Recent astrophysical measurements of distant quasar spectra indicate that the fundamental constants may be changing with time. The dimensionless fine structure constant α, which scales the energy in electromagnetic interactions, might have been smaller at early times in the universe: the difference compared with today’s value is a fraction of 10-5. Assuming that the drift is linear, this would be a change of around 10-15 per year.
In general, such astrophysical measurements probe a drift of constants over extremely long time periods. Laboratory experiments on the other hand are limited to short timescales of some years. However, this can be compensated for by a higher accuracy. The recent dramatic evolution of techniques for measuring the frequency of light with a precision of a few parts in 1015 means that laser spectroscopy of atoms and ions has now reached a level of accuracy where a search for a drift of constants is feasible.
In particular, there is an effective magnification between a drift of α and the drift of an atomic transition frequency compared with the SI second provided by a caesium clock, leading to a frequency drift at an estimated level of 10-14 per year. A remeasurement of atomic transition frequencies previously measured some time ago is therefore of great interest, and earlier this year researchers at MPQ in Munich carried out such an experiment, measuring the 1S-2S transition in atomic hydrogen.
In 1999 the absolute frequency of this transition had been phase coherently compared with a caesium fountain clock, as a primary frequency standard, using a novel frequency comb technique. The second harmonic of laser light from an ultrastable continuous-wave dye laser emitting near 486 nm was used to excite cold hydrogen atoms, Doppler-free, from the ground state to the metastable 2S state. A selection of slow atoms by time-resolved spectroscopy allowed the reduction of systematic effects, mainly the second-order Doppler shift (A Huber et al. 1999). This measurement has demonstrated a precision of 1.9 parts in 1014 (M Niering et al. 2000).
Since then, the techniques used for frequency measurement, as well as the hydrogen spectrometer, have been improved, and data gathered in 12 days of measurement in February this year are now being evaluated. So far, eight days of data have been analysed, and the preliminary result for the difference of the two measurements is -48(60) Hz. Assuming that the measurements performed in 1999 and 2003 are equivalent, this implies a possible drift of the 1S-2S frequency of -5.4(6.8) x 10-15 Hz per year.
Further reading
A Huber et al. 1999 Phys. Rev. A 59 1844.
M Niering et al. 2000 Phys. Rev. Lett. 84 5496.
