Shanghai Huali Microelectronics Corporation

The present application discloses a method for preparing a pixel cell of a CMOS image sensor. Process optimization and adjustment for multiple times of ion implantation in a pixel area of the CMOS image sensor are carried out. That is, photodiode N-type ion implantation is performed before formation of a polysilicon gate of each MOS transistor of a CMOS pixel readout circuit, a photoresist open area is enlarged by means of a photoresist dry etching descum process such that an N-type ion implantation area is enlarged, and second photodiode N-type ion implantation is performed on the basis of the photoresist dry etching descum process, so as to achieve two times of photodiode N-type ion implantation using the same mask, with different depths and different pattern sizes, thereby forming an N-type area of the photodiode that tapers to the bottom, saving a mask layer, and reducing photolithography steps.

The present disclosure provides a CMOS image sensor and a pixel structure thereof, and a method for manufacturing a deep trench isolation grid structure in the pixel structure. The method for manufacturing the deep trench isolation grid structure comprises: depositing a first isolation layer and a second isolation layer sequentially on the side walls and bottom surface of each deep trench; and depositing a third isolation layer that fills each deep trench on the upper surface of the second isolation layer, so that the first isolation layer, the second isolation layer and the third isolation layer in the plurality of deep trenches constitute the grid. The deep trench isolation grid structure formed by the method can effectively reduce electrical crosstalk between adjacent grid lines, thereby improving the device performance of the CMOS image sensor which is built upon the deep trench isolation grid structure and the pixel structure thereof.

The present application discloses a CMOS image sensor. A pixel cell circuit comprises a photodiode and a CMOS pixel readout circuit. The pixel cell circuit is formed on an SOI substrate, and the photodiode is formed on a bottom semiconductor substrate. The CMOS pixel readout circuit is formed on a top semiconductor substrate. A photo-induced carrier of the photodiode is connected to the CMOS pixel readout circuit by means of an electrotransfer structure passing through a dielectric buried layer. The present application also discloses a method for manufacturing a CMOS image sensor. The present application can increase a pixel cell density without reducing a photodiode area, thus achieving an ultra-high CMOS image sensor density and improving the device quality.

The present application discloses a double-layer stacked CMOS image sensor, photo diode and transfer gate transistor of a pixel cell are formed on the first substrate sequentially along a longitudinal direction, and the other pixel transistors of the pixel cell are formed on the second substrate. The first substrate and the second substrate are packaged separately, and the second substrate is stacked on the top side of the first substrate instead of being in juxtaposition. Since the photo diode and the pixel transistors other than the transfer gate transistor of the pixel cell are located on two separate substrates respectively, the area of a photo diode region may be increased significantly, thereby greatly increasing full well capacitance of the image sensor and increasing a dynamic range, and reduce a dark current and image noise significantly, thereby improving the dark line noise and full well capacitance simultaneously.

The disclosure discloses a global shutter CMOS image sensor, which adopts non-uniform storage diffusion region doping to reduce the junction leakage at storage points, so as to ensure that with the increase of the depth of photodiodes and the increase of pixels, all carriers in rows read subsequently can be transferred to storage diffusion regions, the loss of the carriers in the storage diffusion regions is not caused when a global shutter transistor is turned on, and the carriers can be completely transferred from the storage diffusion regions to floating diffusion regions through second transfer transistors even if the number of rows of pixel units increases during reading-out row by row. The disclosure further discloses a method for making the global shutter CMOS image sensor.

The disclosure discloses a global shutter CMOS image sensor, which adopts non-uniform storage diffusion region doping to reduce the junction leakage at storage points, so as to ensure that with the increase of the depth of photodiodes and the increase of pixels, all carriers in rows read subsequently can be transferred to storage diffusion regions, the loss of the carriers in the storage diffusion regions is not caused when a global shutter transistor is turned on, and the carriers can be completely transferred from the storage diffusion regions to floating diffusion regions through second transfer transistors even if the number of rows of pixel units increases during reading-out row by row. The disclosure further discloses a method for making the global shutter CMOS image sensor.

The present invention provides a semiconductor structure for forming a CMOS image sensor. The semiconductor structure includes at least a photodiode formed in the substrate for collecting photoelectrons, and the photodiode has a pinning layer, a first doped region and a second doped region in order from top to bottom in a height direction of the substrate. The semiconductor structure further includes a third doped region located in the substrate corresponding to a laterally extending region of the second doped region. The first doped region has an ion doping concentration greater than the ion doping concentration of the second doped region, the ion doping concentration of the second doped region is greater than the ion doping concentration of the third doped region, and the third doped region is in contact with the second doped region after diffusion. The present invention also provides a method of manufacturing the above-described semiconductor structure.

The present invention provides a semiconductor structure for forming a CMOS image sensor. The semiconductor structure includes at least a photodiode formed in the substrate for collecting photoelectrons, and the photodiode has a pinning layer, a first doped region and a second doped region in order from top to bottom in a height direction of the substrate. The semiconductor structure further includes a third doped region located in the substrate corresponding to a laterally extending region of the second doped region. The first doped region has an ion doping concentration greater than the ion doping concentration of the second doped region, the ion doping concentration of the second doped region is greater than the ion doping concentration of the third doped region, and the third doped region is in contact with the second doped region after diffusion. The present invention also provides a method of manufacturing the above-described semiconductor structure.

The disclosure discloses a global shutter CMOS image sensor, which adopts non-uniform storage diffusion region doping to reduce the junction leakage at storage points, so as to ensure that with the increase of the depth of photodiodes and the increase of pixels, all carriers in rows read subsequently can be transferred to storage diffusion regions, the loss of the carriers in the storage diffusion regions is not caused when a global shutter transistor is turned on, and the carriers can be completely transferred from the storage diffusion regions to floating diffusion regions through second transfer transistors even if the number of rows of pixel units increases during reading-out row by row. The disclosure further discloses a method for making the global shutter CMOS image sensor.

The present application discloses a method for preparing a pixel cell of a CMOS image sensor. Process optimization and adjustment for multiple times of ion implantation in a pixel area of the CMOS image sensor are carried out. That is, photodiode N-type ion implantation is performed before formation of a polysilicon gate of each MOS transistor of a CMOS pixel readout circuit, a photoresist open area is enlarged by means of a photoresist dry etching descum process such that an N-type ion implantation area is enlarged, and second photodiode N-type ion implantation is performed on the basis of the photoresist dry etching descum process, so as to achieve two times of photodiode N-type ion implantation using the same mask, with different depths and different pattern sizes, thereby forming an N-type area of the photodiode that tapers to the bottom, saving a mask layer, and reducing photolithography steps.