Enhancing the fill-factor of CMOS SPAD arrays using microlens integration

Arrays of single-photon avalanche diode (SPAD) detectors were fabricated, using a 0.35 μm CMOS technology process, for use in applications such as time-of-flight 3D ranging and microscopy. Each 150 x 150 μm pixel comprises a 30 μm active area diameter SPAD and its associated circuitry for counting, timing and quenching, resulting in a fill-factor of 3.14%. This paper reports how a higher effective fill-factor was achieved as a result of integrating microlens arrays on top of the 32 x 32 SPAD arrays. Diffractive and refractive microlens arrays were designed to concentrate the incoming light onto the active area of each pixel. A telecentric imaging system was used to measure the improvement factor (IF) resulting from microlens integration, whilst varying the f-number of incident light from f/2 to f/22 in one-stop increments across a spectral range of 500-900 nm. These measurements have demonstrated an increasing IF with fnumber, and a maximum of ~16 at the peak wavelength, showing a good agreement with theoretical values. An IF of 16 represents the highest value reported in the literature for microlenses integrated onto a SPAD detector array. The results from statistical analysis indicated the variation of detector efficiency was between 3-10% across the whole f-number range, demonstrating excellent uniformity across the detector plane with and without microlenses.