Work and heat differ not in the kind of energy but in whether you are keeping the books ── and quantum mechanics shows up as stairs in the heat capacity
Heat and work are both transfers of energy. So what is the difference? Textbooks say "ordered transfer is work, disordered transfer is heat" ── which just offloads the job onto the word disordered. There is a more honest phrasing. Work is energy into degrees of freedom you are tracking; heat is energy into degrees of freedom you stopped tracking. Pushing a piston is work because you are following it; the motion of individual molecules is heat because you are not. Exactly the structure of Episode 1 of the sister series "Renormalization That Clicks." On top of that, energy delivered gets shared equally among degrees of freedom (equipartition) ── except that it breaks at low temperature. The culprit is quantum mechanics, and the evidence is plainly visible as a staircase in the heat capacity. That staircase, which tormented nineteenth-century physicists, was one of the very first pieces of evidence for quantum theory.
Push a piston to compress a gas and its internal energy rises. Heat it with a burner and its internal energy rises. The result is the same ── yet we call one "work" and the other "heat."
| Work | Heat | |
|---|---|---|
| Where the energy goes | degrees of freedom you are tracking (the piston's position) | degrees of freedom you stopped tracking (10²³ molecules) |
| Recoverable? | in principle, all of it | only part of it |
| Entropy | unchanged | increased (\(\Delta S=Q/T\)) |
"Heat" is not a kind of energy; it is the name of a destination whose books we stopped keeping. The same kinetic energy is work if you follow it and heat if you stop.
So how is energy that arrives as heat shared out? Within classical mechanics the answer is startlingly democratic.
"Quadratic" means a form like \(\frac12 mv^2\) or \(\frac12 kx^2\). Whatever the thing is, such a degree of freedom gets a flat \(\frac12 k_BT\). Molecular mass and spring stiffness are irrelevant. In Episode 2's language of \(\beta\), it is the mathematical fact that a Gaussian \(e^{-\beta ax^2}\) always has mean energy \(1/2\beta\).
This rule gives the heat capacity of a monatomic gas ── three directions of motion, so \(\langle E\rangle=\frac32 k_BT\) and \(C_V=\frac32 R\) per mole. That matches experiment exactly.
But it fails for diatomic molecules (\(\mathrm{H_2}\), \(\mathrm{N_2}\), \(\mathrm{O_2}\)). With 3 translational + 2 rotational + 2 vibrational (kinetic and potential) = 7 degrees of freedom, \(C_V\) should be \(\frac72 R\). Measure at room temperature and you get \(\frac52 R\).
This was serious. Energy is not going where it should. The vibrational degree of freedom behaves as though it did not exist. Maxwell and Kelvin both acknowledged it as a grave defect of classical physics.
The answer is quantum mechanics. Vibrational energy is not continuous; it can only be accepted in steps of \(\hbar\omega\). And then ──
Compare the minimum unit that can be received, \(\hbar\omega\), with the typical amount on offer, \(k_BT\) (another dimensionless ratio).
$$\frac{\hbar\omega}{k_BT}\ \begin{cases}\ll 1 & \text{fine steps → classical, receives }\tfrac12k_BT\times2\\[4pt] \gg 1 & \text{steps too coarse → excited only as }e^{-\hbar\omega/k_BT}\text{ = }\textbf{frozen}\end{cases}$$In short, a degree of freedom that can only accept notes too large to give change for declines the payment. Episode 1's 25 meV ruler applies directly ── hydrogen's vibration has \(\hbar\omega\approx0.54\) eV, 21 times room-temperature \(k_BT=0.025\) eV. With \(e^{-21}\approx10^{-9}\), it is fast asleep at room temperature.
| Degree of freedom in H₂ | Characteristic temperature | At 300 K | Contribution |
|---|---|---|---|
| translation (3 directions) | ≈ 0 K | fully awake | \(\tfrac32 R\) |
| rotation (2 axes) | \(\theta_{\text{rot}}\approx85\) K | awake | \(R\) |
| vibration (1 mode) | \(\theta_{\text{vib}}\approx6300\) K | frozen | ≈ 0 |
| total at room temperature | \(\tfrac52 R\) ✓ | ||
The figure plots the heat capacity of hydrogen against temperature (the rotational partition function is actually summed in your browser). The horizontal axis is logarithmic in temperature.
There are two steps. Around 100 K rotation wakes up, \(\frac32R\to\frac52R\); at a few thousand K vibration wakes up, \(\frac52R\to\frac72R\). Move the cursor and you can read \(\hbar\omega/k_BT\) for each degree of freedom at that temperature and whether it is awake. This is quantum mechanics reflected in a thermometer ── the position of each step tells you the spacing of the energy levels.
The same thing happens in solids. The Dulong–Petit law (molar heat capacity \(3R\)) holds at high temperature but the heat capacity dies away as \(T^3\) at low temperature ── Einstein and Debye explained this with quanta between 1907 and 1912. Heat capacity was among the earliest experimental evidence for quantum mechanics.
Episode 1: temperature acts only through the ratio \(E/k_BT\).
Episode 2: that \(k_BT\) comes out of the slope of the sharing.
Episode 3: only degrees of freedom with small \(\hbar\omega/k_BT\) accept delivery.
So "hot" means a state where \(k_BT\) is large enough to wake various degrees of freedom. Cooling means putting them back to sleep one at a time. As absolute zero approaches, fewer and fewer are awake ── and eventually nothing happens at all. That is the intuition behind the third law of thermodynamics (entropy approaching a constant as \(T\to0\)).
Established: that the work/heat distinction corresponds to which degrees of freedom you track; the equipartition theorem of classical statistical mechanics (\(\frac12k_BT\) per quadratic degree of freedom); \(C_V=\frac32R\) for a monatomic ideal gas; \(C_V\approx\frac52R\) at room temperature for diatomics with vibration frozen; the characteristic temperatures of \(\mathrm{H_2}\) (rotation ≈ 85 K, vibration ≈ 6300 K); and the Dulong–Petit law with its low-temperature failure (Einstein and Debye models, the \(T^3\) law) ── all standard physics.
What is simplified: (1) The figure idealises the molecule as a rigid rotor plus harmonic oscillator, ignoring rotation–vibration coupling, anharmonicity and centrifugal stretching. (2) Hydrogen's nuclear spin statistics (ortho/para) actually make the low-temperature rotational heat capacity considerably more complicated ── the figure computes distinguishable rigid rotors and ignores spin statistics. (3) On the high-temperature side the molecule dissociates well before 6300 K, so the right-hand edge of the figure is a what-if-it-survived curve. (4) "Heat = degrees of freedom you stopped tracking" is a physically correct organisation, but what you choose to track involves the observer (a point treated in Bonus ② of "Renormalization That Clicks").
Work and heat differ not in the kind of energy but in whether you are tracking. "Heat" is the name of a destination whose books we stopped keeping (the same structure as Episode 1 of "Renormalization That Clicks").
Delivered energy goes \(\frac12k_BT\) into every quadratic degree of freedom ── except that quantum mechanics interferes. A degree of freedom whose delivery unit \(\hbar\omega\) exceeds \(k_BT\) refuses the payment because no change can be given ── it freezes. So the heat capacity of a diatomic climbs in stairs, and H₂ at room temperature sits at \(\frac52R\) with vibration asleep. The position of each step tells you the level spacing ── quantum mechanics reflected in a thermometer, and the staircase that troubled nineteenth-century physics was among the first pieces of evidence for quantum theory.
Print / make a PDF: ⌘+P (Ctrl+P on Windows). On screen the temperature cursor tells you which degrees of freedom are awake and the value of ℏω/k_BT. Buttons jump to the steps. "See the answer" opens each solution.