Chemical Potential

Let us now turn to look at the situation where two systems can exchange particles. What is the condition of equilibrium in this case? Let two systems S1 and S2 be in thermal contact and in diffusive contact. By diffusive contact we mean that particles can flow from Sl to S2 and vice versa. We maintain a constant by placing both systems in thermal contact with a large reservoir, R. We saw earlier that for a system in thermal equilibrium with a reservoir, the Helmholtz free energy F will be a minimum. Thus for the two systems in contact with R, we must have that F1 + F2 must be a minimum. The change in F = F1 + F2 as the number of particles varies is given by
F F dF = 1 dN 1 + 2 dN 2 N N 1 2

But if we treat Sl + S2 as a closed system, we have that N1 + N2 = N is a constant. Thus dN1 = -dN2 and the requirement that dF = 0 becomes
F1 F2 N = N 1 2

(10.1)

We define the chemical potential of a system to be

=
F N

(10.2)

Then the condition of equilibrium reduces to 1=

2

(10.3)

Strictly speaking, the true definition of is in terms of the difference = F( ,V,N) - F( ,V,N-1) since particles are a discrete set. However we are usually dealing with such large quantities of particles that we are able to treat as a continuous function. Example: The chemical potential of an ideal gas is found from F = - ln(Zn) with Zn = Z1N / N! and Z1 = nq V. So F = - [N ln(nqV) - N ln(N) + N]

and thus
=
F N = ln ( nqV ) ln ( N )

]
(10.4)

n V = ln q N n = ln n q

Connection to Potential Energy

To better understand what a chemical potential is, recall that the Helmhotz free energy is defined to be F=U- where U is the total energy of the system. If the system is subjected to an external force, this will give rise to either a potential energy or dissipative work. In either case, the Helmholtz free energy can be written as F = Uint + NUext - where the N comes from the fact that the external energy acts on each particle in the system. Now consider two systems S1 and S2 which are at the same temperature and are capable of exchanging particles, but which are not yet in diffusive equilibrium. We can assume that initially 2 > 1, and we can denote the initial, non-equilibrium difference in chemical potential as i = 2i - 1i. Now let a difference in potential energy be established between the systems, such that the potential energy of each particle in S1 is raised exactly by i. Such a difference could come from gravitational potential energy, or electrodynamical potential energy, for example. By establishing a potential energy step, we are adding a potential energy N i to the energy of each state in Sl. Thus the final chemical potential energy of S1 is
1 f = 1i +
= 1i + 2i 1i = 2i = 2 f

(10.5)

since we didn't change the energy in S2. But now lf = 2f, so the systems are in diffusive equilibrium. Thus, we see that the chemical potential is equivalent to a true potential energy in that the difference in chemical potential between two systems is

equal to the potential barrier that will bring the two systems into diffusive equilibrium. Since the chemical potential has the same properties as potential energy, we see that we cannot define an absolute chemical potential. Only differences in chemical potential have any physical meanings. From the argument above, we see that if an external potential step exists, the total chemical potential of a system is the sum
tot

int

ext

where ext is the potential energy per particle in the external potential, and int is the chemical potential that would be present if the external potential were zero. Finally, since we can only physically discuss differences in chemical potential, the physical condition of diffusive equilibrium becomes
ext

int

(10.6)

Example: What is the condition for diffusive equilibrium of a column of ideal gas at temperature ?

Let n(0) be the number of particles in the bottom chamber, and n(h) be the number of particles in the top chamber. We want (0) = (h), where we must have = ext + int. Then from (10.4), we have that

ln n( 0 ) ln nq + 0 = ln n( h ) ln nq + mgh
n( h ) = n ( 0 ) e

mgh

(10.7)

Once we have the number of particles at a height h, we can determine the pressure by recalling that the pressure of an ideal gas is proportional to the concentration of particles, p = n Thus the pressure of an ideal gas at height h is

p ( h ) = p( 0 ) e
Example:

mgh

(10.8)

What is the chemical potential of an ideal gas in a magnetic field B?

In a magnetic field we can view the gas as being made up of two types, or species, of particles; those with spin up, and those with spin down. If the spin is up, the potential energy is -mB, while the potential energy of the spin down particles is mB. For the container out of the magnetic field the chemical potentials are
n ( 0) ( 0 ) = ln n q n ( 0) ( 0 ) = ln n q

while those in the magnetic field have

n ( B ) = ln mB n q n ( B ) = ln + mB n q

we also have that

n( 0 ) = n ( 0 ) + n ( 0 ) = n ( 0 ) = n ( 0 ) = n( 0 ) 2

and
n( B ) = n ( B ) + n ( B )

Using the condition of diffusive equilibrium, we get

( B ) = ( 0)
= n = 1 n( 0 ) e 2
and
mB

( B ) = ( 0)
= n = 1 n( 0 ) e 2
so n(B) = n(0) cosh(mB/ ) Example: For a lead acid battery, the relationship between chemical potentials and potential steps is even more apparent. In a normal lead acid battery, the negative electrode consists of metallic lead, Pb, and the positive electrode is a layer of lead oxide, Pb02, on a Pb substrate. The electrodes are immersed in diluted sulfuric acid, H2SO4, which is partially ionized into H+ ions and S04-- ions. In the discharge process, both the metallic Pb of the negative electrode and the Pb02 of the positive electrode are converted to lead sulfate, PbS04, via the two reactions Negative electrode:
Pb + SO -4- PbSO 4 + 2e
m B

Positive electrode:
PbO 2 + 2H + + H 2SO 4 + 2e PbSO 4 + 2H 2O

Thus, the negative electrode acts as a sink for the S04-- ions, and the positive electrode acts as a sink for the H+ ions. If the battery terminals are not connected, electrons are depleted from the positive electrode and accumulate in the negative electrode, thereby charging them both. As a result, electrochemical potential steps develop at the electrode-electrolyte interfaces, steps of exactly the correct magnitude to equalize the chemical potential steps and stop the diffusion of ions, which stops the chemical reactions from proceeding further. If an external current is permitted to flow, the reactions resume. If we denote by V- and V+ the differences in electrostatic potential of the negative and positive electrodes relative to the common electrolyte, diffusion of the sulfate ions will stop when (remembering that the sulfate carries two negative charges)

-2e V- = (SO4--) Similarly, diffusion of the H+ ions will stop when e V+ = (H+)

The electrostatic potential as a function of position is seen to be

The two potentials, V- and V+, are called half-cell potentials, and have magnitudes of V- = -0.4 V and V+ = 1.6 V. Thus, the total electrostatic potential that develops across the cell in order to stop the diffusion reaction is V = V+ - V- = 2 V This is the open-circuit voltage, or EMF, of the battery.

Chemical Potential and Entropy

Recall that we defined the chemical potential as
=
F N ,V

Substituting in the definition of F, we get

=
U N ,V N ,V

We can replace the right hand side of this by considering the following argument. Let the entropy be a function of the independent variables U, V and N, = (U,V,N).

Then the differential of is given by

d = dU + dN + dV U V , N N U ,V V U , N

Let dV = 0 for the process under consideration, and require the changes in d , dU and dN be inter-related in such a way that d = 0. Thus we get that

( d )

= ( dU ) + ( dN ) U N N U

or, after dividing by (dN) and using the definition of 1/ ,

U = + N ,V N ,V N U ,V

(10.9)

Returning to , we can solve (10.9) for

( U N )

,V

N ,V

to yield

(U , V , N ) =

N U ,V

(10.10)

What is the difference between this expression and the one involving F? Recall that F is a function of the variables , V and N, so = ( ,V,N) in this form. In (10.10), = (U,V,N), i.e. it has a dependence on a different set of variables. A third formulation that we will not prove in class is
(U , V , N ) =
U N ,V

(10.11)

We can now derive a more general form of the thermodynamic identity we encountered earlier. Once again consider an infinitesimal change in entropy, where now the entropy depends on U, V and N
d = dU + dV + dN U V , N V U , N N U ,V

but, as before, we recall that

the change in entropy becomes

( U ) = 1 , ( V ) = p
d = 1 p dU + dV dN

and

) = . So N

or
dU =d pdV + dN

(10.12)