A new method for detecting velocity shifts and distortions between optical spectra

Abstract

Recent quasar spectroscopy from the VLT and Keck telescopes suggests that fundamental constants may not actually be constant. To better confirm or refute this result, systematic errors between telescopes must be minimized. We present a new method to directly compare spectra of the same object and measure any velocity shifts between them. This method allows for the discovery of wavelength-dependent velocity shifts between spectra, i.e. velocity distortions, that could produce spurious detections of cosmological variations in fundamental constants. This "direct comparison" method has several advantages over alternative techniques: it is model-independent (cf. line-fitting approaches), blind, in that spectral features do not need to be identified beforehand, and it produces meaningful uncertainty estimates for the velocity shift measurements. In particular, we demonstrate that, when comparing echelle-resolution spectra with unresolved absorption features, the uncertainty estimates are reliable for signal-to-noise ratios >7 per pixel. We apply this method to spectra of quasar J2123-0050 observed with Keck and the VLT and find no significant distortions over long wavelength ranges (~1050 Angstroms) greater than ~180 m/s. We also find no evidence for systematic velocity distortions within echelle orders greater than 500 m/s. Moreover, previous constraints on cosmological variations in the proton--electron mass ratio should not have been affected by velocity distortions in these spectra by more than 4.0 +/- 4.2 parts per million. This technique may also find application in measuring stellar radial velocities in search of extra-solar planets and attempts to directly observe the expansion history of the Universe using quasar absorption spectra.