Entropy：A concept that is not Physical Quantity

shufeng-zhang
College of Physics, Central-South University, China

Abstract

This paper demonstrates that entropy is not a physical quantity.

We define heat engine efficiency η as: η= W/W1, that is, replacing Q1 in the original definition η=W/Q1 with W1, W still
is the net work of the heat engine applied to the outside in one cycle, W1 is the work the heat engine applied to the
outside in the cycle, then, we use Stirling cycle as the element reversible cycle, if ∮dQ/T =0 is tenable, we can prove
∮dW/T =0 and ∮dE/T =0.
If the formula ∮dQ/T=0,∮dW/T=0 and ∮dE/T=0 can really define new system state variables, it should come to the
absurd result of such a definition.
In fact, during the process of obtaining entropy, ∑[(ΔQ)/T)] becoming ∫dQ/T is untenable, therefore, the formula
∮dQ/T=0, ∮dW/T=0 and ∮dE/T=0 are all untenable.
The entropy defined by Boltzmann is used to interpret “entropy” by Clausius, so, it is at the same time denied.

Keywords：physics, thermodynamics, statistical physics, entropy

PACS numbers: 05.70.-a, 05.70.ce
I. Introduction
What is entropy ? This is a controversial question debated for more than 100 years.
In 1865, Mr.Clausius, based on the result ∫ dQ / T = 0 in any reversible cycling process of arbitrary thermodynamic
system, concluded that there existed a new status quantity entropy ( remarked as S). And people believed that the
entropy difference between any two arbitrary balanced states was：
2
ΔS = S 2 − S1 = ∫ dQ / T
1
moreover, only this differential value can be used for calculation in thermodynamics. Mr. Clausius also put forward the
well-known entropy increase law accordingly.
In 1872 Boltzmann put forward the formula of absolute entropy S = k ln Ω , of which, k was Boltzmann’s constant and
Ω is the probability of thermodynamics. He held that entropy was the degree of system disorder or the measuring
symbol of consequence . This was considered as the best interpretation of entropy which is still been used by people
today.
The above conclusion is still broadly accepted and applied, and the above contents can be found in any text books of
thermodynamics and statistical physics. Entropy has been regarded as an important physical quantity and broadly used
by people though people can not ascertain what entropy is.
There exist many unsolved problems or contradictions in the above conclusion, this predicts that there must be
something wrong with entropy.

II. Entropy is not a physical quantity

2.1 Origin of Entropy

To demonstrate entropy being not a physical quantity, we first review the origin of entropy.

We define the efficiency of the heat engine as η = W / Q1 , namely taking the ratio of the net work W done outside by
the heat engine in circulation and the heat quantity Q1 absorbed by the system from outside as efficiency of the heat

( Email: [email protected] The corresponding author Tel.:86-020-13602836958 fax:86-020-87254523 )

-1-
engine, then get the Carnot cycle η = W / Q 1 = 1 − Q2
Q1 , here η has no relation with the working medium but
only has relation with the temperature of the two constant heat sources, and then in view of the above, the
thermodynamics temperature scale θ is defined as: θ 2 / θ1 = Q2 / Q1 ; when working medium of the system is ideal
gas, it can prove：

Q2 / Q1 = T2 / T1 , namely θ 2 / θ1 = T2 / T1
still use sign T to express thermodynamics temperature scale θ, namely Q2 / Q1 = T2 / T1 ⇒ Q1 / T1 + Q2 / T2 = 0 ，
here Q2 means release of heat and is in itself a negative value. according to this, use infinite Carnot cycles to

∫
approximate and replace arbitrary reversible cycles,then, get dQ / T = 0 ，therfore, people believe that dQ/T is a

complete differential and ∫ dQ / T = 0 is used to determine a system status quantity, that is entropy.
2.2 Entropy is not a physical quantity

∫
Entropy originated from dQ / T = 0 , therefore, to prove entropy is not a physical quantity, it is necessary and it is only

∫ ∫
necessary to prove that dQ / T = 0 can not define physical quantity or dQ / T = 0 , itself, is untenable.

∫
As we know that dQ / T = 0 originated from Q2 / Q1 = T2 / T1 in Carnot cycles, and this is a way to define
thermodynamics temperature scale, its existing base is the combination of heat engine efficiency definition mode and
Carnot cycles. It should be known that the heat engine efficiency formula is a definition mode and Carnot cycle,
comparing with other cycles, has only difference in form and should not occupy positions higher than other cycles, its
function to define thermodynamics temperature scale is not unique.

∫
Let us prove that dQ / T = 0 can not define physical quantity and itself is untenabal as follows.
Above all, redefining the heat engine efficiency.
Because heat engine efficiency is significant only to observers, how we define heat engine efficiency has nothing to do
with the objective process of the heat engine system, therefore we can reasonably define heat engine efficiency according
to other mode. Now we redefine heat engine efficiency as: ratio of the net work done to outside by heat engine system in
one cycle to the work done to outside by heat engine system, namely

W
η= ……… (2.2.1)
W1
That is to use W1 done to outside by heat engine system in cycle to replace Q1 heat quantity absorbed by system from
outside in original definition η = W / Q1 , because W1 done to outside by system in cycle can not totally be transferred
into net W , this is just as heat quantity absorbed by system from outside can not totally be used to do net work in
cycle–statement of Kelvin of the Second Law. It is obvious that this two kind of definitions are equal rights. We provide
another statement of the Second Law: there can not be such a heat engine by which all work done to outside can be
transferred into net work in cycle process. Obviously the formulation is equivalent to Kelvin’s formulation.
There exists a heat engine that uses certain work medium to do work W1 to outside in one cycle, and the outside does
work W2 towards the system, and the system is restored, thus get W = W1 − W2 .

From (2.2.1 ), we can determine:

-2-
 W W
η= = 1− 2 ………(2.2.2)
W1 W1
Using the reversible cycle (stirling cycle) shown in Figure 1. as element cycle, which plays the role of Carnot cycle in
∫
the process of reducing dQ / T = 0 .

P
a
``
b

d
c

V
Figure 1. TV(stirling) cycle
abcda is composed of two reversible constant volume processes bc and da and two reversible constant temperature
processes ab and cd.

Here, we define the heat sources of system energy exchange in constant temperature process as work source so as to
understand the following contents, name in short such cycles (stirling cycle) as TV cycles, and name the doing TV
cycling heat engines as TV engines.

We prove as follows: The efficiencies of all reversible engines (namely TV engines) that work between the two constant
temperature work sources are equal, and the efficiency of irreversible engine is less than that of reversible engine.

Take any of two reversible engine E and E＇, and make them to work between constant work sources θ1 and θ2, they are
θ1 and θ2 express respectively the temperatures of high
inevitably TV engines, their working medium is arbitrary; Use
temperature work source and low temperature work source, θ1＞θ2 ; here θ can take any thermodynamics temperature
scale. Let E and E＇do equal net work towards outside in one cycle, as ΔW1 and ΔW2, ΔW1=ΔW2=W, this can always be
done (similar to the situation of Carnot cycle in textbooks), use W1 and W1＇to express the work done towards outside by
E and E＇in one cycle, and W2 and W2＇to express the work done by the outside towards E and E＇in the cycle, η and η＇to
express the efficiency of E and E＇, first prove η=η＇, here we adopt contradiction proof,

Suppose η＇> η
Because E and E＇are both reversible, therefore E can be made to operate reversibly, that is to make E do work W2
towards outside, and the outside does work W1 towards E , the net work done by the outside is W=W1 -W2, W is provided
by engine E＇which operates positively and the heat quantity ΔQ=W(=ΔW1=ΔW2 ) absorbed by engine E＇in cycling is
provided by engine E,
W W
then get > ⇒ W1 > W1 '
W1 ' W1
and because W2 = W1- W
W2＇= W1＇- W
then get W2 ＞W2＇

-3-
Combine E＇and the counter-operating E into one heat engine, after they cycle together one time, the system is restored,
the only result is that the system absorbs work ΔW= W2- W2＇from the low temperature work source (heat source θ2 )
and automatically does work towards the high temperature work source ( heat source θ1) ΔW= W1- W1＇= W2 - W2＇, this
means that there is heat quantity that equals to ΔW= W2 - W2＇= W1 - W1＇automatically transmitting from the low
temperature work source θ（heat
2 source）to the high temperature work source（heat source）θ1, This directly contradicts
Clausius formulation of Second Law, that is to say that η＇> η is untenable. The same, make engine E＇operates reversibly,
it is also proved that η > η＇is untenable, thus, there inevitably occurs

η = η＇ ……… (2.2.3)
If E＇is an irreversible engine, it is not a TV engine, then E＇can not be made to operate reversibly, from here, following
will inevitably occurs
η＇≤ η ……… (2.2.4)
Because there is already a reversible engine E＇, and after this heat engine cycling one time together with another
counter-operating reversible engine E, they can make the system and the outside restore, therefore, if E＇is an irreversible
engine, that means the equality sign in η＇≤η is untenable, because, if it is η=η＇, then it is obvious that the
counter-operating E and the forward operating E＇cycling together for one time can make the system and the outside
completely restore, in that case, E＇can only be a reversible engine which contradicts with E＇being an irreversible engine,
therefore, if E＇is irreversible engine, it must inevitably have

η＇＜ η ……… (2.2.5)

Thus proving that under the definition of formula (2.2.1), the efficiencies of all reversible engines (TV engines) which
working between two constant temperature sources are equal, and that the efficiencies of irreversible engines are smaller
than that of reversible engines, this result have nothing to do with working media.
Because the efficiency of TV engines has nothing to do with the working media, therefore the thermodynamics
temperature scale (absolute thermometric scale) can be defined as

θ1 W1
= ………(2.2.6)
θ 2 W2
That is to say that the ratio of the two thermodynamic temperatures is the ratio of the work W1 done by the TV engine
that work between the two temperature work sources (heat source) to outside. and the work W2 that the outside did to the
TV engine
.
When working medium is ideal gas and the system do TV cycles, the following formula is obtained:

V2
V1
υRT 2 ln
W2 ∫ V2
PdV V1 T2
η =1− =1− =1− =1− ………(2.2.7)
W1 V1
V2 T1
∫V2
P ' dV ' υRT 1 ln
V1
Comparing (2.2.6) with （ 2.2.7), it is known that using ideal gas can get θ 2 / θ1 = T2 / T1 , namely that the
thermodynamics temperature scale defined by (2.2.6) and the thermodynamics temperature scale defined by
θ 2 / θ1 = Q2 / Q1 is of equal to each other. Due to the convention, we still use sign T to express thermodynamics
temperature scale in the following, that is:
T1 W1
= ……………(2.2.8)
T2 W2

-4-
from formula (2.2.8), we get
W1 W2
+ = 0 ……………(2.2.9)
T1 T2
As implied, W2 means the work done by the outside towards the system and is negative value. From here we can know
that when any system making TV cycles, the sum of the ratio of the work (positive or negative) exchanged by the system
with each work source (heat source) and the thermodynamics temperature scale of the work source (heat source) equals
zero.

This process is entirely the same as the process of ∫ dQ / T = 0 , use a serial element TV cycles to separate and replace
any reversible cycles of the system, please see Figure 2.

V
Figure 2 A serial TV cycle to replace any reversible cycling process

Since in two constant volume process of TV cycling, there is always ΔW=0, namely dW=0, when element process is
infinite, namely when the system exchanges work (equivalent heat quantity) with infinite work sources (heat sources),
there occurs
dW
∫ T
= 0 …………… (2.2.10)
The same, this result has nothing to do with the working medium.
dW
Obviously, to irreversible cycle, formula
T ∫
＜0 can be obtained.

We obtain the conclusion ∫ dW / T = 0 which paralleling with ∫ dQ / T = 0 , and using the first law, we can also get
formula dE = dQ + dW , and making reversible cycle of any thermodynamic systems and get

dE dQ dW
∫ ∫ T +∫ T =0
T
=

For over one hundred years, people believed that ∫ dQ / T = 0 defined a system state variable, namely entropy, if it is
right, then,
dW dE
∫ T
=0 and ∫ T
=0

-5-
would inevitably defined new system state variable, moreover, such as in the process of system reversible heat isolation,
when the system from balance state 1 reaches another balance state 2
2 dQ dW 2
∫1 T
≠
T ∫1
……………(2.2.11)
2 dQ 2 dE
∫1 T ≠ ∫1 T ……………(2.2.12)

In the reversible constant volume process, from balance state 3 reaches another balance state 4

4 dW 4 dE
∫3 T
≠ ∫ T3
……………(2.2.13)

From formula (2.2.11)~(2.2.13), we know that if ∫ dQ / T = 0 ﹑ ∫ dW / T = 0 and ∫ dE / T = 0 have defined new

state variables of the system, then they are inevitably different from each other, but their dimensions are the same, all are
J/K, there must be also that a system state variable concurrently have different numerical values, or there concurrently
exist three different system state variables with the same unit in the system, although there can concurrently exist
different system state variable with the same unit in one system, but one S has already made people unaware of what it is,
now we have to define three，this obviously should be absurd.

∫
Therefore, the conclusion is that dQ / T = 0 should be a wrong result.

The same, ∫ dW / T = 0 and ∫ dE / T = 0 should be wrong results either.

III. ∫ dQ / T = 0 is gained based on a calculus error

From the above-mentioned, we can know that ∫ dQ / T = 0 should be a wrong result., then, what is ∫ dQ / T = 0 ?
Textbooks stress that ∫ dQ / T = 0 is a result of physics, but not a conclusion of mathematics，that is to say that we can
not use mathematics to reduce out ∫ dQ / T = 0 ，but if there is really a result of ∫ dQ / T = 0 , then it inevitably defined

a system state variable, and the paper has already demonstrate the absurdity of using ∫ dQ / T = 0 to define physical

quantity, this means that there should not exist such conclusion as ∫ dQ / T = 0 .

The key point of this problem rests on the process in which ∫ dQ / T = 0 is reduced，∑[(ΔQ)/T)] becoming ∫dQ/T is
wrong .

As elementary calculus, here, when ΔQ→0 , we only can obtain that ∑(ΔQ) become∫dQ is tenable; or, if it is Δ(Q/T)
→ 0, then ∑[Δ (Q/T)] becoming ∫d (Q/T) is tenable.

But, if Q is not single-valued function of T, then ∑[(ΔQ)/T] becoming ∫dQ/T is untenable at all, just as we know,Q is not
single-valued function of T, so, ∑[(ΔQ)/T] can not become ∫dQ/T, ∫dQ/T is meaningless, dQ/T itself is meaningless.

What is wrong in textbooks is that there emerges function and calculus confusion.

∫
The conclusion is that dQ / T = 0 is neither a mathematic conclusion, nor a physical result, that is to say that there does

not exists this formula ∫ dQ / T = 0 , and there don’t exist the formula ∫ dW / T = 0 and ∫ dE / T = 0 either.

-6-
 IV. Concerning Boltzmann “entropy”

Then, what is Boltzmann entropy ?

On one hand, entropy from Boltzmann was to construe entropy from Clausius, the unit J/K from Boltzmann entropy
was also transplanted from Clausius entropy. But this paper expounds and proves entropy by Clausius does not exist,
therefore entropy defined by Boltzmann is denied accordingly.

On the other hand, in S=kInΩ, Ω is a so-called thermodynamic probability, but calculation of Ω needs to use phase-cell
division method to separate hyperspace μ, phase cell is two dimensions 2i, and i is the total free degree of molecules
within the system. Calculation of Ω, in essence, is to separate the continuous space μ and makes it produce objective
meaning. In fact, this method won’t work, no matter how much work is being done by the people, a conclusion with
objective-results won’t be able to be obtained. The primary reason is that there is not an objective standard for phase cell
division, that is to say that Ω does not possess the meaning of objective value. And again combining with Liouville
Theorem and the conclusion of this paper, we can conclude that Boltzmann entropy is a skill to display irreversibility.

V. Concerning the second law of thermodynamics

The second law of thermodynamics is formulated by several mutual-equivalent statements, although they have been
testified by endless facts, it is impossible to cover the nature of them only being phenomenological laws, just as
formulating the law of universal gravitation as “any object can not spontaneously move from low place to high place”,
the formulations of the second law are only the statements of the concrete phenomena and can not reveal the unified
nature and law that lead to these phenomena.
The second law of thermodynamics will be re-expounded by new methods.

References

Except textbooks, this paper doesn’t need references, and the content in textbooks is wellknown, so, I don’t list references,
for it is unnecessary.

-7-