362
J. Eur. Opt. Society-Rapid Publ. 22, 36( 2026)
Figure 9. Optical setup of FrObSy. Left: The exposure beams interfere in the focal plane of the microscope objective lens. Inlet: An image of the writing spot is shown. Right: Experimental integration of the observation system, mounted on the platform of the NPMM-200. The distance between the objective of the lithography head and the objective of FrObSy is about 3 mm.
Figure
10. Evaluation steps of the fringe observation system. The carrier-frequency method is applied to a region of interest in the camera image( indicated by the red rectangle). The main frequency peak of the 2D FFT is shifted to the center of the Fourier space and all other frequencies are set to zero. After inverse 2D FFT, the phase can be calculated from the imaginary part and the real part. The wavefront offset, which corresponds to the fringe movement of the fringe pattern, then can be calculated by the mean value.
4 Alignment of an SBIL system
For the alignment of the SBIL system in Section 3.1, we created a well defined alignment protocol, which covers all necessary degrees of freedom to align the setup and to characterize the positioning errors for the scan-and-stitch process. The alignment protocol is visualized in Figure 11. The different steps are explained in detail in the following sections.
4.1 Alignment of the fringe pattern orientation
At first, the orientation c of the fringe pattern is aligned. The NPMM-200 axis will move along the x-axis with a linear trajectory Dx. During movement, the fringes on the camera shift. As illustrated in Figure 12, this leads to a measurable phase offset D /( see Eq.( 15)) depending on c. Figure 13 shows the measurement data of the offset over time.
The measurement of D / over Dx is shown in Figure 14. A linear function is fitted to D /( Dx) to determine the
gradient m x. The fringe pattern orientation c can be calculated by
c ¼ arctan / p
¼ arctan m x p
; ð17Þ
2px 2p
where p is the period of the fringe pattern.
At this point, the fringe pattern period p and the pattern rotation by the rotary drive are still not known with sufficient accuracy, which prevents a direct determination of the orientation angle c. To overcome this, the lithography head is incrementally rotated by dc while monitoring the phase D /( Dx). The rotation measurement is continued until the measured gradient m x, representing the phase change per displacement Dx, vanishes. A zero gradient indicates that the scan direction is orthogonal to the fringe orientation, and thus the residual orientation error of the lithography head has been eliminated. Figure 15 shows a measurement with sufficiently minimized gradient m x = 0.018 mrad / mm with a fitting uncertainty of ± 0.015 mrad / mm.