New metrology techniques improve the production of silicon diffractive optics
New metrology techniques improve the production of silicon diffractive optics
复制标题
新计量技术提高了硅衍射光学器件的生产
DOI:
--
复制
发表时间:
2014
期刊:
影响因子:
--
通讯作者:
D. Jaffe
中科院分区:
文献类型:
--
作者:
C. Brooks;M. Gully;M. Grigas;D. Jaffe
Silicon immersion gratings and grisms offer significant advantages in compactness and performance over frontsurface gratings and over grisms made from lower-index materials. At the same time, the high refractive index of Si (3.4) leads to very stringent constraints on the allowable groove position errors, typically rms < 20 nm over 100 mm and repetitive error of <5 nm amplitude. For both types of devices, we produce grooves in silicon using photolithography, plasma etching, and wet etching. To date, producers have used contact photolithography to pattern UV sensitive photoresist as the initial processing step, then transferred this pattern to a layer of silicon nitride that, in turn, serves as a hard mask during the wet etching of grooves into silicon. For each step of the groove production, we have used new and sensitive techniques to determine the contribution of that step to the phase non-uniformity. Armed with an understanding of the errors and their origins, we could then implement process controls for each step. The plasma uniformity was improved for the silicon nitride mask etch process and the phase contribution of the plasma etch step was measured. We then used grayscale lithography, a technique in which the photoresist is deliberately underexposed, to measure large-scale nonuniformities in the UV exposure system to an accuracy of 3-5%, allowing us to make corrections to the optical alignment. Additionally, we used a new multiple-exposure technique combined with laser interferometry to measure the relationship between UV exposure dose and line edge shift. From these data we predict the contribution of the etching and photolithographic steps to phase error of the grating surface. These measurements indicate that the errors introduced during the exposure step dominate the contributions of all the other processing steps. This paper presents the techniques used to quantify individual process contributions to phase errors and steps that were taken to improve overall phase uniformity.