Reduction of White Spot Defects in CMOS Image Sensors Fabricated Using Epitaxial Silicon Wafer with Proximity Gettering Sinks by CH2P Molecular Ion Implantation

Abstract

Using a new implantation technique with multielement molecular ions consisting of carbon, hydrogen, and phosphorus, namely, CH2P molecular ions, we developed an epitaxial silicon wafer with proximity gettering sinks under the epitaxial silicon layer to improve the gettering capability for metallic impurities. A complementary metal-oxide-semiconductor (CMOS) image sensor fabricated with this novel epitaxial silicon wafer has a markedly reduced number of white spot defects, as determined by dark current spectroscopy (DCS). In addition, the amount of nickel impurities gettered in the CH2P-molecular-ion-implanted region of this CMOS image sensor is higher than that gettered in the C3H5-molecular-ion-implanted region; and this implanted region is formed by high-density black pointed defects and deactivated phosphorus after epitaxial growth. From the obtained results, the CH2P-molecular-ion-implanted region has two types of complexes acting as gettering sinks. One includes carbon-related complexes such as aggregated C–I, and the other includes phosphorus-related complexes such as P4–V. These complexes have a high binding energy to metallic impurities. Therefore, CH2P-molecular-ion-implanted epitaxial silicon wafers have a high gettering capability for metallic impurities and contribute to improving the device performance of CMOS image sensors. (This manuscript is an extension from a paper presented at the 6th IEEE Electron Devices Technology & Manufacturing Conference (EDTM 2022)).

Deep Learning Based Compression with Classification Model on CMOS Image Sensors

“…The succeeding main function was minimizing to a finetune the network to softmax state in Eq. ( 14) [16,17]. where, W and b refers the weight and bias of entire deep network, 𝐽(𝑥, 𝑧) represents the LR cost amongst the classifier attained with input feature x and unsupervised outcome 𝑧̂ and W smc implies the weight and λ smc stands for the weight decomposed parameter [18,19].…”