Example Optimum Aberration Targets

Three radially symmetric and two first order rotational test targets were systematically tested for sensitivity and orthogonality by observing the change in the onaxis image intensity from the unaberrated cases in the presence of 0.05 waves peak of the corresponding 5 aberrations. The results are shown in Figure 4. Here each column corresponds to one of the test targets and each row corresponds to an aberration condition. The entries for the “no aberration” row are the reference levels of the peak value at the center of the target normalized to a clear field intensity of 1. The entries for the aberration rows show the change in this peak value as a fraction of the clear field value due to the introduction of 0.05λ peak of that aberration. The results are grouped by rotational symmetry of the aberrations to emphasize the radial effects within each symmetry class. The goal for the unaberrated peak value was 20 to 25% of the clear field intensity and three of the targets meet this goal. Unfortunately the targets with the lowest order radial variation are considerably higher (at 0.555 for coma and 0.933 for defocus), owing to their very limited on-axis null. The diagonal shows the sensitivity of each target to the aberration for which it was designed. In all cases the targets have been phased such that the aberration increases the intensity at the probe. The magnitudes of the increases are impressively large at 0.390 to 0.823. When normalized to rms, all five cases show a sensitivity of between 0.27 and 0.36 per 0.01λ rms of the associated aberration. Most of the off diagonal terms are negligible and even the most similar aberrations are in the worst case only 1/7 as large as the main effect. We propose to quantify aberration levels in lenses by measuring the central intensity peaks of these targets directly with optical detectors or through inspecting photoresist images printed at various exposure doses. Simpler targets may be used in practice with some but not a large loss of sensitivity and orthogonality. Mask making tolerance requirements have been examined for simpler targets elsewhere [13] and are generally within the capability of mask making. The intensity variation of the central node for the five targets studied was impressive at 1/3 of the full clear field intensity per 0.01 λ of rms aberration. The cross contamination by closely related aberrations was about 6 times smaller and it is thus possible to measure spherical aberration while using a single defocus setting.