MIT OpenCourseWare
http://ocw.mit.edu
5.60 Thermodynamics & Kinetics
Spring 2008
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.
5.60 Spring 2008 Lecture #11 page 1
Fundamental Equations, Absolute Entropy,
and The Third Law
• Fundamental Equations relate functions of state to each other using
1st and 2nd Laws
1st law with expansion work: dU = đq - pextdV
need to express đq in
terms of state variables
because đq is path dependent
Use 2nd law: đqrev = TdS
For a reversible process pext = p and đq = đqrev =TdS
So…… ** dU = TdS – pdV **
This fundamental equation only contains state variables
Even though this equation was demonstrated for a reversible
process, the equation is always correct and valid for a closed (no
mass transfer) system, even in the presence of an irreversible
process. This is because U, T, S, p, and V are all functions of
state and independent of path.
AND The “best” or “natural” variables for U are S and V,
** U(S,V) **
5.60 Spring 2008 Lecture #11 page 2
** U(S,V) **
⎛ ∂U ⎞ ⎛ ∂U ⎞
From dU = TdS – pdV ⇒ ** ⎜ ⎟ =T ; ⎜ ⎟ = −p **
⎝ ∂S ⎠V ⎝ ∂V ⎠S
We can write similar equations for enthalpy
H = U + pV ⇒ dH = dU + d(pV) = dU + pdV + Vdp
inserting dU = TdS – pdV
⇒ ** dH = TdS + Vdp **
The natural variables for H are then S and p
** H(S,p) **
⎛ ∂H ⎞ ⎛ ∂H ⎞
From dH = TdS + Vdp ⇒ ** ⎜ ⎟ =T ; ⎜ ⎟ = V **
⎝ ∂S ⎠ p ⎝ ∂p ⎠S
_______________
We can use these equations to find how S depends on T.
From dU = TdS – pdV ⇒ ⎛ ∂S ⎞ = 1 ⎛ ∂U ⎞ = CV
⎜ ⎟ ⎜ ⎟
⎝ ∂T ⎠V T ⎝ ∂T ⎠V T
From dH = TdS + Vdp ⇒ ⎛ ∂S ⎞ = 1 ⎛ ∂H ⎞ = Cp
⎜ ⎟ ⎜ ⎟
⎝ ∂T ⎠ p T ⎝ ∂T ⎠ p T
5.60 Spring 2008 Lecture #11 page 3
• Absolute Entropies
Absolute entropy of an ideal gas
dU + pdV
From dU = TdS – pdV ⇒ dS =
T
pdV
At constant T, dU=0 ⇒ dST =
T
nRdV
For an ideal gas, pV = nRT ⇒ dST =
V
At constant T d(pV) = d(nRT) = 0 ⇒ pdV = -Vdp
nRdp
So dST = −
p
For an arbitrary pressure p,
p nRdp ⎛ p⎞
S(p, T) = S(p o , T) − ∫po = S(p o , T) − nRln⎜⎜ o ⎟⎟
p ⎝p ⎠
where po is some reference pressure which we set at 1 bar.
⇒ S(p,T) = So(T) – nR lnp (p in bar)
S (p, T) = S o (T) − Rln p (p in bar)
But to finish, we still need S o (T) !
Suppose we had S o (0K) (standard molar entropy at 0 Kelvin)
∂S ⎞ Cp
Then using ⎛⎜ ⎟ = we should be able to get S o (T)
⎝ ∂T ⎠ p T
5.60 Spring 2008 Lecture #11 page 4
Consider the following sequence of processes for the substance A:
A(s,0K,1bar) = A(s,Tm,1bar) = A(ℓ,Tm,1bar) = A(ℓ,Tb,1bar)
= A(g,Tb,1bar) = A(g,T,1bar)
Tm Cp (s)dT ΔHfus Tb Cp (A)dT ΔHvap T Cp (g)dT
S (T,1bar) = S o (OK) + ∫ + +∫ + +∫
0 T Tm Tm T Tb Tb T
So(T) CpdT
ΔS = ∫
T
ΔH vap
Liquid boils, ΔS =
T
ΔH fus
Solid melts, ΔS =
T
0
T
Since ΔS0 is positive for each of these processes, the entropy
must have its smallest possible value at 0 K. If we take S o (0K) =
zero for every pure substance in its crystalline solid state, then
we could calculate the entropy at any other temperature.
This leads us to the Third Law !!!
THIRD LAW:
First expressed as Nernst's Heat Theorem:
• Nernst (1905): As T → 0 K , ΔS → 0 for all isothermal
processes in condensed phases
5.60 Spring 2008 Lecture #11 page 5
More general and useful formulation by M. Planck:
• Planck (1911): As T → 0 K , S → 0 for every chemically
homogeneous substance in a perfect crystalline state
Justification:
c It works!
d Statistical mechanics (5.62) allows us to calculate the
entropy and indeed predicts S o (0K) = 0.
This leads to the following interesting corollary:
It is impossible to decrease the temperature of any system to
T = 0 K in a finite number of steps
How can we rationalize this statement?
Recall the fundamental equation, dU = T dS – p dV
dU = Cv dT For 1 mole of ideal gas, P = RT/V
so Cv dT = T dS – (RT/V) dV
dS = Cv d (ln T) + R d (ln V)
For a spontaneous adiabatic process which takes the system
from T1 to a lower temperature T2,
ΔS = Cv ln (T2/T1) + R ln (V2/V1) ≥ 0
but if T2 = 0, Cv ln (T2/T1) equals minus infinity !
Therefore R ln (V2/V1) must be greater than plus infinity, which
is impossible. Therefore no actual process can get you to T2 = 0 K.
But you can get very very close!
5.60 Spring 2008 Lecture #11 page 6
In Prof. W. Ketterle's experiments on "Bose Einstein
Condensates" (MIT Nobel Prize), atoms are cooled to nanoKelvin
temperatures (T = 10-9 K) … but not to 0 K !
Another consequence of the Third Law is that
It is impossible to have T=0K.
How can we rationalize the alternate statement?
Consider the calculation of S starting at T=0K
T Cp (s)dT
S(s, T,1bar) = ∫
0 T
to prevent a singularity at T=0, Cp → 0 as T → 0 K
in fact, experimentally Cp = γT + AT 3 + ...
That is, the heat capacity of a pure substance goes to zero as T goes
to zero Kelvin and this is experimentally observed.
Combining the above with dT = đqp/Cp , at T=0 any infinitesimally
small amount of heat would result in a finite temperature rise.
In other words, because Cp → 0 as T → 0 K, the heat đqp needed to
achieve a temperature rise dT, (đqp=CpdT) also goes to zero at 0 K. If
you somehow manage to make it to 0 K, you will not be able to maintain
that temperature because any stray heat from a warmer object
nearby will raise the temperature above zero, unless you have perfect
thermal insulation, which is impossible.
5.60 Spring 2008 Lecture #11 page 7
• Some apparent violations of the third law (but which are not !)
Any disorder at T = 0 K gives rise to S > 0
• For example in mixed crystals
ΔSmix = −nR[XA ln XA + XB ln XB ] > 0 Always !!! Even at T=0K
But a mixed crystal is not a pure substance, so the third law
is not violated.
• Any impurity or defect in a crystal also causes S > 0 at
0K
• Any orientational or conformational degeneracies such is
in a molecular crystal causes S > 0 at 0 K, for example in
a carbon monoxide crystal, two orientations are possible:
C O C O C O CO C O CO C O
C O C O C O CO C O O C C O
C O C O C O O C C O CO C O
C O C O C O CO C O CO C O
