Estimate of tilt instability of mesa-beam and Gaussian-beam modes for advanced LIGO
Annotatsiya
Sidles and Sigg have shown that advanced LIGO interferometers will encounter a serious tilt instability, in which symmetric tilts of the mirrors of an arm cavity cause the cavity's light beam to slide sideways, so its radiation pressure exerts a torque that increases the tilt. Sidles and Sigg showed that the strength $T$ of this torque is 26.2 times greater for advanced LIGO's baseline cavities---nearly flat spherical mirrors which support Gaussian beams (FG cavities), than for nearly concentric spherical mirrors which support Gaussian beams (CG cavities) with the same diffraction losses as the baseline case: ${T}^{\mathrm{FG}}/{T}^{\mathrm{CG}}=26.2$. This has motivated a proposal to change the baseline design to nearly concentric, spherical mirrors. In order to reduce thermal noises in advanced LIGO, O'Shaughnessy and Thorne have proposed replacing the spherical mirrors and their Gaussian beams by ``Mexican-Hat'' (MH) shaped mirrors which support flat-topped, mesa shaped beams. In this paper, we compute the tilt-instability torque for advanced-LIGO cavities with nearly flat MH mirrors and mesa beams (FM cavities) and nearly concentric MH mirrors and mesa beams (CM cavities), with the same diffraction losses as in the baseline FG case. We find that the relative sizes of the restoring torques are ${T}^{\mathrm{CM}}/{T}^{\mathrm{CG}}=0.91$, ${T}^{\mathrm{FM}}/{T}^{\mathrm{CG}}=96$, ${T}^{\mathrm{FM}}/{T}^{\mathrm{FG}}=3.67$. Thus, the nearly concentric MH mirrors have a weaker tilt instability than any other configuration. Their thermoelastic noise is the same as for nearly flat MH mirrors, and is much lower than for spherical mirrors.