School of Physics and Astronomy Junior Honours Thermodynamics

Lecture TOPIC 10
1

GJA 2013-2014

(Finn: 6.1, 6.2, 6.3, 6.4, 6.5, 6.6, 6.8, 6.9)

Synopsis: The thermodynamic potentials: internal energy U , enthalpy H , Helmholtz function F and the Gibbs function G. Maxwells relations. dG = 0 as the condition for equilibrium.

Equilibrium and the thermodynamic potentials

The second Law tells us that entropy increases in the universe as a whole. So it must increase in any isolated system. But for a system in contact with surrounds, the Second Law tell us only that the entropy of system+surrounds must increase. A system can reduce its entropy, provided the entropy of the surroundings increases by more. Main postulate of thermodynamics tells us that systems tend towards equilibrium. However, the equilibrium state of a non-isolated system depends on its boundary conditions. We can dene state functions, U,H,F and G which are useful for determining thermodynamic equilibrium under dierent external conditions, and the maximum work that can be extracted from a system when those boundary conditions are changed. They all involve state functions and are independent of the process applied to the system.

Maxwell Relations
The four Maxwells relations are general thermodynamic relations involving P , V , T and S which establish links between dierent measurable properties of substances; specic heat capacities, compressibilities etc. They are most conveniently derived from the requirement that the second derivatives of U,H,F,G with respect to their natural variables do not depend on the order of dierentiation; this is required since the potentials are state functions. dU = TdS -PdV implies that: 2 U/V S = (T /V )s = 2 U/SV = (P/S )v To recall the Maxwell relations from their derivation it can be seen that: 1. The independent (natural) variables of the potential from which each Maxwell relation is derived appear in the denominators of the relation. 2. Cross multiplication of numerators and denominators yields products of pairs of conjugate variables, S T and P V . 3. The sign can be deduced by recourse to the appropriate potential function.

le topic10.pdf

potential internal energy U enthalpy H H = U + PV Helmholtz free energy F F = U TS the Gibbs free energy G G = H TS

dierential dU = T dS P dV

Maxwell relation
T V S

natural variables S, V

P S V

dH = T dS + V dP

T P S

V S P

S, P

dF = P dV SdT

P T V

S V T

T, V

dG = V dP SdT

V T P

S P T

T, P

Enthalpy H = U + P V Internal Energy U From the central equation of thermodynamics. dU = T dS P dV suggests putting U = U (S, V ). U Then dU = U S V dS + V S dV Equate coecients to get: 1. T = 2. P =
U S V U V S

dH = dU + P dV + V dP = T dS + V dP suggests putting H = H (S, P ). H Then dH = H S P dS + P S dP Equate coecients to get: 1. T = 2. V =

H S P H P S

For changes in contact with constant P reservoir at pressure P0 , if no work other than P V work is done: H = Q (Pf = Pi = P0 , no W except PV)

If volume constant & no other work is done, for ANY changes: U = Q (V constant, no work)

for isobaric (constant P) processes. note d qR = T dS for any REVERSIBLE changes with no restriction on dP .

note d qR = T dS for any REVERSIBLE change with no restriction on dV dU = T dS (|V , no work, REV)

dH = T dS

(|P , no W except PV, REV)

Heat capacity For reversible, isochoric (constant volume) QV dUV U heat ows: CV = d dT = dT = T V S dS = T T and CV = T dT V V From partial derivatives for T and P . 2U 2U P T V S = V S = SV = S V , a Maxwells relation.

Heat capacity QP dHP CP = d dT = dT = P and CP = T dS dT = T

H T P S T P

From partial derivatives for T and V. T V P S = S P , another Maxwells relation.

Helmholtz free energy F = U T S dF = dU T dS SdT = T dS P dV T dS SdT = P dV SdT suggests putting F = F (V, T ). F Then dF = V dV + F T V dT T Equate coecients to get: 1. P = 2. S =
F V T F T V

The Gibbs free energy G = H T S dG = dH T dS SdT = dU + P dV + V dP T dS SdT = V dP SdT suggests putting G = G(P, T ). G Then dG = G P T dP + T P dT Equate coecients to get: 1. V = 2. S =
G P T G T P

Finally, from partial derivatives for P and S . P S T V = V T , another Maxwells relation.

Finally, from partial derivatives for V and S . V S T P = P T , another Maxwells relation.

Equilibrium with dierent boundary conditions free diathermal wall

@ @

(i) A system in contact with a T & P reservoir Proof that the condition for equilibrium is dG = 0 Certain processes occur in systems which are constrained to operate or evolve when they are open to the atmosphere, and so are subject to (very nearly) constant pressure and temperature at their boundaries.

system
@ @ Q 6 TP R @

reservoir

T0 P0
@ I @ adiabatic wall

Suppose that the system is initially at the temperature (T0 ) and pressure (P0 ) of the heat and pressure reservoir. If the system evolves, possibly by chemical reaction, its temperature and pressure will, in general, both vary with time. If the system cools, heat Q (assumed measurable) will ow into it from the reservoir. The entropy of the reservoir will change by S0 = Q/T0 . If the system expands, it will do work P0 V on the reservoir. Suppose that the system eventually returns to a new equilibrium state with temperature T0 and pressure P0 , but with its composition somehow changed (possibly its chemical composition). Then, in general, the values of some of its state functions will have changed, possibly the Gibbs free energy and the entropy, amongst others. The change in value of the Gibbs free energy can be found using the principle of increasing entropy, and the 1st law of thermodynamics. In what follows, symbols with the subscript 0 denote properties of the reservoir and no subscript is used for properties of the system. We examine the ensemble of the system and a suciently large volume of reservoir such that the ensemble can be considered to be thermally isolated from the rest of the universe. entropy increase S + S0 Q S + T0 T0 S Q Q T0 S First Law U = Q P0 V Q = U + P0 V

Combining the above,

0 0 0 0

U + P0 V T0 S 0 U + P0 (Vf Vi ) T0 (Sf Si ) 0

Since P0 = Pi = Pf and T0 = Ti = Tf U + (Pf Vf Pi Vi ) (Tf Sf Ti Si ) 0 U + (P V ) (T S ) 0 (U + P V T S ) 0 G 0 A system in thermal and mechanical contact with a heat and pressure reservoir is in equilibrium when the Gibbs free energy is a minimum This means that spontaneous changes in a system in contact with a heat and pressure reservoir are accompanied by a decrease in the value of its Gibbs free energy. Further if any innitesimal changes are somehow induced in a system already in an equilibrium state, the changes are reversible and dG = 0. 4

Equilibrium condition for dierent system boundaries

With reference to the boxed EQNs below, identify the natural variables for each potential. U H = U + PV F = U TS G = U TS + PV dU = T dS P dV U (S, V ) dH = T dS + V dP H (S, P ) dF = SdT P dV F (T, V ) dG = SdT + V dP G(T, P )

Be aware that if there are additional thermodynamic variables and conjugate forces more potentials can be dened eg. for dU = T dS P dV + xdX it may be useful to dene a new potential G (T, P, x) = U T S + P V xX If the boundary constraints x the natural variables for a particular potential, that potential is at an extremum at equilibrium with respect to changes of all internal degrees of freedom (eg xing P = P0 & T = T0 requires dG = 0 at equilm for all possible internal changes of composition). For U, H, F, G the extremum is a minimum. For S (U, V ) the extremum is a maximum. The proof in each case follows the same line of argument as was given on the previous page for G: a generic proof will be given in topic 11. Example Derivation of a Maxwell relation eg. Suppose we need to re-express relation might help.
S P T

in terms of other derivatives and expect a Maxwell

1. The relevant independent variables can be identied as P, T . Since the Maxwell relations just reverse the order of dierentiation, the above derivative can be related to a partial derivative with respect to T with P held constant. (We could also work with the reciprocal P and relate this to a partial derivative with respect to T with S held constant.) S
T

2. The cross multiplication rule tells us that the derivative it is related to is V T P . (At this stage it should be apparent whether the application of a Maxwell relation is a useful step in solving a particular problem or not). 3. We know that G has natural variables (T, P ) therefore look at dG = SdT + V dP : S P = 2G P T = 2G T P
S P T

V T
V T P

Thus a negative sign is required in this case:

This Maxwell relation relates a derivative that does not correspond to a common type of measurement to the isobaric thermal expansivity, = (V /V T )P , which is easily measured or can be written down directly from the equation of state if the latter is known.

Legendre Transformations - optional background material

Legendre transformation, like Fourier transformation acting on f (x) creates a new function of a transformed variable. L[f (x)] fL (f /x) without loss of information: L[L[f (x)]] = f (x) You have already used a Legendre Transformation in classical mechanics to change a Lagrangian L(x, x ) into a Hamiltonian, H (x, p) where p = (L/ x ). In thermodynamics the transformation can be used to transform the internal energy expressed as function of S and V to give dierent potentials that depend on (i) (U/S )V = T instead of S (ii) (U/V )S = P instead of V (iii) (U/x)S,V = X instead of x (iv) Any combination of these. The transformations are achieved by writing fL = f x.(f /x) f X.x and replacing x by x[(f /x)] on the RHS to give the new function fL [(f /x)]. Geometrically for a given value of (f /x) the function fL [(f /x)] gives the value at x=0 of the tangent of f(x) taken at the point, x, at which (f /x) has the specied value. It is necessary to assume that the internal energy is an increasing (convex) function of entropy, volume, etc, for all this to work. The new potentials generated are useful as a short cut for deriving dierent Maxwell relations and because they are found to be minimized when equilibrium is achieved under dierent external conditions. Knowledge about the formal properties of Legendre transforms is not necessary for understanding G,H,F in thermodynamics; we can derive all the useful properties of these potentials more easily on a case by case basis.