Design of Extremely Proximity Gettering Using Hydrogen Ion Implantation for Si CMOS Image-Sensor

As the pixel size in CMOS image-sensor has scaled down less than 1 μm, photo sensitivity which is a ratio of photo and dark current has been important. To improve photo sensitivity in CMOS image-sensor, the achievement of lower dark current is essential. Leakage current is influenced by crystallographic, trapped charge at the SiO₂/Si interface, structural defects, and especially metallic contaminant impurities. Also, these contaminant impurities degrade the minority-carrier recombination lifetime.

 Thus, we investigated the dependency of the properties of CMOS image-sensor on metallic contamination. Standard aqueous solutions of metallic ion were dropped on the surface of silicon wafers and the wafers were spin dry. They were heated to drive the metallic ion contaminants into the silicon bulk. After that, we fabricated 4-Tr CMOS image-sensor cells. In silicon bulk wafer, the minority-carrier recombination lifetime decreased with increasing a metallic contaminant concentration. These contaminations result in degradation of dark current, sensitivity of photodiode and sensing margin of CMOS image-sensor. These results indicate that metallic contaminants are critical for increasing dark current because metallic contaminants remain inside the photo-diode region, ~3 μm.

 In our study, we applied a proximity relaxation-type gettering method. We formed the nano-cavities close to the depletion region using hydrogen ion implantation, as shown in Fig 1. These cavities getter metallic impurities on the cavities wall. Sensing margin of CMOS image-sensor implanted by hydrogen ion was dramatically improved, as shown in Fig 2. In addition, although metallic contaminant concentration increased, the device performances of CMOS image-sensor were constant compare with that of bulk wafer which were not implanted (Fig. 3). These results indicates that proximity gettering by using hydrogen ion implantation was effective in metal gettering for image-sensor. We investigated the correlation of metallic contaminant and sensing margin of CMOS image-sensor. Finally, we will suggest an extremely proximity gettering method for removing metallic contaminants of CMOS image-sensor.

* This work was financially supported by the Brain Korea 21 Plus Program in 2013 and SiWEDS (Silicon Wafer Engineering and Defect Science).