Einstein's photoelectric paper is the one that won him the Nobel Prize, and it's the only one of the 1905 papers he described as revolutionary. It's also the one people summarize most carelessly. The usual version is that Einstein discovered the photon. What he actually wrote was more careful than that, and the caution is the interesting part.

The Problem

By 1905, Maxwell's equations had been enormously successful. Light was a continuous electromagnetic wave, and this explained interference, diffraction, and polarization with no loose ends. The wave picture was not a hypothesis under review. It was settled physics.

But it failed at the edges. Blackbody radiation was the loudest failure: classical theory predicted that a hot object should radiate infinite energy at short wavelengths, which is obviously wrong. Planck had patched this in 1900 by assuming that the walls of a cavity emit and absorb energy in discrete chunks proportional to frequency, but he treated this as a bookkeeping trick about the emitters, not a claim about light itself. The photoelectric effect was a quieter failure. When light hits a metal, electrons come off, and the details did not match what waves should do.

Background Science

The wave model makes a specific prediction about ejected electrons. Energy in a wave is carried by amplitude, so brighter light should mean more energetic electrons. Dim light should mean sluggish electrons, and if you wait long enough, energy should accumulate until an electron finally works loose. Frequency should not matter much.

None of that is what happens. Below a threshold frequency, no electrons come out at all, no matter how bright the light or how long you wait. Above it, electrons appear essentially instantly, even at very low intensity. Increasing brightness produces more electrons but not faster ones. The energy of each electron depends on frequency alone.

What He Did

Einstein's move was not to run an experiment. He worked from thermodynamics, comparing the entropy of low-density blackbody radiation to the entropy of an ideal gas, and found that the radiation entropy depended on volume in the same mathematical way a gas of independent particles does. That is a strange result to get from a wave. It suggested that radiation, at least in this regime, is statistically indistinguishable from a collection of discrete, independent energy packets of size hν.

He then applied this to the photoelectric effect in one line of reasoning. If light arrives in packets, one packet gives its energy to one electron. The electron spends some fixed amount escaping the metal — the work function — and keeps whatever is left:

E = hν − φ

This explains everything at once. The threshold is where hν equals φ. The instantaneity follows because no accumulation is needed. Brightness controls the number of packets, so more light means more electrons, not faster ones. Maximum electron energy rises linearly with frequency, with slope h and intercept −φ.

Why It Mattered

The prediction was falsifiable and specific, and the slope was Planck's constant, showing up in a completely unrelated experiment. Millikan spent roughly a decade trying to disprove it, confirmed it precisely by 1916, and remained unconvinced of the underlying picture even then. That reaction was typical. The equation was accepted long before the physics behind it was.

The Nobel citation in 1921 is careful about this. It credits the law of the photoelectric effect, not the light quantum hypothesis. Einstein got the prize for the part that had been measured, not the part he thought was revolutionary.

Some Thoughts

What strikes me about this paper is how narrow Einstein's claim actually is. His title says the work is about a heuristic viewpoint, and in the introduction he restricts the argument to interactions where light is emitted or absorbed. He is not claiming the wave theory is wrong, and he is not claiming light is made of particles. He is claiming that in this particular corner of physics, treating light as though it were quantized produces the right answer, and that this is worth taking seriously as a clue.

That is a much stranger position to hold than "light is particles." It means holding two incompatible pictures at once and knowing which one to use where, without any account of why. Einstein had no resolution to offer, and said so. The resolution took another twenty years and, arguably, did not so much resolve the tension as formalize it.

I think this is a better model of how physics moves than the one we usually tell. The story we get is that an anomaly appears, someone proposes the correct new picture, and the field updates. What happened here is that someone noticed a mathematical coincidence in an entropy formula, followed it into an unrelated experiment, got a specific number right, and then declined to explain what it meant. The explanation came later and from other people. The paper's value was in showing that the wave theory had a boundary, not in saying what was on the other side of it.