Null lenses are designed for testing of aspherical surfaces, which are mirrors of the off-axis Three Mirror Anastigmat(TMA) camera designed for remote sensing. Most of null lenses are designed in the autostigmatic type of which null lens produces a reference aspheric wave front which is compared interferometrically with the aspheric surface under test. It has the advantage that the null lens can be consisted of small lenses. Some such as parabolic surfaces are tested in the autocollimation type. This type of null lens makes the collimated wave front in combination with the aspheric surface under test. It is twice as sensitive as the autostigmatic type because the test beam is reflected twice on the test surface by a reference flat. However the size of the reference flat should be larger than that of the aspheric surface under test. Therefore, when the aspheric mirror to be tested is large, it is difficult to adapt this type of null lens.
To eliminate the need for a large reference flat, a new type of null lens system, the mixed type, is presented in this thesis. It has the advantage of both the autosigmatic and the autocollimation type of null lenses. The reflected wave front from the test surface is collimated after inversely passing through a null lens. After reflected from the reference flat, the beam goes back along the same path. As a result, the beam is reflected twice on the test surface so that it has 4 times configuration error of the test surface. Therefore, the null lens can be consisted of small lenses and has the same sensitivity as the autocollimation type even if it has a small flat mirror.
Two aspheric mirrors of TMA has the hyperbolic surface and the elliptical surface respectively. The former has a radius of curvature of 3056.65 mm, conic constant of -1.59658 and diameter of 560 mm. The latter has a radius of curvature of 1420.834 mm, conic constant of 0.19646 and diameter of 468 mm.
The null lens is designed by using an optical design tool, SIGMA 2000. In null lens design, the four considerations has been kept as a design goal. First, it should test aspheric surface on axis. Second, it should measure up to a semi-aperture on an aspheric surface. Third, it should be consisted of two components for easy alignment. Forth, null lens should test the aspheric surfaces with required accuracy.
Except the autocollimation type of null lens which needs a large reference flat, both the autostigmatic and the mixed type of null lenses are designed and compared in the sensitivity of the lens parameters. Through the detailed analyses of its relative sensitivity with respect to the change of the surface parameter of optical elements, more proper null lenses to test each aspheric surfaces under test are chosen.
In case of hyperbolic surface, the autostigmatic type is similar to the null lens system of Offner. It is consisted of two lenses and has a point source at infinity. It has OPD about λ/100. The mixed type is consisted of a concave lens and a donut mirror. It has an intermediate image point between a concave lens and test mirror. The position of donut mirror is located at the position of point source to minimize the portion that can not be measured due to the hole of the donut mirror. It has OPD about λ/30. Investigation of relative sensitivity with respect to the change of lens surface parameters gives the result that the autostigmatic type is better than the mixed type in fabrication and measurement setup.
In case of elliptical surface, the autostigmatic type is similar to a zoom null lens system of Shafer. It is consisted of two negative lenses and has a separation 12 mm between lenses for easy alignment. It has OPD less than λ/100. The mixed type is consisted of a concave lens and a donut mirror. It has the similar configuration with the previous mixed type of null lens, but It has not an intermediate image point between a concave lens and test mirror. It has OPD near λ/100. In this case, the mixed type is better than the conventional autostigmatic type in fabrication and measurement setup.
This new type can be applied to design the null lens to measure the other aspheric surfaces. Also, the four considerations as a design goal and the relative sensitivity as a criteria for choosing the proper null lens can generally be used in the null lens design for testing of aspherical surfaces.
