Search

Metasurface coatings on a free-standing SiN thin film membrane are fabricated on a Si substrate using masked lithography and CMOS-compatible surface micromachining. The result is a band-limited absorber for the mid-IR, which is part of a microstructure that is attached to the substrate by long and slender suspension beams to provide thermal isolation. As a residual of the fabrication, the regular pattern of sub-wavelength unit cells of 2.6 μm side length, which defines the metasurface, is interrupted by an equally regular array of sub-wavelength holes of 1–2 μm diameter and at 7.8–15.6 μm of pitch. This array of holes is essential for enabling access of the etchant and attack of the underlying layer during fabrication, which ultimately results in the sacrificial release of the membrane from the underlying substrate. As the plasmonic responses of the two patterns interfere, a maximum is imposed on the hole diameter and a minimum on the hole-to-hole pitch. However, the hole diameter should be sufficiently large to allow access of the etchant, while the maximum spacing between holes is set by the limited selectivity of the different materials to the etchant during sacrificial release. The effect of the parasitic hole pattern on the spectral absorption of a metasurface design is analyzed by simulations of the responses of combined holes–metasurface structures. Arrays of 300 × 180 μm2 Al-Al2O3-Al MIM structures are mask-fabricated on suspended SiN beams. The results show that the effect of the array of holes can be disregarded for a hole-to-hole pitch larger than 6 times the side length of the metamaterial until cell, while the diameter of the hole should remain smaller than about 1.5 μm, and their alignment is critical.

The use of masked UV (i-line) lithography in a MEMS foundry for CMOS-compatible fabrication of large-area metasurface-based absorbers for the mid-infrared is demonstrated. The challenges are in: (a) the limited number of acceptable metals, (b) the thickness tolerance of the layers used in the CMOS process, and (c) the imaging capabilities of i-line lithography as compared to e-beam. The claimed throughput advantage with manageable distortion of masked lithography and the suitability of the layers used in a CMOS-compatible process in the fabrication of mid-IR absorbers was tested. Issues investigated are: (a) the impact of aluminium as the preferred metal in the MIM patch on the plasmonic response, (b) the influence of SiO2 as preferred dielectric material, (c) the effect of corner rounding and horizontal-vertical bias of the masked lithography and (d) measures that can be taken during the design phase to mitigate any such detrimental effect. Based on the findings, disk-shaped patches are identified as the most suitable for shape-tolerant design. Metasurfaces with a unit-cell side-length of 3μm were fabricated over a chip area larger than 105μm2. Measurements do confirm that (a) aluminium is a suitable CMOS-compatible material for mid-IR metamaterial absorber fabrication, (b) a large surface roughness results in widening of the absorption peaks and (c) the typical layer thickness tolerance used in a MEMS foundry is also acceptable for mid-IR metasurface fabrication. Masked lithography limits the minimum design wavelength to about 3.5μm, while the surface roughness Rq ~ 5nm results in a bandwidth up to FWHM = 400nm.

Apps have been used for providing additional user functionality on demand to the smartphone. Several highly promising medical apps have been introduced that rely on the build-in camera. However, some of these apps fall short of the expectations, due to the inherent limitation of the camera to operation in the visible spectral range. The more detailed spectral information that an optical microspectrometer can offer is needed. Although CMOS-compatible MEMS technologies enable the on-chip integration of the microspectrometer with the CMOS camera, implementing such a step would result in a device that would be considered a customized unit within a universal platform. Consequently, the economic viability would depend on the added value of the microspectrometer in popular applications versus the intrusion of the microspectrometer-specific processing on the imager fabrication. The LVOF-based microspectrometer is introduced as a CMOS-compatible concept to meet the challenges imposed by the high production volume that is characteristic for the smartphone. The state-of-the-art of both the interactive optical app and the LVOF-based microspectrometer are presented.

Conventional thermal conductivity detectors (TCDs) demonstrate a flow dependence. The approach presented here to reduce the flow dependence is based on the on-line flow compensation using two thin-film sensors on membranes in parallel on the same chip that are differentially operated. These are laterally identically, but with a different depth of the detection chamber, resulting in different quasi-static sensitivities to the thermal conductivity of the sample gas. The effects of conduction and convection in the structure have been studied using COMSOL Multiphysics. First prototypes have been fabricated and are presently tested.