= Millikan photoelectric experiment
{c}
<Robert Andrews Millikan>'s 1916 experiment used a magnetically turned wheel and rotary knife to expose sodium or lithium in <vacuum>. A surrounding copper-gauze/solid-copper collector caught <Electrons>; an <electrometer> recorded accumulated <electric charge> over thirty seconds at each applied voltage. He located endpoints graphically from <photocurrent> curves. Mercury <spectral lines>, selected by a monochromator and filters, controlled <frequency>; rejecting shorter-<wavelength> contamination protected the endpoint. An oxidized collector suppressed reverse emission for selected lines. A copper test plate measured the <contact potential difference> by a displacement-null method; a certified <voltmeter> set the electrical scale. He compared stable-surface sweeps with rapid alternating-<frequency> measurements to control drift. The slope tested $h/e$; the threshold required the contact correction. His gauge did not measure mercury vapor. Confirmation of the equation coexisted with his continued skepticism about localized light quanta.<A Direct Photoelectric Determination of Planck's h by Robert Andrews Millikan>[§§2–3, pp. 361–367, Figs. 1–3; §4, pp. 372–375, Figs. 5–6; §5, p. 380; Theories of photo emission, pp. 383–385; publication verified by Caltech record 45t7b-bvc23: March 1916, Physical Review 7(3), 355–388]{ref}
\Image[/millikan-photoelectric-apparatus.png]
{title=What <Millikan photoelectric experiment> controls and reads out}
{description=Original functional reconstruction of the 1916 apparatus. Successive operations are separated and geometry rearranged; this is not a scale drawing.}
{source=millikan-photoelectric-apparatus.py}
For a modern repetition of the logic, first establish electrical zero with the illumination blocked, characterize leakage, and verify the <voltmeter> scale independently of the presumed <Planck constant>. Scan several voltages near each endpoint rather than calling one noisy zero a cutoff. Hold the illuminated area and surface state fixed when varying <light intensity>; vary <frequency> independently when testing the slope. Repeating a reference <frequency> between other readings reveals time-dependent offsets. Agreement of a slope with $h/e$ tests an <energy> relation; it does not alone identify every microscopic step of <photon absorption> and escape.
Back to article page
