A 1.5Mpixel RGBZ image sensor that simultaneously captures color (RGB) and time-of-flight (ToF) range (Z) images is presented. While ToF sensors are well documented [1,2], few, if any, monolithic sensors have been reported that capture both range and color. In one sensor [3], a combined pixel structure captures color and range, but presumably in sequential fields. This approach does not allow simultaneous capturing of range with color, and the pixel performance cannot be optimized for each mode. In thie paper, we introduce a sensor that is designed to capture color and range simultaneously with individually optimized pixels.It is expected that future mobile-phone cameras and compact DSCs will be able to capture 3D still and video images, and permit gesture control of certain functions. ToF range sensing has several advantages over stereo-imaging or motioncomputed 3D imaging, including single-lens implementation, less-intensive computing requirements, and freedom from occlusion artifacts. Monolithic integration of color and ToF range sensing for a single-lens camera system is important for those consumer applications that require small-form-factor cameras with lowest-possible cost.In our sensor, color (RGB) and range (Z) pixels are tiled across the image plane. Our RGBZ sensor is modified from a 1920(H)×1080(V) 2Mpixel array of standard 2.5T pixel RGBG color pixels with 2.25µm pitch. Every 3 rd row of color pixels is replaced with a row of Z pixels. Each Z pixel is 2.25×9.0µm2 . Thus the RGBZ kernel consists of two RGBG color kernels and one Z pixel. There are, in essence, two interlaced arrays: a 1920(H)×720(V) physical array of color pixels with an embedded 480(H)×360(V) array of range pixels for a total of 1.55Mpixels. A block diagram of the experimental sensor is shown in Fig. 22.7.1. The color pixels are driven from the left and read out at the top of the array. Range pixels are driven from the right and read out at the bottom. Column signal lines are separate for color and range pixels. ADC and ToF signal processing is performed off-chip.Each Z pixel is a pinned photodiode with two transfer gates and two output ports [4,5]. For ranging, the scene is illuminated with 850nm NIR LEDs modulated at 20MHz. The two transfer gates TG1 and TG2 in the Z pixel are modulated 180°o ut of phase with each other at 20MHz. On one phase the optical signal is integrated on a floating diffusion (FD), and on the other phase the signal is discarded using a drain diffusion (DD). This single-tap mode of operation requires four fields to be collected, corresponding to the four quadrature phases. Each phase field is integrated and then read out. The LED modulation phase is shifted for each field. Our single-tap operation eases layout design and reduces fixed-pattern noise compared to traditional two-tap operation, but SNR is reduced by a factor of √2.Organization of the sensor into range-pixel "stripes" allows for easier routing of modulation-control lines to the range pixels and leads to less high-frequencyclock contamination of th...