Historical research guide · 1114 · 1958
Measurements of attractive forces between flat plates
M. J. Sparnaay
In one minute
In 1948 Hendrik Casimir predicted that two flat metal plates facing each other in empty space would be pushed together by the vacuum itself, because the gap between them can hold fewer light waves than the room outside. Ten years later M. J. Sparnaay, at the Philips research laboratories in Eindhoven, built the apparatus that went looking for it — and this short Physica paper is the first time anyone measured the force Casimir described. Sparnaay reports what he found carefully: the plates do attract, and the attraction does not contradict Casimir’s prediction. He is equally careful about what went wrong earlier. His own previous runs, with clean and with silvered glass and quartz plates in an ionised atmosphere, had been reading a different effect altogether — a voltage difference between the two facing surfaces, pulling them together electrically. Naming that artefact is the second gift of the paper, because every precision Casimir experiment since has had to cancel exactly that voltage before it can see the vacuum.
Connection
Chapter 2 says the vacuum is a real medium a laboratory can push on, and this is the first time a laboratory pushed on it. Chapter 1 gets the other half: a first measurement that establishes the sign and the order of magnitude, states its own limits honestly, and hands the next generation the systematic error to beat — which is what Lamoreaux settled in 1997 and Bressi and colleagues repeated in Casimir’s own parallel-plate geometry in 2002.
From the source
Results of experiments concerning attractive forces between flat metal plates, together with a description of the apparatus used, are reported — and the observed attractions do not contradict Casimir’s theoretical prediction. That sentence, written in 1958, is the first experimental statement about the Casimir force in the literature.Abstract, first two sentences; Physica 24, pages 751 to 764
The force Sparnaay measured, in the units Casimir’s formula uses, is one to four times ten to the minus eighteen dyne centimetres squared divided by the fourth power of the gap, for separations between two and ten micrometres. Casimir’s prediction in the same units is 1.30 times ten to the minus eighteen. Sparnaay’s band contains it.Figures as reported by Steven Weinberg, The cosmological constant problem, Reviews of Modern Physics 61, 1 (1989), Section III, footnote on the Casimir effect — on this site at /library/stm-5610822bc8
The size of the achievement and the size of its error bar belong in the same sentence. This measurement showed an attractive force not inconsistent with the prediction, but with effectively one hundred per cent uncertainty: it settles the sign and the order of magnitude of the effect, not the coefficient in front of it.As characterised in the introduction of Steve K. Lamoreaux, Demonstration of the Casimir Force in the 0.6 to 6 micrometre range, Physical Review Letters 78, 5 (1997) — on this site at /library/stm-e8c853485e
The second half of the paper is a correction of the author’s own earlier work. Attractions previously measured between clean plates, and between silvered glass and quartz plates, all brought into an ionised atmosphere, were due to differences in the surface potentials of the opposing plates — an electrostatic effect, not a vacuum effect.Abstract, third and fourth sentences
That artefact became a permanent part of the method. Contact-potential differences between two facing metal surfaces are the dominant systematic in every Casimir measurement that followed, and cancelling the residual voltage before the force is read is now a standard step of the experiment rather than an optional check.Compensation of the residual electrostatic potential as practised in Bressi, Carugno, Onofrio and Ruoso, Measurement of the Casimir Force between Parallel Metallic Surfaces, Physical Review Letters 88, 041804 (2002) — on this site at /library/stm-208d347532
The configuration Casimir actually calculated stayed the hard one for forty-four years. Before Bressi and colleagues repeated it in 2002, the parallel-plate geometry had been attempted exactly once — here — with large systematic errors and uncontrolled electrostatic forces, while the easier sphere-against-plate geometry had already reached one per cent.Bressi, Carugno, Onofrio and Ruoso, Physical Review Letters 88, 041804 (2002), introduction, survey of previous attempts — on this site at /library/stm-208d347532
The way in
https://doi.org/10.1016/S0031-8914(58)80090-7How to cite it
M. J. Sparnaay (1958) Measurements of attractive forces between flat plates doi:10.1016/S0031-8914(58)80090-7