Frequency fluctuations of lasers cause a broadening of their line shapes. Although the relation between the frequency noise spectrum and the laser line shape has been studied extensively, no simple expression exists to evaluate the laser linewidth for frequency noise spectra that does not follow a power law. We present a simple approach to this relation with an approximate formula for evaluation of the laser linewidth that can be applied to arbitrary noise spectral densities.

We have manufactured more than 250 nominally identical paraffin-coated Cs vapor cells (30 mm diameter bulbs) for multi-channel atomic magnetometer applications. We describe our dedicated cell characterization apparatus. For each cell we have determined the intrinsic longitudinal, Γ 01 , and transverse, Γ 02 , relaxation rates. Our best cell shows Γ 01 /2π ≈ 0.5 Hz, and Γ 02 /2π ≈ 2 Hz. We find a strong correlation of both relaxation rates which we explain in terms of reservoir and spin exchange relaxation. For each cell we have determined the optimal combination of rf and laser powers which yield the highest sensitivity to magnetic field changes. Out of all produced cells, 90% are found to have magnetometric sensitivities in the range of 9 to 30 fT/ √ Hz. Noise analysis shows that the magnetometers operated with such cells have a sensitivity close to the fundamental photon shot noise limit.

The frequency noise properties of commercial distributed feedback quantum cascade lasers emitting in the 4.6 μm range and operated in cw mode near room temperature (277 K) are presented. The measured frequency noise power spectral density reveals a flicker noise dropping down to the very low level of <100 Hz(2)/Hz at 10 MHz Fourier frequency and is globally a factor of 100 lower than data recently reported for a similar laser operated at cryogenic temperature. This makes our laser a good candidate for the realization of a mid-IR ultranarrow linewidth reference.

No abstract