6 Summary

We reconstructed point spread functions of the ADONIS adaptive optics system from its wavefront sensor measurements. For very bad seeing (at least 1.5 arcsec), the PSF reconstruction is not satisfying. The OTF is always underestimated by 20 to 30% at low and medium frequencies. This discrepancy may be due to the presence of speckles on the WFS image which induce biases in the calculation of the center of gravity.

For good seeing and good correction, we estimate the OTF within 10% accuracy at low and medium frequencies. In this case, the major limitation is the presence of a residual non-modeled aberration visible on the first diffraction ring. This aberration is non-stationary and cannot be due to non-common path aberrations which are static.

The temporal spectra of the turbulent phase modes (for i>2) clearly prove the existence of a slowly evolving and non fully developed turbulence. This turbulence is probably due to temperature gradients above the primary mirror. It could be the cause of the residual aberration through the contamination of the estimated residual mirror phase by the high-order non-corrected modes.

This residual aberration limits the photometric precision we can expect on the ADONIS system. When the aberration is variable, the error on the flux ratio of a binary system is important. For flux ratios larger than 4, it is of the order of 10-40% at the location of the first diffraction ring and 5-15% for larger distances. When it is stable, the error is much smaller, of the order of 3-10%.

Reconstructing the PSF from AO data obtained simultaneously to the object image allows to optimize telescope time and to reach higher photometric precisions. It is therefore important to pursue the PSF reconstruction method. However, until the source of the local turbulence has not been identified and removed, doing accurate photometry and detecting faint structures without ambiguity on images obtained with ADONIS will be difficult.

Acknowledgements

We wish to thank F. Marchis, D. Le Mignant, O. Marco, D. Bonaccini and Jean-Luc Beuzit for providing us with the observational data. We thank G. Rousset for his useful comments, E. Gendron for his help to get the necessary information about ADONIS and J.P. Véran for clarifying discussions about the PSF reconstruction.