In this paper we report the theoretical and experimental characterization of a pulsed optically pumped vapor-cell frequency standard based on the detection of the free-induction decay microwave signal. The features that make this standard similar to a pulsed passive maser are presented. In order to predict and optimize the frequency stability, thermal and shot noise sources are analyzed, as well as the conversions of the laser and microwave fluctuations into the output frequency. The experimental results obtained with a clock prototype based on ${}^{87}\mathrm{Rb}$ in buffer gas are compared with the theoretical predictions, showing the practical possibility to implement a frequency standard limited in the medium term only by thermal drift. The achieved frequency stability is ${\sigma }_y\left(\tau \right)=1.2\times {10}^{-12}{\tau }^{-1∕2}$ for measurement times up to $\tau \approx {10}^5\mathrm{s}$. It represents one of the best results reported in literature for gas cell frequency standards and is compliant with the present day requirements for on board space applications.

We reconsider the idea of a pulsed optically pumped frequency standard conceived in the early 1960s to eliminate the light-shift effect. The development of semiconductor lasers and of pulsed electronic techniques for atomic fountains and new theoretical findings allow an implementation of this idea which may lead to a frequency standard whose frequency stability is limited only by the thermal noise in the short term and by the temperature drift in the long term. We shall also show both theoretically and experimentally the possibility of doubling the atomic quality factor with respect to the classical Ramsey technique approach.

We report on the development and characterization of novel 4.596 GHz and 6.834 GHz microwave frequency synthesizers devoted to be used as local oscillators in high-performance Cs and Rb vapor-cell atomic clocks. The key element of the synthesizers is a custom module that integrates a high spectral purity 100 MHz oven controlled quartz crystal oscillator frequency-multiplied to 1.6 GHz with minor excess noise. Frequency multiplication, division, and mixing stages are then implemented to generate the exact output atomic resonance frequencies. Absolute phase noise performances of the output 4.596 GHz signal are measured to be −109 and −141 dB rad2/Hz at 100 Hz and 10 kHz Fourier frequencies, respectively. The phase noise of the 6.834 GHz signal is −105 and −138 dB rad2/Hz at 100 Hz and 10 kHz offset frequencies, respectively. The performances of the synthesis chains contribute to the atomic clock short term fractional frequency stability at a level of 3.1 × 10−14 for the Cs cell clock and 2 × 10−14 for the Rb clock at 1 s averaging time. This value is comparable with the clock shot noise limit. We describe the residual phase noise measurements of key components and stages to identify the main limitations of the synthesis chains. The residual frequency stability of synthesis chains is measured to be at the 10−15 level for 1 s integration time. Relevant advantages of the synthesis design, using only commercially available components, are to combine excellent phase noise performances, simple-architecture, low-cost, and to be easily customized for signal output generation at 4.596 GHz or 6.834 GHz for applications to Cs or Rb vapor-cell frequency standards.