Detection of Ultralow Light Power Back-Reflected or Back-Scattered by Optical Components Using Balanced Low-Coherence Interferometry

We use balanced optical low-coherence interferometry to measure the ultralow light power back-reflected and/or back-scattered by optical interfaces. Indeed, backscattered light from optical surfaces can be a critical source of noise in interferometric gravitational-wave detectors, such as the Laser Interferometer Gravitational-Wave Observatory (LIGO), Virgo, or the Laser Interferometer Space Antenna (LISA), and therefore, it is required to characterize, with high sensitivity, the scattering properties of elementary components involved in their optical design. The use of a broadband light source makes an easy identification of the scattering interface possible, while the implementation of balanced interferometric detection allows very low light levels to be detected, thanks to the simultaneous use of coherent gain and source intensity noise suppression. Experimental results are reported, first on bare glass optical windows (N-BK7 and S-LAH66) and then on a silver-coated mirror, that demonstrate the ability of this setup to quantify specular- or diffuse-reflectance coefficients as low as ${10}^{-11}$.

Figure 1

Optical layout of the experimental setup.

Figure 2

Multiple reflections within a bare glass window.

Figure 3

Theoretical time dependence of the voltage provided by the output channel of the Nirvana receiver.

Figure 4

Frequency filtering. (a) Power spectrum of voltage $V(t)$. (b) Enlarged view of this power spectrum around the carrier frequency. $f_c$ (red curve), and profile of the filter (green curve).

Figure 5

SLD power spectral density for both driving currents (cyan curve, experimental data at 160 mA; blue curve, Gaussian fit at 160 mA; green curve, experimental data at 180 mA; red curve, Gaussian fit at 180 mA).

Figure 6

Experimental measurement of the time dependence of voltage, $V$, recorded during translation of the hollow retroreflector at constant speed, $v$, of 0.5 mm/s for the 1-mm-thick N-BK7 window.

Figure 7

Enlarged views of the five detected echoes for the 1-mm-thick N-BK7 window.

Figure 8

Time dependence of voltage, $V$, for the 2-mm-thick S-LAH66 window.

Figure 9

Enlarged views of the five detected echoes for the 2-mm-thick S-LAH66 window.

Figure 10

Signal envelope extraction and modeling. (a) First echo recorded on a 1-mm-thick N-BK7 window (violet curve); data envelope obtained using a Hilbert transform (black curve). (b) Comparison between the normalized data envelope (black curve), a Gaussian fit (red curve), and the result of a Fourier transform of the SLD PSD Gaussian fit (blue circles).

Figure 11

Synthetic comparison between experimental results and theoretical data for both glass windows (for N-BK7: theoretical data, black line; experimental data, colored disks; for S-LAH66, theoretical data, gray line; experimental data, colored squares; noise floor, dashed line).

Figure 12

Tunable-splitting-ratio scheme.

Figure 13

Simulation of the angular dependence of the light back-scattered by a protected silver mirror (rms roughness $\approx 4$ nm) for three states of polarization.

Figure 14

Simulation of the angular dependence of the back-reflectance, $\rho$, of a protected silver mirror (magenta curve, retro-reflection; red curve, back-scattering).

Figure 15

Back-scattered signal, $V_{\mathrm{ac}}$, measured from a protected silver mirror at different tilt angles: (a) raw data, (b) filtered data.

Figure 16

Experimental measurement of the angular dependence of the back-reflectance, $\rho$, of a protected silver mirror (log-log units).