Measuring parameters of large-aperture crystals used for generating optical harmonics

The purpose of this project was to develop tools for understanding the influence of crystal quality and crystal mounting on harmonic-generation efficiency at high irradiance. Measuring the homogeneity of crystals interferometrically, making detailed physics calculations of conversion efficiency, performing finite- element modeling of mounted crystals, and designing a new optical metrology tool were key elements in obtaining that understanding. For this work, we used the following frequency-tripling scheme: type I second- harmonic generation followed by type II sum-frequency mixing of the residual fundamental and the second harmonic light. The doubler was potassium dihydrogen phosphate (KDP), and the tripler was deuterated KDP (KD*P). With this scheme, near-infrared light (1053 nm) can be frequency tripled (to 351 nm) at high efficiency (theoretically >90%) for high irradiance (>3 GW/cm²). Spatial variations in the birefringence of the large crystals studied here (37 to 41 cm square by about 1 cm thick) imply that the ideal phase-matching orientation of the crystal with respect to the incident laser beam varies across the crystal. We have shown that phase-measuring interferometry can be used to measure these spatial variations. We observed transmitted wavefront differences between orthogonally polarized interferograms of {lambda}/50 to {lambda}/100, which correspond to index variations of order 10^-6. On some plates that we measured, the standard deviation of angular errors is 22-23 µrad; this corresponds to a 1% reduction in efficiency. Because these conversion crystals are relatively thin, their surfaces are not flat (deviate by k2.5 urn from flat). A crystal is mounted against a precision-machined surface that supports the crystal on four edges. This mounting surface is not flat either (deviates by +2.5 µm from flat). A retaining flange presses a compliant element against the crystal. The load thus applied near the edges of the crystal surface holds it in place. We performed detailed finite-element modeling to predict the resulting shape of the mounted crystal. The prediction agreed with measurements of mounted crystals. We computed the physics of the frequency-conversion process to better quantify the effects on efficiency of variation in the crystal� s axis, changes in the shape of the crystal, and mounting-induced stress. We were able to accurately predict the frequency-conversion performance of 37-cm square crystals on Beamlet, a one-beam scientific prototype of the NIF laser architecture, using interferometric measurements of the mounted crystals and the model. In a 2{omega} measurement campaign, the model predicted 64.9% conversion efficiency; 64.1% was observed. When detuned by 640 µrad, the model and measurement agreement is even better (both were 10.4%). Finally, we completed the design and initial testing of a new optical metrology tool to measure the spatial variation of frequency conversion. This system employs a high-power subaperture beam from a commercial laser oscillator and rod amplifier. The beam interrogates the crystal� s aperture by moving the crystal horizontally on a translation stage and translating the laser beam vertically on an optical periscope. Precision alignment is maintained by means of a full-aperture reference mirror, a precision-machined surface on the crystal mount, and autocollimators (the goal for angular errors is 10 µrad). The autocollimators track the mounting angle of the crystal and the direction of the laser beam with respect to the reference mirror. The conversion efficiency can be directly measured by recording l{omega}, 2{omega}, 3{omega} energy levels during the scan and by rocking (i.e., tilting) the crystal mount over an angular range