Material Insights · Understanding the Material World

Waves: sound, light, colour and the wave-particle puzzle: The photoelectric effect and the wave-particle puzzle

All courses / Understanding Science / Module 7: Waves: sound, light, colour and the wave-particle puzzle Part 6 of 6

The photoelectric effect and the wave-particle puzzle

Every test so far says light is a wave. Then a different experiment says it arrives in lumps. This part tells how the photoelectric effect broke the wave picture, how Einstein's 1905 answer was confirmed, and how electrons turned out to show interference too. It ends with a map of which picture to use when. The next module uses the photon to explain atoms.

6.1 The photoelectric effect defeated every wave prediction, and Einstein's 1905 answer was that light comes in lumps of energy hf

Shine ultraviolet light on a clean metal plate and electrons come off. Hertz saw its first hint in 1887, J. J. Thomson showed the particles were electrons, and in 1902 Philipp Lenard measured how the effect depends on the light. A wave picture predicted three things: brighter light should eject faster electrons, any colour should work if bright enough, and dim light should need time for the electrons to gather energy. Experiment showed the opposite. Brighter light ejected more electrons but not faster ones. Faster electrons came with higher frequency. Below a certain frequency, no electrons came out however strong the light. And there was no measurable delay.

In 1905 Einstein, extending Planck's 1900 idea of energy exchanged in packets, proposed that light itself arrives in quanta, each carrying energy E = hf, where h is Planck's constant, 6.626 x 10^-34 J s. One quantum is absorbed by one electron, all at once. The electron spends a fixed work function W to leave the metal, and keeps the rest: maximum kinetic energy = hf - W. Each prediction is now automatic: brightness changes the number of quanta, frequency changes each one's energy, and a quantum smaller than W does nothing.

Millikan set out to disprove this. After about ten years, in 1916, he confirmed it and measured h to within 0.5%. Einstein's Nobel Prize for 1921, awarded in 1922, was for this law, not relativity. Solar cells and light sensors rest on related effects (see 'Photovoltaic effect').

Data Worked estimate: potassium (work function about 2.3 eV) lit by different colours (photon energy in eV = 1,240 / wavelength in nm)
LightPhoton energyElectrons released?Maximum kinetic energy
Red, 700 nm1,240 / 700 = 1.77 eV (calculated)none, below the threshold0
Green, 540 nm1,240 / 540 = 2.30 eV (calculated)just at the thresholdabout 0
Violet, 400 nm1,240 / 400 = 3.10 eV (calculated)yes3.10 eV - 2.30 eV = 0.80 eV (calculated)
Ultraviolet, 250 nm1,240 / 250 = 4.96 eV (calculated)yes4.96 eV - 2.30 eV = 2.66 eV (calculated)
1 mW green laser pointer, 532 nm2.33 eV x 1.602 x 10^-19 J/eV = 3.73 x 10^-19 J (calculated)0.001 J/s / 3.73 x 10^-19 J = 2.7 x 10^15 photons per second (calculated)the lumps are too many to notice

Calculated. Work function 2.3 eV is a typical handbook value for potassium. 1 eV = 1.602 x 10^-19 J.

Source: University of Virginia, Physics 252, The Photoelectric Effect, https://galileo.phys.virginia.edu/classes/252/photoelectric_effect.html (Lenard 1902: electron energy independent of intensity; Einstein 1905: E = hf - W; Millikan confirmed by 1916, h to 0.5%)

6.2 Light and matter both act as waves in flight and as particles on arrival, and the double-slit experiment shows both at once

Light gives interference, diffraction and polarisation, which are wave behaviour. It also gives the photoelectric effect and, in the Compton scattering of X rays (1923), photons with momentum h / wavelength, which are particle behaviour. Neither picture alone is enough.

The resolution comes from Young's experiment with weak light. Turn the source down until, on average, only one photon is in the apparatus at a time. A sensitive screen records single dots, each at one place, as for a particle. Yet as thousands of dots build up, they form the same bright and dark fringes. Each photon seems to go through both slits, and the pattern shows where it is likely to land.

Matter does the same. In 1924 de Broglie proposed that a particle of momentum p has wavelength h / p. In 1927 Davisson and Germer fired electrons at a nickel crystal and found peaks at particular angles, as with X rays. In 1974 Merli, Missiroli and Pozzi in Bologna, and in 1989 Tonomura and colleagues in Japan, sent electrons one at a time through an electron biprism, the electron version of two slits, and watched fringes grow from single dots. Even molecules of dozens of atoms have done so.

The modern account is that a wave describes the probability of detecting a particle at each place. It is a wave in motion and a particle on arrival. Detect which slit it used, and the fringes vanish.

Data Worked estimate: de Broglie wavelength = h / (mass x speed)
ObjectMomentumWavelength
Electron accelerated through 100 Vp = square root of (2 x 9.11 x 10^-31 kg x 1.60 x 10^-17 J) = 5.4 x 10^-24 kg m/s (calculated)6.626 x 10^-34 J s / 5.4 x 10^-24 kg m/s = 0.123 nm (calculated), close to atom spacing in a crystal
Electron in a 100 kV electron microscoperelativistic valueabout 0.0037 nm (calculated)
Cricket ball, 0.16 kg at 40 m/s0.16 kg x 40 m/s = 6.4 kg m/s1.0 x 10^-34 m (calculated), about 10^-19 of a nucleus
Person, 70 kg at 1.4 m/s70 kg x 1.4 m/s = 98 kg m/s6.8 x 10^-36 m (calculated)

Calculated with h = 6.626 x 10^-34 J s. See also Rosa (2012) for the electron experiments.

Source: American Physical Society, Davisson-Germer experiment historic site, https://www.aps.org/programs/honors/history/historicsites/davisson-germer.cfm (1927: peaks in electrons scattered from nickel; electrons behave as waves)

6.3 Rays, waves and quanta are three tools: the size of the object against the wavelength, and the energy of each exchange, decide which one fits

A model is judged by the questions it answers. Where objects are far bigger than the wavelength, light acts like rays that travel straight, reflect at mirrors and bend at lenses. Where sizes are close to the wavelength, as in slits, thin films, gratings and the edge of a shadow, the wave picture is needed. Where light is exchanging energy with single atoms or electrons, in detectors, solar cells, LEDs and lasers, the photon picture is needed. Newton's rays did not go wrong; they are the wave picture when the wavelength is tiny.

The table lists the evidence in time order. For lenses, fibres, lasers, LEDs and solar cells, go to 'Light, semiconductors and information' in Understanding Engineering.

Data How we know: the key experiments in order
YearWhoWhat was found
1676RømerLight takes time: Io's eclipses drift
1801 to 1807YoungInterference of light, then the two-slit test: light is a wave
1808MalusReflected light is polarised: the wave is transverse
1849FizeauLight speed measured on the ground
1887HertzRadio waves made and detected; also the first photoelectric hint
1905 and 1916Einstein and MillikanLight quanta proposed, then confirmed
1927Davisson and GermerElectrons diffract
1974 and 1989Merli et al., Tonomura et al.Single electrons build fringes
1983General Conference on Weights and MeasuresThe metre is defined from the speed of light

Dates from the sources cited in this module.

Source: Feynman, Leighton and Sands, The Feynman Lectures on Physics, Vol. I (Caltech), further reading only: https://www.feynmanlectures.caltech.edu/I_28.html (chapter 28 'Electromagnetic Radiation'), I_30 'Diffraction', I_37 'Quantum Behavior'
  • Feynman, Leighton and Sands, The Feynman Lectures on Physics, Vol. I (Caltech) · further reading only: https://www.feynmanlectures.caltech.edu/I_28.html (chapter 28 'Electromagnetic Radiation'), I_30 'Diffraction', I_37 'Quantum Behavior'
  • Hewitt, Lyons, Suchocki and Yeh, Conceptual Integrated Science, 2nd edition (Pearson, 2013) · ch. 8, 'Waves, Sound and Light'

See it in the table

Ni Nickel

Check your understanding

Violet light of 400 nm (3.1 eV) falls on a metal with a work function of 2.3 eV. What is the maximum kinetic energy of an emitted electron?

Maximum kinetic energy = photon energy - work function = 3.1 - 2.3 = 0.8 eV.

Which result of the photoelectric effect could the wave picture not explain?

A wave spreads energy over the surface, so electrons would need time to gather enough. Instant emission by single lumps of energy fits the photon picture.

A cricket ball has a de Broglie wavelength of about 10^-34 m. What does this mean for seeing its wave nature?

The wavelength is tiny compared with anything it meets, so no interference could be detected. Wave nature is real for it but invisible.
← Part 5: Waves meeting waves: interference, diffraction, Young's experiment and polarisationMark read · Wrap-up →