Quantum Thermodynamics of Single Particle Systems: Md. Manirul Ali, Wei Ming Huang & Wei Min Zhang

www.nature.
com/scientificreports
OPEN Quantum thermodynamics

of single particle systems
Md. Manirul Ali, Wei‑Ming Huang & Wei‑Min Zhang*
Thermodynamics is built with the concept of equilibrium states. However, it is less clear how
equilibrium thermodynamics emerges through the dynamics that follows the principle of quantum
mechanics. In this paper, we develop a theory of quantum thermodynamics that is applicable for
arbitrary small systems, even for single particle systems coupled with a reservoir. We generalize
the concept of temperature beyond equilibrium that depends on the detailed dynamics of quantum
states. We apply the theory to a cavity system and a two-level system interacting with a reservoir,
respectively. The results unravels (1) the emergence of thermodynamics naturally from the exact
quantum dynamics in the weak system-reservoir coupling regime without introducing the hypothesis
of equilibrium between the system and the reservoir from the beginning; (2) the emergence of
thermodynamics in the intermediate system-reservoir coupling regime where the Born-Markovian
approximation is broken down; (3) the breakdown of thermodynamics due to the long-time non-
Markovian memory effect arisen from the occurrence of localized bound states; (4) the existence of
dynamical quantum phase transition characterized by inflationary dynamics associated with negative
dynamical temperature. The corresponding dynamical criticality provides a border separating classical
and quantum worlds. The inflationary dynamics may also relate to the origin of big bang and universe
inflation. And the third law of thermodynamics, allocated in the deep quantum realm, is naturally
proved.
In the past decade, many efforts have been devoted to understand how, starting from an isolated quantum system
evolving under Hamiltonian dynamics, equilibration and effective thermodynamics emerge at long times1–5. On
the other hand, the investigations of open quantum systems initiate interests on the issue of quantum thermody-
namics taking place under the quantum evolution of open s ystems6–20. The questions of how thermodynamics
emerges from quantum dynamics, how do quantum systems dynamically equilibrate and thermalize, and whether
thermalization is always reachable in quantum regime, are central and fundamental to research for quantum
thermodynamics. However, a general theory of quantum thermodynamics that has conceptually no ambiguity in
answering the above questions has not yet been obtained, because investigations in addressing above questions
inevitably take various assumptions and approximations. In this paper, we will attempt to answer these questions
by rigorously solving the quantum dynamics based on the exact master equation we developed recently for a
large class of open quantum s ystems21–25.
Recall that thermodynamics is built with the hypothesis of equilibrium26. Macroscopic systems at equilib-
rium are fully described by the relation between the internal energy E and a set of other extensive parameters,
the entropy S, the volume V, and the particle number Ni of different components i = 1, 2, ···, magnetic moment
M, etc.,
E = E(S, V , N1 , N2 , ···, M, ···), (1)
under the extremum principle of either maximizing the entropy or minimizing the internal energy. The relation
of Eq. (1) is called the fundamental equation of thermodynamics. The thermodynamical intensive parameters,
such as temperature T, the pressure P, the chemical potential µ1 , µ2 , ··· and magnetic field B, etc. are defined by
the first derivative of energy with respect to entropy, the volume and the particle number, etc.
∂E ∂E ∂E ∂E
T= ,P= , µi = , B= , ···, (2)
∂S ∂V ∂Ni ∂M
such that
Department of Physics, National Cheng Kung University, Tainan 70101, Taiwan. *email: [email protected]
Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 1
Vol.:(0123456789)
www.nature.com/scientificreports/

dE = TdS + PdV + µi dNi + B · dM + ···. (3)
i
These first derivatives in Eq. (2) provide the complete set of equations of state in thermodynamics.
Microscopically (statistical mechanically), all thermodynamic parameters at equilibrium (and therefore the
thermodynamic fundamental equation) can be computed from the probability distribution
{pi } over all possible
microstates, which can be determined by maximizing Shannon entropy S = −k i pi ln pi under the condition
of fixed average energy (canonical ensemble) or fixed both average energy and average particle number (grand
canonical ensembles) in the thermodynamic limit, where the system contains a large number of particles (∼ 1023)
and couples very weakly to a reservoir. Here we have multiplied the information theoretic Shannon entropy with
the Boltzmann constant k to make it to be the thermal entropy. The resulting equilibrium probability distribution
is given by the equilibrium density matrix
 1
 Z e−βH , canonical ensemble;
ρE = � (4)
 1 −β(H− i µi Ni )
Ze , grand canonical ensemble,
where Z is the partition function determined by the normalized condition trρE = 1, β = 1/kT is the inverse
temperature of the reservoir, and H is the Hamiltonian of the system. Thus, the thermodynamic fundamental
equation can be manifested from all the thermodynamic extensive parameters calculated from the thermal
equilibrium state ρE:
�A� = Tr [AρE ], A = H, Ni , ···, M, ··· , and S = −kB Tr[ρE ln ρE ]. (5)
In the energy basis, the entropy in Eq. (5) is identical to the Shannon entropy. Equations (1)–(5) provide the
axiomatic description of thermodynamics and statistical mechanics, from which all the thermodynamical laws
can be obtained naturally, as presented in the classic textbook by Callen26.
The left unsolved problem in thermodynamics is the hypothesis of equilibrium itself, i.e. over a sufficiently
long time period, a given macroscopic system can always reach thermal equilibrium with its environment (reser-
voir), and the corresponding equilibrium statistical distribution does not depend on its initial s tate27,28. This pro-
cess is simply called thermalization in the literature. Solving the problem of thermalization within the framework
of quantum mechanics is one of the big challenges in physics that has been an old dream for many physicists29,
and it is also the foundation in the investigation of quantum thermodynamics and non-equilibrium statistical
mechanics30. No doubt, the solution of thermalization relies on a deep understanding of the dynamics of sys-
tems interacting with their environments, i.e. the quantum solution of non-equilibrium time evolution of open
systems. In the 19th century, physicists had already developed the classical kinetic theory, or more specifically
the Boltzmann’s H-theorem to address the p roblem31,32. Unfortunately, it was not until the 20th century to be
realized that microscopic particles in thermodynamical systems, such as atoms, electrons, and photons, followed
the rules of quantum mechanics which provides the underlying principle for thermalization.
In fact, Caldeira and Leggett explored the thermalization problem from the principle of quantum mechanics
by solving the quantum Brownian motion with the Feynman-Vernon influence functional approach in 1980’s.
They recovered both equilibrium statistical mechanics and linear response theory in the asymptotical time limit
without taking Markovian approximation29. Furthermore, in the beginning of 2000’s, Ford and O’Connell solved
exactly the exact master equation of quantum Brownian motion in the Wigner representation derived by Hu,
Paz and Zhang33 and obtained the asymptotic thermal state from initial Gaussian wavepackets34. Zurek explored
the nontrivial problem of quantum-to-classical transition and revealed the thermalization as a consequence of
entanglement between the system and the reservoir that is realized through decoherence dynamics35. In all these
investigations, thermalization is demonstrated mainly in quantum Brownian motion with initial Gaussian wave
packets at high temperature limit29,34,35. In the last decade, we have also derived the exact master equation for
a large class of open quantum systems21–25 and solved the exact master equation with arbitrary initial states at
arbitrary initial temperature of the reservoir such that the thermalization process is generally p rovided36–38. From
such exact solutions for a large class of open quantum systems, we will attempt to understand thermodynamics
with the detailed thermalization process determined from the principle of quantum mechanics in this paper.
Meanwhile, the aim of quantum thermodynamics is also to develop a thermodynamics formulation far from
equilibrium for nanoscale or atomic-scale quantum systems, where the particle number is much less than the
order of 1023. If one can develop quantum thermodynamics for arbitrary small quantum systems, even for single
particle quantum systems, and prove the consistency with equilibrium thermodynamics, such quantum thermo-
dynamics should be generally valid for arbitrary dynamics of quantum systems. Therefore, in this paper, we shall
focus on the quantum thermodynamics far from equilibrium for single particle systems coupled strongly or weakly
to the reservoir at arbitrary temperature, including the zero-temperature at the deep quantum realm. Quantum
mechanically, a closed system in a pure state remains always in pure states so that no concept of thermalization
can be addressed unless the system is quenched5. However, when the system is (even for a single particle system)
coupled with a reservoir, the system must be entangled with the reservoir so that the system can evolve into mixed
states. Thus, entropy is generated and thermalization naturally emerges even for single particle systems coupled
strongly or weakly to reservoirs. Hence, we will start from the full quantum dynamics evolution of the system
in contact to reservoirs. We investigate the emerging thermodynamics through the non-equilibrium quantum
evolution of thermodynamic parameters far from equilibrium, i.e., energy and entropy etc., much before the
system approaches to the steady state. More specifically, we shall build quantum thermodynamics completely
with the first principle of quantum mechanics.
Vol:.(1234567890)
Figure 1. Schematic illustration of an open quantum system and thermalization. An open quantum system
described by Hamiltonian HS is coupled to a reservoir with Hamiltonian HR, and HI describes the interaction
between the system and the reservoir. The reservoir is initially in a thermal equilibrium state with an initial
temperature T0 and the system is initially in an arbitrary state so that the initial total density matrix is
ρtot (t0 ) = ρS (t0 ) ⊗ ρR (t0 ). Then both the system and the reservoir evolve into non-equilibrium quantum states
through the interaction Hamiltonian HI , from which one can compute the reduced density matrix of the system
and describe quantum thermodynamics under the principle of quantum mechanics.
Results
Formulation of quantum thermodynamics far from equilibrium. As we have pointed out in the
introduction, the big challenge in thermodynamics is the mechanism of thermalization. The solution to this
problem relies on the deep understanding of the non-equilibrium time evolution of open quantum systems. A
standard open quantum system coupled to reservoirs is described by the Hamiltonian,
Htot = HS + HR + HI , (6)
which consists of the Hamiltonians of the system and the reservoirs, denoted respectively as HS and HR , and
the interaction between them, given by HI , as shown schematically in Fig. 1. The dynamics of open quantum
systems can be fully determined by the reduced quantum density matrix ρS (t), which is defined by tracing
over all the reservoir degrees of freedom from the total density matrix ρtot (t) of the system plus the reservoirs:
ρS (t) = Tr R [ρtot (t)]. The total density matrix ρtot (t) obeys the von Neumann equation in the von Neumann
formulation of quantum m echanics39 (sometime it is also called the quantum Liouville equation in the literature):
iρ̇tot (t) = [Htot , ρtot (t)], (7)
from which we have the formal solution of the reduced density matrix:
ρS (t) = Tr R [U(t, t0 )ρtot (t0 )U † (t, t0 )]. (8)

t
Here U(t, t0 ) =T exp{− i t0dt ′ Htot (t ′ )} is the standard time-evolution operator in quantum mechanics which
describes the full quantum dynamical evolution of both the system and reservoirs, where T is the time-ordering
operator, and ρtot (t0 ) is the initial state of the total system (the system plus reservoirs). Solving Eq. (8) exactly
and taking the asymptotic limit in time, one can find the answer to thermalization, as has been done using the
Feynman-Vernon influence f unctional40 with initially decoupled system-reservoir state ρtot (t0 ) = ρS (t0 ) × ρR (t0 )
in the previous works for quantum Brownian motion29,34,35 and also in the recent works for a large class of open
quantum systems36–38, which we have summarized in Introduction.
Notice that von Neumann reformulated and extended quantum mechanics in terms of the density matrix
for quantum systems whose states can be either pure states or mixed states, motivated to develop both non-
equilibrium quantum statistical mechanics and the theory of quantum m easurements39. The two fundamental
non-equilibrium quantum theories in the literature, the Schwinger-Keldysh’s nonequilibrium Green function
approach41–43 and the Feynman-Vernon influence functional approach40 are established from the von Neumann
formulation of quantum mechanics. The von Neumann formulation of quantum mechanics has now become the
basis for open systems (systems coupled to reservoirs) in which the reservoir states are not pure states so that
the Schrödinger formulation of quantum mechanics is inadequate to begin with.
Vol.:(0123456789)
With the previous achievements on t hermalization29,34–38, we can develop quantum thermodynamics as fol-
lows. Within the von Neumann formalism of quantum mechanics, all physical observables of the system, such
as energy, particle number, magnetic moment, etc. can be computed straightforwardly from the reduced density
matrix ρS (t) at any time t far from equilibrium:
E(t) = Tr S [ρS (t)HS ], N(t) = Tr S [ρS (t)NS ], M(t) = Tr S [ρS (t)MS ] , ···. (9a)
Also, the entropy which characterizes the information of the state ρS (t) in the von Neumann formulation of
quantum mechanics is given by
S(t) = −k Tr S [ρS (t) ln ρS (t)]. (9b)
The von Neumann entropy is indeed the extension of Gibbs entropy to quantum mechanics, as von Neumann
originally introduced to develop quantum statistical m echanics39. In the thermodynamical limit, the von Neu-
mann entropy is identical to the Shannon entropy in the energy basis, and can correctly reproduce the thermo-
dynamic entropy when the system reaches the equilibrium with the reservoirs.
Equation (9) establishes the fundamental relation among all the physical observables of the system through
the quantum density matrix ρS (t). This provides a general framework for the study of non-equilibrium quantum
thermodynamics from non-equilibrium quantum statistical mechanics, in analogy to the equilibrium statistical
mechanics of Eq. (5) which provides the microscopic basis for the equilibrium thermodynamics of Eqs. (1)–(3)
summarized in Introduction. Now we can generalize the concepts of temperature, chemical potential, and mag-
netic moment, etc. into the regime far from equilibrium. We introduce the dynamical temperature T (t), dynami-
cal chemical potential µ(t), and external magnetic field B(t), etc. as the first derivatives of the dynamical energy
of the system with respect to the dynamical entropy, particle number, and magnetic moment, etc. calculated
from Eq. (9):
∂E(t) ∂E(t) ∂E(t)
T (t) ≡ , µ(t) ≡ , B(t) ≡ , ··· (10)
∂S(t) ∂N(t) ∂M(t)
These dynamical thermal parameters evolve through various quantum states much before they approach to the
steady state or thermal equilibrium with reservoirs. Conditionally, the dynamical thermal parameters defined by
Eq. (10) can describe quantum thermodynamics if they are coincident to these equilibrium thermal parameters
in the steady-state limit when the system dynamically approaches the thermal equilibrium with the reservoirs, as
we will show in the next two sections. In fact, in the practical applications given in the next two sections, we will
demonstrate explicitly that the dynamical temperature defined in Eq. (10) will always approach a non-negative
thermodynamic temperature in the equilibrium limit for the weak system-reservoir couplings, as one expected.
Furthermore, in analogy to the axiomatic description of t hermodynamics26, using Legendre transformation
of E(t) with respect to S(t) that replaces the dynamical quantum entropy by the dynamical temperature as an
independent variable, it natural gives the dynamical Helmholtz free energy far from equilibrium,
F(t) = E(t) − T (t)S(t), (11)
which is a dynamical thermal potential resulting from the time evolution of the system. Equations (8–11) com-
pletes our quantum thermodynamics formalism far from equilibrium, in which all the physical observables are
directly computable from the reduced density matrix which is determined fully from the principle of quantum
mechanics. This novel theory of quantum thermodynamics is built on the non-equilibrium dynamics of open
quantum systems.
In the literature, many efforts have been devoted to the problem of thermalization starting from an isolated
quantum system under unitary e volution1–5,20. In conventional thermo-statistics, there are three equivalent statis-
tical ensemble descriptions: the micro-canonical, canonical and grand-canonical ensembles. Both the canonical
ensemble (which has the energy exchange with the reservoirs) and grand-canonical ensemble (has the energy
and particle exchanges with the reservoirs) deal with open systems, as we have discussed in Introduction. There-
fore, the steady-state limit of Eqs. (8–11) must reproduce the equilibrium statistical mechanics of canonical and
grant-canonical ensemble statistics when the system dynamically approaches equilibrium with reservoirs, which
guarantees the consistency of our formulation with the standard equilibrium thermodynamics.
On the other hand, the micro-canonical ensemble is defined for isolated systems with the hypothesis of equal
probability for all the available microstates in which the energy, the volume and the particle number are fixed.
The micro-canonical ensemble is usually considered as a pure mathematical model but it is fully equivalent to
other two statistical ensemble description in practical statistical analysis. Physically, this is because for a mac-
roscopic system, the energy level spacings of the system almost approach to zero due to the huge number of
particles (∼ 1023). On the other hand, the very weak but inevitable random gravitational and electromagnetic
fields existing everywhere in nature. These very weak gravitational and electromagnetic interactions between
all the particles, which may be negligible in dealing with individual particle dynamics, make enormously rapid
random transition among various nearly degenerated quantum states in macroscopic systems. This results in the
realization of equal probability hypothesis in reality, as analyzed in details in Ref.44. In other words, no physical
system is, or even can be, truly i solated44.
Meantime, for an ideal isolated quantum system, if its quantum state is initially in a pure state, the unitary
evolution makes the system evolving always in pure states so that no thermalization can happen. In order to
make thermalization possible in reality, it is necessary to prepare the isolated system initially in a mixed state or
drive the system out of equilibrium by quench. This is just a special case of Eq. (7) in which no trace should be
taken (because of the isolated system) so that the time-evolution of the quantum density matrix remains unitary
Vol:.(1234567890)
but the quantum state must be expressed in terms of the density matrix and evolution in a controllable way.
Thus, our formulation of quantum thermodynamics covers the studies of thermalization for quenched isolated
systems as well.
Furthermore, in equilibrium thermodynamics, only for isothermal process one can define the thermodynamic
work as the change of the Helmholtz free e nergy26. The thermodynamic heat can be obtained thereby as a con-
sequence of the energy conservation including the energy transfer between the system and reservoirs. However,
for quantum systems under dynamical evolution, no process is isothermal and therefore the extension of the
concept of work far from equilibrium is not straightforward. In fact, in the framework of quantum mechanics, no
concepts of work and heat have been introduced because work and heat are not quantum mechanical observables.
We expect that by coupling the system with the environment, it is possible to explore the microscopical quantum
origins of work and heat from the dynamics of open quantum systems, which are manifested by dissipation
dynamics and fluctuation dynamics, respectively. We will explore these fundamental concepts in the further
research within our exact master equation t heory21–25.
It is also worth mentioning that currently there is also no consensus on the definition of the internal energy
for systems strongly coupled with reservoirs45–48. Some inconsistency has been found in calculating the internal
energy in terms of the local system Hamiltonian but usually such inconsistency is mainly due to the approxima-
tions used in dealing with the dynamics of open quantum s ystems10, namely, lack of the exact solution of Eq. (8).
It has also been shown that by identifying the thermodynamic internal energy with the equilibrium expectation
value of the local system Hamiltonian HS for a damped particle, the resulting specific heat approaches the clas-
sical result for high temperatures and goes to zero for vanishing temperature in accordance with the third law
at equilibrium45.
Nevertheless, the key to solve quantum thermodynamics in this new formulation is the determination of
the time-dependent reduced quantum density matrix of Eq. (8) from the non-equilibrium dynamical evolution
of both the system and reservoirs under the total Hamiltonian of Eq. (6), and from which the non-equilibrium
quantum thermodynamics of the system can be unambiguously investigated. To obtain unambiguous conse-
quences, in the following, we shall apply the theory to a single-mode cavity system and a two-level atomic system
interacting with a reservoir, which are described by the Fano-Anderson m odel49,50 and the multimode Jaynes-
Cummings (JC) model51,52, respectively. The Fano-Anderson model gives the Fano resonance and Anderson
localization observed as universal phenomena throughout nuclear, atomic, molecular, and optical physics, as
well as in various condensed matter systems53,54, while the JC model is the basis of cavity QED, describing the
light-matter interaction in quantum optics55,56. These are the two most general examples that can be solved
unambiguously in open quantum systems.
In these applications, we shall show rigorously and exactly (i) how thermodynamics emerges from quantum
dynamics of a single particle system coupled to a reservoir in the weak system-reservoir coupling regime, without
introducing the hypothesis of equilibrium between the system and the reservoir from the beginning; (ii) how
thermodynamics also emerges in the intermediate system-reservoir coupling regime where the Born-Markovian
approximation is broken down; (iii) how the thermodynamics breaks down in the strong coupling regime, due
to the long-time non-Markovian memory effect arisen from the occurrence of localized bound states; and (iv)
how a dynamical quantum phase transition occurs associated with negative temperature when the reservoir is
initially in vacuum state or its initial thermal energy kT0 is less than the system initial energy E(t0 ). The dynami-
cal criticality in this dynamical quantum phase transition provides an inflationary dynamics that may reveal a
simple physical picture to the origin of universe inflation57. We also show how the third law of thermodynamics,
as a true quantum effect, is obtained in this new formulation of quantum thermodynamics.
Application to cavity systems with Fano‑Anderson model. We first consider the system as a single-
mode bosonic system, such as a single-mode cavity, with HS = ωs a† a, coupled with a heat bath (it is also
applicable to fermionic system in a similar way22,23). Here ωs is the frequency of a single-mode boson. The bath
is modeled as a collection of infinite bosonic modes with a continuous frequency spectrum, and the system-bath
interaction is described by the Fano-Anderson model that has wide applications in nuclear, atomic, molecular,
and optical physics, as well as in various condensed matter s ystems53,54. Thus, the total Hamiltonian of the system
plus the bath is given by the Fano-Anderson Hamiltonian,

Htot = ωs a† a + ωk bk† bk + (Vk a† bk + Vk∗ bk† a).
(12)
k k
The reduced quantum density matrix of the system obey the following exact master e quation21,58,59:
d
ρS (t) = − iωs′ (t, t0 ) a† a, ρ(t) + γ (t, t0 ) 2aρS (t)a† −a† aρS (t)−ρS (t)a† a
dt
(13)
+ γ(t, t0 ) a† ρS (t)a + aρS (t)a† − a† aρS (t)−ρS (t)aa† ,
in which the renormalized system frequency, the dissipation (relaxation) and fluctuation (noise) coef-
ficients are determined simply by ωs′ (t, t0 ) = −Im[u̇(t, t0 )/u(t, t0 )] , γ (t, t0 ) = −Re[u̇(t, t0 )/u(t, t0 )] and
γ(t, t0 ) = v̇(t, t) − 2v(t, t)Re[u̇(t, t0 )/u(t, t0 )], where the non-equilibrium Green’s’ functions satisfy the follow-
ing equations
t
d
u(t, t0 ) + iωs u(t, t0 ) + dτ g(t, τ )u(τ , t0 ) = 0. (14a)
dt t0
Vol.:(0123456789)
t t
v(t, t) = dτ1 g (τ1 , τ2 )u∗ (t, τ2 ),
dτ2 u(t, τ1 ) (14b)
t0 t0
∞ ∞
The integral kernels g(t, τ ) = 0 dωJ(ω)e
−iω(t−τ ) and g (τ1 , τ2 ) = 0 dωJ(ω)n(ω, T0 )e−iω(τ1 −τ2 ) are determined
by the spectral density J(ω) ≡ k |Vk |2 δ(ω − ωk ) of the reservoir, and n(ω, T0 ) = eω/kT 1
0 −1
is the initial particle
number distribution of the reservoir. The exact analytic solution of the non-equilibrium Green’s function u(t, t0 )
has been g iven21,
∞
1 −iωb (t−t0 ) J(ω)e−iω(t−t0 )
u(t, t0 ) = e + dω , (15)
′
1 − � (ωb ) 0 [ω − ωs − �(ω)]2 + π 2 J 2 (ω)
where ωb is the localized photon mode ′ frequency determined
by thepole condition: ωb − ωs − �(ωb ) = 0 with
the self-energy �(z) = dω ′ J (ω ) , from which we have � ω ± i0+ = �(ω) ∓ iπJ(ω), and the principle value
J (ω′ ) z−ω ′
�(ω) = P dω′ z−ω ′ is the frequency shift. The first term in the above solution is the contribution of dissipa-
tionless localized bound state. The second term is the continuous part of the spectra (corresponding to the level
broadening induced by the reservoir) that characterizes the relaxation of the system. The Green’s function v(t, t)
which characterizes the non-equilibrium thermal fluctuations gives the non-equilibrium fluctuation-dissipation
theorem21. Without loss of generality, we consider an Ohmic spectral density J(ω) = ηω exp (−ω/ωc ), where η is
the coupling strength between the system and the thermal reservoir, and ωc is the frequency cutoff of the reservoir
spectra60. In this case, the system has a localized bound state when the coupling strength η exceeds the critical
value ηc = ωs /ωc21. Then, all quantum thermodynamical parameters can be calculated systematically and exactly.
Explicitly, let the system be prepared in a Fock state |n0 �, the corresponding initial density matrix
ρS (t0 ) = |n0 ��n0 |, with initial system energy E0 = n0 ωs and n0 is an integer. The reservoir is initially in thermal
equilibrium state with an initial temperature T0, as a condition for searching the emergence of thermal equilib-
rium of the system. After the initial time t0, both the system and the reservoir evolve into non-equilibrium states
due to the interaction between the system and the reservoir. The system state at arbitrary later time t can then
be obtained rigorously by solving the exact master equation of Eq. (13). The result i s37:

∞
ρS (t) = pnn0 (t)|n��n|, (16a)
n=0
[v(t, t)]n
pnn0 (t) = [1 − A(t, t0 )]n0
[1 + v(t, t)]n+1
0 ,n}n n 1
min{n
A(t, t0 )
k (16b)
0
× ,
k k v(t, t) 1 − A(t, t0 )
k=0
|u(t,t0 )|2
where A(t, t0 ) = 1+v(t,t) , and the explicit solutions of u(t, t0 ) and v(t, t) are given by Eqs. (15) and (14b). It shows
that the system in an initial photon Fock state will evolve into a mixed state of different Fock states due to the
coupling to the reservoir, and the function pnn0 (t) is the probability of finding the system in the Fock state |n� at
time t. The probability pnn0 (t) is fully determined by the initial state of the system and the non-equilibrium Green’s
functions u(t, t0 ) and v(t, t). The different initial state will lead to the different probability distribution in this fully
quantum mechanical non-equilibrium evolution. On the other hand, the dependence of u(t, t0 ) and v(t, t) fully
characterizes the dissipation and fluctuation dynamics between the system and the reservoir during the non-
equilibrium evolution. From these results, we can calculate straightforwardly all the quantum thermodynamic
parameters, namely the dynamical energy, entropy, temperature and free energy, etc. Explicitly, the internal
energy of the system at time t, i.e. Eq. (9a), is given by E(t) = n pnn0 (t)En with En = nω s . Thenvon Neumann
entropy of the reduced density operator ρ(t), i.e., Eq. (9b), is simply reduced to S(t) = −k ∞ n=0 pn (t) log pn (t).
0 n0
These results provide explicitly the fundamental relation between the quantum energy and the quantum entropy
of the system, from which one can determine the dynamical temperature T (t) defined by Eq. (10), and all other
related quantum thermodynamical parameters. The results are given as follows.
Emergence of thermodynamics and thermo‑statistics from quantum dynamics in the weak system‑reservoir cou-
pling limit. As it is well-known, classical statistical mechanism is valid for relatively high temperatures. This
is because at relatively high temperatures, the Fermi-Dirac distribution and the Bose-Einstein distribution are
reduced to the classical Boltzmann distribution. At low temperatures, the Fermi-Dirac distribution and the
Bose-Einstein distribution can be significantly different from the classical Boltzmann distribution, and then
classical statistical mechanics fails. The specific heat of metal at low temperatures provides a prototype example
for such an invalidity of classical statistical mechanics26. However, in general, it is not clear what a temperature
value can be regarded as high temperature. Here we define quantitatively this classical (i.e. the high temperature)
regime: kT0 > E0, namely the initial thermo-energy of the reservoir should be larger than the initial energy of
the system. In this classical regime, it is indeed straightforward to see how the system is dynamically thermalized
in the weak system-reservoir coupling limit (η ≪ ηc ). Initially the system is at zero temperature (because it is in
the pure state |n0 �) with zero entropy, and the reservoir is in thermal equilibrium at a relatively high temperature
T0 with kT0 > E0 = n0 ωs . Then the system and the reservoir undergo a non-equilibrium process due to their
contact with each other. The non-equilibrium quantum dynamics is fully determined by the non-equilibrium
Green functions u(t, t0 ) and v(t, t) of Eq. (14). One can indeed show explicitly that in the long time steady-state
Vol:.(1234567890)
Figure 2. Non-equilibrium quantum thermodynamics in different system-reservoir coupling regimes. The

dynamics of non-equilibrium thermal parameters: the energy, the entropy, the temperature and the free energy
(in the unit of ωs, except for the entropy which is dimensionless when the temperature takes the same unit as
the energy) as a function of the time that calculated from Eqs. (9)–(11) with different initial Fock states: |n0 �
of n0 = 1 (blue), n0 = 5 (green) and n0 = 10 (red). (a) The weak coupling regime η = 0.01ηc ≪ ηc (the first
column): (b) The intermediate coupling range 0.01ηc < η = 0.5ηc < ηc (the second column): (c) The strong
coupling regime η = 1.5ηc > ηc (the third column). The initial temperature of the thermal reservoir is taken at
kT0 = 20ωs, the cut-off frequency ωc = 5ωs.
limit t → ts , we have u(ts , t0 ) → 0, and the average photon number in the cavity obeys the Bose-Einstein statisti-
cal distribution,
1
�a† (ts )a(ts )� = v(ts , ts ) → n(ωs , T0 ) = . (17)
eωs /kT0 − 1
The probability pnn0 (t → ts ) of occupying the n-th energy state is then solely determined by the Bose-Einstein
statistical distribution, n(ωs , T0 ), and the state of the system ultimately approaches unambiguously to the thermal
equilibrium with the r eservoir36,37

∞
[n(ωs , T0 )]n 1 − HS
ρS (t → ts ) = |n��n| = e kT0 , (18)
[1 + n(ωs , T0 )]n+1 Z
n=0
where Z = Tr S [exp(−HS /kT0 )], and T0 is the initial temperature of the reservoir. Without taking any approxi-
mation in the above calculations, we provide for the first time a rigorous proof of the dynamical thermalization
within the quantum mechanical framework.
One can then calculate the dynamical quantum internal energy and the quantum entropy. In Fig. 2a, we show
the dynamics of internal energy E(t), entropy S(t), temperature T (t) and free energy F(t) as a function of time in
the weak system-reservoir coupling regime for different initial states of the system. As we see, both the energy and
entropy of the system increases with time due to its contact with the thermal reservoir through a non-equilibrium
process in the classical regime (kT0 > E0), see Fig. 2a(i) and 2a(ii). The dynamical temperature T (t) of the sys-
tem increases gradually (see Fig. 2a(iii)) and finally approaches to the temperature that is identical to the initial
temperature of the reservoir, T (ts ) = T0, namely the system reaches the thermal equilibrium with the reservoir.
Vol.:(0123456789)
The results in Fig. 2a show precisely how thermalization is dynamically realized in the weak system-reservoir
coupling limit for a single cavity system, independent of its initial states. As it is expected, the entropy always
increases in non-equilibrium processes, see Fig. 2a(ii), demonstrating the second law of thermodynamics. The
second law places constraints on state t ransformations61, according to which the system evolving non-equilib-
riumly from the initial state ρS (t0 ) makes the free energy F(t) goes down, as shown in Fig. 2a(iv). Meantime,
the increasing internal energy with the decreasing free energy always makes the heat transfer into the system
(non-equilibrium extension of the first law of thermodynamics), resulting in the increase of the dynamical
temperature T (t) of the system, as shown in Fig. 2a(iii). In the steady-state limit, the free energy approaches to
a minimum value with maximum entropy in the equilibrium state, which justifies the fundamental equation
of the equilibrium thermodynamics26. Therefore, this solution in the weak system-reservoir coupling regime
provides the foundation of the equilibrium thermodynamics, namely how thermalization is dynamically realized
within the framework of a fully dynamical evolution of quantum mechanics. Indeed, producing the equilibrium
thermodynamics in the classical regime from the exact quantum dynamics with arbitrary initial states of the system
should also be considered as a criterion for an unambiguous formulation of quantum thermodynamics.
Emergence of thermodynamics in the intermediate system‑reservoir coupling regime that goes beyond the
Born‑Markov limit. It may be worth pointing out that the above thermalization in the weak-coupling limit
can also be obtained approximately from the Born-Markov master equation. However, the Born approximation
requires the reservoir to remain in the unchanged equilibrium state all the time which is indeed an inconsistent
assumption. Specifically, if one consider an interesting situation in which the system interacts very weakly with
two same thermal reservoirs with different initial temperatures, say T1 and T2, respectively. Note that here the two
reservoirs plus the systems form a closed system, namely no external thermal sources in contact with to maintain
the two reservoirs with their initial temperatures T1 and T2, respectively. Then after a sufficient long time, the
system and the two reservoirs must eventually reach the same equilibrium state with a same new temperature
T. In other words, both reservoirs cannot be kept unchanged in time so that the basic assumption of the Born-
Markov approximation is no longer valid, even in the very weak-coupling regime. However, our theory describes
exactly the non-equilibrium dynamical evolution of the whole system (the system and also the two reservoirs).
The exact solution of our formulation naturally shows how the system and the two reservoirs evolve into a new
equilibrium state at the same time, with the new equilibrium temperature T = (T1 + T2 )/2, which also demon-
strates the zero-th law of thermodynamics.
Furthermore, we show next that in the intermediate system-reservoir coupling regime, thermodynamics
can still be reached where the Born-Markov approximation is completely broken down. In fact, we find that
the thermalization does not crucially rely on the assumption of the very weak interaction between the system
and the reservoir. In the literature, an overall thermalization condition on the system-reservoir coupling was
proposed62: the strength of the coupling η induces the effect that has to be negligible in comparison to initial
thermal energy kT0. Recently, many efforts have been made to formulate quantum thermodynamics for a strongly
coupled system36,37,63–69. However, quantum thermodynamics has not been fully solved from a quantum dynamics
description in the strong system-reservoir coupling regime.
From our solution of the exact master equation, we can provide an exact quantum dynamics description in
the strong system-reservoir coupling regime, from which we can give a quantitative condition on thermaliza-
tion. The result is that thermalization occurs when η < ηc , where ηc is the critical value of the system-reservoir
coupling for the occurrence of localized bound s tates21,36,37,58,59. In particular, our numerical calculation shows
that the Born-Markov thermalization condition is valid only for the very weak coupling η 0.01ηc . In Fig. 3a,
we plot the non-equilibrium thermodynamical parameters in the very weak-coupling regime for several different
system-reservoir coupling strengths with the same initial system state. It shows that the system reaches the same
equilibrium state in the steady-state limit. In other words, in the very weak-coupling regime, the change of the
coupling strength η induces the effect that is negligible in comparison to initial thermal energy kT0, as proposed
in Ref.62. However, we find that this is only a sufficient but not a necessary condition for thermalization.
We show below that quantum thermodynamics in the strong system-reservoir coupling regime is indeed quite
distinct. As it is shown in Fig. 2b, when the system-reservoir coupling strengths are not so weak (the intermediate
range: 0.01ηc < η < ηc ), the system quickly approaches to the same equilibrium state in the steady-state limit
for different initial states but the same coupling strength. In other words, the system reaches the equilibrium
state which is independent of the initial state of the system as well in the intermediate coupling strength regime.
However, for different intermediate coupling strengths, the system approaches to different equilibrium states
in the steady-state limit, as shown in Fig. 3b. We called these different equilibrium states as thermal-like states
in Ref.36) because these thermal states are characterized by the different steady-state temperatures that are all
different from the initial reservoir temperature T0 (see the blue, green and red curves in Fig. 3b(iii), which cor-
respond to different system-reservoir coupling strengths). In this intermediate coupling strength region, different
equilibrium states are reached by different steady-state minima of the free energy F(t), see Fig. 3b(iv).
These results already show the deviations of quantum thermodynamics away from the equilibrium thermo-
dynamics prediction, namely after a long enough time, the system must equilibrate with the reservoir that is
characterized by the initial reservoir temperature T0. In fact, due to the relatively stronger couplings between
the system and the reservoir, both the system and the reservoir undergo a non-equilibrium process, and then
approach to a new common equilibrium state characterized by the new equilibrium temperature T (t → ts )
that is different from the initial reservoir temperature T0, as our theory predicted. In Figs. 2b and 3b, it shows
that ts ≃ 25/ωs , and T (ts ) changes as η varies. But the principle of thermodynamics, namely the existence of
equilibrium state and the associated thermodynamic laws in the steady state, is still valid, and the results are inde-
pendent of the initial states of the system. In other words, thermalization can still be reached in the intermediate
Vol:.(1234567890)
30 40 40
20
22 22
(i) (i) (i)
5 5 5
0 750 1500 0 50 100 0 250 500
6 6 6
3 3 3
(ii) (ii) (ii)

0 0 0
0 750 1500 0 50 100 0 250 500
30 40 40
20
20 20
10
(iii) (iii) (iii)
0 0 0
0 750 1500 0 50 100 0 250 500
5 5 5
-75 -75
-40
(iv) -150 (iv) (iv)

-150
-80
0 750 1500 0 50 100 0 250 500
Figure 3. Non-equilibrium quantum thermodynamics in different system-reservoir coupling regimes. The

dynamics of non-equilibrium thermal parameters: the energy, the entropy, the temperature and the free energy
(in the unit of ωs, except for the entropy which is dimensionless when the temperature takes the same unit as
the energy) as a function of the time that calculated from Eqs. (9)–(11) for different system-reservoir coupling
strengths with the same initial Fock state |n0 � = |5�. (a) The weak coupling range η ≤ 0.01ηc (the first column):
three different curves represent the three different coupling strengths are η = 0.003ηc (blue), η = 0.005ηc
(green), η = 0.01ηc (red). (b) The intermediate coupling range 0.01ηc < η < ηc (the second column): three
different curves represent the different coupling strengths η = 0.1ηc (blue), η = 0.3ηc (green), η = 0.5ηc (red).
(c) The strong coupling regime η > ηc (the third column): three different curves represent the different coupling
strengths η = 1.2ηc (blue), η = 1.3ηc (green), and η = 1.5ηc (red). The initial temperature of the thermal
reservoir is taken at kT0 = 20ωs, the cut-off frequency ωc = 5ωs.
coupling regime, although the final equilibrium temperature of both the system and the reservoir is different
from the initial temperature of the reservoir, namely the Born-Markovian approximation has been completely
broken down here. This solution also shows explicitly that the thermalization can be realized far beyond the
condition proposed i n62.
Breakdown of thermodynamics and thermo‑statistics due to the occurrence of localized bound states. We consider
next the strong system-reservoir coupling regime (going beyond the critical coupling value η > ηc , where local-
ized bound states occur in the system). In such strong coupling regime, the system indeed does not approach to
thermal equilibrium36, because all different initial states result in different steady states, as shown explicitly in
Fig. 2c. Furthermore, the average energy E(t) of the system initially increases, and then it starts decreasing with
time, indicating a backflow of energy from the system into the reservoir, and finally it goes a stationary oscilla-
tion in time, see Figs. 2c(i) and 3c(i). Therefore, these steady states cannot be equilibrium states because of their
permanent time dependence. A similar dynamical behavior of the system entropy S(t) is seen in Figs. 2c(ii)
and 3c(ii) in this strong coupling regime, namely, although the entropy increases in the beginning, it is decreased
and oscillates in the later time. Thus, the occurrence of localized bound states in the strong-coupling regime
makes the steady state of the system deviated significantly away from the equilibrium state, and thermaliza-
tion cannot be reached. Also, one does not see from Figs. 2c(iii) and 3c(iii) a monotonous increase of dynami-
cal temperature T (t) that can eventually approach to a constant value. In other words, in the strong coupling
Vol.:(0123456789)
5
2
2.5
1
a b
0 0
0 400 800 0 400 800
100 100
0
10
50 0
50
-20
80 800 80 800
0 0
-50 -50
c d
-100 -100
0 50 100 150 0 50 100 150
Figure 4. Inflationary dynamics and dynamical quantum phase transition. The dynamics of (a) energy E(t), (b)
entropy S(t), (c) dynamical temperature T (t), and (d) free energy F(t) (in the unit of ωs, except for the entropy
as specified in Figs. 2 and 3) are shown for the reservoir initially in vacuum state (blue curves) and for the
reservoir with initial thermal energy kT0 = 3ωs (red curves). The system is initially in the state |n0 � = |5� with
the initial energy E0 = 5ωs, and a weak system-reservoir coupling strength η = 0.01ηc is taken.
regime with the occurrence of localized bound states, the system cannot reach to a thermal equilibrium state,
it indeed approaches to a more complex non-equilibrium (oscillatory) steady state36. Under this situation, the
non-equilibrium free energy F(t) also shows nontrivial oscillations with local minima, instead of having a global
minimum, as shown in Figs. 2c(iv) and 3c(iv).
The physical origin for the breakdown of thermodynamics in the strong system-reservoir coupling regime
with the occurrence of localized bound states comes from the important effect of long-time non-Markovian
memory dynamics, as we have discovered earlier21,36,58,59,70,71. Explicitly, the general solution of the non-equilib-
rium Green function u(t, t0 ) in the long time limit is reduced to a partially dissipationless state
1
u(t → ∞, t0 ) = e−iωb (t−t0 ) , (19)
1 − � ′ (ωb )
which is very different from the solution in the long time limit for the case of η < ηc where u(t → ∞, t0 ) = 0
(complete dissipation). Thus, the reduced density matrix of the system, given by Eq. (16) cannot be reduced to
a thermal state, it will maintain the same form of Eq. (16) in the long time limit so that its steady-state limit is
always dependent on the initial state. In this situation, the system can memorize forever some of its initial state
information, i.e. initial state information cannot be completely wiped out through the thermal fluctuations from
the reservoir. Therefore, the basic hypothesis of statistical mechanics, namely after a sufficient long enough time,
the system will evolve into an equilibrium state with the reservoir that its distribution is independent of its initial
state, breaks down completely and the system is unable to be thermalized. The oscillation behavior in Figs. 2c
and 3c is a manifestation of non-Markovian memory effect originating from the localized bound state of the
system, a general non-Markovian dynamic feature in open quantum s ystems21. The localized bound states of
the system produces dissipationless dynamics, see Eq. (19), that do not allow the system to be thermalized with
its environment, which was indeed noticed earlier by Anderson49,54 and is justified recently by one of us36. The
localized bound state can appear when the system-environment coupling strength exceeds some critical value
ηc that is determined by the structure of spectral density J(ω). For the Ohmic spectral density, ηc = ωs /ωc . This
solution also reveals the fact that the transition from thermalization to localization can occur not only in many-
body systems72, but dynamically also in single particle quantum systems coupled with reservoirs73. However, it
must be pointed out that for some spectral densities, such as the Lorentz spectral density considered in the next
section, there will exist no localized bound state no matter how strong the coupling strength is21. In these cases,
the system can always reach to an equilibrium state with the reservoir even in the strong coupling regime, and
the results show the same behavior as demonstrated in Sec. IIIB.
Dynamical quantum phase transition and quantum criticality with quantum‑classical border. Now we discuss a
very interesting situation occurring in the weak-coupling quantum limit in the quantum regime. The quantum
regime is defined as the situation in which the initial thermal energy of the reservoir is less than the system initial
energy: kT0 < E0. In this situation, the energy dissipation dominates the non-equilibrium process so that the
system energy always decreases in time (see Fig. 4a), which is very different from the classical regime in Figs. 2a
and 3a where the reservoir is in a relatively high temperature and the average energy flows between the system
and the reservoir in an opposite direction.
Vol:.(1234567890)
Figure 5. Dynamical quantum phase transition in quantum thermodynamics. Contour color-code plots
of (a) the dynamical energy E(t) and (b) the dynamical temperature T (t) are presented as varying initial
thermal energy kT0 of the reservoir. The other parameters are taken as η = 0.01 ηc , ωc = 5ωs, and the system
is considered to be in an initial state |n0 � = |5�. A dynamical quantum phase transition occurs when initial
thermal energy kT0 of the reservoir is very close to the initial system energy E0 = 5ωs.
On the other hand, as shown in Fig. 4b the system entropy (beginning with zero) will increase to a maximum
value at an intermediate time, then the entropy turns to decrease and eventually approaches to zero (the blue
curve in Fig. 4b for kT0 = 0). This phenomenon can happen because the system is initially in a pure state (zero
entropy), then the system takes dissipation (relaxation) and goes eventually into the vacuum state (back to a pure
state) when all energy of the (small) system is dissipated into the (large) reservoir. Thus, at the beginning and in
the end of the time-evolution process, the entropy of the system is zero. Therefore starting from the beginning,
the entropy must go up (increase) in time and reach the maximum at some intermediate time (corresponding
to the maximizing mixed state of the system), and henceforth go down (decrease) to zero again. On the other
hand, because the system energy always decreases, the dynamical temperature which starts from zero, must
decrease to negative infinity as the entropy approaches to the maximum. It then jumps from the negative infinite
temperature to the positive infinite temperature at the maximum entropy point, after that the temperature goes
down as the entropy decreases, and eventually reaches to zero temperature at the steady state, as shown by the
blue curve in Fig. 4c.
This dramatic change of the dynamical temperature shows that in the quantum regime the time-evolution
of the system could break the second law of thermodynamics for the system alone (because of the appearance
of a negative temperature and the decrease of the entropy in the non-equilibrium regime), but the system plus
the reservoir together still obey the second law. Also at the end, the entropy and the temperature both approach
to zero, as a manifestation of the third law of thermodynamics in equilibrium state. Note that the third law of
thermodynamics is hypothesized in equilibrium thermodynamics, it cannot be demonstrated within equilibrium
thermodynamics, namely the zero temperature is not reachable as the entropy cannot approach to zero in the
classical regime. In fact, zero temperature locates in the very deep quantum realm (corresponding to the situa-
tion of the system staying in the ground state). Therefore as a criteria, the theory of quantum thermodynamics
must be able to capture the generalized third law of thermodynamics, namely the entropy becomes zero as the
temperature approaches to zero in the deep quantum regime. Here we show explicitly how the third law of ther-
modynamics at equilibrium can be obtained dynamically from quantum mechanics.
On the other hand, the negative temperature in open quantum systems we find here only appears in the non-
equilibrium region. We find that this negative dynamical temperature phenomenon occurs as long as the system
is initially in a pure state with the condition that the system initial energy E0 is large than the initial thermal
energy kT0 of the reservoir, namely in the quantum regime. More importantly, we find that the occurrence of
the negative non-equilibrium temperature is accompanied with a nontrivial behavior of the entropy changing
from increasing to decreasing, as shown in Fig. 4b, c. Here we would like to make a conjecture that this unu-
sual dynamics may relate to the origin of the universe inflation57. The experimental observations show that the
universe is still accelerating expansion and the average temperature of the universe today is approximately 2.73
K74. Thus the universe can be considered as an open s ystem75. As an open system, we may propose that in the
very beginning (before the big bang), the universe is a single quantum system staying in a pure quantum s tate76
and nothing else. In other words, one can assume that there is a vacuum reservoir for the universe in the very
beginning. Then the universe might evolve through a non-equilibrium process similar to that given in Fig. 4.
The sudden jump from the negative infinite temperature to the positive infinite temperature at the transition
point in Fig. 4c (corresponding to the entropy varies from increase to decrease in time) allows the system to
suddenly generate an inflationary amount of heat energy, as a result of the maximizing information (entropy)
of the system. At the same time, the free energy of the system is always increased during the non-equilibrium
process, but only at the transition point it is suddenly dropped. This sudden change delivers an inflationary
energy to the reservoir, see Fig. 4d. This may provide a simple physical process for the origin of big bang and the
universe inflation57,76 which is a long-standing problem that has not been solved so far in quantum cosmology.
The continuous decrease of the entropy in the rest of the non-equilibrium process may explain how the uni-
verse evolves to flatness. Interestingly, an inflationary dynamics in Bose-Einstein condensates (BEC) crossing a
quantum critical point is recently observed77. Of course, the description of the inflationary dynamics of the early
Vol.:(0123456789)
universe in terms of such a dynamical picture from the perspective of open quantum systems is only a conjecture.
How to make this inflationary behavior of open quantum systems to become a more realistic cosmology model
remains for further investigation.
Now if we change the initial state of the reservoir from vacuum to a thermal state with a low thermal energy,
kT0 < E0, the dynamical entropy and the temperature of the system show the same behavior, except for the steady
state of the system which becomes now a thermal state with temperature T0 equilibrating to the reservoir, as
shown by the red curves in Fig. 4. Thus, when we change the initial thermal energy (kT0) of the reservoir from
a lower value to a higher value crossing the system initial energy E0 (or alternatively change the system initial
energy from E0 < kT0 to E0 > kT0 which is experimentally more feasible), we observe (see Fig. 5) a nontrivial
dynamical quantum phase transition when the thermal energy kT0 of the reservoir is very close to the initial
system energy E0. This quantum criticality provides indeed a dynamical border separating the classical regime
from quantum regime, namely the classical regime quantitatively corresponds to a rather high temperature
regime: kT0 > E0, while in the quantum regime: kT0 < E0, quantum thermodynamics can capture the occur-
rence of negative temperature and the decrease of entropy in the non-equilibrium process. To check whether
this nontrivial quantum thermodynamical transition is generally true, we consider an another application, a
two-level atomic system coupled to a bosonic reservoir in the next section.
Application to two‑level atomic systems. To demonstrate the universality of the above finding in the
simple cavity system described by the Fano-Anderson model, we investigate next the quantum thermodynamics
of another widely studied open quantum system, namely a two-level atomic system (or a generalized spin-1/2
particle) interacting with a bosonic reservoir, to see how thermodynamics emerges from quantum dynamics of
spontaneous decay process. The total Hamiltonian of the system plus reservoir is given by

Htot = ωs σ+ σ− + ωk bk† bk + gk σ+ bk + gk∗ σ− bk† , (20)
k k
which is also the basis of cavity QED in quantum optics that describes the matter-light interaction, where
σ+ = |1��0| and σ− = |0��1| are the raising and lowering operators of the Pauli matrix with ground state |0�,
excited state |1�, and transition frequency ωs . The reservoir Hamiltonian is described by a collection of infinite
bosonic modes with the bosonic operators bk† and bk , and gk is the coupling strength between the system and
k-th mode of the reservoir. The general exact master equation for two-level system with the above Hamiltonian
is still unknown. Here we consider that the system is prepared initially in the excited state |1� and the reservoir
in a vacuum state. Then the time evolution of the reduced density matrix of the two-level system is given by the
exact master e quation78,79
d
ρS (t) = − iωs′ (t, t0 )[σ+ σ− , ρS (t)]
dt (21)
+ γ (t, t0 )[2σ− ρS (t)σ+ − σ+ σ− ρS (t) − ρS (t)σ+ σ− ],
where the renormalized transition frequency ωs′ (t, t0 ) = −Im(u̇(t, t0 )/u(t, t0 )) and the dissipation coefficient
γ (t, t0 ) = −Re(u̇(t, t0 )/u(t, t0 )), which is the same as that given in the general exact master equation of Eq. (13).
The Green function u(t, t0 ) satisfies also the same integro-differential equation as Eq. (14a)
t
d
u(t, t0 ) + iωs u(t, t0 ) + dτ g(t, τ )u(τ , t0 ) = 0, (22)
dt t0
where the time non-local integral kernel g(t, τ ) is related to the spectral density J(ω) of the reservoir as
g(t, τ ) = dωJ(ω)e−iω(t−τ ). Here we consider a Lorentzian spectral density,
1 γ0 2
J(ω) = , (23)
2π (ω0 − ω)2 + 2
for which the Green function u(t, t0 ) of Eq. (22) can easily be solved with the exact solution

Ŵ(t − t0 ) Ŵ(t − t0 )
u(t, t0 = 0) = e(−iωs − 2 )(t−t0 ) cosh + sinh , (24)
2 Ŵ 2

where Ŵ = 2 − 2γ0 , which has no localized bound state because the spectral density covers the whole range
of the frequency domain21. The spectral width of the reservoir is connected to the reservoir correlation time
τB by the relation τB ≈ −1. The relaxation time scale τR over which the state of the system changes is related
to the coupling strength γ0 by the relation τR ≈ γ0−1. The reduced density matrix ρS (t) depicts the dissipative
relaxation process of the system at far from equilibrium. Once we have the density matrix ρS (t), the dynamical
thermal parameters are determined by the relations (9–10). In Fig. 6, we show the dynamics of these thermal
parameters, namely, the dynamical internal energy E(t), entropy S(t), temperature T (t) and the free energy F(t).
We consider two different regime of the system-reservoir parameters: (i) > 2γ0, for which the reservoir cor-
relation time is small compared to the relaxation time (τB < τR ) of the system and the behavior of u(t, t0 ) shows
a Markov exponential decay (ii) < 2γ0, in this regime the reservoir correlation time τB is large or comparable
to the relaxation time scale τR where non-Markovian effects come into p lay80.
We investigate the non-equilibrium thermodynamics of the system in the above two physically distinct
regimes. In the weak coupling case with τR = 5τB, the energy E(t) shows a monotonous decay due to dissipation,
Vol:.(1234567890)
1.0 0.8
a (i) λ > 2γ0 (τR = 5τB ) b
0.8 (ii) λ < 2γ0 (τR = 0.2τB ) 0.6
0.6
E(t)
S(t)
0.4
0.4
0.2 0.2
0.0 0.0
0 50 100 150 200 250 300 0 50 100 150 200 250 300
50 50
c d
25 25
kT (t)
F (t)
0 0
−25 −25
−50 −50
0 20 40 ω t 60 80 100 0 20 40 ω t 60 80 100
s s
Figure 6. Non-equilibrium quantum thermodynamics of a two-level atomic system. The dynamics of non-
equilibrium thermal parameters at two physically distinct regime (i) > 2γ0 and (ii) < 2γ0. We plot (a)
average energy E(t) (b) entropy S(t) (c) non-equilibrium temperature T (t) and (d) free energy F(t) (in the unit
of ωs except for the entropy), where the blue curves represent weak coupling regime (i) with τR = 5τB, and the
red curves indicate the strong coupling regime (ii) with τR = 0.2τB. The system is initially in the excited state,
and the reservoir is initially in vacuum state.
as shown in Fig. 6a (the blue curve). On the other hand, Fig. 6b (the blue curve) shows that the entropy S(t)
increases due to the contact of the system to the reservoir, and attains a maximum value at an intermediate time,
henceforth the entropy starts decreasing and eventually approaches to zero. This happens because the system is
initially in a pure state (zero entropy), then the entropy increases during the non-equilibrium process and reach
the maximum corresponding to maximum mixing state of the system, finally the entropy goes to zero again when
the system relaxes into the vacuum state (pure state) in order to equilibrate with the reservoir. The decrease in
energy with increasing entropy results in a negative dynamical temperature of the system, as shown in Fig. 6c.
The dynamical temperature T (t) decreased to negative infinity as the entropy approaches to the maximum, it
then jumps from negative infinite temperature to positive infinite temperature at the maximum entropy point,
eventually the system reaches to zero temperature at the steady state as shown in Fig. 6c, the exactly same as what
we find in the cavity system in the weak-coupling quantum regime.
The physical origin of negative dynamical temperature for this two-level atomic system and the associated
inflationary behavior are also the same as that in the cavity system: The free energy increases instead of decreas-
ing (see Fig. 6d). Hence, for this two-level system, the second law of thermodynamics also cannot be maintained
when the system evolves from a pure state to a mixed state and then back to a pure state in the whole spontane-
ous decay processes. Nevertheless, the entropy and the temperature both dynamically approach to zero in the
steady-state limit, as a manifestation of the third law of thermodynamics in equilibrium state in the quantum
regime. In the strong coupling case with τR = 0.2τB (see the red curves of Fig. 6), we show the non-equilibrium
thermodynamics in the regime < 2γ0. In this regime, the average energy E(t) initially decreases, then it shows
a small oscillation, demonstrating a backflow of energy (red curve of Fig. 6a) from the system into the reservoir,
as a non-Markovian memory effect. We see a similar dynamical behavior in the entropy S(t) under this strong
coupling, namely, the entropy increases in the short time, then decreases and oscillates in later time (red curve
of Fig. 6b). The dynamical temperature and free energy in the strong coupling regime (red curves of Fig. 6c, d)
also show similar behaviors except a small shift in the time-scale. The dynamical transition associated with the
inflationary dynamics is the same as we found in the single cavity system.
The concept of negative temperatures has indeed been discussed in the literature81,82 which demands the
system to be in the equilibrium and also to have a finite spectrum of energy states, bounded both from the top
and from the bottom. Such negative equilibrium temperature has been realized in localized spin systems83–85,
where the discrete finite spectrum naturally provides both lower and upper energy bounds. A similar situation
is also found recently where a cloud of potassium atoms is tuned to negative temperatures via quantum phase
transition86. However, the physical origin or the condition under which the negative temperature phenomenon
occurs in our cases is quite distinct from the previous studies81–86. In previous studies81–86, it was emphasized
that in thermal equilibrium, the probability for a particle to occupy a state with energy En is proportional to the
Boltzmann factor exp(−En /kT). For negative temperatures, the Boltzmann factor increases exponentially with
increasing En and the high-energy states are then occupied more easily than the low-energy states. Contrary to
that, the emergence of negative temperature here we see is dynamical, and it has a clear physical origin in the
non-equilibrium process that the entropy increases and then decreases as the system starts from a pure state,
evolves into a maximum mixed state and then returns back to a pure state (or a less-maximized mixed state)
dominated by the spontaneous emissions when the reservoir is initially in vacuum state (or an initial state has a
Vol.:(0123456789)
very low thermal energy). As long as one can prepare the system in a pure state with energy E0 and the reservoir
in a thermal state with initial thermal energy kT0 such that kT0 < E0 (the quantum regime), the occurrence of
a negative dynamical temperature and the associated inflationary behavior are universal behaviors in the ther-
malization processes of open quantum systems.
Conclusions and outlook

Thermodynamics is built with the hypothesis of equilibrium, where dynamics in time is absent and all thermo-
dynamic parameters are time-independent. In this paper, we develop a non-equilibrium theory for quantum
thermodynamics validating even for single particle quantum systems coupled with reservoirs far from equilib-
rium. We generalize the concept of temperature and other thermodynamic parameters to non-equilibrium regime
that depends on the details of quantum states of the system and their time evolution. We show that the theory
unravels the intimate connection between the laws of thermodynamics and their quantum origins. We explicitly
and exactly solved the dynamical evolution of two single quantum systems interacting with a thermal reservoir
described by the Fano-Anderson model and multimode Jaynes-Cummings (JC) model, respectively. These are
the two most general examples that can be solved unambiguously in open quantum systems. We demonstrated
the emerging thermodynamics of the system through the time evolution of various thermodynamic parameters,
namely the dynamical energy, entropy, temperature, free energy, in the weak and intermediate system-reservoir
coupling regimes under the situation of the system far from equilibrium. We showed how the system dynamically
approaches to a thermal equilibrium state asymptotically in the steady-state limit. In the strong system-reservoir
coupling regime if there exist localized bound states21, a revival dynamics in the thermodynamic parameters is
observed, indicating back-reactions between the system and the environment as a long-time non-Markovian
memory effect70. Consequently, equilibrium thermodynamics is unreachable when the system has localized
mode or localized bound state, as noticed early by Anderson49 and justified recently by one of us36. However,
if the spectral density is continuous and covers the whole energy range so that no localized bound state can be
generated by the couplings between the system and reservoirs, then thermalization is always reachable.
The other remarkable findings in the quantum thermodynamics appear when the reservoir thermal energy kT0
being less than the system initial energy, which is defined as the quantum regime. Under this situation, we observe
a dramatic change in the non-equilibrium thermodynamic behavior of the system. This dramatic change induces
an inflationary dynamics which may provide a simple picture for the origin of big bang and universe inflation.
The ending steady state in this situation manifests the third law of thermodynamics in the deep quantum realm.
We also observed a nontrivial dynamical quantum phase transition when the thermal energy of the reservoir
is very close to the system initial energy. The corresponding dynamical criticality provides a border separating
quantum and classical regimes. This dynamical quantum phase transition is manifested through non-equilibrium
quantum processes that cannot be captured by equilibrium thermodynamics itself.
These new findings for quantum thermodynamics could be examined through BEC experiments or cavity
QED experiments. More specifically, one can experimentally explore the photon dynamics or boson dynamics
via quantum non-demolition m easurement87,88 to justify these findings. For example, by preparing the cavity in
a Fock state, and sending sequences of circular Rydberg atoms through the photonic cavity in different environ-
ment, which carry the cavity state information without destroying the cavity photon state. The cavity photon
state can be measured using an experimental setup similar to that given in Refs.87,89. Another way to observe the
nontrivial quantum thermodynamics we obtained is to measure the energy loss and the state evolution of a two-
level atom in cavity QED system. It may be also possible to use the BEC systems77 to find these new features. On
the other hand, it is also straightforward to apply this quantum thermodynamics theory to electronic systems in
nanostructures where the exact master equation with quantum transport has been d eveloped21,90.
Received: 25 February 2020; Accepted: 30 June 2020
References
1. Cazalilla, M. A. & Rigol, M. Focus on dynamics and thermalization in isolated quantum many-body systems. New J. Phys. 12,
055015 (2010).
2. Trotzky, S. et al. Probing the relaxation towards equilibrium in an isolated strongly correlated one-dimensional Bose gas. Nat. Phys.
8, 325 (2012).
3. Gring, M. et al. Relaxation and prethermalization in an isolated quantum system. Science 337, 1318 (2012).
4. Langen, T., Geiger, R., Kuhnert, M., Rauer, B. & Schmiedmayer, J. Local emergence of thermal correlations in an isolated quantum
many-body system. Nat. Phys. 9, 640 (2013).
5. Kaufman, A. M. et al. Quantum thermalization through entanglement in an isolated many-body system. Science 353, 794 (2016).
6. Gemmer, J., Michel, M. & Mahler, G. Quantum Thermodynamics: Emergence of Thermodynamic Behavior Within Composite
Quantum Systems 2nd edn. (Springer, Berlin, 2009).
7. Campisi, M., Hänggi, P. & Talkner, P. Colloquium: quantum fluctuation relations: foundations and applications. Rev. Mod. Phys.
83, 771 (2011).
8. Jarzynski, C. Equalities and inequalities: irreversibility and the second law of thermodynamics at the nanoscale. Annu. Rev. Condens.
Matter Phys. 2, 329 (2011).
9. Scully, M. O., Chapin, K. R., Dorfman, K. E., Kim, M. B. & Svidzinsky, A. Quantum heat engine power can be increased by noise-
induced coherence. PNAS 108, 15097 (2011).
10. Kosloff, R. Quantum thermodynamics: a dynamical viewpoint. Entropy 15, 2100 (2013).
11. Jurcevic, P. et al. Quasiparticle engineering and entanglement propagation in a quantum many-body system. Nature 511, 202
(2014).
12. Eisert, J., Friesdorf, M. & Gogolin, C. Quantum many-body systems out of equilibrium. Nat. Phys. 11, 124 (2015).
13. Esposito, M., Ochoa, M. A. & Galperin, M. Quantum thermodynamics: a nonequilibrium Greens function approach. Phys. Rev.
Lett. 114, 080602 (2015).
Vol:.(1234567890)
14. Pekola, J. P. Towards quantum thermodynamics in electronic circuits. Nat. Phys. 11, 118 (2015).
15. Gelbwaser-Klimovsky, D., Niedenzu, W. & Kurizki, G. Thermodynamics of quantum systems under dynamical control. Adv. Atom.
Mol. Opt. Phys. 64, 329 (2015).
16. Roßnagel, J. et al. A single-atom heat engine. Science 352, 325 (2016).
17. Millen, J. & Xuereb, A. Perspective on quantum thermodynamics. New J. Phys. 18, 011002 (2016).
18. Anders, J. & Esposito, M. Focus on quantum thermodynamics. New J. Phys. 19, 010201 (2017).
19. Ochoa, M. A., Zimbovskaya, N. & Nitzan, A. Quantum thermodynamics for driven dissipative bosonic systems. Phys. Rev. B 97,
085434 (2018).
20. Deffner, S. & Campbell, S. Quantum Thermodynamics, (https://iopscience.iop.org/book/978-1-64327-658-8).
21. Zhang, W. M., Lo, P. Y., Xiong, H. N., Tu, M. W. Y. & Nori, F. General non-Markovian dynamics of open quantum systems. Phys.
Rev. Lett. 109, 170402 (2012).
22. Tu, M. W. Y. & Zhang, W. M. Non-Markovian decoherence theory for a double-dot charge qubit. Phys. Rev. B 78, 235311 (2008).
23. Jin, J., Tu, M. W. Y., Zhang, W. M. & Yan, J. Y. Non-equilibrium quantum theory for nanodevices based on the Feynman-Vernon
influence functional. New. J. Phys. 12, 083013 (2010).
24. Lei, C. U. & Zhang, W. M. A quantum photonic dissipative transport theory. Ann. Phys. 327, 1408 (2012).
25. Zhang, W. M. Exact master equation and general non-Markovian dynamics in open quantum systems, arXiv:1807.01965 (2018).
26. Callen, H. B. Thermodynamics and an Introduction to Thermostatistics (Wiley, New York, 1985).
27. Landau, L. D. & Lifshitz, E. M. Statistical Physics Vol. 5 of Course of Theoretical Physics, (2nd Ed., Pergamon Press, 1969).
28. Kubo, R., Toda, M. & Hashitsume, N. Statistical physics II: Nonequilibrium statistical mechanic (Springer Series in Solid-State Sci-
ences, Vol. 31, 2nd ed., 1991).
29. Caldeira, A. O. & Leggett, A. J. Path integral approach to quantum Brownian motion. Physica 121A, 587 (1983).
30. Huang, K. Statistical Mechanics (2nd ed., John Wiley and Sons, 1987), Chap. 8.7, p. 189–191.
31. De Groot, S. R. & Mazur, P. Non-equilibrium Thermodynamics (Dover, New York, 2013).
32. Demirel, Y. Nonequilibrium Thermodynamics: Transport and rate Processes in Physical, Chemical and Biological Systems 3rd edn.
(Elsevier Science, Amsterdam, 2014).
33. Hu, B.-L., Paz, J. P. & Zhang, Y. Quantum Brownian motion in a general environment: exact master equation with nonlocal dis-
sipation and colored noise. Phys. Rev. D 45, 2843 (1992).
34. Ford, G. W. & Connell, R. F. Exact solution of the Hu-Paz-Zhang master equation. Phys. Rev. D 64, 105020 (2001).
35. Zurek, W. H. Decoherence, einselection, and the quantum origins of the classical. Rev. Mod. Phys. 75, 715 (2003).
36. Xiong, H. N., Lo, P. L., Zhang, W. M., Feng, D. F. & Nori, F. Non-Markovian complexity in the quantum-to-classical transition.
Sci. Rep. 5, 13353 (2015).
37. Lo, P. Y., Xiong, H. N. & Zhang, W. M. Breakdown of Bose–Einstein distribution in photonic crystals. Sci. Rep. 5, 9423 (2015).
38. Xiong, F. L. & Zhang, W. M. Exact dynamics and thermalization of open quantum systems coupled to reservoir through particle
exchanges, arXiv :2003.09598.
39. von Neumann, J. Mathematische Grundlagen der Quantenmechanik. (Springer, Berlin, 1932), also Mathematical foundations of
quantum mechanics. translated by Beyer, R. T. (Princeton university press, 1955).
40. Feynman, R. P. & Vernon, F. L. The theory of a general quantum system interacting with a linear dissipative system. Ann. Phys. 24,
118 (1963).
41. Schwinger, J. Brownian motion of a quantum oscillator. J. Math. Phys. 2, 407 (1961).
42. Keldysh, L. V. Diagram technique for nonequilibrium processes. Sov. Phys. JETP 20, 1018 (1965).
43. Kadanoff, L. P. & Baym, G. Quantum Statistical Mechanics (Benjamin, New York, 1962).
44. See the more detailed discussion on pages 329-331 in the classical textbook of thermodynamics, Ref. [26].
45. Hänggi, P., Ingold, G.-L. & Talkner, P. Finite quantum dissipation: the challenge of obtaining specific heat. New J. Phys. 10, 115008
(2008).
46. Ingold, G.-L., Hänggi, P. & Talkner, P. Specific heat anomalies of open quantum systems. Phys. Rev. E. 79, 061105 (2009).
47. Ochoa, M. A., Bruch, A. & Nitzan, A. Energy distribution and local fluctuations in strongly coupled open quantum systems: the
extended resonant level model. Phys. Rev. B 94, 035420 (2016).
48. Carrego, M., Solinas, P., Sassetti, M. & Weiss, U. Energy exchange in driven open quantum systems at strong coupling. Phys. Rev.
Lett. 116, 240403 (2016).
49. Anderson, P. W. Absence of diffusion in certain random lattices. Phys. Rev. 109, 1492 (1958).
50. Fano, U. Effects of configuration interaction on intensities and phase shifts. Phys. Rev. 124, 1866 (1961).
51. Jaynes, E. T. & Cummings, F. W. Comparison of quantum and semiclassical radiation theories with application to the beam maser.
Proc. IEEE 51, 89 (1963).
52. Allen, L. & Eberly, J. H. Optical Resonance and Two-Level Atoms (Dover, New York, 1987).
53. Miroshnichenko, A. E., Flach, S. & Kivshar, Y. S. Fano resonances in nanoscale structures Rev. Mod. Phys. 82, 2257 (2010).
54. Mahan, G. D. Many-Body Physics (Kluwer Academic/Plenum Publishers, New York, 2000).
55. Cohen-Tannoudji, C., Dupont-Roc, J. & Grynberg, G. Atom-Photon Interactions (Wiley, New York, 1992).
56. Scully, M. O. & Zubairy, M. S. Quantum Optics (Cambridge University Press, Cambridge, 1997).
57. Guth, A. H. Inflationary universe: a possible solution to the horizon and fatness problems. Phys. Rev. D 23, 347 (1981).
58. Wu, M. H., Lei, C. U., Zhang, W. M. & Xiong, H. N. Non-Markovian dynamics of a microcavity coupled to a waveguide in photonic
crystals. Opt. Express 18, 18407 (2010).
59. Lei, C. U. & Zhang, W. M. Decoherence suppression of open quantum systems through a strong coupling to non-Markovian
reservoirs. Phys. Rev. A 84, 052116 (2011).
60. Leggett, A. J. et al. Dynamics of the dissipative two-state system. Rev. Mod. Phys. 59, 1 (1987).
61. Brandão, F., Horodecki, M., Ng, N., Oppenheim, J. & Wehner, S. The second laws of quantum thermodynamics. Proc. Natl. Acad.
Sci. 112, 3275 (2015).
62. Riera, A., Gogolin, C. & Eisert, J. Thermalization in nature and on a quantum computer. Phys. Rev. Lett. 108, 080402 (2012).
63. Seifert, U. First and second law of thermodynamics at strong coupling. Phys. Rev. Lett. 116, 020601 (2016).
64. Xiong, H. N., Zhang, W. M., Wang, X. & Wu, M. H. Exact non-Markovian cavity dynamics strongly coupled to a reservoir. Phys.
Rev. A 82, 012105 (2010).
65. Gallego, R., Riera, A. & Eisert, J. Thermal machines beyond the weak coupling regime. New J. Phys. 16, 125009 (2014).
66. Esposito, M., Ochoa, M. A. & Galperin, M. Nature of heat in strongly coupled open quantum systems. Phys. Rev. B 92, 235440
(2015).
67. Campisi, M., Talkner, P. & Hänggi, P. Fluctuation theorem for arbitrary open quantum systems. Phys. Rev. Lett. 102, 210401 (2009).
68. Jarzynski, C. Stochastic and macroscopic thermodynamics of strongly coupled systems. Phys. Rev. X 7, 011008 (2017).
69. Perarnau-Llobet, M., Wilming, H., Riera, A., Gallego, R. & Eisert, J. Strong coupling corrections in quantum thermodynamics.
Phys. Rev. Lett. 120, 120602 (2018).
70. Ali, M. M., Lo, P. Y., Tu, M. W. Y. & Zhang, W. M. Non-Markovianity measure using two-time correlation functions. Phys. Rev. A
92, 062306 (2015).
71. Ali, M. M. & Zhang, W. M. Nonequilibrium transient dynamics of photon statistics. Phys. Rev. A 95, 033830 (2017).
Vol.:(0123456789)
72. Nandkishore, R. & Huse, D. A. Many-body localization and thermalization in quantum statistical mechanics. Ann. Rev. Condens.
Matter Phys. 6, 15 (2015).
73. Lin, Y. C., Yang, P. Y. & Zhang, W. M. Non-equilibrium quantum phase transition via entanglement decoherence dynamics. Sci.
Rep. 6, 34804 (2016).
74. Astier, P. & Pain, R. Observational evidence of the accelerated expansion of the universe. C. R. Phys. 13, 521 (2012).
75. Hooft, G. Quantum gravity as a dissipative deterministic system. Class. Quantum Grav. 16, 3263 (1999).
76. Hartle, J. B. & Hawking, S. W. Wave function of the Universe. Phys. Rev. D 28, 2960 (1983).
77. Feng, L., Clark, L. W., Gaj, A. & Chin, C. Coherent inflationary dynamics for Bose–Einstein condensates crossing a quantum
critical point. Nat. Phys. 14, 269 (2018).
78. Garraway, B. M. Nonperturbative decay of an atomic system in a cavity. Phys. Rev. A 55, 2290 (1997).
79. Vacchini, B. & Breuer, H. P. Exact master equations for the non-Markovian decay of a qubit. Phys. Rev. A 81, 042103 (2010).
80. Bellomo, B., Franco, R. L. & Compagno, G. Non-Markovian effects on the dynamics of entanglement. Phys. Rev. Lett. 99, 160502
(2007).
81. Ramsey, N. F. Thermodynamics and statistical mechanics at negative absolute temperatures. Phys. Rev. 103, 20 (1956).
82. Klein, M. J. Negative absolute temperatures. Phys. Rev. 104, 589 (1956).
83. Purcell, E. M. & Pound, R. V. A nuclear spin system at negative temperature. Phys. Rev. 81, 279 (1951).
84. Oja, A. S. & Lounasmaa, O. V. Nuclear magnetic ordering in simple metals at positive and negative nanokelvin temperatures. Rev.
Mod. Phys. 69, 1 (1997).
85. Medley, P. et al. Spin gradient demagnetization cooling of ultracold atoms. Phys. Rev. Lett. 106, 195301 (2011).
86. Braun, S. et al. Negative absolute temperature for motional degrees of freedom. Science 339, 52 (2013).
87. Guerlin, C. et al. Progressive field-state collapse and quantum non-demolition photon counting. Nature 448, 889 (2007).
88. Deléglise, S. et al. Reconstruction of non-classical cavity field states with snapshots of their decoherence. Nature 455, 510 (2008).
89. Sayrin, C. et al. Real-time quantum feedback prepares and stabilizes photon number states. Nature 477, 73 (2011).
90. Yang, P. Y. & Zhang, W. M. Master equation approach to transient quantum transport in nanostructures. Front. Phys. 12, 127204
(2017).
Acknowledgements
WMZ acknowledges the support from the Ministry of Science and Technology of Taiwan under Contract
Nos. MOST-105-2112-M-006-008-MY3 and MOST-108-2112-M-006-009-MY3. MMA acknowledges the sup-
ports from the Ministry of Science and Technology of Taiwan and the Physics Division of National Center for
Theoretical Sciences, Taiwan.
Author contributions
MMA performed the detailed calculations and WMH performed the calculation for the two-level systems, WMZ
proposed the ideas, developed the formulation and interpreted the physics, MMA also contributed to the discus-
sion on the early version of the physical formulation, and WMZ wrote the manuscript.
Competing interests
The authors declare no competing interests.
Additional information
Correspondence and requests for materials should be addressed to W.-M.Z.
Reprints and permissions information is available at www.nature.com/reprints.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the
Creative Commons license, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2020
Vol:.(1234567890)

Quantum Thermodynamics of Single Particle Systems: Md. Manirul Ali, Wei Ming Huang & Wei Min Zhang

Uploaded by

Copyright:

Available Formats

Quantum Thermodynamics of Single Particle Systems: Md. Manirul Ali, Wei Ming Huang & Wei Min Zhang

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Quantum Thermodynamics of Single Particle Systems: Md. Manirul Ali, Wei Ming Huang & Wei Min Zhang

Uploaded by

Copyright:

Available Formats

www.nature.

OPEN Quantum thermodynamics

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 1

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 2

ρS (t) = Tr R [U(t, t0 )ρtot (t0 )U † (t, t0 )]. (8)

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 3

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 4

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 5

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 6

Figure 2. Non-equilibrium quantum thermodynamics in different system-reservoir coupling regimes. The

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 7

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 8

(ii) (ii) (ii)

(iv) -150 (iv) (iv)

Figure 3. Non-equilibrium quantum thermodynamics in different system-reservoir coupling regimes. The

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 9

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 10

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 11

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 12

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 13

Conclusions and outlook

Received: 25 February 2020; Accepted: 30 June 2020

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 14

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 15

© The Author(s) 2020

Scientific Reports | (2020) 10:13500 | https://doi.org/10.1038/s41598-020-70450-y 16

You might also like