Leon, M. A. Fogarty, J. C. Hwang, [14] X. Mi, M. Benito, S. Putz, D. M. Zajac, J. M. Taylor, F. E. Hudson, K. M. Itoh, A. Morello, A. Laucht, and G. Burkard, and J. R. Petta, A Coherent Spin-Photon A. S. Dzurak, Fidelity Benchmarks for Two-Qubit Gates Interface in Silicon, Nature 555, 599 (2018). in Silicon, Nature 569, 532 (2019). [15] N. Samkharadze, G. Zheng, N. Kalhor, D. Brousse, [8] M. F. Gonzalez-Zalba, S. de Franceschi, E. Char- A. Sammak, U. C. Mendes, A. Blais, G. Scappucci, and bon, T. Meunier, M. Vinet, and A. S. Dzurak, Scal- L. M. K. Vandersypen, Strong Spin-Photon Coupling in 20

Silicon, Science 359, 1123 (2018). [30] X. Li, E. Barnes, J. P. Kestner, and S. Das Sarma, In- [16] S. Hermelin, S. Takada, M. Yamamoto, S. Tarucha, trinsic errors in transporting a single-spin qubit through A. D. Wieck, L. Saminadayar, C. Baüerle,and T. Me- a double quantum dot, Phys. Rev. A 96, 012309 (2017). unier, Electrons surfing on a sound wave as a platform [31] F. Ginzel, A. R. Mills, J. R. Petta, and G. Burkard, for quantum optics with flying electrons, Nature 477, Spin shuttling in a silicon double quantum dot, Phys. 435 (2011). Rev. B 102, 195418 (2020). [17] R. P. G. McNeil, M. Kataoka, C. J. B. Ford, C. H. W. [32] M. Friesen, S. Chutia, C. Tahan, and S. N. Copper- Barnes, D. Anderson, G. A. C. Jones, I. Farrer, and smith, Valley splitting theory of sige/si/sige quantum D. A. Ritchie, On-demand single-electron transfer be- wells, Phys. Rev. B 75, 115318 (2007). tween distant quantum dots, Nature 477, 439 (2011). [33] D. Culcer, X. Hu, and S. Das Sarma, Interface rough- [18] B. Bertrand, S. Hermelin, S. Takada, M. Yamamoto, ness, valley-orbit coupling, and valley manipulation in S. Tarucha, A. Ludwig, A. D. Wieck, C. Bäuerle,and quantum dots, Phys. Rev. B 82, 205315 (2010). T. Meunier, Fast spin information transfer between dis- [34] F. A. Zwanenburg, A. S. Dzurak, A. Morello, M. Y. tant quantum dots using individual electrons, Nature Simmons, L. C. L. Hollenberg, G. Klimeck, S. Rogge, Nanotechnology 11, 672 (2016). S. N. Coppersmith, and M. A. Eriksson, Silicon quan- [19] S. Takada, H. Edlbauer, H. V. Lepage, J. Wang, P.-A. tum electronics, Rev. Mod. Phys. 85, 961 (2013). Mortemousque, G. Georgiou, C. H. W. Barnes, C. J. B. [35] X. Zhao and X. Hu, Toward high-fidelity coherent elec- Ford, M. Yuan, P. V. Santos, X. Waintal, A. Lud- tron spin transport in a gaas double quantum dot, Sci. wig, A. D. Wieck, M. Urdampilleta, T. Meunier, and Rep. 8, 13968 (2018). C. Bäuerle,Sound-driven single-electron transfer in a [36] X. Zhao and X. Hu, Coherent electron trans- circuit of coupled quantum rails, Nature Communica- port in silicon quantum dots, arXiv:1803.00749 tions 10, 4557 (2019). arXiv:1803.00749v1 (2018). [20] P.-A. Mortemousque, B. Jadot, E. Chanrion, V. Thiney, [37] E. Cota and S. E. Ulloa, Spin-orbit interaction and con- C. Bäuerle,A. Ludwig, A. D. Wieck, M. Urdampil- trolled singlet-triplet dynamics in silicon double quan- leta, and T. Meunier, Enhanced spin coherence while tum dots, J. Phys.: Condens. Matter 30, 295301 (2018). displacing electron in a 2D array of quantum dots, [38] S. N. Shevchenko, A. I. Ryzhov, and F. Nori, Low- arXiv:2101.05968 (2021). frequency spectroscopy for quantum multi-level sys- [21] B. Jadot, P.-A. Mortemousque, E. Chanrion, V. Thiney, tems, Phys. Rev. B 98, 195434 (2018). A. Ludwig, A. D. Wieck, M. Urdampilleta, C. Bäuerle, [39] J. A. Krzywda and L. Cywiński,Adiabatic electron and T. Meunier, Distant spin entanglement via fast and charge transfer between two quantum dots in presence coherent electron shuttling, arXiv:2004.02727 (2020). of 1/f noise, Phys. Rev. B 101, 035303 (2020). [22] T. A. Baart, M. Shafiei, T. Fujita, C. Reichl, [40] B. Buonacorsi, B. Shaw, and J. Baugh, Simulated co- W. Wegscheider, and L. M. K. Vandersypen, Single-Spin herent electron shuttling in silicon quantum dots, Phys. CCD, Nature Nanotechnology 11, 330 (2016). Rev. B 102, 125406 (2020). [23] T. Fujita, T. A. Baart, C. Reichl, W. Wegscheider, and [41] R. K. Malla, V. Y. Chernyak, and N. A. Sinitsyn, Nona- L. M. K. Vandersypen, Coherent shuttle of electron-spin diabatic transitions in Landau-Zener grids: Integrability states, npj Quantum Information 3, 22 (2017). and semiclassical theory, arXiv:2101.04169 (2021). [24] H. Flentje, P.-A. Mortemousque, R. Thalineau, A. Lud- [42] M. Yamaguchi, T. Yuge, and T. Ogawa, Markovian wig, A. D. Wieck, C. Bäuerle,and T. Meunier, Coherent quantum master equation beyond adiabatic regime, long-distance displacement of individual electron spins, Phys. Rev. E 95, 012136 (2017). Nature Communications 8, 501 (2017). [43] L. Arceci, S. Barbarino, R. Fazio, and G. E. Santoro, [25] A. R. Mills, D. M. Zajac, M. J. Gullans, F. J. Schupp, Dissipative Landau-Zener problem and thermally as- T. M. Hazard, and J. R. Petta, Shuttling a Sin- sisted Quantum Annealing, Phys. Rev. B 96, 054301 gle Charge across a One-Dimensional Array of Sili- (2017). con Quantum Dots, Nature Communications 10, 1063 [44] A. J. Leggett, S. Chakravarty, A. T. Dorsey, M. P. A. (2019). Fisher, A. Garg, and W. Zwerger, Dynamics of the dissi- [26] T. Nakajima, M. R. Delbecq, T. Otsuka, S. Amaha, pative two-state system, Rev. Mod. Phys. 59, 1 (1987). J. Yoneda, A. Noiri, K. Takeda, G. Allison, A. Ludwig, [45] S. Javanbakht, P. Nalbach, and M. Thorwart, Dissipa- A. D. Wieck, X. Hu, F. Nori, and S. Tarucha, Coherent tive Landau-Zener quantum dynamics with transversal transfer of electron spin correlations assisted by dephas- and longitudinal noise, Phys. Rev. A 91, 052103 (2015). ing noise, Nature Communications 9, 2133 (2018). [46] Z. Huang and Y. Zhao, Dynamics of dissipative Landau- [27] C. J. van Diepen, T.-K. Hsiao, U. Mukhopadhyay, Zener transitions, Phys. Rev. A 97, 013803 (2018). C. Reichl, W. Wegscheider, and L. M. K. Vandersypen, [47] D. Zueco, P. Hänggi,and S. Kohler, Landau–Zener tun- Electron cascade for distant spin readout, Nature Com- nelling in dissipative circuit QED, New J. Phys. 10, munications 12, 77 (2021). 115012 (2008). [28] J. Yoneda, W. Huang, M. Feng, C. H. Yang, K. W. [48] K. Saito, M. Wubs, S. Kohler, Y. Kayanuma, Chan, T. Tanttu, W. Gilbert, R. C. C. Leon, F. E. Hud- and P. Hänggi,Dissipative Landau-Zener transitions son, K. M. Itoh, A. Morello, S. D. Bartlett, A. Laucht, of a qubit: Bath-specific and universal behavior, A. Saraiva, and A. S. Dzurak, Coherent spin qubit trans- Phys. Rev. B 75, 214308 (2007). port in silicon, arXiv:2008.04020 . [49] M. Wubs, K. Saito, S. Kohler, P. Hänggi, and [29] S. N. Shevchenko, S. Ashhab, and F. Nori, Landau- Y. Kayanuma, Gauging a Quantum Heat Bath with Dis- Zener-Stückelberg Interferometry, Phys. Rep. 492, 1 sipative Landau-Zener Transitions, Phys. Rev. Lett. 97, (2010). 200404 (2006). 21

[50] P. Ao and J. Rammer, Quantum dynamics of a two- [70] V. Langrock et al., Proposal of a scalable quantum bus state system in a dissipative environment, Phys. Rev. B device for coherent mid-range qubit transfer in disor- 43, 5397 (1991). dered si/sige/sio2, unpublished (2021). [51] V. L. Pokrovsky and D. Sun, Fast quantum noise in [71] M. Demirplak and S. A. Rice, Adiabatic Population the Landau-Zener transition, Phys. Rev. B 76, 024310 Transfer with Control Fields, J. Phys. Chem. A 107, (2007). 9937 (2003). [52] Y. Kayanuma and H. Nakayama, Nonadiabatic Transi- [72] S. Shevchenko, S. Ashhab, and F. Nori, Lan- tion at a Level Crossing with Dissipation, Phys. Rev. B dau–Zener–Stückelberg interferometry, Phys. Rep. 492, 57, 13099 (1998). 1 (2010). [53] P. Nalbach, Adiabatic-Markovian bath dynamics at [73] R. J. Schoelkopf, A. A. Clerk, S. M. Girvin, K. W. avoided crossings, Phys. Rev. A 90, 042112 (2014). Lehnert, and M. H. Devoret, Qubits as spectrometers [54] A. Benseny and K. Mølmer, All’s well that ends well: of quantum noise, in Quantum Noise in Mesoscopic The unexpectedly good performance of adiabatic pas- Physics, edited by Y. V. Nazarov (Kluwer, Dordrecht, sage, arXiv:2010.05093 (2020). 2003) pp. 175–203, (cond-mat/0210247). [55] H. Chen and D. A. Lidar, Why and When Pausing is [74] E. Paladino, Y. M. Galperin, G. Falci, and B. L. Alt- Beneficial in Quantum Annealing, Phys. Rev. Appl. 14, shuler, 1/f noise: Implications for solid-state quantum 014100 (2020). information, Rev. Mod. Phys. 86, 361 (2014). [56] Y. Kayanuma, Nonadiabatic Transitions in Level Cross- [75] F. Marquardt and V. A. Abalmassov, Spin relaxation in ing with Energy Fluctuation. I. Analytical Investiga- a quantum dot due to Nyquist noise, Phys. Rev. B 71, tions, Journal of the Physical Society of Japan 53, 108 165325 (2005). (1984). [76] P. Huang and X. Hu, Spin Relaxation due to Charge [57] Y. Kayanuma, Stochastic Theory for Nonadiabatic Noise, Phys. Rev. B 89, 195302 (2014). Level Crossing with Fluctuating Off-Diagonal Coupling, [77] U. Weiss, Quantum Dissipative Systems (World Scien- Journal of the Physical Society of Japan 54, 2037 tific, Singapore, 1999). (1985). [78] A. Hollmann, T. Struck, V. Langrock, A. Schmidbauer, [58] A. Dodin, S. Garmon, L. Simine, and D. Segal, Landau- F. Schauer, T. Leonhardt, K. Sawano, H. Riemann, Zener transitions mediated by an environment: Popu- N. V. Abrosimov, D. Bougeard, and L. R. Schreiber, lation transfer and energy dissipation, J. Chem. Phys. Large, tunable valley splitting and single-spin relaxation 140, 124709 (2014). mechanisms in a Si/sixge1 x quantum dot, Phys. Rev. [59] V. L. Pokrovsky and N. A. Sinitsyn, Fast noise in the Applied 13, 034068 (2020).− Landau-Zener theory, Phys. Rev. B 67, 144303 (2003). [79] P. Szańkowski, G. Ramon, J. Krzywda, D. Kwiatkowski, [60] Z. X. Luo and M. E. Raikh, Landau-Zener transition andL. Cywiński, Environmental noise spectroscopy driven by slow noise, Phys. Rev. B 95, 10.1103/Phys- with qubits subjected to dynamical decoupling, Jour- RevB.95.064305 (2017). nal of Physics: Condensed Matter 29, 333001 (2017). [61] R. K. Malla, E. G. Mishchenko, and M. E. Raikh, Sup- [80] Y.-C. Yang, S. N. Coppersmith, and M. Friesen, Achiev- pression of the Landau-Zener Transition Probability by ing high-fidelity single-qubit gates in a strongly driven Weak Classical Noise, Phys. Rev. B 96, 075419 (2017). charge qubit with 1/f charge noise, npj Quantum Infor- [62] N. A. Sinitsyn and N. Prokof’ev, Nuclear spin bath mation 5, 12 (2019). effects on Landau-Zener transitions in nanomagnets, [81] X. Mi, S. Kohler, and J. R. Petta, Landau-Zener In- Phys. Rev. B 67, 134403 (2003). terferometry of Valley-Orbit States in Si/SiGe Double [63] J. I. Vestgarden, J. Bergli, and Y. M. Galperin, Non- Quantum Dots, Phys. Rev. B 98, 161404 (2018). linearly driven Landau-Zener transition with telegraph [82] T. Struck, A. Hollmann, F. Schauer, O. Fedorets, noise, Phys. Rev. B 77, 014514 (2008). A. Schmidbauer, K. Sawano, H. Riemann, N. V. Abrosi- [64] M. H. S. Amin and D. V. Averin, Macroscopic Reso- mov,L. Cywiński,D. Bougeard, and L. R. Schreiber, nant Tunneling in the Presence of Low Frequency Noise, Low-frequency spin qubit energy splitting noise in Phys. Rev. Lett. 100, 197001 (2008). highly purified 28Si/SiGe, npj Quantum Information 6, [65] X. You, A. A. Clerk, and J. Koch, Positive- and 40 (2020). negative-frequency noise from an ensemble of two-level [83] B. M. Freeman, J. S. Schoenfield, and H. Jiang, Com- fluctuators, Phys. Rev. Research 3, 013045 (2021). parison of Low Frequency Charge Noise in Identically [66] T. Albash, S. Boixo, Daniel A Lidar, and Paolo Zanardi, Patterned Si/SiO2and Si/SiGe Quantum Dots, Appl. Quantum adiabatic Markovian master equations, New Phys. Lett. 108, 253108 (2016). J. Phys. 14, 123016 (2012). [84] E. J. Connors, J. J. Nelson, H. Qiao, L. F. Edge, and [67] M. Vogelsberger, D. A. Garanin, and R. Schilling, J. M. Nichol, Low-frequency charge noise in Si/SiGe Butterfly hysteresis curve is a signature of adiabatic quantum dots, Phys. Rev. B 100, 165305 (2019). Landau-Zener transition, Phys. Rev. B 73, 092412 [85] L. Petit, J. M. Boter, H. G. Eenink, G. Droulers, M. L. (2006). Tagliaferri, R. Li, D. P. Franke, K. J. Singh, J. S. Clarke, [68] P. Haikka and K. Molmer, Dissipative Landau-Zener R. N. Schouten, V. V. Dobrovitski, L. M. Vandersypen, level crossing subject to continuous measurement: Ex- and M. Veldhorst, Spin Lifetime and Charge Noise in citation despite decay, Phys. Rev. A 89, 052114 (2014). Hot Silicon Quantum Dot Qubits, Phys. Rev. Lett. 121, [69] A. A. Clerk, M. H. Devoret, S. M. Girvin, F. Marquardt, 076801 (2018). and R. J. Schoelkopf, Introduction to quantum noise, [86] J.-S. Kim, T. M. Hazard, A. A. Houck, and S. A. Lyon, measurement, and amplification, Rev. Mod. Phys. 82, A low-disorder metal-oxide-silicon double quantum dot, 1155 (2010). Applied Physics Letters 114, 043501 (2019). 22

[87] L. Kranz, S. K. Gorman, B. Thorgrimsson, Y. He, [102] F. K. Malinowski, F. Martins,L.Cywiński,M. S. Rud- D. Keith, J. G. Keizer, and M. Y. Simmons, Exploiting ner, P. D. Nissen, S. Fallahi, G. C. Gardner, M. J. a Single-Crystal Environment to Minimize the Charge Manfra, C. M. Marcus, and F. Kuemmeth, Spectrum Noise on Qubits in Silicon, Advanced Materials 32, of the Nuclear Environment for GaAs Spin Qubits, 2003361 (2020). Phys. Rev. Lett. 118, 177702 (2017). [88] E. J. Connors, J. J. Nelson, and J. M. Nichol, [103] V. Srinivasa, K. C. Nowack, M. Shafiei, L. M. K. Van- Charge-noise spectroscopy of Si/SiGe quantum dots dersypen, and J. M. Taylor, Simultaneous Spin-Charge via dynamically-decoupled exchange oscillations, Relaxation in Double Quantum Dots, Phys. Rev. Lett. arXiv:2103.0244 (2021). 110, 196803 (2013). [89] R. M. Jock, N. T. Jacobson, M. Rudolph, D. R. [104] K. Wang, C. Payette, Y. Dovzhenko, P. W. Deelman, Ward, M. S. Carroll, and D. R. Luhman, A sili- and J. R. Petta, Charge Relaxation in a Single-Electron con singlet-triplet qubit driven by spin-valley coupling, Si / SiGe Double Quantum Dot, Phys. Rev. Lett. 111, arXiv:2102.12068 (2021). 046801 (2013). [90] O. E. Dial, M. D. Shulman, S. P. Harvey, H. Bluhm, [105] J. Medford, J. Beil, J. M. Taylor, E. I. Rashba, H. Lu, V. Umansky, and A. Yacoby, Charge Noise Spectroscopy A. C. Gossard, and C. M. Marcus, Quantum-dot-based Using Coherent Exchange Oscillations in a Singlet- resonant exchange qubit, Phys. Rev. Lett. 111, 050501 Triplet Qubit, Phys. Rev. Lett. 110, 146804 (2013). (2013). [91] P. Cerfontaine, T. Botzem, J. Ritzmann, S. S. Humpohl, [106] K. Ono, T. Mori, and S. Moriyama, High-temperature A. Ludwig, D. Schuh, D. Bougeard, A. D. Wieck, and operation of a silicon qubit, Scientific Reports 9, 469 H. Bluhm, Closed-loop control of a GaAs-based singlet- (2019). triplet spin qubit with 99.5% gate fidelity and low leak- [107] L. Petit, M. Russ, H. G. J. Eenink, W. I. L. Lawrie, age, Nat. Comm. 11, 4144 (2020). J. S. Clarke, L. M. K. Vandersypen, and M. Veldhorst, [92] L. Petit, H. G. J. Eenink, M. Russ, W. I. L. Lawrie, High-fidelity two-qubit gates in silicon above one Kelvin, N. W. Hendrickx, S. G. J. Philips, J. S. Clarke, L. M. K. arXiv:2007.09034 (2020). Vandersypen, and M. Veldhorst, Universal quantum [108] C. H. Yang, R. C. C. Leon, J. C. C. Hwang, A. Saraiva, logic in hot silicon qubits, Nature 580, 355 (2020). T. Tanttu, W. Huang, J. Camirand Lemyre, K. W. [93] K. W. Chan, W. Huang, C. H. Yang, J. C. C. Hwang, Chan, K. Y. Tan, F. E. Hudson, K. M. Itoh, A. Morello, B. Hensen, T. Tanttu, F. E. Hudson, K. M. Itoh, M. Pioro-Ladrière,A. Laucht, and A. S. Dzurak, Oper- A. Laucht, A. Morello, and A. S. Dzurak, Assessment of ation of a silicon quantum processor unit cell above one a silicon quantum dot spin qubit environment via noise kelvin, Nature 580, 350 (2020). spectroscopy, Phys. Rev. Appl. 10, 044017 (2018). [109] M. Abramowitz and I. A. Stegun, Handbook of Mathe- [94] T. Nakajima, M. R. Delbecq, T. Otsuka, S. Amaha, matical Functions with Formulas, Graphs, and Mathe- J. Yoneda, A. Noiri, K. Takeda, G. Allison, A. Ludwig, matical Tables, ninth dover printing, tenth gpo printing A. D. Wieck, X. Hu, F. Nori, and S. Tarucha, Coherent ed. (Dover, New York City, 1964). transfer of electron spin correlations assisted by dephas- [110] E. A. Chekhovich, M. N. Makhonin, A. I. Tartakovskii, ing noise, Nat. Comm. 9, 2133 (2018). A. Yacoby, H. Bluhm, K. C. Nowack, and L. M. K. Van- [95] Z. Shi, C. B. Simmons, D. R. Ward, J. R. Prance, R. T. dersypen, Nuclear spin effects in semiconductor quan- Mohr, T. S. Koh, J. K. Gamble, X. Wu, D. E. Savage, tum dots, Nature Materials 12, 494 (2013). M. G. Lagally, M. Friesen, S. N. Coppersmith, and M. A. [111] L. V. C. Assali, H. M. Petrilli, R. B. Capaz, B. Koiller, Eriksson, Coherent quantum oscillations and echo mea- X. Hu, and S. D. Sarma, Hyperfine interactions in silicon surements of a Si charge qubit, Phys. Rev. B 88, 075416 quantum dots, Phys. Rev. B 83, 165301 (2011). (2013). [112] B. M. Maune, M. G. Borselli, B. Huang, T. D. [96] M. Raith, P. Stano, F. Baruffa, and J. Fabian, Theory of Ladd, P. W. Deelman, K. S. Holabird, A. A. Kise- spin relaxation in two-electron lateral coupled quantum lev, I. Alvarado-Rodriguez, R. S. Ross, A. E. Schmitz, dots, Phys. Rev. Lett. 108, 246602 (2012). M. Sokolich, C. A. Watson, M. F. Gyure, and A. T. [97] P. Y. Yu and M. Cardona, Fundamentals of Semicon- Hunter, Coherent singlet-triplet oscillations in a silicon- ductors, Graduate Texts in Physics (Springer Berlin based double quantum dot, Nature 481, 344 (2012). Heidelberg, Berlin, Heidelberg, 2010). [113] E. Kawakami, P. Scarlino, D. R. Ward, F. R. Braak- [98] G. Burkard, D. Loss, and D. P. DiVincenzo, Coupled man, D. E. Savage, M. G. Lagally, M. Friesen, S. N. quantum dots as quantum gates, Phys. Rev. B 59, 2070 Coppersmith, M. A. Eriksson, and L. M. K. Vander- (1999). sypen, Electrical control of a long-lived spin qubit in [99] Q. Li, L. Cywiński,D. Culcer, X. Hu, and S. Das Sarma, a si/sige quantum dot, Nature Nanotechnology 9, 666 Exchange coupling in silicon quantum dots: Theoretical (2014). considerations for quantum computation, Phys. Rev. B [114] J. Yoneda, K. Takeda, T. Otsuka, T. Nakajima, M. R. 81, 085313 (2010). Delbecq, G. Allison, T. Honda, T. Kodera, S. Oda, [100] Z.-H. Wang and S. Takahashi, Spin decoherence and Y. Hoshi, N. Usami, K. M. Itoh, and S. Tarucha, electron spin bath noise of a nitrogen-vacancy center in A quantum-dot spin qubit with coherence limited by diamond, Phys. Rev. B 87, 115122 (2013). charge noise and fidelity higher than 99.9%, Natute [101] H. G. J. Eenink, L. Petit, W. I. L. Lawrie, J. S. Clarke, Nanotechnology 13, 102 (2018). L. M. K. Vandersypen, and M. Veldhorst, Tunable Cou- [115] P. Szańkowski, M. Trippenbach, andL.Cywiński,Spec- pling and Isolation of Single Electrons in Silicon Metal- troscopy of cross correlations of environmental noises Oxide-Semiconductor Quantum Dots, Nano Letters 19, with two qubits, Phys. Rev. A 94, 012109 (2016). 8653 (2019).