Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Characterizing Adiabaticity in Quantum Many-Body Systems at Finite Temperature

Characterizing Adiabaticity in Quantum Many-Body Systems at Finite Temperature A. H. Skelt Department of Physics, University of York, York YO10 5DD, United Kingdom I. D’Amico Department of Physics, University of York, UK and International Institute of Physics, Federal University of Rio Grande do Norte, Natal, Brazil (Dated: April 14, 2020) The quantum adiabatic theorem is fundamental to time dependent quantum systems, but being able to characterize quantitatively an adiabatic evolution in many-body systems can be a chal- lenge. This work demonstrates that the use of appropriate state and particle-density metrics is a viable method to quantitatively determine the degree of adiabaticity in the dynamic of a quantum many-body system. The method applies also to systems at finite temperature, which is impor- tant for quantum technologies and quantum thermodynamics related protocols. The importance of accounting for memory effects is discussed via comparison to results obtained by extending the quantum adiabatic criterion to finite temperatures: it is shown that this may produce false readings being quasi-Markovian by construction. As the proposed method makes it possible to characterize the degree of adiabatic evolution tracking only the system local particle densities, it is potentially applicable to both theoretical calculations of very large many-body systems and to experiments. INTRODUCTION are usually taken as the ground and first excited states. When it comes to accurately characterizing an adia- Adiabatic evolutions are important in many areas of batic evolution, there are though many challenges, such quantum physics, such as quantum computation, quan- as the complexity of calculations involving many-body tum thermodynamics, and quantum field theory [1–9]. systems and defining the criterion at finite temperature. One particularly important application of adiabatic evo- Recently, the validity and sufficiency of this QAC for cer- lutions is achieving specific (target) states (e.g. in adi- tain systems have been questioned [16–18], and new ap- abatic quantum computation, where the target state is proaches to characterizing adiabaticity have also come known to be the ground state of the final Hamiltonian). to light [19, 20] where they look at comparing the time Other important applications of adiabatic evolutions are evolved state of the system with the “adiabatic” state, i.e. for quantum thermodynamic cycles, where, for example, the instantaneous ground state. It was demonstrated in they may yield the highest extractable quantum work reference [19] that metrics can be used to characterize [10, 11]. Indeed, the relevance of adiabaticity has even adiabaticity through a variety of approaches to best suit given rise to new subfields, such as shortcuts to adia- the quantities one has at hand. The issue of tracking adi- baticity [12, 13] abaticity both at finite temperature and for many-body The quantum adiabatic theorem [14] defines an adia- systems remains outstanding. batic evolution as one in which no transitions between In this work the adiabatic theorem is written in terms energy levels occurs, and is a fundamental concept for of two distance measures (metrics), namely the Bures any time dependent quantum system. It was first pro- and the trace distances. The Bures distance is connected posed in 1928 by Born and Fock, and demonstrates that to the fidelity (which is though not a proper measure), for a quantum system to be considered adiabatic, it must and, at zero temperature, one can use the “adiabatic fi- be evolved slowly enough that it remains in an instanta- delity” as a figure of merit for adiabaticity in time de- neous eigenstate, with a gap between its eigenenergy and pendent systems [20]. Importantly, the Bures distance the rest of the Hamiltonian’s spectrum [14]. Later, Avron can be derived from conservation laws [21, 22] so that and Elgart relaxed this gap condition through a reformu- it can provide relevant information on the physics of the lation of the theorem [15]). At zero temperature, this is many-body system [21–24]. Within quantum informa- often interpreted mathematically with the quantum adi- tion processing, the trace distance is considered the best abatic criterion (QAC) [16–18]: measure to operationally distinguish two quantum states [25], so we also look at using the trace distance in place of hm(t)| H (t)|n(t)i the Bures, and find that both the trace and Bures can be 1, (1) used to determine the degree of adiabaticity. To provide a (|E (t)− E (t)|) n m comparison with a somewhat more familiar quantity, we where H is the time derivative of the Hamiltonian, |mi propose an extension of the QAC to finite temperatures, and |ni are the instantaneous eigenstates of H with in- and discuss its limitations. All of the above broadens stantaneous eigenenergies E and E respectively, and the choice of measures of adiabaticity to best suit one’s m n arXiv:2004.05842v1 [quant-ph] 13 Apr 2020 2 needs. with At zero temperature, the Bures distance for pure states E (t)− E (t) < sk T, (3) n 0 B has already been demonstrated to characterize adiabatic- E (t)− E (t) < s k T, (4) m 0 B ity in single electron systems [19], and was seen to have potential for characterizing adiabaticity in two-electron s > s ≥ 1. (5) systems [26]. Here we consider many-body systems, for- Here E ≤ E ≤ . . . ≤ E and m 6= n. In the calculations 0 1 n mally of any size, continuous and described over a lattice, presented here, we use s = 1 and do not cap s . at zero and finite temperatures. Computationally, as sys- For adiabaticity to hold, we still required that (t)  1. tem size increases exponentially, we apply the method to In T-QAC, the criterion is adapted for degenerate states many-body systems up to 6 electrons on a discrete lat- following Rigolin and Ortiz [29], so that the maximum tice. We take inspiration from density-functional theory distance between the degenerate subspaces and other lev- to ask: “can metrics based on the particle density alone els is considered when calculating (t). give quantitative guidance to the level of adiabaticity of a system?” Particle density is in principle experimentally observable and much easier to estimate than the corre- Metrics for density, n, and quantum state, ρ sponding many-body state, e.g. by density functional methods; by demonstrating that this question has a pos- Metrics provide a quantitative measure – the distance itive answer, we provide a manageable way to measure – to differentiate between two elements of a set [30], and and track adiabaticity of many-body systems, even if the must obey three axioms: positivity, D(x, y) ≥ 0 and temperature is finite. This work aims to help guide those wanting an adi- D(x, y) = 0 iff y = x; (6) abatic evolution (either experimentally or computation- ally) in many-body systems towards achieving an under- symmetry, D(x, y) = D(y, x); and the triangle inequality, standing of the degree of adiabaticity of their system. A D(x, z) ≤ D(x, y) + D(y, z). (7) guideline threshold for considering an evolution adiabatic is then presented with discussion of the factors which im- The use of metrics for investigating the relationship be- pact this threshold and the important quantities to con- tween wavefunctions and corresponding particle densities sider when deciding a threshold for one’s system. was developed in [21–23, 31] where the chosen metrics were derived from conservation laws (’natural’ metrics [22]) to ensure that they could provide physical insights. THEORY AND PROPOSED METHODS Reference [19] introduced a method of using these metrics for characterizing adiabaticity in single electron systems Temperature-dependent quantum adiabatic criterion at zero temperature; other works [21, 22, 26, 31] support the possibility of developing this metric-based method As mentioned previously, the QAC based on the quan- to characterize adiabaticity in many-body systems. All tum adiabatic theorem [14] has been under scrutiny re- these works considered pure states; since here the focus cently, and new approaches have come to light. However is also on finite temperature, the ‘natural’ metrics must all of these approaches only consider quantum adiabatic- be extended to mixed states. ity for pure states at zero temperature. A new expression The metric for the wavefunction developed in [21] is in is introduced here for characterizing adiabaticity in sys- fact the limit for zero temperature of the Bures metric, tems at finite temperature and so described by mixed which for mixed states reads states. First, however, one needs to define what is meant q 1/2 by being adiabatic at finite temperature. The require- √ √ D (σ, ρ) = 2 1− Tr ρσ ρ , (8) ment for quantum adiabaticity at finite temperature [27] is that there are no transitions between eigenstates of the where σ and ρ are quantum system states (density ma- system as it evolves [28]. Practically this implies that the trices) [32] population of the various eigenstates should not change A metric for quantum states which is widely used by with time. the quantum technology community as measure of distin- As a comparison between the metrics and a more fa- guishability between quantum states is the trace distance miliar quantity, we propose the following extension of [25, 33], which is defined as the QAC (1), valid at any temperature T (T-QAC) and which will include degeneracies q 1 1 D (ρ, σ) = Tr [|ρ− σ|] = Tr (ρ− σ) (ρ− σ). (9)   2 2   hm(t)| H (t)|n(t)i This metric will also be considered in this work. Bu- (t) = max (2) n,m   (|E (t)− E (t)|) n m res’ and trace distances are related by bounds [25] and, 3 at least for the systems and dynamics discussed in this The adiabatic threshold work, they provide very similar diagnostic; it is suggested then that the decision of which to use be made on which System quantum states quantities are more readily accessible, e.g. if the fidelity is easy to obtain, the Bures distance should be chosen. In practice, when can a system be considered adia- batic? A reasonable answer is ‘when, during the dy- The ‘natural’ density metric is unaffected by the type namics, the system remains close enough to its adiabatic of state, and remains limit’; in this section we will quantify the concept of ‘close enough’ using the tool of ‘adiabatic threshold’ [19]. D (n , n ) = |n (r)− n (r)| d r, (10) We exploit the fact that the chosen metrics for quantum n 1 2 1 2 max states have well-defined maximum values D , and so it is possible to quantify an adiabatic threshold as a per- with n (r) the particle density of system j at position centage of these maxima: we consider a state ρ(t) to r. In (10) we use that for the present purposes systems behave adiabatically for all practical purposes (f.a.p.p.) 1 and 2 will have the same number of particles N and when consequently rescale the metric with respect to [21]. D (ρ (t), ρ(t)) ≤ Δ , (11) GS/Th ρ All the metrics considered have a maximum value. For the particle density metric the maximum distance is 2; where GS indicates the ground state, and Th the refer- with the system states normalized to 1, the Bures dis- ence state at finite temperature, which is specified in the tance maximum is 2 and the trace distance maximum present case in section “Numerical Results”, and which is 1. reduces to GS for T → 0. In this paper we choose B,max Δ = D /10. This threshold can of course be ad- justed, depending on the accuracy/constraints of the ex- periment or calculation being performed. We note that, as the temperature increases, k T becomes the dominat- Method for measuring and dynamically tracking ing energy scale so that the same external driving will af- adiabaticity fect the system less. This implies that, for the same drive but increased temperature, dynamical states will remain The key questions we face are: could suitable metrics closer to adiabaticity, suggesting that tighter adiabatic be used to characterize adiabaticity at finite tempera- thresholds could be chosen in this case. ture and even for complex interacting many-body sys- tems? Could an easy-to-calculate method to measure adiabaticity and its time evolution be provided even for System particle densities and the adiabatic line complex many-body systems, which are notoriously dif- ficult to treat? Here we propose an operative definition In [19, 21–23, 26] it was shown that there is a mono- of adiabaticity based on the ’adiabaticity threshold’ and tonic relationship between ground state distances and suggest to answer the above questions by tracking the their corresponding particle densities’ distances, and that instantaneous distance between the time-dependent sys- this relationship is quasi-linear up to relatively large dis- tem state and its adiabatic counterpart using the Bures B,max tances ≈ (2/3)D [38], with [23] indicating this re- (8) and trace metrics (9). We also propose, and justify lationship to hold also for higher order eigenstates and below, that the much-simpler-to-calculate distance (10) corresponding particle densities. Results from this study between the corresponding particle densities (continuum) show the same behaviour at finite temperatures (see fig- or site occupations (discrete systems) can be use in al- ure 1). We refer to this quasi-linear relationship as the ternative. ‘adiabatic line’ [19]: this would be the region, in metric The motivation to use the density distance rely on space, populated by adiabatic systems and hence by a the theorem by Runge and Gross [34, 35] for continu- system evolving adiabatically. The adiabatic line for a ous systems, and on its extension to lattice Hamiltonians certain time-dependent process can then be written as [36, 37]. These, at least at zero temperature, provide a one-to-one correspondence between the driven system D (n (0), n (t)) ≈ mD (ρ (0), ρ (t)). n GS/Th GS/Th GS/Th GS/Th many-body state and the corresponding particle density. (12) This allows to shift the attention from the system’s quan- For adiabatic-enough systems, we can always assume tum states to the corresponding particle densities (con- that D (n(t), n (t)) ≤ D (n (0), n (t)), n GS/Th n GS/Th GS/Th tinuum) and site occupations (lattice Hamiltonians), ob- see Supporting Information, section 2, so that using (12), jects much simpler to calculate, e.g. by density functional we can write an upper bound for the adiabatic threshold methods [35, 37]. for the density distance, Δ , in terms of the correspond- n 4 ing threshold for the state as sites, with nearest-neighbour hopping is Δ = mΔ . (13) n ρ † † H = −J c ˆ c ˆ + c ˆ c ˆ i+1,σ i,σ i,σ i+1,σ The gradient m, will depends on N , U , and T , as well as i,σ on the type of driving potential. N N X X In principle a more accurate (and more + U n ˆ n ˆ + v n ˆ , (14) i,↑ i,↓ i i computationally expensive) estimate of Δ i i could be achieved by using a polynomial fit- ting to the curve D (n (0), n (t)) = n GS/Th GS/Th where J is the hopping parameter for an electron with f (D (ρ (0), ρ (t))), but we find that the GS/Th GS/Th spin σ, with σ =↑ or ↓, U is the on-site electron- linear approximation (12) and the simple method electron repulsion strength, and v is the external po- described above is sufficient for achieving good results † tential at site i. Also, c ˆ and c ˆ are the usual cre- i,σ i,σ (see Figs. 2 and 3). ation and annihilation operators for a spin-σ fermion on site i, and n ˆ = n ˆ + n ˆ is the number operator, with i i,↑ i,↓ n ˆ = c ˆ c ˆ . i,σ i,σ i,σ Estimate for the gradient of the adiabatic line The non-equilibrium dynamics is driven through the application for a time τ of a uniform electric field linearly In practice, the gradient m can be estimated increasing with time from a potential difference along the by calculating D (n (0), n (t)) and GS/Th GS/Th B chain of 1J to a potential difference of 10J . The on-site D (ρ (0), ρ (t)) for 2-3 values of t. For these ρ GS/Th GS/Th 0 τ potential at site i is then written as v (t) = μ + μ t/τ B B,max i i chosen values, D should be less than (2/3)D , ρ ρ 0 0 0 0 τ where μ = 2μ /N × i− μ where μ = 0.5J , and μ = i i and the origin should be included in the fit in virtue of τ τ τ 2μ /N × i− μ with μ = 4.5J . eq. (6). Estimating m requires then exact or approxi- The Hubbard model is used to simulate various phys- mate diagonalization of the system Hamiltonian at 2-3 ical systems of interest to quantum technologies [40, 43– instants in time. Of course at zero temperature only the 47], and the proposed dynamics could represent transient estimate of the GS is necessary. electronic currents along a chain of e.g. nanostructures (for example coupled quantum dots) or of atoms due to the application of a time-dependent electric field across NUMERICAL RESULTS the chain. It is noted that the final Hamiltonian does not depend While the methods proposed can be applied to both on the evolution time τ , and therefore the τ measures continuous and lattice systems, here we will illustrate also the inverse speed of the evolution. Hence for consid- them using the epitome for strongly correlated many- ering adiabatic evolutions, the larger τ is, the closer to body quantum systems, the Hubbard model, firstly at adiabaticity the system is expected to be. zero and then at finite temperatures. We have analysed the dynamics of short non-homogeneous Hubbard chains (N = 2, 4, 6), driven at different rates. In the following, we will discuss explicitly the results for N = 6, corre- Estimate for the density adiabatic threshold sponding to a Hamiltonian of size 400×400 at half-filling. The complexity of its spectrum may be appreciated by looking at the supporting information, figure 3. Curves for D (n (0), n (t)) against n GS/Th GS/Th D (ρ (0), ρ (t)) are shown in figure 1 for GS/Th GS/Th three temperatures (k T = 0, GS, left; k T = 0.2J , B B Hubbard model and system drive Th, middle; k T = 2.5J , Th, right). In figure 1 it can be seen how increasing U (from red to green to blue) or increasing temperature decreases the curves’ gradient. To demonstrate the properties of the methods for In calculating the adiabatic threshold, we have used the characterizing adiabaticity proposed in this work, the linear approximation (12) with m estimated as described out-of-equilibrium dynamics of the inhomogeneous one- in section “Estimate for the gradient of the adiabatic dimensional Hubbard model at half-filling is considered. line”. The approximated values for m can be found in The inhomogeneous Hubbard model is often used as a the supporting information (table 1, final column) for test-bed for developing techniques for strongly correlated all combinations of 3 values each of N , U , and T . many-body systems [10, 39] as it displays non-trivial properties even for the small chains [10, 11, 40–42] for In calculating Δ we have used the method described which it can be solved (numerically) exactly. The corre- in section “Estimate for the gradient of the adiabatic sponding Hamiltonian for a system of N fermions and N line”. 5 1.6 1.6 1.6 (a) (b) (c) 1.4 1.4 1.4 1.2 1.2 1.2 1 1 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.4 0.4 0.2 0.2 0.2 0 0 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 D( (0), (t)) ρ ρ D( (0), (t)) ρ ρ D( (0), (t)) ρ ρ GS Th Th FIG. 1: Curves D (n (0), n (t)) versus D (ρ (0), ρ (t)) for 3 interaction strengths: U = 0J in n GS/Th GS/Th GS/Th GS/Th red, U = 5J in green, and U = 10J in blue; and 3 temperatures, T = 0J/k (GS, left) T = 0.2J/k (Th, middle), B B T = 2.5J/k (Th, right). Note that n(0) = n (0) and ρ(0) = ρ (0) B GS/Th GS/Th Zero temperature A similar pattern occurs for U = 10J , figure 2(c), where though the avoided crossings happen only for t ≈ 95%τ (inset), when the applied potential becomes Predictions from (t) of the order of U . Once more, its Markovianity induces (t) to drop quickly afterwards. At zero temperature, the system initial state is the ground state: (t), as implemented, compares GS to all excited states and includes treatment of degeneracies ac- Adiabatic and non-adiabatic behaviour according to cording to [29]. In figure 2, panels (a)-(c) show (t) from D (ρ (t), ρ(t)) and D (n (t), n(t)) GS n GS eq. (2) with respect to time in units of τ . We consider different rates of dynamics (τ = 0.5/J , Figure 2 displays D (ρ (t), ρ(t)), panels (d)-(f ), and GS red, ‘fast’ dynamics; τ = 5/J , green, ‘intermediate’ dy- D (n (t), n(t)), panels (g)-(i) versus time for the same n GS namics; τ = 50/J , blue, ‘slow’ dynamics, closer to adia- parameters as (t) (panels (a)-(c)). [48]. The horizontal baticity), and three interaction strengths (U = 0J , left, dashed lines indicates the threshold Δ for the states’ dis- no interaction; U = 5J , middle, medium interaction; tances (panels (d)- (f )) and the corresponding threshold U = 10J , right, strong interaction). One would expect Δ for the particle density distances (panels (g)- (i)). that the red curves will demonstrate non-adiabatic be- For U = 0J and intermediate to fast dynamics, the pre- havior, whereas the blue curves should exhibit behavior B dictions from D (ρ (t), ρ(t)) and D (n (t), n(t)) are GS n GS closer to adiabaticity, and the green curves be somewhere in striking contrast with the predictions by (t). At very between the two. For U = 0J , figure 2(a), the initial dy- short times both metrics correctly predict a behaviour namics as described by (t) indeed follows these expec- close to adiabatic: the initial state is the GS and it will tations, though (t) predicts that the dynamics becomes take a finite time to the system state to significantly com- more adiabatic as time progresses, with in particular the bine with higher energy states. At intermediate to long τ = 5/J dynamics becoming adiabatic for t ∼ 0.5τ . times, while (t) would erroneously predict a fast return towards adiabaticity for the red and green dynamics, the For U = 5J , figure 2(b), many-body interactions be- metrics clearly show that the system remains far from come important and the static system would be in the adiabatic: the system dynamics far from equilibrium is process of transitioning between a metal and a quasi- highly affected by the trajectory in phase space at pre- Mott insulator (see e.g. [11, 21, 41]): states with double vious times (memory) and so considering a measure of occupation are ‘pushed up‘ in energy and the dynamics is adiabaticity which is non-Markovian – such as the pro- stiffen for low-enough applied potentials. Then, initially, posed metrics – becomes crucial to avoid false reading. all dynamics satisfy the QAC expressed by (2). However For slow dynamics (τ = 50), the behaviour remains at t ≈ 0.4τ the applied time-dependent potential has in- always at or below the adiabatic threshold. The oscil- creased enough to produce an avoided level-crossing in lations shown by D (ρ (t), ρ(t)) and D (n (t), n(t)) the low-energy spectrum of the instantaneous Hamilto- GS n GS nian (see inset in figure 2(b)) so that (t) predicts non- for τ = 50 were also observed in the single-electron sys- adiabatic behavior for both fast (τ = 0.5/J ) and inter- tems studied in reference [19]. These are explained by the mediate (τ = 5/J ) dynamics. However, as time increases system inertia in adjusting to gradually-applied electric further, according to (t), the dynamics quickly returns field. adiabatic for all dynamic rates. We note here that a tool For finite many-body interaction strengths, both directly derived from the QAC (1), such as (t) eq. (2), is D (ρ (t), ρ(t)) and D (n (t), n(t)) strongly respond GS n GS Markovian by construction, i.e. does not include memory to the (avoided) level crossings at t ≈ 0.4τ (U = 5J ) as it is based on instantaneous quantities. and t ≈ 9.5τ (U = 10J ), but, crucially also signal that D(n(0),n (t)) GS D(n(0),n (t)) Th D(n(0),n (t)) Th 6 10 50 1000 (a) (c) (b) −1.8 −1.1 8 40 −2.4 800 −3 −1.3 6 30 600 −3.6 −1.5 0.3 0.4 0.5 4 20 −1.7 2 10 200 0.8 0.84 0.88 0.92 0.96 1 0 0 0 (e) (f) (d) 1.2 1.2 1.2 1 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.4 0.4 0.2 0.2 0.2 0 0 (h) (i) 0.6 0.3 0.4 0.5 0.3 0.4 0.2 0.3 0.2 0.2 0.1 0.1 0.1 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 τ t/τ t/τ t/ FIG. 2: Zero temperature results. Red lines: τ = 0.5/J dynamics, green lines: τ = 5/J dynamics, blue lines: τ = 50/J dynamics. Panels show: (t) [(a)-(c)], D (ρ (t), ρ(t)) [(d)-(f )], and D (n (t), n(t)) [(g)-(i)] versus t/τ . GS n GS Three interaction strengths are considered: U = 0J (left), U = 5J (middle), and U = 10J (right). In all panels the horizontal dashed lines indicate the corresponding adiabatic threshold. Insets of panel (b) and (c): low energy spectrum of the instantaneous Hamiltonian versus t/τ for U = 5 [panel (b)] and U = 10 [panel (c)]. afterwards the system dynamics remain strongly non- Finite temperature adiabatic, with then important contributions from mem- ory effects. Note that this is the case even for the slow A thermal bath at temperature T is now connected to dynamics (τ = 50): compare blue lines in panels (a), (c), the Hubbard chain, to thermalize the system. Once ther- (h) for t > 0.5τ . malized, at t = 0 , the bath is disconnected, and then the closed system is evolved from t = 0 to τ . Therefore, the initial state is now a thermal state, with a corresponding D (ρ (t), ρ(t)), as distance between the system GS thermal particle density. Because of the closed dynamics, quantum state and its adiabatic counterpart, can be read- we will then consider the distance between the dynami- ily associated to the definition of adiabaticity; this is less cal system state ρ(t) and its finite-temperature adiabatic so for D (n (t), n(t)): particle densities, being just a n GS counterpart function of position and time, could be expected to be much less sensitive to details than the system state, e.g. it might be less sensitive to details of the instantaneous j,0 Hamiltonian spectrum, or less susceptible to dynamical exp− k T ρ (t) = |ψ ihψ | , (15) Th j,t j,t changes of the system and corresponding memory effects. E k,0 exp− j k k T However, because of the theorems in [34, 36], we know that the considered dynamical system state and its cor- where E is the j-th eigenenergy of the Hamiltonian responding particle density contain the same amount of j,0 at t = 0, and |ψ i is the j-th eigenstate of the in- information: we can then conjecture that both the re- j,t stantaneous Hamiltonian at time t. The corresponding lated ‘natural’ [21] metrics can be used successfully as particle density n (t) is used in the density distance measures of adiabaticity. This was confirmed in [19] for Th D (n (t), n(t)). single-particle systems, and here for many-body systems. n Th This leads to the possibility of characterizing adiabatic- Two temperatures are considered in this work; a lower ity using the sole particle density, a quantity much more temperature of k T = 0.2J , and a higher temperature accessible than the corresponding system quantum state. of k T = 2.5J . D(n (t),n(t)) D( (t), (t)) ρ GS GS ε 7 Low temperature the occupation probability of the two relevant eigenstates before the crossing. For t < t , we can estimate each of ac the 20 lower-band eigenstates to have roughly 1/20 oc- For k T = 0.2J , (t) and both metrics show the cupation probability, and the upper band having no oc- systems to behave mostly very similarly to the zero- cupation. The change in the Bures distance across each temperature case. A notable difference occurs for U = 10 B,max crossing would then be about D /20 = 0.07, giving and 0.9 < t/τ < 1, where the inset of fig. 2(c) shows the an overall height for the six steps of D of 0.42, which occurrence of four low-energy avoided crossings. Due to ρ is reasonably close to what we observe in fig. 3(f ). A the finite-temperature state mixing, both metrics signal similar structure is faithfully signalled by D . The other the four crossings with corresponding steps in the dis- n anti-crossings, which occur deeper in the lower band, are tances, while, (t) remains sensitive only to the crossing between eigenstates with very similar occupation proba- occurring at t/τ ≈ 0.96 between ground and first excited bilities, so that the overall state should be expected to state. These results suggest that for low temperatures, change very little at crossings: this is faithfully captured k T  J , the density could be used as a good indicator by the chosen metrics, much less affected by those anti- to characterize adiabaticity. For completeness, we report crossings. The T-QAC measure (t) presents a series of all results for k T = 0.2J in the supporting information, peaks in the region of where the bands cross, but without figure 1. distinguishing between the anti-crossing being at the top, or deeper within, the lower energy band. Importantly we find that the anti-crossings affecting (t) are often not the High temperature ones expected to substantially change the system state. The problem of (t) in signaling inappropriately the For k T = 2.5J and thermal equilibrium, tens of anti-crossings is even more evident (and problematic) for eigenstates of the initial Hamiltonian spectrum are sig- U = 5: here the crossing between lowest and immedi- nificantly populated (initial state). For U = 0, the be- ately upper bands starts already at t/τ ≈ 0.05 [we re- haviour of (t) and of both metrics is qualitatively simi- port the relevant part of the instantaneous spectrum ver- lar to the lower temperatures examined: no anti-crossing sus time in the supporting information, figure 3(b)]. As are observed within ∼ k T of the instantaneous ground there is no substantial gap between them at t = 0, the state, while energy gaps in this part of the spectrum tend top levels of the lowest and the lower levels of this up- to increase with time. We report part of the instanta- per band are fairly similarly populated. This means that neous spectrum versus time in the supporting informa- the corresponding mixed system state does not change tion, figure 3(a). significantly at each of these crossings, as correctly dis- At U = 10, many-body interactions creates distinct played by both metrics. However (t) dramatically sig- bands in the instantaneous Hamiltonian spectrum, shown nals the initial anti-crossings, thus proving a false reading for completeness in the supporting information, figure of non-adiabaticity already at t/τ ≈ 0.05 [see inset of fig- 3(c). The eigenstates in the lowest energy band are linear ure 3(b)]. These spikes in (t) may be related to the combinations of the 20 possible permutation of single-site problem of small denominators for this type of measures, occupations at half-filling. Even at t = 0, the system is see [50]. slightly inhomogeneous, so these eigenstates are not de- Although in this work the degenerate form of QAC generate. The next band contains six eigenstates, com- was adapted for finite temperature, the results show that binations of states which may allow double occupation it is still not well suited for high T . On the other side, in one of the sites, and the bandgap due to this on-site the metrics, which naturally include degeneracy and non- Coulomb repulsion is about 6J at t = 0, substantially Markovianity, can be seen to cope well with the temper- larger than k T , so that initially only the lower band ature increase. is significantly occupied. For all dynamical rates consid- ered, this gap is also much larger than 1/τ , and indeed all measures considered remain below or close to their adia- batic thresholds until the two bands start (anti) crossing Results for the trace distance at t /τ ≈ 0.8, see fig. 3(c), (f ) and (i). We note six- ac < < steps in both metrics for t ∼ t ∼ 0.9τ , see figs. 3(f ) ac and (i). Each step signals one of the upper-band eigen- With respect to its own adiabatic threshold[51], the states starting to anti-cross the lower band. The Bures trace distance results quantitatively close to the Bures distance between two orthogonal pure states is maximal, distance, including signaling with steps relevant anti- so the Bures distance between the system state and its crossings. This means that it can be used as an alterna- adiabatic reference is in principle set up to signal non- tive quantitative measure of adiabaticity[52]. For com- adiabatic behaviour at any avoided crossing[49]. However pleteness, we report the related results in the supporting the strength of the signal will depend on how different is information, figure 2. 8 16 120000 (a) (c) (b) 10 120000 6 80000 4 0 0.05 0.1 0.15 0.2 0.25 0.3 0 0 0 (d) (e) (f) 0.4 0.6 0.5 0.35 0.5 0.3 0.4 0.4 0.25 0.3 0.2 0.3 0.15 0.2 0.2 0.1 0.1 0.1 0.05 0 0 (g) (h) (i) 0.16 0.25 0.1 0.14 0.12 0.2 0.08 0.1 0.15 0.06 0.08 0.06 0.1 0.04 0.04 0.05 0.02 0.02 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 t/τ t/τ t/τ FIG. 3: High temperature (T = 2.5J/k ) results. Red lines: τ = 0.5/J dynamics, green lines: τ = 5/J dynamics, blue lines: τ = 50/J dynamics. Panels show: (t) [(a)-(c)], D (ρ (t), ρ(t)) [(d)-(f )], and D (n (t), n(t)) [(g)-(i)] Th n Th versus t/τ . Three interaction strengths are considered: U = 0J (left), U = 5J (middle), and U = 10J (right). In all panels the horizontal dashed lines indicate the corresponding adiabatic threshold. Inset of panel (b): zoom into short times of main panel. CONCLUSION old for the system state metric to an upper bound for the threshold for the local particle density metric. This up- per bound is tight enough along most parts of the time- We have introduced methods based on appropriate evolutions analysed, and it is relatively easy to estimate, metrics to measure adiabaticity for the dynamics of even for large systems. We aim to refine it as future work. many-body quantum systems at finite temperature, and We discuss an extension to finite temperature of the to track it with time evolution. As system state met- quantum adiabatic criterion which include treatment of rics, Bures and trace distances give consistently similar degeneracies. Comparing the results from this and the predictions, for all dynamics and temperatures analysed. metrics has highlighted the importance of properly in- Additionally, our results support the conjecture – based cluding memory effects when wishing to evaluate and on the fundamental theorems of time-dependent density track the adiabatic level of a many-body dynamics: by functional theory – that the ‘natural’ metric tracking the construction, a measure based on the quantum adiabatic evolution of the local particle density alone would be suf- criterion is basically Markovian, as, at most, the instan- ficient to estimate the level of adiabaticity of the systems’ taneous Hamiltonian derivative is included. Our results dynamics. This is a great simplification as, in general, show that this leads to false readings, as highly out-of- the system local particle density may be estimated more equilibrium dynamics may be pictured as adiabatic. Our accurately and much more simply than the correspond- results have also shown that while the metric-based meth- ing many-body system state. It is also an experimentally ods correctly reflect the amount of change in the system measurable quantity, which opens additional possibilities state at instantaneous eigenenergy anti-crossings, the ex- for the method. tension to finite temperatures of the quantum adiabatic Because these metrics have a finite maximum, they criterion is often sensitive – and sometimes extremely are suitable for the design of practical ’adiabatic thresh- sensitive – to the anticrossing where the actual many- olds’, so that a distance below (above) the corresponding body state does not change significantly. Once more this threshold signals adiabatic (non-adiabatic) behaviour. may lead to false readings, this time predicting the sys- Using the results in this paper and previous results, we tem to be far from adiabaticity while it is actually still have been able to consistently relate the adiabatic thresh- behaving adiabatically. D(n (t),n(t)) D( (t), (t)) ρ ρ Th Th ε 9 We acknowledge fruitful discussions with V. V. Franca ity, especially when considering applications to quantum systems. and thank K. Zawadzki for the code for the Hubbard [28] M. V. Berry, Journal of Physics A 42, 36 (2009). chain dynamics; AHS acknowledges support from EP- [29] G. Rigolin and G. Ortiz, Phys. Rev. A 85, 062111 (2012). SRC. [30] W. Sutherland, Introduction to Metric and Topological Spaces (Oxford University Press, Oxford, 2009). [31] V. V. Fran¸ ca, J. P. Coe, and I. D’Amico, Scientific Re- ports 8 (2018). √ √ 2 [32] F (σ, ρ) = Tr ρσ ρ is known as fidelity, which [1] T. Albash and D. A. Lidar, “Adiabatic quantum com- is also used in the literature for estimating how simi- puting,” arXiv:1611.04471. lar two quantum states are, though cannot be considered [2] E. Farhi, J. Goldstone, S. Gutmann, J. Lapan, A. Lund- a proper distance, as it does not obey all the metrics’ gren, and D. Preda, Science 292, 472 (2001). axioms. [3] M. Gell-Mann and F. Low, Phys. Rev. 84, 350 (1951). [33] I. Bengtsson and K. Zyczkowski, Geometry of Quan- [4] D. Bacon and S. T. Flammia, Phys. Rev. Lett. 103, tum States: An Introduction to Quantum Entanglement 120504 (2009). (Cambridge University Press, 2006). [5] I. Hen, Phys. Rev. A 91, 022309 (2015). [34] E. Runge and E. K. U. Gross, Phys. Rev. Lett. 52, 997 [6] A. C. Santos, R. D. Silva, and M. S. Sarandy, Phys. Rev. (1984). A 93, 012311 (2016). [35] C. A. Ullrich, Time-Dependent Density-Functional The- [7] O. Abah and E. Lutz, Eur. Lett. (EPL) 118, 40005 ory: Concepts and Applications. (Oxford University (2017). Press, 2013). [8] J. He, J. Chen, and B. Hua, Phys. Rev. E 65, 036145 [36] C. Verdozzi, Phys. Rev. Lett. 101, 166401 (2008). (2002). [37] K. Capelle and V. L. Campo, Physics Reports 528, 91 [9] C.-K. Hu, J.-M. Cui, A. C. Santos, Y.-F. Huang, C.-F. (2013). Li, G.-C. Guo, F. Brito, and M. S. Sarandy, Scientific [38] See e.g. [21], figure 2. Reports 9, 10449 (2019). [39] M. Herrera, K. Zawadzki, and I. D’Amico, The European [10] M. Herrera, R. M. Serra, and I. D’Amico, Scientific Re- Physical Journal B 91, 248 (2018). ports 7, 4655 (2017). [40] S. Murmann, A. Bergschneider, V. M. Klinkhamer, [11] A. Skelt, K. Zawadzki, and I. D’Amico, Journal of G. Zurn, ¨ T. Lompe, and S. Jochim, Phys. Rev. Lett. Physics A: Mathematical and Theoretical 52, 48 (2019). 114, 080402 (2015). [12] X. Chen, A. Ruschhaupt, S. Schmidt, A. del Campo, [41] D. J. Carrascal, J. Ferrer, J. C. Smith, and K. Burke, J. D. Gu´ ery-Odelin, and J. G. Muga, Phys. Rev. Lett. 104, Phys. Cond. Mat. 27(39), 393001 (2015). 063002 (2010). [42] K. Zawadzki, I. D’Amico, and L. Oliveira, Brazilian [13] A. del Campo and K. Kim, New Journal of Physics 21, Journal of Physics 47, 488 (2017). 050201 (2019). [43] J. P. Coe, V. V. Fran¸ ca, and I. D’Amico, Phys. Rev. A [14] M. Born and V. A. Fock, Z. Phys. A 51, 165–180 (1928). 81, 052321 (2010). [15] J. E. Avron and A. Elgart, Communications in Mathe- [44] S. Yang, X. Wang, and S. Das Sarma, Phys. Rev. B 83, matical Physics 203, 445–463 (1999). 161301 (2011). [16] K.-P. Marzlin and B. C. Sanders, Phys. Rev. Lett. 93, [45] J. P. Coe, V. V. Fran¸ ca, and I. D’Amico, EPL (Euro- 160408 (2004). physics Letters) 93, 10001 (2011). [17] D. M. Tong, K. Singh, L. C. Kwek, and C. H. Oh, Phys. [46] P. T. Brown, D. Mitra, E. Guardado-Sanchez, Rev. Lett. 95, 110407 (2005). R. Nourafkan, A. Reymbaut, C.-D. H´ ebert, S. Berg- [18] J. Ortigoso, Phys. Rev. A 86, 032121 (2012). eron, A.-M. S. Tremblay, J. Kokalj, D. A. Huse, [19] A. H. Skelt, R. W. Godby, and I. D’Amico, Phys. Rev. P. Schauß, and W. S. Bakr, Science 363, 379 (2019), A 98, 012104 (2018). https://science.sciencemag.org/content/363/6425/379.full.pdf. [20] O. Lychkovskiy, O. Gamayun, and V. Cheianov, Phys. [47] M. A. Nichols, L. W. Cheuk, M. Okan, T. R. Hartke, Rev. B 98, 024307 (2018). E. Mendez, T. Senthil, E. Khatami, H. Zhang, [21] I. D’Amico, J. P. Coe, V. V. Fran¸ ca, and K. Capelle, and M. W. Zwierlein, Science 363, 383 (2019), Phys. Rev. Lett. 106, 050401 (2011). https://science.sciencemag.org/content/363/6425/383.full.pdf. [22] P. M. Sharp and I. D’Amico, Phys. Rev. B 89, 115137 [48] Corresponding results for the trace distance will be dis- (2014). cussed in section “Results for the trace distance” and in [23] P. M. Sharp and I. D’Amico, Phys. Rev. A 92, 032509 the supporting information. (2015). [49] In an anti-crossing, the component of the dynamical state [24] S. Marocchi, S. Pittalis, and I. D’Amico, Phys. Rev. behaving non-adiabatically will be orthogonal to the cor- Materials 1, 043801 (2017). responding component of the adiabatic reference state. [25] M. M. Wilde, Quantum Information Theory (Cambridge [50] M. Kolodrubetz, D. Sels, P. Mehta, and A. Polkovnikov, University Press, 2013) Chap. 9. Physics Reports 697, 1 (2017). [26] A. H. Skelt, R. W. Godby, and I. D’Amico, Brazilian [51] Remember that the maximum value of the trace distance Journal of Physics 48, 467 (2018). is 1 for normalized states, therefore the numerical values [27] Since temperature is being introduced, there are two def- of the distance will be different to those of the Bures initions of adiabaticity; quantum and thermal. Thermal distance. adiabaticity looks at the heat loss of the system, which is [52] A comparison between estimates from the Kullback rel- zero for this investigation as the system is closed. There- ative entropy and results from the trace distance for the fore it only makes sense to look at the quantum adiabatic- 10 distance of mixed states from thermal equilibrium can be different parameter values. found in figure 4 of [53]. There it is found that the be- [53] K. Zawadzki, R. M. Serra, and I. D’Amico, haviour of the two quantities have qualitatively similar arxiv:1908.06488 (2019). features, but, for example, the maxima/minima occur at Characterizing Adiabaticity in Quantum Many-Body Systems at Finite Temperature – Supporting Information A. H. Skelt Department of Physics, University of York, York YO10 5DD, United Kingdom I. D’Amico Department of Physics, University of York, UK and International Institute of Physics, Federal University of Rio Grande do Norte, Natal, Brazil (Dated: April 14, 2020) TABLE OF GRADIENTS FOR ADIABATICITY (t), BURES DISTANCE, AND PARTICLE DENSITY DISTANCE FOR T = 0.2J/k Table I gives the gradients m for the Hubbard model Figure 1 shows the results for (t), the Bures distance, at 3 values of N , of U , and of T . This could be used as and the particle density distance when T = 0.2J/k (low a guide for the expected gradient of the adiabatic line in temperature). different systems. TRACE DISTANCE We look at how the trace distance copes with charac- DEMONSTRATION THAT, FOR terizing the level of adiabaticity. For this we will use ADIABATIC-ENOUGH SYSTEMS, D (n(t), n (t)) ≤ D (n (0), n (t)) q n GS/Th n GS/Th GS/Th 1 1 D (ρ, σ) = Tr [|ρ− σ|] = Tr (ρ− σ) (ρ− σ). (5) 2 2 A key formal property of metrics is that Beginning at zero temperature, we compare figure 2 (a)- (c) here to figure 2 (d)-(f ) in the main text. For low D(A, B) = 0 if and only if A = B (1) temperature we compare figure 2 (d)-(f ) here to figure 1 (d)-(f ) also here. And for the high temperature, we com- For dynamics starting from the ground or thermal state, pare figure 2 (g)-(h) here to figure 3 (d)-(f ) in the main this property implies that, at t = 0, text. Qualitatively the Bures and trace distances agree for all U . The quantitative difference is mainly associated D (n(t = 0), n (t = 0)) = 0 n GS/Th to the difference in the maximum distances. For all temperatures, the same conclusions that were and, of course, D (n (t = 0), n (t = 0)) = 0. n GS/Th GS/Th drawn for the Bures distance can be drawn for the trace This satisfies the claim for t = 0. distance, which can therefore be used to characterize adi- For t > 0 and a generic time-dependent Hamiltonian, abatic evolutions. equation (1) implies that D (n (0), n (t)) > 0, (2) n GS/Th GS/Th INSTANTANEOUS EIGEN-ENERGIES OF THE FERMI-HUBBARD HAMILTONIAN as in general n (0) 6= n (t). [? ] GS/Th GS/Th At the same time, for a perfectly adiabatic dynamics, Figure 3 shows the low-mid section of the instanta- the definition of adiabaticity would impose that, at all neous spectrum for the driven Hubbard Hamiltonian con- times, sidered in this work with respect to time, and for (a) U = 0J , (b) U = 5J , and (c) U = 10J . For U = 10J , D (n(t), n (t)) = 0. (3) n GS/Th the Coulomb repulsion for states which include double occupation of site(s) leads to the formation of energy Equations (2) and (3) allow us to state that, for systems bands. The lowest two bands start crossing when the whose dynamics is close enough to adiabaticity, we can applied external potential is of the order of U , around always assume that t/τ = 0.8. For U = 5J the gap between the lowest two energy bands is just starting to form, and the two bands D (n(t), n (t)) ≤ D (n (0), n (t)). (4) start crossing for t/τ ≈ 0.05. n GS/Th n GS/Th GS/Th arXiv:2004.05842v1 [quant-ph] 13 Apr 2020 2 k T U N Gradient N Gradient N Gradient 0J 0J 2 1.33294 4 0.647511 6 0.450959 0J 5J 2 0.91163 4 0.411435 6 0.2779 0J 10J 2 0.502006 4 0.218927 6 0.237795 0.2J 0J 2 0.924189 4 0.647354 6 0.45059 0.2J 5J 2 0.911343 4 0.409499 6 0.278244 0.2J 10J 2 0.619401 4 0.210794 6 0.205762 2.5J 0J 2 0.924189 4 0.57114 6 0.444165 2.5J 5J 2 0.744725 4 0.473781 6 0.389355 2.5J 10J 2 0.48877 4 0.231146 6 0.179414 TABLE I: Gradients m of the adiabatic line for the driven Hubbard model considered in this paper for three temperatures (k T ), correlation strengths (U ), and electron numbers (N ). 50 1000 (a) (b) (c) 40 800 6 30 600 4 20 400 2 10 200 0 0 0 (d) (e) (f) 1.2 1.2 1 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.4 0.4 0.2 0.2 0.2 0 0 0 (g) (h) (i) 0.6 0.25 0.4 0.5 0.2 0.3 0.4 0.15 0.3 0.2 0.1 0.2 0.1 0.05 0.1 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 t/τ t/τ t/τ FIG. 1: Low temperature (T = 0.2J/k ) results. Red lines: τ = 0.5/J dynamics, green lines: τ = 5/J dynamics, blue lines: τ = 50/J dynamics. Panels show: (t) [(a)-(c)], D (ρ (t), ρ(t)) [(d)-(f )], and D (n (t), n(t)) [(g)-(i)] Th n Th versus t/τ . Three interaction strengths are considered: U = 0J (left), U = 5J (middle), and U = 10J (right). In all panels the horizontal dashed lines indicate the corresponding adiabatic threshold. D( (t), (t)) ρ D(n (t),n(t)) Th Th ε 3 1 1 1 (b) (c) 0.9 0.9 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.6 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.4 0.3 0.3 0.3 0.2 0.2 0.2 0.1 0.1 0.1 1 0.9 (d) (f) (e) 0.9 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.6 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.4 0.3 0.3 0.3 0.2 0.2 0.2 0.1 0.1 0.1 0.4 0.6 0.35 (g) (h) (i) 0.35 0.3 0.5 0.3 0.25 0.4 0.25 0.2 0.2 0.3 0.15 0.15 0.2 0.1 0.1 0.1 0.05 0.05 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 t/τ t/τ t/τ FIG. 2: All figures here show D (ρ (t), ρ(t)) against t/τ . Interaction strengths: U = 0J (left), U = 5J ρ GS(Th) (middle), and U = 10J (right); red lines corresponds to τ = 0.5/J , green lines to τ = 5/J , and blue lines to τ = 50/J . Panels (a)-(c) show the zero temperature results, T = 0J/k ; (d)-(f ) show the low temperature results, T = 0.2J/k ; (g)-(i) show the high temperature results, T = 2.5J/k . B B −2 −4 −6 −8 −10 −2 −12 −4 −14 −16 −6 −18 −1 (a) (b) (c) −20 −8 −2 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 τ τ t/ t/ t/ FIG. 3: Zoom into the low-mid section of the instantaneous spectrum versus t/τ of the time-dependent Hubbard Hamiltonian considered in this work (6 sites, half filling). Panel (a) corresponds to zero on-site Coulomb interaction (U = 0J ), (b) to U = 5J , and (c) to U = 10J . Note the different energy scales on the y-axis. E D( (t), (t)) ρ D( (t), (t)) ρ ρ D( (t), (t)) ρ ρ Th Th GS http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Quantum Physics arXiv (Cornell University)

Characterizing Adiabaticity in Quantum Many-Body Systems at Finite Temperature

Skelt, A. H.; D'Amico, I.
Quantum Physics , Volume 2020 (2004) – Apr 13, 2020
Download PDF
13 pages

Loading next page...
 
/lp/arxiv-cornell-university/characterizing-adiabaticity-in-quantum-many-body-systems-at-finite-n6Uva0BWSA
ISSN
2511-9044
eISSN
ARCH-3342
DOI
10.1002/qute.201900139
Publisher site
See Article on Publisher Site

Abstract

A. H. Skelt Department of Physics, University of York, York YO10 5DD, United Kingdom I. D’Amico Department of Physics, University of York, UK and International Institute of Physics, Federal University of Rio Grande do Norte, Natal, Brazil (Dated: April 14, 2020) The quantum adiabatic theorem is fundamental to time dependent quantum systems, but being able to characterize quantitatively an adiabatic evolution in many-body systems can be a chal- lenge. This work demonstrates that the use of appropriate state and particle-density metrics is a viable method to quantitatively determine the degree of adiabaticity in the dynamic of a quantum many-body system. The method applies also to systems at finite temperature, which is impor- tant for quantum technologies and quantum thermodynamics related protocols. The importance of accounting for memory effects is discussed via comparison to results obtained by extending the quantum adiabatic criterion to finite temperatures: it is shown that this may produce false readings being quasi-Markovian by construction. As the proposed method makes it possible to characterize the degree of adiabatic evolution tracking only the system local particle densities, it is potentially applicable to both theoretical calculations of very large many-body systems and to experiments. INTRODUCTION are usually taken as the ground and first excited states. When it comes to accurately characterizing an adia- Adiabatic evolutions are important in many areas of batic evolution, there are though many challenges, such quantum physics, such as quantum computation, quan- as the complexity of calculations involving many-body tum thermodynamics, and quantum field theory [1–9]. systems and defining the criterion at finite temperature. One particularly important application of adiabatic evo- Recently, the validity and sufficiency of this QAC for cer- lutions is achieving specific (target) states (e.g. in adi- tain systems have been questioned [16–18], and new ap- abatic quantum computation, where the target state is proaches to characterizing adiabaticity have also come known to be the ground state of the final Hamiltonian). to light [19, 20] where they look at comparing the time Other important applications of adiabatic evolutions are evolved state of the system with the “adiabatic” state, i.e. for quantum thermodynamic cycles, where, for example, the instantaneous ground state. It was demonstrated in they may yield the highest extractable quantum work reference [19] that metrics can be used to characterize [10, 11]. Indeed, the relevance of adiabaticity has even adiabaticity through a variety of approaches to best suit given rise to new subfields, such as shortcuts to adia- the quantities one has at hand. The issue of tracking adi- baticity [12, 13] abaticity both at finite temperature and for many-body The quantum adiabatic theorem [14] defines an adia- systems remains outstanding. batic evolution as one in which no transitions between In this work the adiabatic theorem is written in terms energy levels occurs, and is a fundamental concept for of two distance measures (metrics), namely the Bures any time dependent quantum system. It was first pro- and the trace distances. The Bures distance is connected posed in 1928 by Born and Fock, and demonstrates that to the fidelity (which is though not a proper measure), for a quantum system to be considered adiabatic, it must and, at zero temperature, one can use the “adiabatic fi- be evolved slowly enough that it remains in an instanta- delity” as a figure of merit for adiabaticity in time de- neous eigenstate, with a gap between its eigenenergy and pendent systems [20]. Importantly, the Bures distance the rest of the Hamiltonian’s spectrum [14]. Later, Avron can be derived from conservation laws [21, 22] so that and Elgart relaxed this gap condition through a reformu- it can provide relevant information on the physics of the lation of the theorem [15]). At zero temperature, this is many-body system [21–24]. Within quantum informa- often interpreted mathematically with the quantum adi- tion processing, the trace distance is considered the best abatic criterion (QAC) [16–18]: measure to operationally distinguish two quantum states [25], so we also look at using the trace distance in place of hm(t)| H (t)|n(t)i the Bures, and find that both the trace and Bures can be 1, (1) used to determine the degree of adiabaticity. To provide a (|E (t)− E (t)|) n m comparison with a somewhat more familiar quantity, we where H is the time derivative of the Hamiltonian, |mi propose an extension of the QAC to finite temperatures, and |ni are the instantaneous eigenstates of H with in- and discuss its limitations. All of the above broadens stantaneous eigenenergies E and E respectively, and the choice of measures of adiabaticity to best suit one’s m n arXiv:2004.05842v1 [quant-ph] 13 Apr 2020 2 needs. with At zero temperature, the Bures distance for pure states E (t)− E (t) < sk T, (3) n 0 B has already been demonstrated to characterize adiabatic- E (t)− E (t) < s k T, (4) m 0 B ity in single electron systems [19], and was seen to have potential for characterizing adiabaticity in two-electron s > s ≥ 1. (5) systems [26]. Here we consider many-body systems, for- Here E ≤ E ≤ . . . ≤ E and m 6= n. In the calculations 0 1 n mally of any size, continuous and described over a lattice, presented here, we use s = 1 and do not cap s . at zero and finite temperatures. Computationally, as sys- For adiabaticity to hold, we still required that (t)  1. tem size increases exponentially, we apply the method to In T-QAC, the criterion is adapted for degenerate states many-body systems up to 6 electrons on a discrete lat- following Rigolin and Ortiz [29], so that the maximum tice. We take inspiration from density-functional theory distance between the degenerate subspaces and other lev- to ask: “can metrics based on the particle density alone els is considered when calculating (t). give quantitative guidance to the level of adiabaticity of a system?” Particle density is in principle experimentally observable and much easier to estimate than the corre- Metrics for density, n, and quantum state, ρ sponding many-body state, e.g. by density functional methods; by demonstrating that this question has a pos- Metrics provide a quantitative measure – the distance itive answer, we provide a manageable way to measure – to differentiate between two elements of a set [30], and and track adiabaticity of many-body systems, even if the must obey three axioms: positivity, D(x, y) ≥ 0 and temperature is finite. This work aims to help guide those wanting an adi- D(x, y) = 0 iff y = x; (6) abatic evolution (either experimentally or computation- ally) in many-body systems towards achieving an under- symmetry, D(x, y) = D(y, x); and the triangle inequality, standing of the degree of adiabaticity of their system. A D(x, z) ≤ D(x, y) + D(y, z). (7) guideline threshold for considering an evolution adiabatic is then presented with discussion of the factors which im- The use of metrics for investigating the relationship be- pact this threshold and the important quantities to con- tween wavefunctions and corresponding particle densities sider when deciding a threshold for one’s system. was developed in [21–23, 31] where the chosen metrics were derived from conservation laws (’natural’ metrics [22]) to ensure that they could provide physical insights. THEORY AND PROPOSED METHODS Reference [19] introduced a method of using these metrics for characterizing adiabaticity in single electron systems Temperature-dependent quantum adiabatic criterion at zero temperature; other works [21, 22, 26, 31] support the possibility of developing this metric-based method As mentioned previously, the QAC based on the quan- to characterize adiabaticity in many-body systems. All tum adiabatic theorem [14] has been under scrutiny re- these works considered pure states; since here the focus cently, and new approaches have come to light. However is also on finite temperature, the ‘natural’ metrics must all of these approaches only consider quantum adiabatic- be extended to mixed states. ity for pure states at zero temperature. A new expression The metric for the wavefunction developed in [21] is in is introduced here for characterizing adiabaticity in sys- fact the limit for zero temperature of the Bures metric, tems at finite temperature and so described by mixed which for mixed states reads states. First, however, one needs to define what is meant q 1/2 by being adiabatic at finite temperature. The require- √ √ D (σ, ρ) = 2 1− Tr ρσ ρ , (8) ment for quantum adiabaticity at finite temperature [27] is that there are no transitions between eigenstates of the where σ and ρ are quantum system states (density ma- system as it evolves [28]. Practically this implies that the trices) [32] population of the various eigenstates should not change A metric for quantum states which is widely used by with time. the quantum technology community as measure of distin- As a comparison between the metrics and a more fa- guishability between quantum states is the trace distance miliar quantity, we propose the following extension of [25, 33], which is defined as the QAC (1), valid at any temperature T (T-QAC) and which will include degeneracies q 1 1 D (ρ, σ) = Tr [|ρ− σ|] = Tr (ρ− σ) (ρ− σ). (9)   2 2   hm(t)| H (t)|n(t)i This metric will also be considered in this work. Bu- (t) = max (2) n,m   (|E (t)− E (t)|) n m res’ and trace distances are related by bounds [25] and, 3 at least for the systems and dynamics discussed in this The adiabatic threshold work, they provide very similar diagnostic; it is suggested then that the decision of which to use be made on which System quantum states quantities are more readily accessible, e.g. if the fidelity is easy to obtain, the Bures distance should be chosen. In practice, when can a system be considered adia- batic? A reasonable answer is ‘when, during the dy- The ‘natural’ density metric is unaffected by the type namics, the system remains close enough to its adiabatic of state, and remains limit’; in this section we will quantify the concept of ‘close enough’ using the tool of ‘adiabatic threshold’ [19]. D (n , n ) = |n (r)− n (r)| d r, (10) We exploit the fact that the chosen metrics for quantum n 1 2 1 2 max states have well-defined maximum values D , and so it is possible to quantify an adiabatic threshold as a per- with n (r) the particle density of system j at position centage of these maxima: we consider a state ρ(t) to r. In (10) we use that for the present purposes systems behave adiabatically for all practical purposes (f.a.p.p.) 1 and 2 will have the same number of particles N and when consequently rescale the metric with respect to [21]. D (ρ (t), ρ(t)) ≤ Δ , (11) GS/Th ρ All the metrics considered have a maximum value. For the particle density metric the maximum distance is 2; where GS indicates the ground state, and Th the refer- with the system states normalized to 1, the Bures dis- ence state at finite temperature, which is specified in the tance maximum is 2 and the trace distance maximum present case in section “Numerical Results”, and which is 1. reduces to GS for T → 0. In this paper we choose B,max Δ = D /10. This threshold can of course be ad- justed, depending on the accuracy/constraints of the ex- periment or calculation being performed. We note that, as the temperature increases, k T becomes the dominat- Method for measuring and dynamically tracking ing energy scale so that the same external driving will af- adiabaticity fect the system less. This implies that, for the same drive but increased temperature, dynamical states will remain The key questions we face are: could suitable metrics closer to adiabaticity, suggesting that tighter adiabatic be used to characterize adiabaticity at finite tempera- thresholds could be chosen in this case. ture and even for complex interacting many-body sys- tems? Could an easy-to-calculate method to measure adiabaticity and its time evolution be provided even for System particle densities and the adiabatic line complex many-body systems, which are notoriously dif- ficult to treat? Here we propose an operative definition In [19, 21–23, 26] it was shown that there is a mono- of adiabaticity based on the ’adiabaticity threshold’ and tonic relationship between ground state distances and suggest to answer the above questions by tracking the their corresponding particle densities’ distances, and that instantaneous distance between the time-dependent sys- this relationship is quasi-linear up to relatively large dis- tem state and its adiabatic counterpart using the Bures B,max tances ≈ (2/3)D [38], with [23] indicating this re- (8) and trace metrics (9). We also propose, and justify lationship to hold also for higher order eigenstates and below, that the much-simpler-to-calculate distance (10) corresponding particle densities. Results from this study between the corresponding particle densities (continuum) show the same behaviour at finite temperatures (see fig- or site occupations (discrete systems) can be use in al- ure 1). We refer to this quasi-linear relationship as the ternative. ‘adiabatic line’ [19]: this would be the region, in metric The motivation to use the density distance rely on space, populated by adiabatic systems and hence by a the theorem by Runge and Gross [34, 35] for continu- system evolving adiabatically. The adiabatic line for a ous systems, and on its extension to lattice Hamiltonians certain time-dependent process can then be written as [36, 37]. These, at least at zero temperature, provide a one-to-one correspondence between the driven system D (n (0), n (t)) ≈ mD (ρ (0), ρ (t)). n GS/Th GS/Th GS/Th GS/Th many-body state and the corresponding particle density. (12) This allows to shift the attention from the system’s quan- For adiabatic-enough systems, we can always assume tum states to the corresponding particle densities (con- that D (n(t), n (t)) ≤ D (n (0), n (t)), n GS/Th n GS/Th GS/Th tinuum) and site occupations (lattice Hamiltonians), ob- see Supporting Information, section 2, so that using (12), jects much simpler to calculate, e.g. by density functional we can write an upper bound for the adiabatic threshold methods [35, 37]. for the density distance, Δ , in terms of the correspond- n 4 ing threshold for the state as sites, with nearest-neighbour hopping is Δ = mΔ . (13) n ρ † † H = −J c ˆ c ˆ + c ˆ c ˆ i+1,σ i,σ i,σ i+1,σ The gradient m, will depends on N , U , and T , as well as i,σ on the type of driving potential. N N X X In principle a more accurate (and more + U n ˆ n ˆ + v n ˆ , (14) i,↑ i,↓ i i computationally expensive) estimate of Δ i i could be achieved by using a polynomial fit- ting to the curve D (n (0), n (t)) = n GS/Th GS/Th where J is the hopping parameter for an electron with f (D (ρ (0), ρ (t))), but we find that the GS/Th GS/Th spin σ, with σ =↑ or ↓, U is the on-site electron- linear approximation (12) and the simple method electron repulsion strength, and v is the external po- described above is sufficient for achieving good results † tential at site i. Also, c ˆ and c ˆ are the usual cre- i,σ i,σ (see Figs. 2 and 3). ation and annihilation operators for a spin-σ fermion on site i, and n ˆ = n ˆ + n ˆ is the number operator, with i i,↑ i,↓ n ˆ = c ˆ c ˆ . i,σ i,σ i,σ Estimate for the gradient of the adiabatic line The non-equilibrium dynamics is driven through the application for a time τ of a uniform electric field linearly In practice, the gradient m can be estimated increasing with time from a potential difference along the by calculating D (n (0), n (t)) and GS/Th GS/Th B chain of 1J to a potential difference of 10J . The on-site D (ρ (0), ρ (t)) for 2-3 values of t. For these ρ GS/Th GS/Th 0 τ potential at site i is then written as v (t) = μ + μ t/τ B B,max i i chosen values, D should be less than (2/3)D , ρ ρ 0 0 0 0 τ where μ = 2μ /N × i− μ where μ = 0.5J , and μ = i i and the origin should be included in the fit in virtue of τ τ τ 2μ /N × i− μ with μ = 4.5J . eq. (6). Estimating m requires then exact or approxi- The Hubbard model is used to simulate various phys- mate diagonalization of the system Hamiltonian at 2-3 ical systems of interest to quantum technologies [40, 43– instants in time. Of course at zero temperature only the 47], and the proposed dynamics could represent transient estimate of the GS is necessary. electronic currents along a chain of e.g. nanostructures (for example coupled quantum dots) or of atoms due to the application of a time-dependent electric field across NUMERICAL RESULTS the chain. It is noted that the final Hamiltonian does not depend While the methods proposed can be applied to both on the evolution time τ , and therefore the τ measures continuous and lattice systems, here we will illustrate also the inverse speed of the evolution. Hence for consid- them using the epitome for strongly correlated many- ering adiabatic evolutions, the larger τ is, the closer to body quantum systems, the Hubbard model, firstly at adiabaticity the system is expected to be. zero and then at finite temperatures. We have analysed the dynamics of short non-homogeneous Hubbard chains (N = 2, 4, 6), driven at different rates. In the following, we will discuss explicitly the results for N = 6, corre- Estimate for the density adiabatic threshold sponding to a Hamiltonian of size 400×400 at half-filling. The complexity of its spectrum may be appreciated by looking at the supporting information, figure 3. Curves for D (n (0), n (t)) against n GS/Th GS/Th D (ρ (0), ρ (t)) are shown in figure 1 for GS/Th GS/Th three temperatures (k T = 0, GS, left; k T = 0.2J , B B Hubbard model and system drive Th, middle; k T = 2.5J , Th, right). In figure 1 it can be seen how increasing U (from red to green to blue) or increasing temperature decreases the curves’ gradient. To demonstrate the properties of the methods for In calculating the adiabatic threshold, we have used the characterizing adiabaticity proposed in this work, the linear approximation (12) with m estimated as described out-of-equilibrium dynamics of the inhomogeneous one- in section “Estimate for the gradient of the adiabatic dimensional Hubbard model at half-filling is considered. line”. The approximated values for m can be found in The inhomogeneous Hubbard model is often used as a the supporting information (table 1, final column) for test-bed for developing techniques for strongly correlated all combinations of 3 values each of N , U , and T . many-body systems [10, 39] as it displays non-trivial properties even for the small chains [10, 11, 40–42] for In calculating Δ we have used the method described which it can be solved (numerically) exactly. The corre- in section “Estimate for the gradient of the adiabatic sponding Hamiltonian for a system of N fermions and N line”. 5 1.6 1.6 1.6 (a) (b) (c) 1.4 1.4 1.4 1.2 1.2 1.2 1 1 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.4 0.4 0.2 0.2 0.2 0 0 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 D( (0), (t)) ρ ρ D( (0), (t)) ρ ρ D( (0), (t)) ρ ρ GS Th Th FIG. 1: Curves D (n (0), n (t)) versus D (ρ (0), ρ (t)) for 3 interaction strengths: U = 0J in n GS/Th GS/Th GS/Th GS/Th red, U = 5J in green, and U = 10J in blue; and 3 temperatures, T = 0J/k (GS, left) T = 0.2J/k (Th, middle), B B T = 2.5J/k (Th, right). Note that n(0) = n (0) and ρ(0) = ρ (0) B GS/Th GS/Th Zero temperature A similar pattern occurs for U = 10J , figure 2(c), where though the avoided crossings happen only for t ≈ 95%τ (inset), when the applied potential becomes Predictions from (t) of the order of U . Once more, its Markovianity induces (t) to drop quickly afterwards. At zero temperature, the system initial state is the ground state: (t), as implemented, compares GS to all excited states and includes treatment of degeneracies ac- Adiabatic and non-adiabatic behaviour according to cording to [29]. In figure 2, panels (a)-(c) show (t) from D (ρ (t), ρ(t)) and D (n (t), n(t)) GS n GS eq. (2) with respect to time in units of τ . We consider different rates of dynamics (τ = 0.5/J , Figure 2 displays D (ρ (t), ρ(t)), panels (d)-(f ), and GS red, ‘fast’ dynamics; τ = 5/J , green, ‘intermediate’ dy- D (n (t), n(t)), panels (g)-(i) versus time for the same n GS namics; τ = 50/J , blue, ‘slow’ dynamics, closer to adia- parameters as (t) (panels (a)-(c)). [48]. The horizontal baticity), and three interaction strengths (U = 0J , left, dashed lines indicates the threshold Δ for the states’ dis- no interaction; U = 5J , middle, medium interaction; tances (panels (d)- (f )) and the corresponding threshold U = 10J , right, strong interaction). One would expect Δ for the particle density distances (panels (g)- (i)). that the red curves will demonstrate non-adiabatic be- For U = 0J and intermediate to fast dynamics, the pre- havior, whereas the blue curves should exhibit behavior B dictions from D (ρ (t), ρ(t)) and D (n (t), n(t)) are GS n GS closer to adiabaticity, and the green curves be somewhere in striking contrast with the predictions by (t). At very between the two. For U = 0J , figure 2(a), the initial dy- short times both metrics correctly predict a behaviour namics as described by (t) indeed follows these expec- close to adiabatic: the initial state is the GS and it will tations, though (t) predicts that the dynamics becomes take a finite time to the system state to significantly com- more adiabatic as time progresses, with in particular the bine with higher energy states. At intermediate to long τ = 5/J dynamics becoming adiabatic for t ∼ 0.5τ . times, while (t) would erroneously predict a fast return towards adiabaticity for the red and green dynamics, the For U = 5J , figure 2(b), many-body interactions be- metrics clearly show that the system remains far from come important and the static system would be in the adiabatic: the system dynamics far from equilibrium is process of transitioning between a metal and a quasi- highly affected by the trajectory in phase space at pre- Mott insulator (see e.g. [11, 21, 41]): states with double vious times (memory) and so considering a measure of occupation are ‘pushed up‘ in energy and the dynamics is adiabaticity which is non-Markovian – such as the pro- stiffen for low-enough applied potentials. Then, initially, posed metrics – becomes crucial to avoid false reading. all dynamics satisfy the QAC expressed by (2). However For slow dynamics (τ = 50), the behaviour remains at t ≈ 0.4τ the applied time-dependent potential has in- always at or below the adiabatic threshold. The oscil- creased enough to produce an avoided level-crossing in lations shown by D (ρ (t), ρ(t)) and D (n (t), n(t)) the low-energy spectrum of the instantaneous Hamilto- GS n GS nian (see inset in figure 2(b)) so that (t) predicts non- for τ = 50 were also observed in the single-electron sys- adiabatic behavior for both fast (τ = 0.5/J ) and inter- tems studied in reference [19]. These are explained by the mediate (τ = 5/J ) dynamics. However, as time increases system inertia in adjusting to gradually-applied electric further, according to (t), the dynamics quickly returns field. adiabatic for all dynamic rates. We note here that a tool For finite many-body interaction strengths, both directly derived from the QAC (1), such as (t) eq. (2), is D (ρ (t), ρ(t)) and D (n (t), n(t)) strongly respond GS n GS Markovian by construction, i.e. does not include memory to the (avoided) level crossings at t ≈ 0.4τ (U = 5J ) as it is based on instantaneous quantities. and t ≈ 9.5τ (U = 10J ), but, crucially also signal that D(n(0),n (t)) GS D(n(0),n (t)) Th D(n(0),n (t)) Th 6 10 50 1000 (a) (c) (b) −1.8 −1.1 8 40 −2.4 800 −3 −1.3 6 30 600 −3.6 −1.5 0.3 0.4 0.5 4 20 −1.7 2 10 200 0.8 0.84 0.88 0.92 0.96 1 0 0 0 (e) (f) (d) 1.2 1.2 1.2 1 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.4 0.4 0.2 0.2 0.2 0 0 (h) (i) 0.6 0.3 0.4 0.5 0.3 0.4 0.2 0.3 0.2 0.2 0.1 0.1 0.1 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 τ t/τ t/τ t/ FIG. 2: Zero temperature results. Red lines: τ = 0.5/J dynamics, green lines: τ = 5/J dynamics, blue lines: τ = 50/J dynamics. Panels show: (t) [(a)-(c)], D (ρ (t), ρ(t)) [(d)-(f )], and D (n (t), n(t)) [(g)-(i)] versus t/τ . GS n GS Three interaction strengths are considered: U = 0J (left), U = 5J (middle), and U = 10J (right). In all panels the horizontal dashed lines indicate the corresponding adiabatic threshold. Insets of panel (b) and (c): low energy spectrum of the instantaneous Hamiltonian versus t/τ for U = 5 [panel (b)] and U = 10 [panel (c)]. afterwards the system dynamics remain strongly non- Finite temperature adiabatic, with then important contributions from mem- ory effects. Note that this is the case even for the slow A thermal bath at temperature T is now connected to dynamics (τ = 50): compare blue lines in panels (a), (c), the Hubbard chain, to thermalize the system. Once ther- (h) for t > 0.5τ . malized, at t = 0 , the bath is disconnected, and then the closed system is evolved from t = 0 to τ . Therefore, the initial state is now a thermal state, with a corresponding D (ρ (t), ρ(t)), as distance between the system GS thermal particle density. Because of the closed dynamics, quantum state and its adiabatic counterpart, can be read- we will then consider the distance between the dynami- ily associated to the definition of adiabaticity; this is less cal system state ρ(t) and its finite-temperature adiabatic so for D (n (t), n(t)): particle densities, being just a n GS counterpart function of position and time, could be expected to be much less sensitive to details than the system state, e.g. it might be less sensitive to details of the instantaneous j,0 Hamiltonian spectrum, or less susceptible to dynamical exp− k T ρ (t) = |ψ ihψ | , (15) Th j,t j,t changes of the system and corresponding memory effects. E k,0 exp− j k k T However, because of the theorems in [34, 36], we know that the considered dynamical system state and its cor- where E is the j-th eigenenergy of the Hamiltonian responding particle density contain the same amount of j,0 at t = 0, and |ψ i is the j-th eigenstate of the in- information: we can then conjecture that both the re- j,t stantaneous Hamiltonian at time t. The corresponding lated ‘natural’ [21] metrics can be used successfully as particle density n (t) is used in the density distance measures of adiabaticity. This was confirmed in [19] for Th D (n (t), n(t)). single-particle systems, and here for many-body systems. n Th This leads to the possibility of characterizing adiabatic- Two temperatures are considered in this work; a lower ity using the sole particle density, a quantity much more temperature of k T = 0.2J , and a higher temperature accessible than the corresponding system quantum state. of k T = 2.5J . D(n (t),n(t)) D( (t), (t)) ρ GS GS ε 7 Low temperature the occupation probability of the two relevant eigenstates before the crossing. For t < t , we can estimate each of ac the 20 lower-band eigenstates to have roughly 1/20 oc- For k T = 0.2J , (t) and both metrics show the cupation probability, and the upper band having no oc- systems to behave mostly very similarly to the zero- cupation. The change in the Bures distance across each temperature case. A notable difference occurs for U = 10 B,max crossing would then be about D /20 = 0.07, giving and 0.9 < t/τ < 1, where the inset of fig. 2(c) shows the an overall height for the six steps of D of 0.42, which occurrence of four low-energy avoided crossings. Due to ρ is reasonably close to what we observe in fig. 3(f ). A the finite-temperature state mixing, both metrics signal similar structure is faithfully signalled by D . The other the four crossings with corresponding steps in the dis- n anti-crossings, which occur deeper in the lower band, are tances, while, (t) remains sensitive only to the crossing between eigenstates with very similar occupation proba- occurring at t/τ ≈ 0.96 between ground and first excited bilities, so that the overall state should be expected to state. These results suggest that for low temperatures, change very little at crossings: this is faithfully captured k T  J , the density could be used as a good indicator by the chosen metrics, much less affected by those anti- to characterize adiabaticity. For completeness, we report crossings. The T-QAC measure (t) presents a series of all results for k T = 0.2J in the supporting information, peaks in the region of where the bands cross, but without figure 1. distinguishing between the anti-crossing being at the top, or deeper within, the lower energy band. Importantly we find that the anti-crossings affecting (t) are often not the High temperature ones expected to substantially change the system state. The problem of (t) in signaling inappropriately the For k T = 2.5J and thermal equilibrium, tens of anti-crossings is even more evident (and problematic) for eigenstates of the initial Hamiltonian spectrum are sig- U = 5: here the crossing between lowest and immedi- nificantly populated (initial state). For U = 0, the be- ately upper bands starts already at t/τ ≈ 0.05 [we re- haviour of (t) and of both metrics is qualitatively simi- port the relevant part of the instantaneous spectrum ver- lar to the lower temperatures examined: no anti-crossing sus time in the supporting information, figure 3(b)]. As are observed within ∼ k T of the instantaneous ground there is no substantial gap between them at t = 0, the state, while energy gaps in this part of the spectrum tend top levels of the lowest and the lower levels of this up- to increase with time. We report part of the instanta- per band are fairly similarly populated. This means that neous spectrum versus time in the supporting informa- the corresponding mixed system state does not change tion, figure 3(a). significantly at each of these crossings, as correctly dis- At U = 10, many-body interactions creates distinct played by both metrics. However (t) dramatically sig- bands in the instantaneous Hamiltonian spectrum, shown nals the initial anti-crossings, thus proving a false reading for completeness in the supporting information, figure of non-adiabaticity already at t/τ ≈ 0.05 [see inset of fig- 3(c). The eigenstates in the lowest energy band are linear ure 3(b)]. These spikes in (t) may be related to the combinations of the 20 possible permutation of single-site problem of small denominators for this type of measures, occupations at half-filling. Even at t = 0, the system is see [50]. slightly inhomogeneous, so these eigenstates are not de- Although in this work the degenerate form of QAC generate. The next band contains six eigenstates, com- was adapted for finite temperature, the results show that binations of states which may allow double occupation it is still not well suited for high T . On the other side, in one of the sites, and the bandgap due to this on-site the metrics, which naturally include degeneracy and non- Coulomb repulsion is about 6J at t = 0, substantially Markovianity, can be seen to cope well with the temper- larger than k T , so that initially only the lower band ature increase. is significantly occupied. For all dynamical rates consid- ered, this gap is also much larger than 1/τ , and indeed all measures considered remain below or close to their adia- batic thresholds until the two bands start (anti) crossing Results for the trace distance at t /τ ≈ 0.8, see fig. 3(c), (f ) and (i). We note six- ac < < steps in both metrics for t ∼ t ∼ 0.9τ , see figs. 3(f ) ac and (i). Each step signals one of the upper-band eigen- With respect to its own adiabatic threshold[51], the states starting to anti-cross the lower band. The Bures trace distance results quantitatively close to the Bures distance between two orthogonal pure states is maximal, distance, including signaling with steps relevant anti- so the Bures distance between the system state and its crossings. This means that it can be used as an alterna- adiabatic reference is in principle set up to signal non- tive quantitative measure of adiabaticity[52]. For com- adiabatic behaviour at any avoided crossing[49]. However pleteness, we report the related results in the supporting the strength of the signal will depend on how different is information, figure 2. 8 16 120000 (a) (c) (b) 10 120000 6 80000 4 0 0.05 0.1 0.15 0.2 0.25 0.3 0 0 0 (d) (e) (f) 0.4 0.6 0.5 0.35 0.5 0.3 0.4 0.4 0.25 0.3 0.2 0.3 0.15 0.2 0.2 0.1 0.1 0.1 0.05 0 0 (g) (h) (i) 0.16 0.25 0.1 0.14 0.12 0.2 0.08 0.1 0.15 0.06 0.08 0.06 0.1 0.04 0.04 0.05 0.02 0.02 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 t/τ t/τ t/τ FIG. 3: High temperature (T = 2.5J/k ) results. Red lines: τ = 0.5/J dynamics, green lines: τ = 5/J dynamics, blue lines: τ = 50/J dynamics. Panels show: (t) [(a)-(c)], D (ρ (t), ρ(t)) [(d)-(f )], and D (n (t), n(t)) [(g)-(i)] Th n Th versus t/τ . Three interaction strengths are considered: U = 0J (left), U = 5J (middle), and U = 10J (right). In all panels the horizontal dashed lines indicate the corresponding adiabatic threshold. Inset of panel (b): zoom into short times of main panel. CONCLUSION old for the system state metric to an upper bound for the threshold for the local particle density metric. This up- per bound is tight enough along most parts of the time- We have introduced methods based on appropriate evolutions analysed, and it is relatively easy to estimate, metrics to measure adiabaticity for the dynamics of even for large systems. We aim to refine it as future work. many-body quantum systems at finite temperature, and We discuss an extension to finite temperature of the to track it with time evolution. As system state met- quantum adiabatic criterion which include treatment of rics, Bures and trace distances give consistently similar degeneracies. Comparing the results from this and the predictions, for all dynamics and temperatures analysed. metrics has highlighted the importance of properly in- Additionally, our results support the conjecture – based cluding memory effects when wishing to evaluate and on the fundamental theorems of time-dependent density track the adiabatic level of a many-body dynamics: by functional theory – that the ‘natural’ metric tracking the construction, a measure based on the quantum adiabatic evolution of the local particle density alone would be suf- criterion is basically Markovian, as, at most, the instan- ficient to estimate the level of adiabaticity of the systems’ taneous Hamiltonian derivative is included. Our results dynamics. This is a great simplification as, in general, show that this leads to false readings, as highly out-of- the system local particle density may be estimated more equilibrium dynamics may be pictured as adiabatic. Our accurately and much more simply than the correspond- results have also shown that while the metric-based meth- ing many-body system state. It is also an experimentally ods correctly reflect the amount of change in the system measurable quantity, which opens additional possibilities state at instantaneous eigenenergy anti-crossings, the ex- for the method. tension to finite temperatures of the quantum adiabatic Because these metrics have a finite maximum, they criterion is often sensitive – and sometimes extremely are suitable for the design of practical ’adiabatic thresh- sensitive – to the anticrossing where the actual many- olds’, so that a distance below (above) the corresponding body state does not change significantly. Once more this threshold signals adiabatic (non-adiabatic) behaviour. may lead to false readings, this time predicting the sys- Using the results in this paper and previous results, we tem to be far from adiabaticity while it is actually still have been able to consistently relate the adiabatic thresh- behaving adiabatically. D(n (t),n(t)) D( (t), (t)) ρ ρ Th Th ε 9 We acknowledge fruitful discussions with V. V. Franca ity, especially when considering applications to quantum systems. and thank K. Zawadzki for the code for the Hubbard [28] M. V. Berry, Journal of Physics A 42, 36 (2009). chain dynamics; AHS acknowledges support from EP- [29] G. Rigolin and G. Ortiz, Phys. Rev. A 85, 062111 (2012). SRC. [30] W. Sutherland, Introduction to Metric and Topological Spaces (Oxford University Press, Oxford, 2009). [31] V. V. Fran¸ ca, J. P. Coe, and I. D’Amico, Scientific Re- ports 8 (2018). √ √ 2 [32] F (σ, ρ) = Tr ρσ ρ is known as fidelity, which [1] T. Albash and D. A. Lidar, “Adiabatic quantum com- is also used in the literature for estimating how simi- puting,” arXiv:1611.04471. lar two quantum states are, though cannot be considered [2] E. Farhi, J. Goldstone, S. Gutmann, J. Lapan, A. Lund- a proper distance, as it does not obey all the metrics’ gren, and D. Preda, Science 292, 472 (2001). axioms. [3] M. Gell-Mann and F. Low, Phys. Rev. 84, 350 (1951). [33] I. Bengtsson and K. Zyczkowski, Geometry of Quan- [4] D. Bacon and S. T. Flammia, Phys. Rev. Lett. 103, tum States: An Introduction to Quantum Entanglement 120504 (2009). (Cambridge University Press, 2006). [5] I. Hen, Phys. Rev. A 91, 022309 (2015). [34] E. Runge and E. K. U. Gross, Phys. Rev. Lett. 52, 997 [6] A. C. Santos, R. D. Silva, and M. S. Sarandy, Phys. Rev. (1984). A 93, 012311 (2016). [35] C. A. Ullrich, Time-Dependent Density-Functional The- [7] O. Abah and E. Lutz, Eur. Lett. (EPL) 118, 40005 ory: Concepts and Applications. (Oxford University (2017). Press, 2013). [8] J. He, J. Chen, and B. Hua, Phys. Rev. E 65, 036145 [36] C. Verdozzi, Phys. Rev. Lett. 101, 166401 (2008). (2002). [37] K. Capelle and V. L. Campo, Physics Reports 528, 91 [9] C.-K. Hu, J.-M. Cui, A. C. Santos, Y.-F. Huang, C.-F. (2013). Li, G.-C. Guo, F. Brito, and M. S. Sarandy, Scientific [38] See e.g. [21], figure 2. Reports 9, 10449 (2019). [39] M. Herrera, K. Zawadzki, and I. D’Amico, The European [10] M. Herrera, R. M. Serra, and I. D’Amico, Scientific Re- Physical Journal B 91, 248 (2018). ports 7, 4655 (2017). [40] S. Murmann, A. Bergschneider, V. M. Klinkhamer, [11] A. Skelt, K. Zawadzki, and I. D’Amico, Journal of G. Zurn, ¨ T. Lompe, and S. Jochim, Phys. Rev. Lett. Physics A: Mathematical and Theoretical 52, 48 (2019). 114, 080402 (2015). [12] X. Chen, A. Ruschhaupt, S. Schmidt, A. del Campo, [41] D. J. Carrascal, J. Ferrer, J. C. Smith, and K. Burke, J. D. Gu´ ery-Odelin, and J. G. Muga, Phys. Rev. Lett. 104, Phys. Cond. Mat. 27(39), 393001 (2015). 063002 (2010). [42] K. Zawadzki, I. D’Amico, and L. Oliveira, Brazilian [13] A. del Campo and K. Kim, New Journal of Physics 21, Journal of Physics 47, 488 (2017). 050201 (2019). [43] J. P. Coe, V. V. Fran¸ ca, and I. D’Amico, Phys. Rev. A [14] M. Born and V. A. Fock, Z. Phys. A 51, 165–180 (1928). 81, 052321 (2010). [15] J. E. Avron and A. Elgart, Communications in Mathe- [44] S. Yang, X. Wang, and S. Das Sarma, Phys. Rev. B 83, matical Physics 203, 445–463 (1999). 161301 (2011). [16] K.-P. Marzlin and B. C. Sanders, Phys. Rev. Lett. 93, [45] J. P. Coe, V. V. Fran¸ ca, and I. D’Amico, EPL (Euro- 160408 (2004). physics Letters) 93, 10001 (2011). [17] D. M. Tong, K. Singh, L. C. Kwek, and C. H. Oh, Phys. [46] P. T. Brown, D. Mitra, E. Guardado-Sanchez, Rev. Lett. 95, 110407 (2005). R. Nourafkan, A. Reymbaut, C.-D. H´ ebert, S. Berg- [18] J. Ortigoso, Phys. Rev. A 86, 032121 (2012). eron, A.-M. S. Tremblay, J. Kokalj, D. A. Huse, [19] A. H. Skelt, R. W. Godby, and I. D’Amico, Phys. Rev. P. Schauß, and W. S. Bakr, Science 363, 379 (2019), A 98, 012104 (2018). https://science.sciencemag.org/content/363/6425/379.full.pdf. [20] O. Lychkovskiy, O. Gamayun, and V. Cheianov, Phys. [47] M. A. Nichols, L. W. Cheuk, M. Okan, T. R. Hartke, Rev. B 98, 024307 (2018). E. Mendez, T. Senthil, E. Khatami, H. Zhang, [21] I. D’Amico, J. P. Coe, V. V. Fran¸ ca, and K. Capelle, and M. W. Zwierlein, Science 363, 383 (2019), Phys. Rev. Lett. 106, 050401 (2011). https://science.sciencemag.org/content/363/6425/383.full.pdf. [22] P. M. Sharp and I. D’Amico, Phys. Rev. B 89, 115137 [48] Corresponding results for the trace distance will be dis- (2014). cussed in section “Results for the trace distance” and in [23] P. M. Sharp and I. D’Amico, Phys. Rev. A 92, 032509 the supporting information. (2015). [49] In an anti-crossing, the component of the dynamical state [24] S. Marocchi, S. Pittalis, and I. D’Amico, Phys. Rev. behaving non-adiabatically will be orthogonal to the cor- Materials 1, 043801 (2017). responding component of the adiabatic reference state. [25] M. M. Wilde, Quantum Information Theory (Cambridge [50] M. Kolodrubetz, D. Sels, P. Mehta, and A. Polkovnikov, University Press, 2013) Chap. 9. Physics Reports 697, 1 (2017). [26] A. H. Skelt, R. W. Godby, and I. D’Amico, Brazilian [51] Remember that the maximum value of the trace distance Journal of Physics 48, 467 (2018). is 1 for normalized states, therefore the numerical values [27] Since temperature is being introduced, there are two def- of the distance will be different to those of the Bures initions of adiabaticity; quantum and thermal. Thermal distance. adiabaticity looks at the heat loss of the system, which is [52] A comparison between estimates from the Kullback rel- zero for this investigation as the system is closed. There- ative entropy and results from the trace distance for the fore it only makes sense to look at the quantum adiabatic- 10 distance of mixed states from thermal equilibrium can be different parameter values. found in figure 4 of [53]. There it is found that the be- [53] K. Zawadzki, R. M. Serra, and I. D’Amico, haviour of the two quantities have qualitatively similar arxiv:1908.06488 (2019). features, but, for example, the maxima/minima occur at Characterizing Adiabaticity in Quantum Many-Body Systems at Finite Temperature – Supporting Information A. H. Skelt Department of Physics, University of York, York YO10 5DD, United Kingdom I. D’Amico Department of Physics, University of York, UK and International Institute of Physics, Federal University of Rio Grande do Norte, Natal, Brazil (Dated: April 14, 2020) TABLE OF GRADIENTS FOR ADIABATICITY (t), BURES DISTANCE, AND PARTICLE DENSITY DISTANCE FOR T = 0.2J/k Table I gives the gradients m for the Hubbard model Figure 1 shows the results for (t), the Bures distance, at 3 values of N , of U , and of T . This could be used as and the particle density distance when T = 0.2J/k (low a guide for the expected gradient of the adiabatic line in temperature). different systems. TRACE DISTANCE We look at how the trace distance copes with charac- DEMONSTRATION THAT, FOR terizing the level of adiabaticity. For this we will use ADIABATIC-ENOUGH SYSTEMS, D (n(t), n (t)) ≤ D (n (0), n (t)) q n GS/Th n GS/Th GS/Th 1 1 D (ρ, σ) = Tr [|ρ− σ|] = Tr (ρ− σ) (ρ− σ). (5) 2 2 A key formal property of metrics is that Beginning at zero temperature, we compare figure 2 (a)- (c) here to figure 2 (d)-(f ) in the main text. For low D(A, B) = 0 if and only if A = B (1) temperature we compare figure 2 (d)-(f ) here to figure 1 (d)-(f ) also here. And for the high temperature, we com- For dynamics starting from the ground or thermal state, pare figure 2 (g)-(h) here to figure 3 (d)-(f ) in the main this property implies that, at t = 0, text. Qualitatively the Bures and trace distances agree for all U . The quantitative difference is mainly associated D (n(t = 0), n (t = 0)) = 0 n GS/Th to the difference in the maximum distances. For all temperatures, the same conclusions that were and, of course, D (n (t = 0), n (t = 0)) = 0. n GS/Th GS/Th drawn for the Bures distance can be drawn for the trace This satisfies the claim for t = 0. distance, which can therefore be used to characterize adi- For t > 0 and a generic time-dependent Hamiltonian, abatic evolutions. equation (1) implies that D (n (0), n (t)) > 0, (2) n GS/Th GS/Th INSTANTANEOUS EIGEN-ENERGIES OF THE FERMI-HUBBARD HAMILTONIAN as in general n (0) 6= n (t). [? ] GS/Th GS/Th At the same time, for a perfectly adiabatic dynamics, Figure 3 shows the low-mid section of the instanta- the definition of adiabaticity would impose that, at all neous spectrum for the driven Hubbard Hamiltonian con- times, sidered in this work with respect to time, and for (a) U = 0J , (b) U = 5J , and (c) U = 10J . For U = 10J , D (n(t), n (t)) = 0. (3) n GS/Th the Coulomb repulsion for states which include double occupation of site(s) leads to the formation of energy Equations (2) and (3) allow us to state that, for systems bands. The lowest two bands start crossing when the whose dynamics is close enough to adiabaticity, we can applied external potential is of the order of U , around always assume that t/τ = 0.8. For U = 5J the gap between the lowest two energy bands is just starting to form, and the two bands D (n(t), n (t)) ≤ D (n (0), n (t)). (4) start crossing for t/τ ≈ 0.05. n GS/Th n GS/Th GS/Th arXiv:2004.05842v1 [quant-ph] 13 Apr 2020 2 k T U N Gradient N Gradient N Gradient 0J 0J 2 1.33294 4 0.647511 6 0.450959 0J 5J 2 0.91163 4 0.411435 6 0.2779 0J 10J 2 0.502006 4 0.218927 6 0.237795 0.2J 0J 2 0.924189 4 0.647354 6 0.45059 0.2J 5J 2 0.911343 4 0.409499 6 0.278244 0.2J 10J 2 0.619401 4 0.210794 6 0.205762 2.5J 0J 2 0.924189 4 0.57114 6 0.444165 2.5J 5J 2 0.744725 4 0.473781 6 0.389355 2.5J 10J 2 0.48877 4 0.231146 6 0.179414 TABLE I: Gradients m of the adiabatic line for the driven Hubbard model considered in this paper for three temperatures (k T ), correlation strengths (U ), and electron numbers (N ). 50 1000 (a) (b) (c) 40 800 6 30 600 4 20 400 2 10 200 0 0 0 (d) (e) (f) 1.2 1.2 1 1 0.8 0.8 0.8 0.6 0.6 0.6 0.4 0.4 0.4 0.2 0.2 0.2 0 0 0 (g) (h) (i) 0.6 0.25 0.4 0.5 0.2 0.3 0.4 0.15 0.3 0.2 0.1 0.2 0.1 0.05 0.1 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 t/τ t/τ t/τ FIG. 1: Low temperature (T = 0.2J/k ) results. Red lines: τ = 0.5/J dynamics, green lines: τ = 5/J dynamics, blue lines: τ = 50/J dynamics. Panels show: (t) [(a)-(c)], D (ρ (t), ρ(t)) [(d)-(f )], and D (n (t), n(t)) [(g)-(i)] Th n Th versus t/τ . Three interaction strengths are considered: U = 0J (left), U = 5J (middle), and U = 10J (right). In all panels the horizontal dashed lines indicate the corresponding adiabatic threshold. D( (t), (t)) ρ D(n (t),n(t)) Th Th ε 3 1 1 1 (b) (c) 0.9 0.9 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.6 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.4 0.3 0.3 0.3 0.2 0.2 0.2 0.1 0.1 0.1 1 0.9 (d) (f) (e) 0.9 0.9 0.8 0.8 0.8 0.7 0.7 0.7 0.6 0.6 0.6 0.5 0.5 0.5 0.4 0.4 0.4 0.3 0.3 0.3 0.2 0.2 0.2 0.1 0.1 0.1 0.4 0.6 0.35 (g) (h) (i) 0.35 0.3 0.5 0.3 0.25 0.4 0.25 0.2 0.2 0.3 0.15 0.15 0.2 0.1 0.1 0.1 0.05 0.05 0 0 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 t/τ t/τ t/τ FIG. 2: All figures here show D (ρ (t), ρ(t)) against t/τ . Interaction strengths: U = 0J (left), U = 5J ρ GS(Th) (middle), and U = 10J (right); red lines corresponds to τ = 0.5/J , green lines to τ = 5/J , and blue lines to τ = 50/J . Panels (a)-(c) show the zero temperature results, T = 0J/k ; (d)-(f ) show the low temperature results, T = 0.2J/k ; (g)-(i) show the high temperature results, T = 2.5J/k . B B −2 −4 −6 −8 −10 −2 −12 −4 −14 −16 −6 −18 −1 (a) (b) (c) −20 −8 −2 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 τ τ t/ t/ t/ FIG. 3: Zoom into the low-mid section of the instantaneous spectrum versus t/τ of the time-dependent Hubbard Hamiltonian considered in this work (6 sites, half filling). Panel (a) corresponds to zero on-site Coulomb interaction (U = 0J ), (b) to U = 5J , and (c) to U = 10J . Note the different energy scales on the y-axis. E D( (t), (t)) ρ D( (t), (t)) ρ ρ D( (t), (t)) ρ ρ Th Th GS

Journal

Quantum Physics – arXiv (Cornell University)

Published: Apr 13, 2020

There are no references for this article.