We present a fully ab initio model and calculations of the spectral shapes of absorption lines in a pure molecular gas under conditions where the influences of collisions and of the Doppler effect are significant. Predictions of the time dependence of dipole autocorrelation functions (DACFs) are made for pure CO 2 at room temperature using requantized classical molecular dynamics simulations. These are carried, free of any adjusted parameter, on the basis of an accurate anisotropic intermolecular potential. The Fourier-Laplace transforms of these DACFs then yield calculated spectra which are analyzed, as some measured ones, through fits using Voigt line profiles. Comparisons between theory and various experiments not only show that the main line-shape parameters (Lorentz pressure-broadening coefficients) are accurately predicted, but that subtle observed non-Voigt features are also quantitatively reproduced by the model. These successes open renewed perspectives for the understanding of the mechanisms involved (translational-velocity and rotational-state changes and their dependences on the molecular speed) and the quantification of their respective contributions. The proposed model should also be of great help for the test of widely used empirical line-shape models and, if needed, the construction of more physically based ones.

Precision spectroscopy of the hydrogen molecule is a test ground of quantum electrodynamics (QED), and it may serve for the determination of fundamental constants. Using a comb-locked cavity ring-down spectrometer, for the first time, we observed the Lamb-dip spectrum of the R(1) line in the overtone of hydrogen deuteride (HD). The line position was determined to be 217 105 182.79±0.03_{stat}±0.08_{syst}  MHz (δν/ν=4×10^{-10}), which is the most accurate rovibrational transition ever measured in the ground electronic state of molecular hydrogen. Moreover, from calculations including QED effects up to the order m_{e}α^{6}, we obtained predictions for this R(1) line as well as for the HD dissociation energy, which are less accurate but signaling the importance of the complete treatment of nonadiabatic effects. Provided that the theoretical calculation reaches the same accuracy, the present measurement will lead to a determination of the proton-to-electron mass ratio with a precision of 1.3 parts per billion.

The very weak S(0)-S(3) electric quadrupole transitions of the second overtone band of molecular hydrogen have been recorded in the laboratory by continuous-wave cavity ring-down spectroscopy near 0.8 μm. The ultrahigh sensitivity of the spectrometer (α min ∼ 1 × 10 −10 cm −1 ) allows us to detect the considered transitions at a relatively low sample pressure (50-750 torr). The line positions, intensity, and pressure-shift coefficients are derived from a fit of the line shape using a Galatry profile. Compared with literature values, the relative differences between the experimental and theoretical transition intensities are reduced by one order of magnitude, reaching a value of about 2% mainly dependent of the line-shape function adopted for the profile fitting. The thermal equilibrium relative intensity of the S(1) to S(0) line is determined with an accuracy of 0.4%, which can be used to probe the ortho-to para-H 2 concentration ratio. Our measurements confirm the quality of the high-level ab initio calculations, including the relativistic and quantum electrodynamics corrections.

No abstract

No abstract

No abstract

We present a new method of comb-locked cavity ring-down spectroscopy for the Lamb-dip measurement of molecular ro-vibrational transitions. By locking both the probe laser frequency and a temperature-stabilized high-finesse cavity to an optical frequency comb, we realize saturation spectroscopy of molecules with kilohertz accuracy. The technique is demonstrated by recording the R(9) line in the υ = 3 - 0 overtone band of CO near 1567 nm. The Lamb-dip spectrum of such a weak line (transition rate 0.0075 s) is obtained using an input laser power of only 3 mW, and the position is determined to be 191 360 212 770 kHz with an uncertainty of 7 kHz (δν/ν∼3.5×10), which is currently limited by our rubidium clock.

No abstract