Skip to content
Tutorials Center
My learning

Topics

Lesson 8 of 39 ยท Microstates, Macrostates and Entropy

Entropy and the Second Law

Watch two Einstein solids share energy and see the second law emerge from counting: the most probable split has the highest entropy and fluctuations vanish for large systems.

The second law of thermodynamics says that the entropy of an isolated system never decreases. It sounds like a new law of nature, yet statistical mechanics shows it is a consequence of counting. In this lesson you will build the argument with two Einstein solids that can exchange energy.

Two solids in thermal contact

Take solid A with N_A oscillators and solid B with N_B oscillators. Together they hold q quanta in total. They are isolated from everything else, but energy may flow between them. Let q_A be the number of quanta in A; then B has q โˆ’ q_A.

For each division q_A, the multiplicity of the combined system is the product

Text
ฮฉ_total(q_A) = ฮฉ_A(q_A) ยท ฮฉ_B(q โˆ’ q_A)

where ฮฉ(N, q) = C(q + N โˆ’ 1, q) as derived in the lesson on counting the Einstein solid. Each q_A is a macrostate of the combined system. By the fundamental postulate every microstate is equally likely, so the most probable q_A is the one with the largest ฮฉ_total.

Python Multiplicity against energy sharing

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The peak sits near q_A = q N_A/(N_A + N_B): energy is shared in proportion to the number of oscillators. That means equal energy per oscillator, which is the microscopic content of "equal temperature".

Entropy is largest at the peak

Taking logarithms, S_total = S_A + S_B. The macrostate with the most microstates has the largest total entropy. A system started in an unlikely division, say all the energy in A, will almost certainly end up in a division near the peak, because there are vastly more microstates there. Energy flowing from "hot" to "cold" is just the system moving toward the sharp maximum of S_total.

flowchart LR
  A["Start: unusual division<br/>few microstates"] -->|"random energy exchange"| B["Likely division<br/>near the peak"]
  B --> C["Stays near the peak<br/>(small fluctuations)"]

Energy exchange drifts the system to the macrostate with the most microstates.

The peak sharpens with size

The width of the peak shrinks, relative to the total energy, as the system grows. Compare several sizes with equal solids and q = N (so each oscillator has one quantum on average).

Python Relative width of the peak

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The fraction std / q_total shrinks roughly as 1/โˆšN. For macroscopic objects the energy division is effectively fixed at the peak.

The second law, statistically

Put the argument in words:

  1. Macrostates with larger ฮฉ are more probable.
  2. For large systems the probability is so strongly concentrated that only the highest-entropy macrostates (and those very near it) ever occur.
  3. Therefore an isolated system moves toward, and stays at, macrostates of maximum entropy.

The second law is a statement of overwhelming probability, not absolute certainty. Decreases of entropy are possible, but their probability is like exp(โˆ’ฮ”S/k_B); for a macroscopic ฮ”S such as 1 J/K, that is exp(โˆ’7.2 ร— 10^22).

Entropy can decrease in a part of a larger system. A refrigerator cools its contents and lowers their entropy, but it dumps more entropy into the room. The total of the isolated whole does not decrease.

Python Probability of an entropy decrease

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

An entropy dip of 10^-23 J/K (below one k_B) is quite possible; a dip of a milli-joule per kelvin is hopelessly improbable.

Irreversibility

The microscopic laws are time-reversible, yet we see an arrow of time. Statistical mechanics resolves this: the reverse of a typical process is not forbidden, just astronomically unlikely, because it requires the system to land on a tiny set of special microstates.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Two Einstein solids with equal numbers of oscillators exchange energy. What is the most probable division of the total energy?

    The product Omega_A Omega_B is largest at the equal split for identical solids.
  2. 2. How does the relative width of the probability peak for energy sharing change as the solids get larger?

    Relative fluctuations are of order 1/sqrt(N).
  3. 3. What is the statistical meaning of the second law?

    The second law expresses overwhelming probability that isolated systems end up in the highest-multiplicity macrostates.
  4. 4. A refrigerator lowers the entropy of its contents. Does that violate the second law?

    The second law applies to the total (isolated) system; the surroundings gain at least as much entropy as the contents lose.

Summary

  • For two systems sharing a fixed total energy, ฮฉ_total(q_A) = ฮฉ_A(q_A) ฮฉ_B(q โˆ’ q_A).
  • The most probable division maximizes total entropy and shares energy in proportion to the number of oscillators.
  • The peak becomes sharp as 1/โˆšN, so large systems appear to follow a deterministic law.
  • The second law is overwhelming probability, with the chance of an entropy decrease ฮ”S of order exp(โˆ’ฮ”S/k_B).

Next: Temperature from Entropy