Statistical The Interpretation of Entropy
Physical meaning of entropy Microstates and macrostates Statistical interpretation of entropy and Boltzmann equation Configurational entropy and thermal entropy Calculation of the equilibrium vacancy concentration
Reading: Chapter 4 of Gaskell Optional reading: Chapter 1.5.8 of Porter and Easterling
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
What is the physical meaning of entropy?
Entropy is introduced in phenomenological thermodynamics based on the analysis of possible and impossible processes. We know that heat flows from a hot region to a cold region of a system and that work can be irreversibly transferred into heat. To describe the observations, the entropy, and the 2nd law stating that entropy is increasing in an isolated system, have been introduced. The problem with phenomenological thermodynamics is that it only tells us how to describe the empirical observations, but does not tell us why the 2nd law works and what is the physical interpretation of entropy.
In statistical thermodynamics entropy is defined as a measure of randomness or disorder. Intuitive consideration: In a crystal atoms are vibrating about their regularly arranged lattice sites, in a liquid atomic arrangement is more random Sliquid > Ssolid. Atomic disorder in gaseous state is greater than in a liquid state - Sgas > Sliquid.
S
L
G
Does this agrees with phenomenological thermodynamics? Melting at constant pressure requires absorption of the latent heat of melting, q = Hm, therefore Sm = Hm /Tm - the increase in the entropy upon melting correlates with the increase in disorder.
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
How to quantify "disorder"? Microstates and Macrostates A macroscopic state of a system can be described in terms of a few macroscopic parameters, e.g. P, T, V. The system can be also described in terms of microstates, e.g. for a system of N particles we can specify coordinates and velocities of all atoms. The 2nd law can be stated as follows: The equilibrium state of an isolated system is the one in which the number of possible microscopic states is the largest. Example for making it intuitive (rolling dice)
Macrostate the total of the dice. Each die have 6 microstates, the system of 2 dices has 6 6=36 microstates, a system of N dice has 6N microstates. For two dice there are 6 ways/microstates to get macrostate 7, but only one microstate that correspond to 2 or 12. The most likely macrostate is 7. For a big number N of dice, the macrostate for which the number of possible microstates is a maximum is 3.5 N If you shake a large bag of dice and roll them it is likely that you get the total close to 3.5 N for which the number of ways to make it from individual dice is maximum. An isolated thermodynamic system is similar thermal fluctuations do the shaking, the macrostate corresponds to the largest number of microstates. Actually, the system of dice is closer to a quantum system with discrete states. In the classical case the states form a continuum and we have to replace the sum over states by integrals over phase space.
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Statistical interpretation of entropy If we combine two systems, the number of microstates multiply (remember 6 6=36 for two dice). At the same time we know that entropy is an extensive quantity, SA+B = SA + SB, and if we want to relate enthalpy to the number of microstates, we have to make sure that this equation is satisfied. If we take logarithm of the number of microstates, the logarithms adds when we put systems together. The quantity maximized by the second law can be defined then by equation written on Ludwig Boltzmann's tombstone in Vienna:
S = kB ln
where is the number of microstates, k is the Boltzmann constant (it is the same constant that relates kinetic energy to temperature, but it was first introduced in this equation), and S is the entropy. The 2nd law can be restated again: An isolated system tends toward an equilibrium macrostate with maximum entropy, because then the number of microstates is the largest. The entropy is related to the number of ways the microstate can rearrange itself without affecting the macrostate.
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Configurational entropy and thermal entropy Thermal entropy In the example on the heat transfer, the number of microstates can be thought as the number of ways in which the thermal energy can be divided between the atoms. The thermal entropy of a material Increases as the temperature increases Decreases as the cohesive energy increases The thermal entropy plays important role in many polymorphic transitions. With increasing T, the polymorphic transition is from a phase with lower entropy to the one with higher entropy. Configurational entropy Entropy can be also considered in terms of the number of ways in which particles themselves can be distributed in space. Mixing of elements in two crystals placed in physical contact or gases in two containers (mass transport) leads to the increase of Sconf and is similar to the heat transfer case when Sth is increasing.
T1 < T2
Sth
Sconf
T1 = T2
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Heat flow and production of entropy Phenomenological thermodynamics: transfer of energy from hot to cold is an irreversible process that leads to the production of entropy. Consideration of probabilities of microstates can lead to the same conclusion. TA, UA,
A
TB, UB,
B
When the thermal contact is made between the two systems, the number of microstates, A B is not in the maximum and heat starts to flow, increasing the value of A B. The heat flows until the increase in A caused by the increase in UA is greater than the decrease in B caused by the decrease in UB. The heat flow stops when A B reaches its maximum value: ln A B = 0 If we only have a heat exchange (no work) q = qA = - qB
ln
A
qA k BTA
B
ln
B
qB k BTB
ln
A
ln
A
ln
q
B
kB
1 TA
1 TB
ln A B = 0 only when TA = TB reversible heat transfer is only possible after temperatures are equal
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Configurational - entropy equilibrium vacancy concentration The configurational entropy of a crystal refers to the distinguishable ways the atoms can be arranged on the lattice sites. For monoatomic crystals the only origin of configurational entropy is the presence of crystal defects such as vacancies lattice sites without atoms. The removal of an atom from its lattice site breaks bonds and increases the internal energy of the material by vf. Increases the randomness of the atomic configuration, and, thus, Sconf. Let's first calculate the configurational entropy: Consider a crystal lattice with N lattice sites and N-1 atoms (one vacancy). Assume that N is large enough so that we can neglect the surface in our consideration and assume that all lattice sites are equivalent. In the system of N lattice sites we can have a vacancy in any of the N sites N different configurations are possible and S1conf = kB lnN. If there are two vacancies, for each location of the first vacation we have N-1 locations of the second one. Therefore, S2conf = kB ln [ N (N-1)] >> S1conf. ( comes from the fact that vacancies are identical and the states where one vacancy is at the site i and the second one at j, are identical to the state (j,i) ). S3conf = kB ln [ 1/6 N (N-1)(N-2)] >> S2conf, ...
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Equilibrium vacancy concentration (2) The number of distinct configurations for n vacancies is
1 N N 1 N 2 ... N n 1 n!
n Sconf
N! n! N n !
k B ln
N! n! N n !
The equilibrium concentration of vacancies is the one that corresponds to the minimum of free energy: A = U TS Neglecting a small change in the thermal entropy due to the change in the vibrational frequencies of atoms around a vacancy (formation entropy), and ignoring the vacancy-vacancy interactions, we can write:
A
n
n
f v
N! k B T ln n! N n !
f v
If N is large enough, introduction of the first vacancy will reduce the free energy for any finite T: A1
k B T ln N
0
The equilibrium concentration of vacancies can be found from:
An n
0
n n eq
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Equilibrium vacancy concentration (3) Let's first consider
n Sconf n
using Stirling formula for big
numbers: ln N! N ln N
n Sconf n
N
kB
n
ln
N! n! N n !
kB
n
ln N! ln n! ln N n !
kB kB
n n
N ln N
N n ln n n
N n ln N n
N n
N ln N n ln n
N n ln N n
kB
ln n 1 ln N n
N n N n
1
k B ln
N n n
k B ln
N 1 n
k B ln
N n
k B ln
n N
For the equilibrium concentration of vacancies:
An n
n n eq
n
n
f v
TS
n conf n n eq
f v
k B T ln
n eq N
0
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Equilibrium vacancy concentration (4)
n eq
N exp
f v
k BT
For metals, vf 9kBTm. We can estimate that at room temperature in copper there is one vacancy pet 1015 lattice atoms, whereas just below the melting point there is one vacancy for every 10,000 atoms very strong temperature dependence.
A U neq n -TS = -kBT ln
The equilibrium concentration of the vacancies is defined by the balance between the internal (potential) energy of vacancy formation and the increase of the configurational entropy.
The formation energy vf in crystals, that we used in the derivation, is practically indistinguishable from the formation enthalpy Hvf which has to be used if the pressure and not the volume is kept constant.
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Summary (I) What is entropy? Entropy is the measure of the "probability" of a given macrostate, or, essentially the same thing, the number of microstates possible for the given macrostate. What is a microstate and a macrostate? Any particular arrangement of atoms where we look only on average quantities is a macrostate. Any individual arrangement defining the properties (e.g. positions and velocities) of all the atoms for a given macrostate is a microstate. For a microstate it matters what individual particles do, for the macrostate it does not. What is the probability of a macrostate? The probability for a macrostate = the number of possible ways/microstates to generate the same macrostate divided by the number of all microstates (all possible combinations of the dice, which is a constant). Remember examples on playing dice and on equilibrium concentration of vacancies. Example with vacancies: we just have to find how many ways (# of microstates) are there to arrange n vacancies (macrostate!) in a crystal of N lattice sites. After we find the probability, we can use the Boltzmann equation to calculate the entropy and we can use the equilibrium condition to select the most likely macrostate - the number of vacancies in equilibrium.
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Summary (II) Just knowing the internal energy U of a system with a constant volume and temperature is not enough to determine what the equilibrium configuration is. There could be many macrostates with the same U. That's why just minimizing U (or H) is not good enough, we have to minimize A = U - TS or G = H - TS to find the equilibrium configuration of the system. Of all the macrostates possible for a given U (or H), the one with the largest entropy at the given temperature will be the one that the system will adopt. Equilibrium is a state in which the best possible balance between a small energy and a large entropy is achieved. High entropies often mean high energies and vice verse - both quantities are opposed to each other. The entropy part in A or G becomes more important at high temperatures. The entropy of a certain macrostate can be calculated by the statistical definition of the entropy S, using the Boltzmann entropy equation:
S = kB ln
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei
Summary (III) Irreversible transition takes a system from a less probable state to a more probable state. The transition in the opposite direction (decrease of the entropy of the "universe") is considered by the statistical thermodynamics as an improbable process, whereas the classical thermodynamics just postulates that such transition is impossible.
MSE 3050, Phase Diagrams and Kinetics, Leonid Zhigilei