Optical confinement methods for continued scaling of CMOS image sensor pixels

The pixels that make up CMOS image sensors have steadily decreased in size over the last decade. This scaling has two effects: first, the amount of light incident on each pixel decreases, making optical efficiency, i.e., the collection of each photon, more important. Second, diffraction comes into play when pixel size approaches the wavelength of visible light, resulting in increased spatial optical crosstalk. To address these two effects, we investigate and compare three methods for guiding incident light from the microlens down to the photodiode. Two of these techniques rely on total internal reflection (TIR) at the boundary between dielectric media of different refractive indices, while the third uses reflection at a metal-dielectric interface to confine the light. Simulations are performed using a finite-difference time-domain (FDTD) method on a realistic 1.75-mum pixel model for on-axis as well as angled incidence. We evaluate the optical efficiency and spatial crosstalk performance of these methods compared to a reference pixel and find significant (10%) improvement for the TIR designs with properly chosen parameters and nearly full spatial crosstalk elimination using metal to confine the light. We also show that these improvements are comparable to those achieved by thinning the image sensor stack.