4 Conclusion

The price to pay for the perfect spatial filtering of a single-mode fiber is a loss of photons. To obtain the best possible coupling efficiency (78%) one needs a diffraction-limited beam with no central obstruction; yet, even in perfect imaging conditions, the central obstruction of the secondary in most 4-meter class telescopes reduces the maximum efficiency down to 42%. When the central obstruction ratio is less than 20%, which is the case for the new generation of 8 m class telescopes, this reduction remains acceptable.

When the fiber is fed with turbulent starlight, then the coupled power displays strong temporal fluctuations. It was shown that the coupling efficiency is then directly linked to the Strehl ratio of the image, or equivalently, to the integral of the pupil autocorrelation. The total amount of energy that can be injected into the fiber is thus limited by the size r₀ of a coherence area on the pupil.

For a large telescope ($d \gg r_0$) the coupling efficiency can be dramatically improved by (at least partially) correcting the incoming wavefront with adaptive optics. Measuring this efficiency is also an excellent means to monitor the real-time performance of an adaptive optics system. This is how the coupling of a single-mode fiber to the ADONIS adaptive optics system on the 3.6m ESO telescope in La Silla enabled us to detect an instrumental periodic perturbation at 25Hz, which was too fast to have been detected otherwise, yet which was strong enough to induce up to a 70% degradation of the average Strehl ratio. Thus it appears that single-mode fibers can be a unique tool to assess the dynamic evolution of the optical quality in modern telescopes.