Towards a Better Understanding of OPD Limitations for Higher Sensitivity and Contrast at the VLTI

TL;DR

This paper investigates the limitations of optical path difference control in the VLTI, combining theoretical analysis and experimental data to identify vibration sources and model their impact on OPD stability, aiming to improve sensitivity and contrast.

Contribution

It provides a detailed vibrational model of the VLTI's OPD residuals, identifies dominant vibration frequencies, and suggests methods to reduce OPD RMS below 100nm.

Findings

Experimental evidence of the $f^{-17/3}$ power law in ATs

Vibration frequencies of 40-50Hz and 60-90Hz dominate OPD residuals

Removing vibrations below 100Hz can achieve sub-100nm OPD RMS

Abstract

Precise control of the optical path differences (OPD) in the Very Large Telescope Interferometer (VLTI) was critical for the characterization of the black hole at the center of our Galaxy - leading to the 2020 Nobel prize in physics. There is now significant effort to push these OPD limits even further, in-particular achieving 100nm OPD RMS on the 8m unit telescopes (UT's) to allow higher contrast and sensitivity at the VLTI. This work calculated the theoretical atmospheric OPD limit of the VLTI as 5nm and 15nm RMS, with current levels around 200nm and 100nm RMS for the UT and 1.8m auxillary telescopes (AT's) respectively, when using bright targets in good atmospheric conditions. We find experimental evidence for the $f^{-17/3}$ power law theoretically predicted from the effect of telescope filtering in the case of the ATs which is not currently observed for the UT's. Fitting a series of vibrating mirrors modelled as dampened harmonic oscillators, we were able to model the UT OPD PSD of the gravity fringe tracker to $<1nm/\sqrt{Hz}$ RMSE up to 100Hz, which could adequately explain a hidden $f^{-17/3}$ power law on the UTs. Vibration frequencies in the range of 60-90Hz and also 40-50Hz were found to generally dominate the closed loop OPD residuals of Gravity. Cross correlating accelerometer with Gravity data, it was found that strong contributions in the 40-50Hz range are coming from the M1-M3 mirrors, while a significant portion of power from the 60-100Hz contributions are likely coming from between the M4-M10. From the vibrating mirror model it was shown that achieving sub 100nm OPD RMS for particular baselines (that have OPD$\sim$200nm RMS) required removing nearly all vibration sources below 100Hz.