Mass measurements of few-electron systems in Penning traps

The accuracy of the SMILETRAP mass spectrometer has been verified by a number of mass comparisons involving well-known masses. Our results for H₂ ⁺,Ne⁶⁺,Ne^{9+ ,10+},Si^{12+ ,13+ ,14+},and Ar^{14+ ,16+} all agree within the statistical errors (0.3–1 ppb) with previous determinations. However, all measurements involving He give a deviation. The combined He^1+,2+ data results in a mass deviation of +1.9 ±0.23 ppb. The uncertainty of the accepted He mass is 0.25 ppb, thus this represents a significant deviation. High statistics comparisons (statistical uncertainty <0.5 × 10^-9utilizing different species (excluding He) and charge states agree within ±0.5 ppb. An analysis estimating the contribution from individual systematic error sources and other auxiliary tests does not allow a systematic error larger than ± 0.85 ppb. We conclude that for now we cannot rule out the presence of an unknown systematic error which in the He comparison results in a near 2 ppb deviation. Thus, as a safety measure we should exclude the He data when calculating the proton mass. The He discrepancy also forces us to give a larger limit of the systematic error of the proton mass than motivated by high statistics comparisons. However, due to the consistency of all other measurements and tests, it appears unlikely that this deviation should be present to the same extent in other comparisons. Thus, for now, after a preliminary analysis we report a proton mass = 1.007 276 466 72 ± 16 ± 85 u, where the errors are the weighted statistical errors and the estimated maximal systematical error, respectively. After a complete analysis we expect the systematic error to be reduced below ±0.5 ppb. This revised version was published online in July 2006 with corrections to the Cover Date.