Comparing the Ghost Frequency Comb to Intensity Fluctuation Correlations
Dr. Yan-hua Shih, Department of Physics.
The recently discovered the ghost frequency comb (GFC) is of both fundamental and practical interest, with demonstrations of having applications in LiDAR as well as non-local time transfer. A
50% contrast GFC has been observed in relation to the second-order temporal correlation of a CW laser with 500,000 cavity modes. This second-order measurement was made by splitting the CW laser beam to two point-like photodetectors, D1 and D2. The output currents of the detectors were digitized by a 50 GHz A-to-D converter which were then used to calculate the temporal correlation, ⟨I1(t)I2(t + τ )⟩. The second-order correlation of classical light is typically considered as the intensity fluctuation correlation; however, the intensity fluctuations of a laser consisting of half a million cavity modes would be extremely low. The averaging of 500,000 modes can be approximated as an ensemble average, resulting in a maximum intensity fluctuation on the order of 10^−3 compared to the mean intensity value, and thus insignificant. Based on the calculated intensity fluctuations, the intensity fluctuation correlation would be on the order of 10^−6. This raises the question: is the GFC the result of intensity fluctuation correlations?
We are seeking to present a numerical simulation of a 500,000 mode CW laser to find the intensity fluctuations relative to the mean intensity, as well as the intensity fluctuation correlations. The simulation will be developed to mimic the experiment that found 50% contrast to show the contribution classical intensity fluctuation correlations make to the GFC contrast. Such a simulation raises the question whether the ghost frequency comb can be considered the result of the intensity fluctuation
correlation of a CW laser beam.
With the simulation we are creating, more computational power and memory is required than we are able to obtain with typical PCs. The next steps of the simulation will be to use real collected data, which will be even more computationally expensive.