A diffractive optical element (DOE) consists essentially of a microstructure with (2D) local variation of grating period and (3D) relief or phase profile. The 2D-design, which can be performed by optical raytracing, determines the optical function of the element, the 3D-design determines its diffraction efficiency. A great number of promising system concepts with DOE are known, but the critical point still remains the masterizing of manufacture methods suitable for industrial applications of DOE. Different DOE manufacture methods are under investigation at CSEM. They have already been applied to the manufacture of test and prototype elements. Well masterized technologies, like the interferometric recording of transmission type holographic optical elements (HOE) in dichromated gelatin or photoresist are currently used to fabricate application specific elements. Typical spatial frequencies are 1700 - 2000 1/mm with an efficiency ranging from 50 to 90%, where the best values were still achieved with DCG elements. First elements were also recorded in photopolymer, a very promising photosensitive material for HOE. For reflection type elements the preference has been given to technologies derived from microelectronics silicon wafer processing. CSEM's electron-beam writer is routinely used for generating basic (2D) DOE patterns in the form of standard chromium-on-glass masks, which are, in optical terminology, binary computer generated holograms (CGH). These 2D patterns were transferred onto several types of substrates (silicon, quartz, glass, copper) by contact copy photolithography and subsequent wet or dry etching in order to obtain a 3D profile. The highest spatial frequency routinely transferable with this standard process is presently limited to 500 1/mm. For higher spatial frequencies (up to 1500 1/mm) the basic pattern is directly written by e-beam on the photoresist coated substrate and subsequently processed. Ion exchange in glass is an interesting technology for DOE because the diffractive structure is imbedded in the substrate and thus protected from dirt and dust. For this process, compatibility with a microelectronic environment is much more difficult to achieve. First results already show that a considerable technology effort would be necessary in order to overcome the observed limited spatial resolution of 200 1/mm.
|