QUANTUM THEORY
45
/
. Law I. (Generalised Principle of Exchanges.) Every process of trans
formation occurring in a system in thermodynamical equilibrium is capable
of direct reversal; and transformations in the two directions occur with
equal frequency*.
No formal proof can be offered, but a little consideration will show how
difficult it is to evade the law—to have the general balance occurring
without the particular balance. The kind of phenomenon that would
upset the balance is a cyclic series of processes. For example, an atom
might become excited, then become ionised by expelling an electron in
the excited state, and finally capture an electron at the normal level (i.e.
without returning via the excited state). This will keep the number n
constant although the individual processes are unbalanced. But that is
not the whole effect. The exciting of the atom involves absorption of a
quantum of radiation of particular frequency; its ionisation involves
absorption of another quantum; and the recapture of the electron (taken
for simplicity to have the same velocity as the expelled electron so as not
to upset the distribution of electronic velocities) involves emission of a
quantum of yet another frequency. Thus the effect of the cycle is to alter
the constitution of the radiation of the field. To preserve the equilibrium of
the radiation we must link with it another cycle occurring to the radiation
and undoing the change. The trouble is—if we provide a separate mechan
ism for changing back the radiation, how are we to prevent it from working
in the absence of the particular atoms we have considered ? The mechanism
will have the same distribution of radiation to operate on whether these
particular atoms are present or not, because we have proved (§ 29) that
the equilibrium distribution of radiation is independent of the chemical
nature of the material present. Apparently the only way of securing that
the mechanism will act when it is required and not when it is not required
is to make the special atoms play an essential part in it; this they must
do by absorbing the radiation previously emitted and liberating the
radiation previously absorbed, with corresponding changes of their own
energy. Except in very special cases this means that the linked cycle is
the exact reverse of the first, so that each of the three processes of the
cycle is now separately balanced by its opposite.
Looking at the question more generally we note that three types of
energy come under consideration—
(1) Radiant energy.
(2) Kinetic (translatory) energy of electrons and molecules.
(3) Internal energy of molecules and atoms.
on enumerations of large assemblies. Entropy is the most convenient statistic, and
the rule is that t must be measured so that dS/dt is positive. Our law asserts that
when this test fails ( dS/dt =0) all other statistical tests fail to determine a direction
of time.
* This is often referred to as the “principle of detailed balancing.”