Wavefront sensing is a process by which optical system errors are deduced from the aberrations in the image of an ideal source. The method has been used successfully in near-normal incidence, but not for grazing incidence systems. This innovation highlights the ability to examine outof- focus images from grazing incidence telescopes (typically operating in the x-ray wavelengths, but integrated using optical wavelengths) and determine the lower-order deformations. This is important because as a metrology tool, this method would allow the integration of high angular resolution optics without the use of normal incidence interferometry, which requires direct access to the front surface of each mirror.

Measuring the surface figure of mirror segments in a highly nested x-ray telescope mirror assembly is difficult due to the tight packing of elements and blockage of all but the innermost elements to normal incidence light. While this can be done on an individual basis in a metrology mount, once the element is installed and permanently bonded into the assembly, it is impossible to verify the figure of each element and ensure that the necessary imaging quality will be maintained. By examining on-axis images of an ideal point source, one can gauge the loworder figure errors of individual elements, even when integrated into an assembly. This technique is known as wavefront sensing (WFS).

By shining collimated light down the optical axis of the telescope and looking at out-of-focus images, the blur due to low-order figure errors of individual elements can be seen, and the figure error necessary to produce that blur can be calculated. The method avoids the problem of requiring normal incidence access to the surface of each mirror segment. Mirror figure errors span a wide range of spatial frequencies, from the lowest-order “bending” to the highest-order “micro-roughness.” While all of these can be measured in normal incidence, only the lowest-order contributors can be determined through this WFS technique.

During integration, typically only the low-order shape changes. The stress introduced does not affect the higher-order ripple or roughness, so one can use the measurements done in normal incidence to characterize the mirror in the mid- and high-frequency domains, and WFS measurements for the low-frequency domain.

By analyzing multiple out-of-focus images at different positions, the path of each photon can be determined, and the figure error necessary to generate that array of photon paths can be deduced. The method is applicable to any wavelength being examined, though the range of spatial periods that can be examined depends on what wavelength of light is being imaged, due to diffraction blurring out the focused image.

The innovation is unique in that it determines physical surface errors using a method that requires neither normal incidence access nor contact of the optical surface. The primary advantage of the technique is the ability to probe surface figure errors when the mirror is in a system that denies access to the front surface of the mirror, such as during xray testing (requiring the mirror to be in a vacuum chamber) or after it has been integrated into a highly nested structure. This software is capable of determining figure errors at the sub-micrometer level for up to 4th order errors.

This work was done by Scott Rohrbach and Timo Saha of Goddard Space Flight Center. GSC-15926-1