Melted (reflow) microlens array formation

The development of an inexpensive, reproducible and highly accurate method for the production of microlens arrays is of vital importance to the next generation of optics based technologies. Currently there are a large number of companies using many different techniques to produce these arrays, however no single technique has proven itself capable of providing the necessary range of lens properties. One technique that has been used to produce arrays is the reflow of photoresist method first suggested by Popovic et al. in 1988. This technique utilises mature fabrication patterning technology, i.e. photolithography, however when lenses are produced they are often found to have large deviations from the spherical case. This limits the range of f-numbers,f/#, that can be produced. There are a number of polynomial models that have been presented in the literature that are used to model the resulting lens profiles, however these models are not based on the physical nature of the fabrication technique and hence provide no information about the processes involved. One model that attempts to model the profile from a physical basis has been reported in the literature, the Curvature Correction Model, (CCM), by Abe and Sheridan in 1999. This model has been shown to provide qualitative fits to the I-D case (cylindrical lenses). In this paper we discuss our current work in generalising the model to the 2-D case (spherical lenses).