A Determination of em for an Electron by a New Deflection Method. II

Abstract

A final report is here made of a deflection determination using a new method described in a preliminary report (Part I). There are two fundamental advantages of this method over other free electron methods: (1) the accelerating voltage is not measured. In its place, the determination of the frequency of a radiofrequency oscillator is made, and this can be done with far greater precision. (2) Errors from all contact potentials are either entirely eliminated or made negligibly small, as is shown in detail. The ever prevalent error arising from surface changes in all free electron determinations was at least greatly reduced by the use of evaporated gold surfaces throughout the deflecting chamber, said surfaces being frequently renewed to cover over the insulating layers formed. The error remaining was eliminated by extrapolation of results to infinite electron energy. The magnetic field was most carefully calibrated in terms of a specially designed standard solenoid. The frequency calibration was based on government standards through the cooperation of the Federal Radio Commission. Standards of voltage, resistance and lengths were calibrated at the National Bureau of Standards. The object throughout the work has been to produce a value as free as possible from errors both experimental and subjective. The result obtained is $\frac{e}{{m}_{0}}=(1.7597\ifmmode\pm\else\textpm\fi{}0.0004)\ifmmode\times\else\texttimes\fi{}{10}^{7} \mathrm{e}.\mathrm{m}.\mathrm{u}.$ The stated probable error is based on allowances for unknown constant errors, the observational probable errors being negligible. This value of $\frac{e}{{m}_{0}}$ is about 1.25 parts in ${10}^{3}$ higher than that found in recent spectroscopic determinations, a definite discrepancy being indicated. With the inclusion of the final values obtained in the last ten years by nine experimenters using both free electron and spectroscopic methods, the present most probable value is found to be $\frac{e}{{m}_{0}}=(1.7584\ifmmode\pm\else\textpm\fi{}0.0003)\ifmmode\times\else\texttimes\fi{}{10}^{7} \mathrm{e}.\mathrm{m}.\mathrm{u}.$ but in view of the discrepancy this is at best only a tentative value.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 22 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.