In LISA Pathfinder and LISA the position fluctuations of drag free test masses will be determined interferometrically to picometer precision. To this end, laser light is brought to interference on an ultra stable optical bench after being reflected on the test mass, which needs to be in an ultra-high vacuum. The present baseline for both missions includes a separate vacuum enclosure for each test mass, so that the sensing laser beam has to pass through an optical window. This window is therefore a transmissive element in the interferometer and its associated pathlength fluctuations are potentially significant. We have selected an athermal glass that should minimize the thermally induced pathlength changes. Several prototype windows, both mounted and unmounted, have been produced and characterized. The pathlength sensitivity to both temperature fluctuations and temperature gradients has been measured with a dedicated interferometer prototype. We have also compared the long-term stability of the LISA Technology Package interferometer when an optical window is present in the optical path to the situation without window. Finally, glass samples have been radiated and the absorption in the glass after the radiation tests has been measured to be negligible at the wavelength of interest (1064 nm). We present here the results of our measurements, which indicate that using a window does not influence the interferometer performance.
