No abstract

Diffractive optical elements, designed by computer optimization techniques, can provide high coupling efficiencies between components in optical communications systems, and can be reproduced accurately in large numbers by the same technology employed to make microcircuits. This applications-led development was for optimized f/0.48 Fresnel lenses to couple light from arrays of semiconductor lasers into monomode optical fibers. The lenses were fabricated as four phase level structures in fused quartz using electron-beam lithography and reactive ion etching. The low f number needed to match between the laser and fiber numerical apertures required 0.2 μm feature sizes in the outer region of the lenses etched 2.1 μm deep. Effective use was made of focused ion beam etching and imaging to obtain cross sections of these high aspect ratio structures during process development. Differing lens designs have achieved 34% and 50% coupling efficiencies between 30° full width half peak 1.55 μm lasers and cleaved, monomode system fibers. In the former case, lens to fiber alignment tolerance was ±6 μm, making passive assembly of the lens fiber arrays feasible.

The step into the production line environment is a quantum leap for physical failure analysis (PFA) and will change its work in the near future. Wafer sacrifice for analysis becomes obsolete. The main benefits are: 1. reduction of wafer costs, 2. more splits per development lot, 3. reduced cycle time of analysis and technology development. Machines needed for that purpose are dual beam SEM/FIB tools. In the following we present solutions how PFA in a broad range can be carried out inside of a production line. The analyzed wafers can be fed back into the production flow which results in lower overall costs and the feedback loop to production engineers is dramatically shortened leading to reduced down times of production tools etc. The highest risk that has kept the majority of semiconductor manufacturers from proceeding into this direction is the contamination of the productive wafer with Ga, the FIB beam particle, that may diffuse into productive parts of the wafer during heat cycles after the analysis step. We show that the risk of contamination by Ga and other materials can be controlled.