Silicon Pore Optics

SPO utilises commercially available Si wafers which have surface figure and roughness quality ideally suited to X-ray optics applications. The process used to create the pore structure from the wafers is illustrated here.

The wafers are diced into rectangles typically 60 mm wide and with varing lengths. A thin wedge is deposited onto the wafer so that when the wafers are stacked the reflecting surfaces are arranged in a radial pattern which provides a common in-plane focus. Regular grooves, with a rectangular profile, are cut leaving a thin membrane of thickness 0.15 mm which supports the entire reflecting surface. The sides of the grooves form parallel ribs which also have a thickness of 0.15 mm. The faces at the tops of the ribs are untouched and when the wafers are pressed together they cold-bond to the surface of the adjacent wafer, without  glueing, forming a rigid block containing an array of very regular, rectangular pores. The reflecting surfaces are coated with high-Z material (e.g. Iridium or Gold) leaving uncoated strips so that the top of the ribs of one wafer match the pristine Si strips in the next wafer allowing the cold-bond to be made securely. Wafers are curved to the appropriate radius of curvature using a precision mandrel so that reflecting surfaces in all the pores match the surface of revolution required in a Wolter I optical system. This curvature provides the out-of-plane focusing between successive pores, while the wedge angle and azimuthal curvature of the wafers ensure that all the pores in a given stack point towards a common vertex.

A SPO module comprises two wafer stacks integrated together. The first stack is a narrow sector of a nest of surfaces of revolution which approximate the paraboloid surfaces (1st reflection) in a nested Wolter I system. Similarly, the second stack provides an approximation to the hyperboloid (2nd reflection) in the nested Wolter I system. The grazing angles of the two reflections are set to be equal so the kink angle between the axis of pores in the 1st stack and the axis of pores in the 2nd stack must be set precisely at twice the grazing angle. Each module acts as an individual lenslet. The full telescope aperture can be populated by arranging the modules in rings.

 

   A recent summary presentation at the SPIE conference is available here