Comment on “ Entanglement and the Thermodynamic Arrow of Time ” and Correct Reply on “ Comment on " Quantum Solution to the Arrow-of-Time Dilemma " ” of David Jennings and Terry Rudolph

Recently David Jennings and Terry Rudolph published two papers as reaction on Maccone’s paper "Quantum Solution to the Arrow-of-Time Dilemma". In these papers, the authors suppose that second law of thermodynamics is not relevant for quantum systems. Unfortunately, these papers did not get relevant reply from Maccone. The reason of this is following. Both Maccone and the above-mentioned authors use thermodynamic law and thermodynamic-like terminology for non-thermodynamic systems, for example, microscopic system of three qubits. However, big size of a system (quantum or classic) is also not an enough condition for a system to be macroscopic. The macroscopic system must also be chaotic and has small chaotic interaction with its environment/observer resulting in decoherence (decorrelation). We demonstrate that for relevant thermodynamic macroscopic quantum systems no objection appears.


Introduction
The paper of David Jennings, Terry Rudolph "Entanglement and the Thermodynamic Arro w of Time" is very interesting. However, the Thermodynamic A rrow of Time is not applicable for microsystems. It is a nice paper about quantum fluctuation, but not a paper about Thermodynamic Arrow of Time . In the Abstract of the paper, "Entanglement and the Thermodynamic Arrow of Time" the authors write: "We examine in detail the case of three qubits, and also propose some simple experimental demonstrations possible with sma ll nu mbers of qubits." Nevertheless, no thermodynamics is possible for such a microsystem. D. Jennings and T. Rudolph (like Maccone) do not understand that category "thermodynamic arrow o f time" is correct only for large macrosystems. Using these categories for sma ll fluctuating systems has no physical sense. They also (like Maccone) use incorrect defin ition of macroscopic thermodynamic entropy.
We also give (instead of Maccone) the correct reply to "Co mment on "Quantu m So lut ion to the A rro w-of-Time Dilemma"". The correct reply is that no contradictions (found in this Co mment) appear for macroscopic systems. Only for a microscopic system, such contradictions e xist. However, the concepts "the Thermodynamic Arro w of Time" and "the entropy growth law" is not relevant for such systems. We illustrate this fact by consideration of a quantum chaotic macrosystem and demonstrate that no contradiction (found by David Jennings, Terry Rudolph for a microscopic system) exists for this correct thermodynamical case. It must be mentioned that big size of a system (quantum or classic) is also not an enough condition for a system to be macroscopic. The macroscopic system (considered in Thermodynamics) must also be chaotic (quantum or classic) and has small chaotic interact ion with its environment/obser ver resulting in decoherence (for quantum mechanics) or decorrelation (for classical mechanics). It should be also mentioned that thermodynamic-like terminology is widely and effectively used in quantum mechanics, quantum computers field, and information theory. The big nu mber of the examp les can be found in the references of Jennings and Rudolph's paper. The other nice examp le is Shannon's entropy in information theory. But usually an author (using such a thermodynamic-like termino logy) does not consider such a paper as analysis of classical Thermodynamics. Contrarily Jennings and Rudolph "disprove" the second law of Thermodynamics on the basis of the irrelevant "Comment on "Quantum Solution to the Arrow-of-Time Dilemma"" of David Jennings and Terry Rudolph microscopic system (in their Co mment) and give (also in this Co mment) the announcement of their next paper «Entanglement and the Thermodynamic Arrow of Time" as a correct consideration and a disproof of the second law. The reason of align ment of thermodynamic time arrows in a quantum mechanics, as well as in the classical mechanics, is small interaction between real chaotic macroscopic systems. This well studied appearance carrying a tit le «decoherence» [2-3, 17, 24-27]. Its result is not only widely known «entangling» states of systems, but also align ment of thermodynamic time arrows. (The direct ion of thermodynamic t ime arro w is defined by a direct ion of the entropy increase.) The reason of align ment of thermodynamic t ime arrows is the same, as in the classical Hamilton mechanics -instability of processes with opposite time arrows with respect to small perturbations. These perturbations exist between the observer/environment and observed system (decoherence).
Similar argu ments in the case of quantum mechanics have been given in Maccone's paper [4]. Ho wever there he formulated, that the similar logic is applicab le only in a quantum mechanics. The incorrectness of this conclusion has been shown in our previous papers [1,5]. The other objection has been formulated in the paper [6]. There are considered small systems with strong fluctuations. Alignment of thermodynamic t ime arrows does not exist for such small systems. It must be mentioned that both Maccone' s replay to this objection and the subsequent paper of objection authors [7] do not explain the true reason of described disagreement. The real solution is very simple. More specifically, the entropy increase law, the concept of thermodynamic time arrows and their align ment are applicable only to nonequilibriu m macroscopic objects. Vio lation of these laws for microscopic systems with strong fluctuations is widely known fact. Nevertheless, though the objection [6] is trivial physically, but it is interesting from purely mathemat ical point of v iew. It g ives good mathematical criterion for macroscopicity of chaotic quantum systems.

Reduction of System at Measurement
This part is based on [22,23] Let's consider a situation when a measuring device was at the beginning in state | 0 ⟩ , and the object was in superposition of states | ⟩ = ∑ | ⟩ , where | ⟩ are experiment eigenstates. The init ial statistical operator is given by expression (1) The partial track of this operator which is equal to statistical operator of the system, including only the object, looks like ( 0 ) = ∑ ⟨ | 0 | ⟩ where | ⟩ -any co mplete set of device eigenstates. Thus, (2) Where the relation ∑ | ⟩ ⟨ | = 1 and normalizat ion condition for | 0 › are used. We have statistical operator correspondent to object state | ›. After measuring there is a correlation between device and object states, so the state of full system including device and object is featured by a state vector (Since various states | ⟩ of device are o rthogonal each other); thus, We have obtained statistical operator including only the object, featuring probabilit ies | | 2 for object states | ⟩ . So, we co me to formu lation of the following theorem.
Theorem 1 (about measuring). If two systems S and A interact in such a manner that to each state | ⟩ systems S there corresponds a certain state | ⟩ of systems A the statistical operator ( ) over full systems (S and A) reproduces wave packet reduction for measuring, yielded over system S, wh ich before measuring was in a state | ⟩ = ∑ | ⟩ .
Suppose that some subsystem is in mixed state but the full system including this subsystem is in pure state. Such mixed state is named as improper mixed state.

The Theorem about Decoherence at Interacti on with the Macroscopic Device
This part is based on [18,84] Let's consider now that the device is a macroscopic system. It means that each distinguishable configuration of the device (for example, position of its arrow) is not a pure quantum state. It states nothing about a state of each separate arrow mo lecule. Thus, in the above-stated reasoning the initial state of the device | 0 ⟩ should be described by some statistical distribution on microscopic quantum states � 0, � ; the initial statistical operator is not given by expression (1), and is equal Each state of the device � 0 , � will interact with each object eigenstate | ⟩ . So, it will be transformed to some other state � , � . It is one of the quantum states of set with macroscopic description correspondent to arrow in position i; more precisely we have the formu la   (7) and (8) fo llows that at | ⟩ = ∑ | ⟩ the statistical operator after measuring will be given by following expression: = ∑ ( , , ) * � , − , � | ⟩ � , � � , � � � (9) As from (9) the sa me result (6) can be concluding. So we see that the statistical operator (9) reproduces an operation of reduction applied to given object. It also practically reproduces an operation of reduction applied to device only ("practically" in the sense that it is a question about "macroscopic" observable variab le). Such observable variable does not distinguish the different quantum states of the device corresponding to the same macroscopic description, i.e. matrix elements of this observable variable correspondent to states | ⟩ � , � and � � � , � do not depend on r and s. Average value of such macroscopic observable variable A is equal to As phases , are distributed randomly, the sum over s are zero at ≠ ; hence, We obtain statistical operator wh ich reproduces operation of reduction on the device. If the device arro w is observed in position i, the device state for some s will be � , � . The probability to find state � , � is equal to probability of that before measuring its state was � , � . Thus, we co me to the following theorem.
Theorem 2. About decoherence of the macroscopic de vice. Suppose that the quantum system interacts with the macroscopic device in such a manner that there is a chaotic distribution of states phases of the device. Suppose that is a statistical operator of the device after the measuring, calculated with the help of Schrodinger equations, and ′ is the statistical operator obtained as a result of reduction application to operator . Then it is impossible to yield such experiment with the macroscopic device which would register difference between and ′ . It is the so-called Daneri-Loinger-Prosperi theorem [21].
For a wide class of devices it is proved that the chaotic character in distribution of phases formulated in the theorem 2 really takes place if the device is macroscopic and chaotic with unstable initial state. Indeed, randomness of phase appears from rando mness of energies (eigenvalues of Hamiltonian) in quantum chaotic systems [8].
It is worth to note that though Eq. (48) is relevant with a split-hair accuracy it is only assumption with respect to (9). There fro m it is often concluded that the given above proof is FAPP. It means that it is only difficult to measure quantum correlations practically. Actually they continue to e xist. Hence, in principle they can be measured. It is, however, absolutely untruly. Really, fro m Poincare's theorem about returns follo ws that the system will not remain in the mixed state (12), and should return to the initial state (7). It is the result of the very sma ll corrections (quantum correlat ion) which are not included to (12). Nevertheless, the system featured here � , � corresponds to the introspection case, and consequently, it is not capable to observe experimentally these returns in principle (as it was shown above in resolution of Poincare and Loshmidt paradoxes). Hence, effects of these small corrections e xist only on paper in the coordinate time of ideal dynamics, but it cannot be observed experimentally with respect to thermodynamic t ime arrow of observable dynamics of the macroscopic device. So, we can conclude that Daneri-Lo inger-Prosperi theorem actually results in a complete resolution (not only FAPP!) of the reduction paradox in principle. It proves impossibility to distinguish experimentally the complete and inco mplete reduction.
The logic produced here strongly reminds Maccone's paper [4]. It is not surprising. Indeed, the pass fro m (7) to (12) corresponds to increasing of microstates number and entropy growth. And the pass from (12) in (7) corresponds to the entropy decrease. Accordingly, our statement aboute xperim ental unobservability to remainder quantum correlation is equivalent to the statement about unobservability of the entropy decrease. And it is proved by the similar methods, as in [4]. The object ion [6] was made against this paper. Unfortunately, Maccone could not give the reasonable reply [28] to this objection. Here we will try to do it ourselves.
Let's define here necessary conditions. Suppose is our device, and is the measured quantum system.
The first value, the mutual entropy ( : ) is the coarsened entropy of ensemble (received by separation on two subsystems) excluding the ensemble entropy. As the second excluding term is constant, so ( : ) describes well the behavior of macroentropy in time: The second value ( : ) is the classical mutual informat ion. It defines which ma ximu m information about measured system � � we can receive fro m indication of instrument ( ) . The more correlation exists between systems, the more informat ion about measured system we can receive: where � : � = ∑ log − ∑ log − ∑ log , = � ⊗ �, = ∑ and = ∑ given POVMs (Positive Operator Va lued Measure) and for A and C, respectively. Maccone [4] proves an inequality ( : ) ≥ ( : ) (13) He concludes from it that entropy decrease results in reduction of the informat ion (me mo ry) about the system + and . But (13) contains an inequality. Correspondingly in [6] an exa mple o f the quantum system of three qubits is supplied. For this system the mutual entropy decrease is accompanied by mutual informat ion increases. It does not contradict to (13) "Comment on "Quantum Solution to the Arrow-of-Time Dilemma"" of David Jennings and Terry Rudolph because mutual entropy is only up boundary for mutual informat ion there.
Let's look what happens in our case of the macroscopic device and the measured quantum system Before measurement (7 Thus, our case corresponds to ( : ) = ( : ) (14) in (13). No problems exist for our case. It is not surprising -the equality case in (13) corresponds to macroscopic chaotic system. The system supplied by the objection [6] is not macroscopic. It demonstrates the widely known fact that such thermodynamic concepts as the thermodynamic time arrows, the entropy increase and the measurement device concern to macroscopic chaotic systems. Both the paper [6] and the subsequent paper [7] describe not thermodynamic time arrows but, main ly, strongly fluctuating sma ll systems. No thermodynamics is possible for such small systems as three cubits. The useful outcome of these papers is equality (14). It can be used as a measure for macroscopicity of chaotic quantum systems. On the other hand, the difference between mutual information and mutual entropy can be a criterion of fluctuations value.

Conclusions
D. Jennings and T. Rudolph (like Maccone) use category "thermodynamic arrow of t ime" for non-macroscopic systems, for examp le, sma ll fluctuating quantum systems. As a result, they get objections with the second law of thermodynamics. We demonstrate that for relevant macroscopic quantum thermodynamical systems no objecti on appears.