Arxiv:2105.14040V1 [Quant-Ph] 28 May 2021 (ACL) Model

Adapted Caldeira-Leggett Model

Andreas Albrecht∗ and Rose Baunach† Center for Quantum Mathematics and Physics and Department of Physics and Astronomy UC Davis, One Shields Ave, Davis CA.

Andrew Arrasmith‡ Theoretical Division, Los Alamos National Laboratory, Los Alamos, NM USA. (Dated: June 1, 2021) We preset a variant of the Caldeira-Leggett (CL) model of a harmonic oscillator coupled to an environment. The CL model is a standard tool for studying the physics of decoherence. Our “adapted Caldeira-Leggett” (ACL) model is built in a ﬁnite Hilbert space which makes it suitable for numerical studies. Taking a numerical approach allows us to avoid the limitations of standard approximation schemes used with the CL model. We are able to evolve the ACL model in a fully reversible unitary manner, without the built-in time asymmetry and other assumptions that come with the master equation methods typically used. We have used the ACL model to study new topics in the ﬁeld of decoherence and einselection where the full unitary evolution is essential to our work. Those results (reported in companion papers) include an examination of the relationship between einselection and the arrow of time, and studies of the very earliest stages of einselection. This paper provides details about the ACL model and our numerical methods. Our numerical approach makes it straightforward to explore and plot any property of the physical system. Thus we believe the examples and illustrations we present here may provide a helpful resource for those wishing to improve their familiarity with standard decoherence results, as well as those looking to probe the underpinnings of our companion papers. We expect the ACL model will be a useful tool for exploring additional phenomena that cannot be studied using traditional approximation schemes.

I. INTRODUCTION construct the ACL model, we present here results from “putting it through its paces” which demonstrate that The Caldeira-Leggett (CL) model is a toy model de- the ACL model does a good job of reproducing physics scribing a particle which moves in its own potential and is phenomena that are an established part of the decoher- also coupled to an environment [1–3]. The environment is ence literature. These cross-checks give us a solid foun- usually treated as an infinite set of harmonic oscillators, dation on which to explore the new directions, which we and the particle is often taken to move in a harmonic po- report in companion papers [6, 7]. We do not expect the tential as well. The particle plus environment describe a phenomena presented in this paper to be new to an ex- closed system which can in principle be treated quantum pert on decoherence. On the other hand, someone learn- mechanically as a system undergoing reversible unitary ing this topic might find our graphical presentation cen- evolution. In practice the CL model is often treated in tered on a specific physical system a useful compliment to the “Markovian limit” where the particle evolution can a more thorough review such as [8] and may even provide be described by an irreversible master equation. Working a helpful starting point. in this limit provides tractable mathematics which can be The development of the ACL model was originally mo- used to study particle-environment interactions in situa- tivated by our interest in exploring the physics of einselec- tions which naturally have an arrow of time. For example tion under equilibrium conditions1. The Markovian limit, the CL model has been used in pioneering explorations with its definite arrow of time, clearly cannot describe of decoherence [4] and einselection [5]. the full fluctuations of an equilibrium system. We also This paper introduces an “adapted Caldeira-Leggett” expect the ACL model will be useful in exploring other arXiv:2105.14040v1 [quant-ph] 28 May 2021 (ACL) model. The adaptations are chosen to reproduce physics outside of the Markovian regime, and we have the essential features of the CL model as fully as possi- already found one such example (which we’ve named the ble within a finite Hilbert space. The goal is to be able “copycat process”) that we mention briefly in Sect. III E to evolve the ACL model easily on a desktop computer and develop further in a companion paper [7]. in its full unitary form, thus enabling the convenient ex- We organize this paper as follows. Section II defines ploration of a more complete range of physical situations the ACL model and demonstrates the robustness of our including those outside the Markovian limit. numerical calculations. Section III explores a variety Aside from describing various technicalities of how we

1 These motivations originate in cosmology where connections be- ∗ [email protected] tween the emergence of classicality (related to einselection) and † [email protected] the arrow of time (which originates with cosmology, as discussed ‡ [email protected] for example in [9]) might lead to useful insights. 2 of standard results from the literature using the ACL model. For example we show how an initial Schrödinger cat state of superposed wavepackets is einselected to a classical mixture of single packets. We also introduce the “copycat process,” a new phenomenon which we explore extensively in [7]. Section IV explores the way the ACL model both approaches and then remains solidly situ- q3 ated in a fluctuating equilibrium state when evolved long enough. The presence of a fully fluctuating equilibrium q2 state is a behavior not accessible through master equation techniques, but one which is very naturally achieved with q our methods. This equilibrium behavior forms a founda- 1 tion for our exploration of the relationship between eins- e Action of H I election and the arrow of time in [6]. In Sect. V we introduce the “reduced Caldeira-Leggett” (RCL) model which i e replaces the SHO with a single qubit. We demonstrate Action of H how the RCL model can access a different set of phenomena. The results from this paper are placed in the context of the existing literature in Sect. VI. Among other things, we relate some of our results to Nuclear Mag- I netic Resonance (NMR) physics and “Loschmidt echos” FIG. 1. The interaction term q ⊗ He moves the initial environment state along a specific path in the e Hilbert space (a concept developed in discussions of the arrow of time). I Section VII presents our conclusions. determined by He , illustrated by the solid curve. The rate of movement along this path is proportional to the value of q, A series of appendices present additional technical and that allows different q states to become entangled with information. Appendix A explores einselection in the I different environment states. In the case where H ,He 6= 0, “quantum limit” of the ACL model. Appendix B presents e the action of He can push the evolution off the original path details of the eigenstates of the truncated SHO, which re- in a variety of different directions depending on the starting veal differences between the truncated and the continuum point (∝ the value of q). These various paths are illustrated by cases. We give a detailed picture of the spectra of the dif- the dashed curves. The non-commuting property can make ferent Hamiltonians (SHO, environment and combined) the process of entanglement much more efficient (especially in Appendix C. Appendix D presents our numerical tech- for the large Ne case, not shown in this sketch). niques.

positions become entangled with different environment I II. THE ACL MODEL states. When He and He don’t commute (the case for both CL and ACL models) the entangling process is much We consider a “world” Hilbert space w = s ⊗ e which more effective, as illustrated heuristically in Fig. 1. is a tensor product of a “system” Hilbert space s and the environment space e. We consider a Hamiltonian of the form A. The SHO H = H ⊗ 1e + q ⊗ HI + 1s ⊗ H . (1) w s s e e For a normal (un-truncated) SHO the matrix elements Equation 1 describes the form of both the CL and ACL of the lowering operator a in the basis given by number models. The differences arise in the specifics of the dif- (or energy) eigenstates is given by ferent ingredients. These are the system Hamiltonian p Hs, the self-Hamiltonian for the environment He, and hi| a |ji = jδi,j−1 (2) the piece of the interaction Hamiltonian in the e sub- I space, He . We focus on the case where s is a simple with j ≥ 1. For our truncated SHO the same formula harmonic oscillator (SHO). The position operator of the is valid for â (where the hat denotes the truncated ver- system, qs = qSHO, is defined in the usual way for the sion) but it only applies for {i, j ∈ 1 : Ns} where Ns is CL model. However for the ACL model s is a “truncated the size of the truncated SHO Hilbert space. The oper- † SHO” (in order to allow a numerical treatment) and the ator â is formed by conjugating â, and ˆq, ˆp and Hˆ SHO † definition of qSHO for that case is nontrivial. Hamilto- are all constructed from â and â using the usual formu- nians of this form have features that enable the system las from the un-truncated case. These operators in the to become entangled with the environment in ways that truncated space don’t have all the usual properties due reflect certain realistic physical situations. The inter- to the truncation. For example action term changes the state of the environment with † a strength proportional to the value of qs, so different â, â = 1 + ∆ (3) 3

0.4 0.4 t = 0 0.2 t=0 0.2 0 0 -10 -5 5 10 -10 -5 0 5 10 q 0.4 0.4 7 0.2 t = 10 0.2 t= /4 0 0 -10 -5 q 5 10 -10 -5 0 5 10 1 0.4 residuals 10-10 0.2 t= /2 10-20 0 q -10 -5 5 10

-10 -5 q 5 10

FIG. 3. A coherent state at t = 0 and t = 107τ. The third panel shows the residuals for the probabilities (solid) and for FIG. 2. A coherent state wavefunction (squared) for the trun- the real (dotted) and imaginary (dot-dashed) part of the am- cated SHO shown at diﬀerent points in its period τ. Despite plitude. These curves illustrate that the numerically evolved certain diﬀerences from the continuum case noted in the text, truncated model reproduces the periodic properties expected the shape and robustness under evolution of this state corre- of the continuum case to an excellent degree of accuracy. sponds to the properties of continuum coherent states.

residuals. The very small sizes of the residuals further where ∆(i, j) = −Nsδi,Ns δj,Ns . We chose these defini- demonstrate the robust nature of the truncated SHO. tions for the truncated operators because they have some The specifics of our numerical approach (including sev- practical advantages over other choices. The main ad- eral additional checks) are discussed in Appendix D. vantage is illustrated in Figs. 2 and 3. Figure 2 shows a coherent state constructed thus: B. The interaction and environment † ∗ ψα (q) = hq | αi = hq| exp αâ − α â |0i (4) self-Hamiltonian where |0i is the ground state of Hˆ and hq| is the q SHO The interaction Hamiltonian has the form q ⊗HI . For eigenstate of ˆq. The x axis gives the eigenvalue of ˆq, s e the ACL model we use q = ˆq. The environment piece, which is really a discrete quantity (ˆq has only N eigen- s s HI , has the form values, which run from −2π to 2π). The discrete sets of e points plotted (shown by markers) are connected only to I e 0 H = EI R + E . (5) reference the continuum of the un-truncated SHO which e I I 2 e this system is intended to approximate . We call the The matrix RI is a random matrix constructed by draw- SHO period τ and in our units τ = 2π. We’ve taken ing each of the real and imaginary parts of each indepen- Ns = 30 here, and in all the examples shown in this dent matrix element of a Ne × Ne Hermitian matrix uni- paper. formly from the interval [−0.5, 0.5] using the computer’s The top two panels of Fig. 3 show the same coherent random number generator. state at t = 0 and t = 107τ. The third panel shows the The environment self-Hamiltonian is given by

e 0 He = EeR + Ee (6)

2 e e The truncated form does lead to some novel features in the eigen- where R is constructed in the same manner as RI , but states of H as discussed in Appendix B. s as a separate realization. In Eqns. 5 and 6, EI and Ee are 4 c-numbers which parameterize the overall energy scales. e e Both RI and R are fixed initially and are not changed during the time evolution. The full Hamiltonian of the ACL model is time independent. All the results in this 0 0 paper use EI = Ee = 0, but we have found nonzero 0.2 t = 0 values for these offset parameters to be helpful for other 0.1 calculations we report elsewhere. I 0 The job of He and He is to move states around in the environment efficiently, so that entanglement between -0.1 the SHO and the environment can emerge as fully as -10 -5 0 5 10 possible despite working within the confines of a finite q 3 system . We find the random form of these operators 0.2 I t = /4 does this job well, and since He ,He is just another random matrix the non-commutivity discussed with Fig. 1 is 0.1 easily achieved. The work presented here uses Ne = 600. 0 This choice, along with Ns = 30, was made via an in- -0.1 formal optimization process to maximize the utility of -10 -5 0 5 10 the ACL model within the constrained resources of our q desktop computer. 0.2 t = /2 There is also a simple way to modify our ACL model 0.1 to create He’s with different spectra. The crucial aspect achieved by the random matrices in Hw is the non- 0 I commutivity of He and He. This aspect is enabled by the -0.1 eigenvectors of independently generated random matri- -10 -5 0 5 10 ces in large spaces having very little overlap. One could I q alternatively create He and He by starting in diagonal form (with a spectrum of eigenvalues of your choosing) and then changing basis using a random unitary to produce a “random matrix” with the specified eigenvalue FIG. 4. Evolution of a “SchrödingerCat” superposition of spectrum. We experimented a bit with this approach to coherent states (specifics similar to Fig. 2). I generating He and He, but did not find that the extra complexity sufficiently changed the quality of the explorations we were doing to be worthwhile for our purposes. B. Generating entanglement

Now we consider the case where system-environment III. SOME ILLUSTRATIVE EXAMPLES interactions are turned on. The interactions will cause an initial product state given by A. Decoupled “Schr¨odingercat”

|ψiw = |ψis|ψie (8) Here we consider the “Schrödingercat” state formed as a coherent superposition of two coherent states: to evolve into an entangled state, where the states of the system and environment are described by the density |ψi = a1 |α1i + a2 |α2i (7) matrices where each |αi is given by Eqn. 4. ρs ≡ T re (|ψiww hψ|) (9) Figures 4 and 5 are of the same form as figs. 2 and√ 3 but showing a state given by Eqn. 7 with a1 = 1/ 3, and p α1 = 3, a2 = 2/3 and α2 = −2.1. Again, these are evolved with system-environment interactions turned ρe ≡ T rs (|ψiww hψ|) . (10) off. These figures show that the evolution of coherent superpositions is also robust for the ACL model, even The Von Neumann entropy, though the discrete nature of the truncated SHO shows up in the jagged features of the wavefunction when the S ≡ tr (ρs ln ρs) = tr (ρe ln ρe) , (11) two packets collide. takes larger values when the degree of entanglement is greater. The maximum possible value for the entropy is given by 3 I The approach to He and He used here is similar to that pioneered in [10], although in that work the “system” was a single qubit. Smax = ln (Nmin) (12) 5

0.2 t = 0

0.1 max 0.5

0 S/S -0.1 Weak coupling -5 0 5 Strong coupling q 0 t = 107 100 105 0.2 t 0.1 0 -0.1 FIG. 6. The evolution of the von Neumann entropy for -5 0 5 EI = 0.03 (“weak coupling,” dashed) and EI = 0.25 (“strong 1 q coupling,” solid). Increasing the interaction strength causes residuals the entanglement to increase more rapidly, and also allows the system to come a bit closer to Smax. 10-10

-20 10 t=2.5 -5 0 5 0.4 q Top 0.2 eigenstate 0 FIG. 5. Evolution of a “SchrödingerCat” superposition of -5 0 5 0.4 coherent states (specifics similar to Fig. 3). q 0.2 2nd eigenstate 0 where Nmin is the smaller of Ns and Ne. Figure 6 shows the evolution of the entropy for two values of EI . -5 0 5 Throughout this work we use units where ~ = ωSHO = 1. We also take Ee = 0.75 throughout. For Fig. 6 the initial state has the product form (Eqn. 8) with |ψis given FIG. 7. The two most probable eigenstates of ρs after einse- by the Schrödingercat state discussed above and |ψie given by the 500th eigenstate of H (indexed from lowest lection has completed. The initial states was the Schrödinger e cat state depicted in Fig. 3. to highest eigenvalues). The choice of |ψie will be discussed further Sect. IV. We consider a “weak coupling” (E = 0.03) case and a “strong coupling” (E = 0.25) I I There is a special case of decoherence called “einse- case. lection” where the initial state and interactions can be set up to favor a special set of eigenstates for ρs called “pointer states.” The CL model has been used in many C. Einselection of the pioneering studies of decoherence and einselection. Here we revisit some of these results using the ACL A generic state for w will be an entangled state with model. non-trivial density matrices, ρ, for system and environ- The Schrödinger cat state depicted in the top panel of ment. Thus, it is not surprising that in the interacting Fig. 5 is a superposition of two coherent states which can case that starts in a product state the entanglement en- be thought of as “classical wavepackets.” Fig. 7 shows tropy will increase from zero. This process is generally what this initial state evolves into by time t = 2.5τ for called decoherence, and it would take place with just the weakly interacting case. The state of s for t > 0 is a about any Hamiltonian for w. For a randomly chosen density matrix, and Fig. 7 shows the two eigenstates of Hw, one would expect the entanglement entropy to be- ρs with the largest eigenvalues. One can see that these come large and the eigenstates of ρs and ρe to evolve look like single classical wavepackets. Figure 8 depicts randomly over time without displaying any regular be- similar information about the state but evolved further havior. in time, to t = 4τ. These eigenstates also look like classi- 6

t=4 0.4 Top 5 0.2 eigenstate 0 -5 0 5 0.4 q 0 2nd q 0.2 eigenstate 1st eigenstate 0 2nd eigenstate -5 -5 0 5 10-1 100 101 t 4

FIG. 8. The top post-einselection eigenstates of ρs shown in Fig. 7, but here shown at a diﬀerent phase in their periodic motion.

rms 2 q cal wavepackets, just caught at a different phase of their 1st eigenstate oscillation. 2nd eigenstate There are a variety of technical tools that are useful 0 in studying einselection. One can anticipate the pointer 10-1 100 101 states and study the decrease in the off diagonal element of ρs in that basis (as per [11]). The consistent histories framework can also be useful. The approach we use here, focusing on the eigenstates of ρs, parallels that developed in [10] (where a comparison with the consistent histories FIG. 9. The evolution hqi and qrms as a function of time for the top two eigenstates of ρs shown in Figs. 7 and 8. (The approach is also presented). We also use the consistent most probable eigenstate is shown with the solid curve, the histories method extensively with the ACL model in [6]. next most probable is dashed.) One can see these attributes One can look at this phenomenon a bit more systemati- evolve from those of the Schrödingercat initial state (oscil- cally by studying how various moments of the eigenstates lating qrms and small oscillating values of hqi) to those of evolve over time. Figure 9 shows the time evolution of hqi individual wavepackets (essentially constant qrms with larger and qrms. One can see how these quantities first exhibit oscillations in hqi). the “Schrödingercat” properties, but over time develop the properties of einselected pointer states. We conclude 4 that the ACL model nicely reproduces the well-known dissipation times , which in turn is connected with the phenomenon of “einselection,” as it should if it is to re- competition between the interaction Hamiltonian (which flect key properties of the CL model. tries to localize the SHO in space) and the SHO Hamil- tonian (which causes localized states to spread). Figure 11 shows a case with more widely separated decoherence and dissipation times. The calculation shown 3 D. Evolution of the eigenvalues of ρ in Fig. 11 uses τSHO = 2π × 10 and the initial state is a superposition of eigenstates of qSHO (in the same proportions and locations as the coherent states used in Figure 10 shows the eigenvalues pi of ρs (for pure states in w, the nonzero ones are always identical to the nonzero Fig. 10). These differences mean the interaction term (∝ qSHO) is not trying to “chop up” the initial wavepack- eigenvalues of ρe). The evolution of the pi’s includes the information reflected in the von Neumann entropy ets, in contrast to the coherent state initial conditions, (Fig. 6), and clearly shows a transient phase during ein- which spread across several eigenstates of qSHO. selection and a subsequent equilibrium phase where the Note that for a while p1 and p2 in Fig. 11 correspond to the probabilities assigned to the wavepackets in the pi values are closer together and hold reasonably steady. One can infer from Fig. 9 that the time to full einselection initial superposed state. This feature means that the is O(20). The dissipation processes that lead to equilibration operate on a time scale roughly 20 times longer. One can see that by the time einselection is complete 4 This is in contrast to more macroscopic systems, where the deco- there are somewhat more than two nonzero pi’s. This is herence and dissipation timescales are typically widely separated related to the relative closeness of the decoherence and (see e.g. [4, 8, 12]). 7

1 1

0.8 0.8

s s

0.6 0.6

0.4 0.4

Eigenvalues of of Eigenvalues of Eigenvalues 0.2 0.2

0 0 10-3 100 103 106 10-3 100 103 106 t t

FIG. 10. The eigenvalues of ρs. The purity of the initial state FIG. 11. Similar to Fig. 10 but with the initial state and is reflected in the fact that only one eigenvalue is nonzero Hamiltonian parameters modified as discussed in the text. initially. The “einselection time” (marked by the left vertical This example shows more strongly separated decoherence line) corresponds to the “collapse” of the Schrödingercat pure times and dissipation times. Note in particular that the state into a mixture of wavepackets. The dissipation time two top eigenvalues spend an extended period of time at the (right vertical line) is about 20 times longer. The dashed probability values (dashed lines) assigned to the initial (su- horizontal lines show the probabilities assigned to the two perposed) wavepackets, indicating that the environment has wavepackets in the initial Schrödingercat state. made a “good measurement” of the SHO.

environment can be thought of as “making a good mea- IV. APPROACH TO EQUILIBRIUM suremnt” of the SHO, in the sense that interactions with the environment have put the SHO in a classical mixture of wavepackets with the right probabilities. Later, this Figure 13 shows the evolution of entropy and energies good measurement comes unraveled as dissipation sets over time for a variety of initial states of the environment in. for the weakly coupled case (EI = 0.03). The strong coupling case is shown in Fig. 14. We start the environment in an eigenstate of He, with values of the index ie chosen from {1, 100, 200, 300, 400, 500, 600} (ordered so the E. The copycat process ie runs from lowest to highest eigenvalues). Each case shows characteristics of equilibration. Each curve corre- Our ACL model allows us to scrutinize the very first sponds to a single realization of the random Hamiltonians I steps of the einselection process. In doing so we’ve be- used in He and He. We have found that the notewor- come intrigued by certain aspects of these early stages. thy features of the curves remain unchanged as different Figure 12 shows the early evolution of the 2nd eigenvalue realizations are chosen, except for the cases at the ends and eigenstate of ρs, in the case where the system starts of the spectrum where the density of the eigenstates of in a pure Schrödingercat state which becomes entangled He is low and the noise from the randomness shows up with the environment. The eigenstate takes an intriguing more strongly. Also note that the timescale for the first form that appears to be a “mirror image” of the initial significant evolution of the entropy up from zero is sim- state, and remains in this form in a transiently stable ilar for all values of ie except the extremal ones, which way over several decades of time evolution (and growth of rise more slowly. This also chimes with what one might p2). We call these mirror image states “copycat” states. expect from the low density of states case. In [7] we systematically investigate this curious behavior The finite sizes of the systems makes standard defini- and argue that it is quite generic for early time evolu- tions of temperature difficult to apply. Still, we have tion of Schrödingercat states. We also discuss how this found some generalized notions of temperature which phenomenon generalizes in the case of larger numbers of describe the observed behavior well and indicate that “cats.” conventional notions of equilibration apply here, as sug- 8

Properties of eigenstates 100 probability of 2nd 1

eigenstate max

10-10 0.5 S/S 10-5 t 100 1st eigenstate 2nd eigenstate 0.2 0.2 0 0 5 -6 10 10 t=10-6 t=10 0.1 0.1 20

0 0 > -10 0 10 -10 0 10

SHO 10

0.2 0.2

0.3 >

t=4 t=4 e 0

0 0 -20 -10 0 10 -10 0 10 100 105 q q t

FIG. 12. Copycats in the early stages of entanglement: The FIG. 13. The evolution of entropy and subsystem energies system is initially taken to be in a Schrödingercat state (2nd over time, choosing the environment initial state from among row, left panel) which becomes entangled with the environ- the eigenstates of He. The dotted curves correspond to the ment as it evolves. The 2nd eigenvalue and |ψ (q)|2 for the very lowest and very highest eigenvalues, and the other curves first two eigenstates of ρs are shown from early stages of the run from lowest to highest index (from the set given in the evolution. The 2nd eigenstate generically takes the mirror im- text) corresponding to the low or high positions on plots. age “copycat” form over several decades of evolution before Each entropy curve stabilizes over time around its highest finally einselecting to a coherent state form. value, and the corresponding energy curves stabilize as well (implying no net energy flow after the initial transient). These are characteristics of equilibration. gested strongly by Figs. 13 and 14. We plan on reporting this thermal analysis in future work. √ p and present results using a1 = 1/ 3 and a2 = 2/3 (as with the SHO Schrödingercat state discussed above). V. THE REDUCED CALDEIRA-LEGGETT Figure 15 shows the evolution of the eigenvalues of ρ . MODEL s The simplified form of the RCL model means there is no self Hamiltonian for the system competing with the The ACL can be reduced by replacing the SHO with a interaction term, and the pointer states are simply the single qubit, and turning off the self-Hamiltonians of both spin states {|↑i , |↓i} determined by the form of the in- the system and the environment. The resulting “reduced teraction Hamiltonian. Thus the “good measurement” Caldeira-Legget” (RCL) model has this Hamiltonian5: 2 behavior (with the pi’s stabilizing at the values |a1| and 2 I |a2| given by the dotted lines) is realized more robustly HRCL = λSz ⊗ He (13) than in the case depicted in Fig. 11. where Sz ≡ |↑i h↑| − |↓i h↓|. We consider an initial Figure 16 shows the (real parts of the) off-diagonal Schrödingercat state of the form elements of ρs in the pointer state basis (a.k.a. hSxi). |ψi = a |↑i + a |↓i (14) From this perspective, the approach of hSxi toward zero s 1 2 reflects the process of einselection. The uneven fluctuations in the approach toward zero reflect inefficiency in the decoherence process. The RCL model has no self- 5 ↑ ↓ The RCL is the same model discussed in [10] with H1 = −H1 Hamiltonian for e and thus the decoherence boosting ef- and E1 = 0. fects depicted in Fig. 1 are not available (Figure 17 shows 9

1 i

P 0.5 max

0.5 S/S

0 0 100 105 0 5 10 15 20 20 t

> 15 SHO 10 FIG. 15. Eigenvalues of ρ (P ) as a function of time for

20 measurement.” >

e 0

10 10 >

t x 0.2

FIG. 14. This figure is constructed the same way as Fig. 13 0 5 10 15 20 except here strong coupling (EI = 0.25) is shown. The behav- t ior is broadly similar in terms of equilibration (with the overall entropies tending to be larger, as mentioned with Fig. 6). In the strong coupling case the backreaction tends to significantly impact the effective potential in which the oscillator FIG. 16. The quantity hSxi, giving the off diagonal elements moves, and can even shift around the location of the mini- of ρs in the spin basis for the RCL model. While the the mum. The additional broad oscillations on the approach to spin basis is nominally the pointer basis, the inefficiencies equilibrium vs Fig. 13 appear to be related to this effect. of einselection in the RCL model allow significant deviations from zero at late times. results comparable to Fig. 16 but with a self-Hamiltonian added, and one can see that the oscillations have essen- teraction term dominates. In that limit the pointer states tially disappeared). In Sec. VI C we discuss how such are eigenstates of the interaction Hamiltonian. The case curves relate to phenomena seen in NMR experiments, we illustrate in Fig. 11 is approaching the quantum mea- and connect these features with a phenomenon known surement limit. Another extreme is the “quantum limit,” as “Loschmidt echos.” And in [7] we explore more sys- where the self-Hamiltonian of the system dominates. The tematically the variety of behaviors possible for the full pointer states in this case are the energy eigenstates of complex values of the off-diagonal elements of ρs. the system. We explore this limit for the ACL model in Appendix A. When the effects of the interaction term and self- VI. COMPARISON WITH OTHER WORK Hamiltonian are similar (the “intermediary regime”), the pointer states tend to be the coherent states. Much of 6 A. Limits of einselection our discussion in Sect. III covers this regime .

As reviewed in [8], Zurek and collaborators have (in the context of CL models) considered various interesting 6 I For the way we have parameterized He , the environment size Ne limits which cause different pointer states to be selected impacts the strength of the interaction term. When that effect I by the decoherence processes. We have reproduced each is taken into account, the effective strengths of He and Hs are of these limits in this paper. similar for the “weakly interacting” parameters chosen in the The “quantum measurement limit” occurs when the in- first parts of Sect. III. 10

UV cutoﬀ scale. One consequence of these assumptions, along with the Markov approximation, is that the CL master equation 0.4 typically predicts exponential decay for the oﬀ-diagonal

> elements of the system density matrix—an exponential x 0.2 rate of decoherence. This exponential result is also found

In [21] Paz and Zurek consider the case where Hs dominates over the other terms in Hw. They call this case the VII. CONCLUSIONS “quantum limit.” We consider the quantum limit briefly here. We have not explored this limit as systematically We have presented a modified version of the clas- as the other topics in this paper, due to our sense that sic Caldeira-Leggett (CL) model which can be studied it is not so closely connected to the main new directions using full unitary evolution in the combined system- we intend to explore with the ACL model. Still, we have environment space. This adapted Caldeira-Leggett noticed several interesting behaviors and we report those (ACL) model enables explorations beyond the various here, choosing an Appendix since we’ve not developed approximation schemes which are usually used with the this part as fully. CL model. Examples of such new explorations are pre- The pointer states in the quantum limit have been sented in companion papers devoted to studying whether shown to be the eigenstates of Hs [21]. To explore this the notion of einselection makes sense under conditions limit with the ACL model we use the “predictability which do not exhibit an arrow of time [6], and examin- sieve” ideas [5, 8, 22–27], which are grounded in the no- ing the very earliest stages of the einselection process [7]. tion that the pointer states should be the states which This paper provides background information, including are most stable against entanglement with the environ- 7 −3 details of how the ACL model is constructed and of our ment . Here we consider the case where EI = 3 × 10 highly accurate numerical techniques. and Ee = 0.015, well below the values considered else- We have also reproduced a number of well-known re- where in this paper, while keeping Hs the same. We sults from the literature on decoherence and einselection. considered initial states of product form (Eqn. 8) where These build our confidence that the ACL model is well |ψis is either an eigenstate of Hs, the Schrödingercat suited for our intended studies, and also help us know (SC) state shown in Fig. 5, or a single (α = 3) coherent its limitations. Our full numerical treatment enables de- state (CS), and compare the evolution in these cases. tailed scrutiny of all aspects of the process of einselection, Figures 18 and 19 show the evolution of the von Neu- and our extensive graphical representations of that phe- mann entropy for these choices of initial state. Iden- nomenon may provide a useful resource for those wishing tifying robustness against entanglement with small val- to learn more about einselection. ues of the entropy at late times, once can conclude that In addition, Sec. III E briefly introduces new results the cat state is least robust, the lower n energy eigen- which anticipate the work presented in [7]. Also, experts states are most robust, and the coherent state comes in versed in the notion of the “quantum limit” of the einse- about the same as n = 11. (We found the larger n val- lection process might enjoy our exploration of that limit ues reach larger late-time entropies but, as discussed in in Appendix A. While such experts would not find those Appendix B, we also expect significant finite size effects results altogether surprising, we appreciate the way the to come in for the higher eigenstates of Hs.) Interest- ACL model allows us to explore interesting intermediate ingly, the cat and the CS states exhibit much lower en- behaviors on the way to the full quantum limit. tropies for several decades of earlier time evolution which We conclude that the ACL model provides a reliable suggests a different (and transient) hierarchy of robust- tool with which to explore decoherence and einselection ness. Furthermore, if one uses the timescale for the early under conditions which cannot be treated using the stan- time onset of entanglement as the measure of robustness, dard approximation schemes. the coherent state is significantly more robust than the other cases considered. The original work on the quantum limit [21] only showed the stability of eigenstates of H at late times, and did not actually compare the rate of VIII. ACKNOWLEDGMENTS s

We thank Wojciech Zurek for numerous inspiring con- versations (over many years in the case of one of us, 7 While we’re not doing a thorough sifting of the entire Hilbert AA) which created the foundation for this work. We also space in our analysis here, we find utilizing“predictability sieve” thank Fabio Anza, Nick Curro and Zhipang Wang for dis- arguments to make comparisons between specific states sufficient cussions of NMR phenomena. This work was supported for our purposes. 12

~~10-5 100 1~~

~~0.8 10-2 Cat 0.6~~

~~n=11~~

~~max max~~

~~-4 n=6 Cat S/S S/S 10 n=2 0.4 n=11 CS n=6 0.2 n=2 10-6 CS~~

~~0 100 105 0 0.02 0.04 0.06 0.08 t t~~

FIG. 18. von Neumann Entropy evolution in a case where Hs FIG. 19. Zooming in on Fig. 18 and showing linear axes. The dominates. The initial state is a product state with |ψis given different initial rates of the onset of entanglement are clearly by a cat state (solid, upper), energy eigenstates with index 11 exhibited here. In this initial period the coherent state (CS) (dashed), 6 (dotted), 2 (dot dashed) or a single coherent state is the most robust against entanglement. (CS, solid, lower). In the idealized “quantum limit” where Hs fully dominates, the energy eigenstates are the pointer states which are expected to be the “most robust” against the onset around t = 106 but the stability does not extend to other of entanglement. In this example we see that there is an early moments of the Schmidts. Those Schmidts are not actual and intermediate period where the coherent state is favored, eigenstates of Hs. and it is only later that the full einselection of the energy Figure 21 shows the full wavefunctions of the Schmidt eigenstates sets in. states for the case where the initial state is an energy eigenstate (the top two of these have moments shown in Fig. 20). These “snapshots” are taken for t >= 104, onset of entanglement. It appears that during the early where the corresponding curves in in Fig. 20 are very and intermediate periods the coherent states exhibit the stable. One can see that these Schmidts are highly stable strongest resistance to entanglement (reflecting the sort in this time period and are very close to true eigenstates of behavior demonstrated in Sect. III E), and only later of H 8. does the long time behavior set in ultimately favoring the s Finally, in Fig 22 we show the evolution several of the energy eigenstates. top eigenvalues of ρs. Not surprising for a case with very Figure 20 shows the evolving properties of the top two weak interactions, the top eigenvalue does not deviate too eigenstates of ρs (aka “Schmidt states”). For the coher- far from unity. We also note the interesting “crossover” ent initial state (CS), these Schmidt states exhibit the behavior, where alternate eigenvalues rise faster and ex- properties of coherent states (steadily oscillating hqi and perience an initial noisy period in equilibrium before set- constant qrms) for an extended period before degrading tling down. We speculate that this behavior is related to into more noisy, unstable behavior. This fits with the the eventual emergence of the other eigenstates of H as narrative we surmised from the entropy curves. For the s eigenstates of ρs and suspect that the two types of be- energy eigenstate initial state the top Schmidt is per- havior are related to the parity of the energy eigenstates fectly stable, maintaining the energy eigenstate features, that emerge. as expected for a pointer state. The second Schmidts All the results reported in this Appendix appear to be (panels 2 and 4) emerge due to the process of decoher- consistent with statements in the literature about the ence (they are ill defined at t = 0, where ρs has only one nonzero eigenvalue) and reflect interesting properties of the decoherence process (also discussed in Sect. III E). For the energy eigenstate initial state, the 2nd Schmidt 8 It is interesting that despite their high degree of stability, the 2nd (4th panel) first reflects some oscillating behavior before and 3rd Schmidts do not match perfectly to eigenstates of Hs. becoming highly stable as well. The curves for CS initial We conjecture that this is due to a small “effective potential” for state case exhibit a transient period of stable behavior the SHO due to the interactions with the environment. 13

~~Coherent state initial state, 1st Schmidt~~

~~5 rms 0~~

~~0.4 Schmidt , q , -5 1 Coherent state initial state, 2nd Schmidt~~

~~5 0.2 rms 0~~

~~, q , -5 0 H eigenstate initial states, 1st Schmidt s~~

~~4 rms 2~~

~~, q , 0~~

~~H eigenstate initial states, 2nd Schmidt s 0.4 Schmidt 4 rms 2 2 0.2 , q , 0 100 105 t 0~~

FIG. 20. Evolution of hqi (dashed) and qrms (solid) for the first and 2nd most probable eigenstates of ρs starting with 0.4 Schmidt different initial states. The coherent state initial state in the top panel initially exhibits the usual oscillatory behavior, but 3 then degrades into noise. The energy eigenstate initial state 0.2 in the third panel is highly stable as expected in the quantum limit. The 2nd Schmidt states (second and fourth panels) 0 are ill defined at t = 0, but they emerge due to the interactions with the environment. Each roughly reflects the behav- -5 0 5 iors of their corresponding 1st Schmidt, although the energy eigenstate initial state case takes a while to get there. The q energy eigenstate initial states, in order descending from the top curve are n = 11, n = 6, and n = 2.

FIG. 21. Energy eigenstates as pointer states: Snapshots of the |ψ(q)| for the three most probable eigenstates of ρ (solid quantum limit case, although we’ve not done a suffi- s curves). Each panel shows the state at t = 104, t = 105, ciently thorough investigation to explicitly demonstrate t = 106 and t = 107. These correspond to the period of time that eigenstates of Hs are the most robust against inter- where all the curves in the 4th panel of Fig. 20 are very sta- actions with the environment out of all possible choices. ble. The wavefunctions at these different times are mostly The behavior of the other eigenstates of ρs noted here is indistinguishable to the eye, indicating that the stability goes intriguing. While it appears broadly consistent with es- well beyond the two moments plotted in Fig. 20. Also plot- tablished ideas about the quantum limit, we’ve not found ted on each panel are (top to bottom) the n = 6, n = 7 any report of these particular effects in the literature. and n = 5 eigenstates of Hs (markers). As discussed in the text, the behaviors depicted here strongly reflect the fact, developed in earlier literature, that the energy eigenstates of H are the pointer states in the quantum limit. We are es- Appendix B: Eigenstates of H s s pecially intrigued by the 2nd and 3rd panels which illustrate that Schmidt states similar to these pointer states are distilled Our form of Hs does a nice job of describing the evolu- out of the messy physics of decoherence by the einselection tion we associate with the continuum SHO using a finite process. (The eigenvalues are 0.98, 0.015 and 0.004.) Hilbert space, as discussed in the body of this paper. Here we provide some further information, focusing especially on the eigenstates of Hs. a finite space. (In these figures the normalization is ad- Figure 23 and 24 depict selected eigenstates of Hs justed for easy cross-comparison.) One can see that the shown along with their continuum counterparts, given lower energy eigenstates (Fig. 23) follow the behavior of in the q basis. In these figures the states of the trun- the continuum states quite nicely. As one approaches cated SHO are shown only as markers (with no connect- higher energies (Fig. 24) the eigenstates reach the edge ing lines) to emphasize the fact that these states exists in of the finite q range and start showing nonzero values 14

~~100 n=0~~

~~2 0.2 |~~

~~| 0.1 0 -5 0 5~~

~~i n=4~~

~~2 P~~

~~| 0.1 | 10-10 0 -5 0 5~~

~~0.1 n=13~~

~~2 |~~

~~| 0.05~~

~~0 10 0 10 10 -5 0 5 t~~

~~n=15 2~~

~~| 0.1 |~~

FIG. 22. The evolution of the top 12 eigenvalues of ρ2 0 for the case where the system starts in its n = 6 energy -5 0 5 eigenstate. The interesting crossing behavior and alternat- q ing “noise buldges” are discussed in the text.

FIG. 23. Energy eigenstates of the truncated SHO (markers) along with the corresponding continuum SHO eigenstates at the q edges. This leads to behaviors at high energies (curves). The two track one another nicely, although the that deviate significantly from the details of the contin- tracking comes under a bit of strain for the n = 15 state uum case, although some broad features remain. Because where the continuum state starts pressing up against the fi- of this behavior, we have avoided studying cases that put nite bounds on q which exist in the truncated case. the SHO in higher energy excitations in this paper as well as in other work using the ACL model, since our inten- tion is to represent a realistic SHO as well as possible. Appendix C: Energy spectra We found for example that coherent states with consider- ably higher amplitudes than those shown here executed interesting combinations of reflection and periodic trans- Here we take a look at the eigenvalue spectrum of Hw, mission at the q boundaries, hardly surprising given the and see how it relates to the spectra of Hs and He. Fig- forms of the higher energy eigenstates. ure 26 shows histograms of the eigenvalues of each of e 0 e these H’s using EI = 0.01, Ee = 0.05, EI = EI and 0 e We also note an exotic feature that appears as an ar- Ee = E . (These are different from the values used in tifact of our finite construction. Figure 25 shows the this paper but match those used in [6], where the spec- same ground state wavefunction shown in the top panel trum of Hw will be relevant for a discussion of our “eigen- of Fig. 23, but here we show ψ(q) both with and without state einselection hypothesis.”) The spectrum of a true the norm. The un-normed values show a jaggedly varying SHO is flat, and so is the spectrum for our SHO shown sign. In continuum terms such jaggedness would result in the top panel of Fig. 26, although this spectrum is in an energy much higher than the ground state energy, truncated at E = 29 reflecting the finite Hilbert space but our Hs has correspondingly complicated off diagonal inhabited by our truncated SHO. The spectrum of He elements coupling certain neighboring points which make (middle panel) reflects the well-known “Wigner semicir- the ψ(q) shown truly the lowest energy state. We’ve also cle” property of random matrices. The eigenvalues of Hw checked that these considerations do not disrupt our use are essentially sums of eigenvalues of Hs and He (with a of continuum intuition with other eigenstates of Hs, at small additional contribution from the interaction term). least for n . Ns/2. The robust behavior of the iso- So it is not surprising that the full spectrum of Hw (lower lated oscillator reported in Figs. 3 and 5 also supports panel) appears to be a combination of the spectra shown our confidence that our truncated SHO is overall a good in the upper and middle panels. For these parameters approximation to the continuum case. the energy of the SHO dominates, and the spectrum of 15

~~n = 0~~

~~n=17 2~~

~~| 0.1 0.4 |~~

~~0 -5 0 5 0.2~~

~~n=22 2~~

~~| 0.1 0 |~~

~~0 -5 0 5 -0.2 0.2 n=27~~

~~2 -0.4 |~~

~~| 0.1~~

~~0 -0.6 -5 0 5 -5 0 5 q~~

~~0.1 n=29~~

~~2 |~~

~~| 0.05~~

0 FIG. 25. The ground state of Hs. Blue: |ψ (q)|, Red: -5 0 5 Re (ψ (q)). The state is defined in an Ns dimensional Hilbert q space, and the discrete nature of that space is expressed by the markers on the plot. The markers are connected by lines in order to reference the continuum SHO case. In the case of |ψ (q)| this correspondence appears to be simple, but Re (ψ (q)) has FIG. 24. Energy eigenstates of the truncated SHO (mark- jagged features not found in the continuum SHO ground state. ers) along with the corresponding continuum SHO eigenstates We discuss the nature of these features in the text and note (curves) shown for larger n values. The tracking behavior that while appearing to be exotic, they do not interfere with noted in Fig. 23 is present here as well, although the edge an intuitive understanding of our truncated SHO, which over- effects are more pronounced. For these n values, taken alone all exhibits behaviors very similar to the continuum case. the markers appear to trace very different curves, but this is only because the discrete grid on which they lie beats in an interesting way off of the frequency exhibited by the continuum the state of w expressed in the eigenbasis of Hw was al- states. ways saved so there was never a need to “rotate back” and thus no associated noise introduced in the evolution.) Regarding numerical accuracy, the critical aspect was Hw roughly takes the form of the SHO spectrum (modu- the ability of our code to accurately evaluate exponentials lated by little semicircles). For cases where He dominates with potentially large imaginary arguments (to rotate the the spectrum of Hw looks more like a single semicircle, phases). The residuals shown in Figs. 3 and 5 give some with “wings” giving a broadening induced by the SHO sense of the capabilities of our code. Note that while spectrum. those figures refer to the case where EI = 0 and focus on the behavior of the SHO, the results were generated e with E = 0.03 and Ne = 600 (and thus Nw = 18, 000) Appendix D: Numerical techniques and tolerances so the residuals reflect a stronger test than one might initially expect. Figure 27 shows several quantities dis- The total Hamiltonian (Hw) was constructed as de- cussed in this paper evolved to later times than previ- scribed in the text and then diagonalized numerically. ously shown. One can see evidence of the breakdown of The initial states were constructed in the appropriate numerical accuracy around t = 1014, when the exponen- subsystem bases and then expanded in the basis of eigen- tial expressions for the (extremely large) phases start fail- states of Hw. Time evolution was performed by rotating ing to compute properly. For example energy conserva- the phases of the coefficients of the eigenstates of Hw action (the constancy of the solid curve in the lower panel) cording to the Schrödingerequation. Density matrices is lost, and the requirement that S ≤ Smax = ln (Ns) for subsystems s and e at a given time were generated by (Eqn. 12) is violated. These, and many other tests of rotating into an s×e product basis and tracing over e and the numerics proved robust up to times just below the s respectively. These density matrices were then used to t ≈ 1014 breakdown point. The availability of accurate extract information about the two subsystems. (Note, numerical computations over such a wide time range pro- 16

~~3 1 2 0.5 max~~

~~SHO 1 S/S 0 0 10 20 30 0 0 10 20 10 10 5~~

~~rms q Environment 0 0 10 20 30 0 100 100 1010 40 50 Total 20~~

~~0 0 10 20 30 Energy 0 100 1010~~

FIG. 26. The eigenvalues spectra of Hw (lower panel) and t its two main components, the SHO (upper) and He (middle). We discuss in the text how these spectra relate to one another and reflect the way the different H’s are defined. FIG. 27. Various quantities are shown evolved over a huge time range to illustrate the point where our numerical computations fail. Top panel: Entropy. Middle panel: q vides excellent latitude for exploring the physics of the rms of the most probable eigenstate of ρs (discussed in Fig. 9). ACL model. (For context, recall that the period of the s Bottom Panel: < HSHO > (dashed), < He > (dotted) and 9 I oscillator is τ = 2π.) < Hw >≡< Hs > + < He > + < H > (solid). All the Our calculations were performed using Matlab on a quantities show the expected physical behavior until t ≈ 1014 64 bit Windows computer with a 3.6GHz Intel i7-4790 where the breakdown of the numerical computation of the processor and 32GB RAM. Each time step, which in- phases sets in. This figure illustrates the very large dynamic range of our numerical computations. (Recall that the SHO cluded calculating a wide variety of information from ρs period is 2π.) and ρe (including the sort reported here), took 20-30 sec- onds. (We noticed a roughly 25% speedup after simultaneously upgrading from Windows 8.1 to 10 and from Matlab R17a to R18b.) The initial construction of all relevant matrices (of which the diagonalization of Hw is the most time consuming) takes around 1.5 hours for the case with Ns = 30 and Ne = 600. We rarely wanted more than 2000 time steps to produce long times views such as shown in Figs. 13 and 14, and for many purposes (such as Fig. 9 and various rough explorations) a lot fewer were sufficient. Much of our code development and testing could be done with smaller environment sizes, for which the time steps were more or less instantaneous. With these sorts of turnaround times we found it possible to work with the ACL model in a reasonably interactive manner.

~~9 For the senior member of this collaboration whose last experience with this kind of calculation was in the 1990’s [10] the comparison of capabilities between then and now is truly remarkable. 17~~

[1] A. O. Caldeira and A. J. Leggett, Phys. Rev. Lett. 46, 2843 (1992). 211 (1981). [15] W. G. Unruh, Phys. Rev. A51, 992 (1995), arXiv:hep- [2] A. O. Caldeira and A. J. Leggett, Physica 121A, 587 th/9406058 [hep-th]. (1983). [16] J. R. Anglin, J. P. Paz, and W. H. Zurek, Physical Re- [3] A. O. Caldeira and A. J. Leggett, Annals Phys. 149, 374 view A 55, 4041–4053 (1997). (1983). [17] F. M. Cucchietti, J. P. Paz, and W. H. Zurek, Phys. [4] W. H. Zurek, “Reduction of the wavepacket: How long Rev. A 72, 052113 (2005). does it take?” in Frontiers of Nonequilibrium Statisti- [18] W. H. Zurek, F. M. Cucchietti, and J. P. Paz, “Gaussian cal Physics, edited by G. T. Moore and M. O. Scully decoherence and gaussian echo from spin environments,” (Springer US, Boston, MA, 1986) pp. 145–149. (2006), arXiv:quant-ph/0611200 [quant-ph]. [5] W. H. Zurek, S. Habib, and J. P. Paz, Phys. Rev. Lett. [19] P. R. Levstein, G. Usaj, and H. M. Pastawski, 70, 1187 (1993). The Journal of Chemical Physics 108, 2718 (1998), [6] A. Albrecht, R. Baunach, and A. Arrasmith, “Einselec- https://doi.org/10.1063/1.475664. tion, equilibrium and cosmology,” Submitted simultane- [20] A. Goussev, R. A. Jalabert, H. M. Pastawski, and ously with this paper to the arXiv. D. A. Wisniacki, Scholarpedia 7, 11687 (2012), revision [7] R. Baunach, A. Albrecht, and A. Arrasmith, “Copycat #127578. process in the early stages of einselection,” Submitted [21] J. P. Paz and W. H. Zurek, Phys. Rev. Lett. 82, 5181 simultaneously with this paper to the arXiv. (1999), arXiv:quant-ph/9811026. [8] M. Schlosshauer, Decoherence and the Quantum-To- [22] W. H. Zurek, Progress of Theoretical Physics 89, Classical Transition, The Frontiers Collection (Springer, 281 (1993), https://academic.oup.com/ptp/article- 2007). pdf/89/2/281/5226677/89-2-281.pdf. [9] A. Albrecht, Phys. Rev. D91, 103510 (2015), [23] J. J. Halliwell, J. P´erez-Mercader, and W. H. Zurek, arXiv:1401.7309 [hep-th]. Physical Origins of Time Asymmetry (1996). [10] A. Albrecht, Phys. Rev. D46, 5504 (1992). [24] M. R. Gallis, Phys. Rev. A 53, 655 (1996). [11] W. H. Zurek, Phys. Rev. D 26, 1862 (1982). [25] M. Tegmark and H. S. Shapiro, Phys. Rev. E 50, 2538 [12] E. Joos and H. Zeh, Z. Physik B - Condensed Matter 59, (1994). 223 (1985). [26] D. A. R. Dalvit, J. Dziarmaga, and W. H. Zurek, Phys- [13] H. P. Breuer and F. Petruccione, The Theory of Open ical Review A 72 (2005), 10.1103/physreva.72.062101. Quantum Systems (Oxford University Press, 2002). [27] W. H. Zurek, Rev. Mod. Phys. 75, 715 (2003). [14] B. L. Hu, J. P. Paz, and Y. Zhang, Phys. Rev. D 45,