Sie sind auf Seite 1von 7

Homework 19 (Due Wednesday, November 13th)

1. Consider monatomic particles with hard core repulsion. (a) Given that all of the energy of the
particles is kinetic energy, show that the pressure may be obtained from
( )
∂S
P =T
∂V T,N

(b) Find the pressure for a gas of such particles by using the micro-canonical ensemble with the
simple construction that the volume of the system may be partitioned into M = V /v0 boxes, where v0
is the volume occupied by a single particle (M ≫ N ). Show that your result may be written as

N kT
P =
V − bN

and determine the constant b in terms of v0 .


(c) Find the heat capacities Cv and Cp .

2. Consider an equilibrium classical fluid of N hard rods confined to move on a line of length L; the
length of each rod is ℓ. The number density is ρ = N/L and ρCP = 1/ℓ is the close packed number
density. Assume that the system is periodic so x = 0 and x = L are the same point.1
(a) Define the pair distribution function g(∆x) as the average density of rods at position ∆x given
that a (tagged) rod is at the origin, normalized by the mean density. Note that ∆x can be positive or
negative and that g → 1 for ∆x ≫ ℓ. Make a hand-drawn sketch showing the high density (ρ ≈ ρCP )
and low density (ρ ≈ 0) limits of g(∆x).
(b) For ρ → ρCP , compute the value of the integral
∫ αℓ
dx g(x)
0

where α is a positive constant.


(c) As a function of N , ρ, and T , determine the average kinetic energy of a rod.
(d) Assume that the pressure may be written as

ρkT
P =
1 − bρ

where b is independent of density. Relate b to the second virial coefficient.


(e) Explain why the Helmloltz free energy for this system may be written as,
[ ∫ L ∫ L ]
1
A = −kT ln dx1 . . . dxN e−βV(x1 ,...,xN )
N !λN 0 0
1 Imagine this one-dimensional system to be like a closed loop.

1
and why the pressure may be written as
( )
∂A
P =−
∂L T,N

(f) Use the results from part (e) to get an expression for pressure in the form given in part (d).

3. Using the approximation that the pair correlation function is g(r) ≈ e−βυ(r) find the second virial
coefficient for:
(a) Hard core potential: 
 ∞ r<r
min
υ(r) =
 0 r≥r
min

(b) Square-well potential: 



 ∞


r < rmin
υ(r) = −ϵ rmin < r < rwell



 0 r ≥ rwell

4. (a) Find an integral expression for the second virial coefficient, B2 (T ), for the Lennard–Jones
potential, [( ) ]
σ 12 ( σ )6
u(r) = 4ϵ −
r r
and show that it may be written as,
∫ ∞ { [ ] }
4
B2 (T ∗ ) = −3B2HC x2 exp − ∗ (x−12 − x−6 ) − 1 dx
0 T

where T ∗ = kT /ϵ and B2HC is the second virial coefficient for hard core exclusion of radius rmin = σ.
(b) Evaluating the integral numerically and produce a graph of B2 (T ) for nitrogen (ϵ/k = 95 K,
σ = 0.375 nm) in the range of temperatures from 50 K to 400 K. From your graph, estimate the Boyle
temperature, TB , defined as B2 (TB ) = 0.

2
Solutions
1. (a) In general, ( ) ( ) ( )
∂A ∂U ∂S
P =− =− +T
∂V T,N ∂V T,N ∂V T,N

Yet since all the energy is kinetic, by equipartition theorem U (T, V, N ) = 32 kN T so (∂U/∂V )T,N = 0.
(b) In the micro-canonical ensemble we need to find Γ(E, V, N ), the number of possible states for
the particles for fixed E = U , V , and N . This number may be factored as Γ = ΓE (E, N )ΓV (V, N )
since the number of microstates for a given energy is independent of the volume (since all the energy
is kinetic) and vice versa. You can also think of this as the product of the number of micro-states in
momentum space times the number of micro-states in position space.
The number of microstates for a given volume is the number of combinations for placing the N
particles into the M available boxes (i.e., choose N occupied boxes out of M total) so

M!
ΓV =
N !(M − N )!

We won’t need ΓE since we’re after the pressure, for which


( ) ( )
∂S ∂
P = T =T [k ln(ΓE (E, N )ΓV (V, N ))]
∂V T,N ∂V T,N
( )

= kT [ln(ΓE (E, N )) + ln(ΓV (V, N ))]
∂V T,N
( ) ( )( )
∂ dM ∂
= kT ln(ΓV (V, N )) = kT ln(ΓV (V, N ))
∂V T,N dV ∂M N

As always, we’ll use Sterling’s approximation ln x! ≈ x ln x − x so


( ) ( )
∂ ∂
ln(ΓV (V, N )) = [M ln M − N ln N − (M − N ) ln(M − N )]
∂M N ∂M N
( )
M ( ) N N2
= ln = ln [1 − N/M ]−1 = − ln (1 − N/M ) ≈ +
M −N M 2M 2

since (1 + x)a ≈ 1 + ax and ln(1 + x) ≈ x − x2 /2 for |x| ≪ 1.


We now put it all together to get
( )( ) ( )( ) ( )
dM ∂(ln ΓV ) 1 N N2 N kT N v0 /2
P = kT = kT + = 1 +
dV ∂M N v0 M 2M 2 V V

To put this into the desired form, we rewrite this as,


( )−1
N kT N v0 /2 N kT
P ≈ 1− =
V V V − N v0 /2

so b = 12 v0 .
(c) Since U (T, V, N ) = 32 N kT , ( )
∂E 3
Cv = = Nk
∂T V,N 2

3
For Cp we use the general thermodynamic result,
( ) [( ) ]2
∂P ∂V
Cp = Cv − T
∂V T,N ∂T P,N

Evaluating
( )
∂P N kT P2
= − = −
∂V T,N (V − bN )2 N kT
( )
∂V Nk
=
∂T p,N P
so ( )[ ]2
3 P2 Nk 5
Cp = Nk − T − = Nk
2 N kT P 2
Notice that we get the same result as with the ideal gas.

2. (a) Your sketch should look something like Figure 1. In general g(∆x) = 0 for |∆x| < ℓ due to the
hard rod exclusion. At high densities, the linear pair correlation function has peaks at ∆x = ℓ, 2ℓ, . . .
with decaying magnitude. At low densities g(∆x) ≈ 1 for |∆x| > ℓ, that is, particles are equally likely
to be found anywhere except overlapping. For comparison, see Figures 10.5 and 10.6 in Pathria and
Beale.

Figure 1: Second virial coefficient for nitrogen.

4
(b) Near close packing the particles touch so the linear pair correlation function is

g(∆x) ≈ δ(|∆x| − ℓ) + δ(|∆x| − 2ℓ) + . . .

where δ(z) is the Dirac delta function. The integral equals ⌊αℓ⌋ where ⌊x⌋ is the floor of x (the nearest
integer when rounding down).
(c) By equipartition, the average kinetic energy of a particle is 21 kT and the total energy equals the
total kinetic energy, 12 N kT since the potential is hard core.
(d) In the limit of low density, (1 − bρ)−1 ≈ 1 + bρ so from the definition of the virial expansion,

P = ρkT (1 + ρB2 + . . .)

so we identify that b ≈ B2 . The second virial coefficient is evaluated to be


∫ ∫
1 L/2 1 ℓ
B2 = − [g(x) − 1] dx = − [−1] dx = ℓ
2 −L/2 2 −ℓ
Notice that B2 is independent of temperature.
(e) For any one-dimensional system the classical canonical partition function may be written as
∫ ∫
1 1 IN
QN = N
dx1 . . . dxN exp[−βV(x1 , . . . , xN )] =
N ! λT N !λN
T
where the formulation is the same as presented in the notes for 3-dimensional systems, simply with
positions and momenta in only the x direction. In general, we have that A = −kT ln QN , which is
the first desired result. For the second we recognize that in one-dimensional systems we just have L in
place of V for all thermodynamic expressions (e.g., dU = T dS − P dL). Note that for one-dimensional
systems the “pressure” is actually just the force (since dW = P dL is the work) but we’ll call it the
pressure anyway.
(f) For hard core exclusion, the integrand in IN is one if none of the particles overlap and zero if
any of them do. So what we have to compute is the total volume minus the occupied volume. Simplest
way to do this is to consider that the first particle can be anywhere between x1 = 0 and L; the second
particle can be anywhere except where the first particle is located, which means that it’s available
“volume” is L − 2ℓ.2 The volume for the third particle is L − 4ℓ and so forth with the volume for the
N th particle being L − 2(N − 1)ℓ. From this,

N −1
1 1
QN = [L × (L − 2ℓ) × . . . × (L − 2(N − 1)ℓ)] = (L − 2iℓ)
N !λN
T N !λN
T i=0

Note that this is the same argument used in the lecture notes for the 3D hard core exclusion.
Finally, to obtain the equation of state in the low density approximation,
( ) ( )
P ∂ ∂
= − (βA) = ln QN
kT ∂L T,N ∂L T,N
( ) ∑
N −1 ∑
N −1 N∑−1

= ln(L − 2iℓ) = (L − 2iℓ)−1 ≈ L−1 (1 + 2iℓ/L)
∂L T,N i=0 i=0 i=0
N −1 [ ]
N 2ℓ ∑ N 2ℓ N (N − 1) N ℓN
= + 2 i= + 2 ≈ 1+ = ρ[1 + ℓρ]
L L i=0 L L 2 L L
2 Note that if N = 2 and L = 2ℓ then the available volume is a zero (i.e., position of the second particle is uniquely
determined by position of the first).

5
which agrees with the result in part (d).

3. We’ll simply do part (b) since the solution for part (a) is then obtained for ϵ = 0. From the notes
we have,
∫ [ ]
1
B2 (T ) = − d⃗r e−βυ(r) − 1
2
{∫ rmin ∫ rwell }
1 [ ]
= − (4π) dr r2 (−1) + dr r2 eβϵ − 1
2 0 rmin
2π [ ]
= (rmin )3 eβϵ + (rwell )3 (1 − eβϵ )
3
For part (a), the hard core model, we have
2π 1
B2HC (T ) = (rmin )3 = v0
3 2
For part (b) we may rewrite the result as,
[ ] [ ( )]
va ( ) ϵ va
B2 (T ) = B2 (T ) e −
HC βϵ
1−e βϵ
≈ B2 (T ) 1 +
HC
1−
v0 kT v0

4. (a) The second virial coefficient is computed from the intermolecular potential as
∫ ∫ ∞
1 1
B2 (T ) = − dr e−βu(r) − 1 = − (4π) dr r2 e−βu(r) − 1
2 2 0

Note that with the change of variable x = r/σ,


4ϵ −12
βu(r) ⇒ (x − x−6 )
kT
so ∫ ∫ ∞ ( )
1 4ϵ −12
B2 (T ) = − dr e−βu(r) − 1 = −2πσ 3 dx x2 exp (x − x−6 ) − 1
2 0 kT
Finally, since the second virial coefficient for hard core exclusion is B2HC = 2πσ 3 /3 (which is half the
excluded volume) we get the desired result.
(b) The MATLAB program below may be used to solve this part; the graph is shown in Fig. 2. Note
that the Boyle temperature obtained (324 K) is close to the accepted value of 335 K (E. Guggenheim,
Thermodynamics, 1957).

% Program to plot second virial coefficient


clear all;
sigma = 0.375e-9; % Lennard-Jones diameter for nitrogen
eps_over_k = 95; % Lennard-Jones "boiling" temperature
nPlot = 100;
T = linspace(50,400,nPlot); % Temperatures to plot
Tstar = T/eps_over_k; % Dimensionless temperature
xInfty = 20; % Graphing the integrand shows 20 is big enough

6
for iPlot = 1:nPlot
intB2(iPlot) = quad(@(x) x.^2.*(exp(-(4/Tstar(iPlot)).* ...
(x.^(-12) - x.^(-6)))-1),0,xInfty);
end
B2HC = 2*pi*sigma^2/3; % Second virial coeff for hard core exclusion
B2 = -3*B2HC*intB2; % Second virial coeff for Lennard-Jones
for iPlot = 2:nPlot
if( B2(iPlot)*B2(iPlot-1) < 0 ) % Find the Boyle temperature
BoyleT = 0.5*( T(iPlot) + T(iPlot-1) ); % by looking for where B2
end % changes sign from - to +
end
plot(T,B2,’-’);
xlabel(’Temperature’); ylabel(’Second viral coeff.’);
title(sprintf(’Estimated Boyle temperature = %g’,BoyleT));

−18 Estimated Boyle temperature = 323.99


x 10
1

0.5

0
Second viral coeff.

−0.5

−1

−1.5

−2

−2.5
50 100 150 200 250 300 350 400
Temperature

Figure 2: Second virial coefficient for nitrogen.

Das könnte Ihnen auch gefallen