This article refers to the address: http://
Abstract: A new solid-state image sensor LBCAST JFET is used in Nikon's D2H SLR digital camera. The device has a great improvement in reading mode and internal structure. Compared with CCD and CMOS image sensors, it has the characteristics of instant start, high sensitivity, high resolution, low energy consumption, high yield and low noise.
Keywords: LBCAST JFET image sensor XY addressing pixel switch color separation noise
introduction
Before the end of 2003, solid-state image sensors were also classified into CCD type and CMOS type, but Nikon Corporation of Japan rewrote this history at the end of 2003, and used a D2H lens-converted anti-single digital camera released in July 2003. A new type of solid LBCAST JFET image sensor (Lateral Buried Charge Accumulator and Sensing Transistor array Junction Field Effect Transistor). It can be said that the product of CCD and CMOS technology fusion, fully embodies the advantages of CMOS low power consumption and CCD high-speed data reading, the size is 23.3mm × 15.5mm, the diagonal length is 28.4mm, the total number of pixels is 4.26 million (2560 × 1664), the effective number of pixels is 4.1 million, and the pixel interval is 9.4 μm.
1 JFET and MOSFET
FETs mainly include junction field effect transistors (JFETs) and insulated gate field effect transistors (IGFETs, which are commonly referred to as MOSFETs due to their metal gates), which have high input impedance, low noise, low power consumption, and heat. High stability and strong radiation resistance. The difference between them is that the conduction mechanism and the current control principle are fundamentally different. The JFET uses the width variation of the depletion region to change the width of the conduction channel to control the drain current. The MOSFET uses the electric field effect of the semiconductor surface and the electric induction charge. Change the conductive groove to control the current. The difference in their properties is that JFETs are often used in the amplifier input stage (previous stage) and MOSFETs are used in the final stage of the amplifier (output stage). However, under some operating conditions, the input resistance of the MOSFET is not high enough to meet the requirements, and at high temperature operation, the resistance of the PN junction increases significantly, and the drain current also increases.
2 Features of LBCAST JFET
In the pursuit of high-bandwidth, low-power image sensor competition, CMOS image sensor has more advantages than CCD in design: small size, low system cost, and low power consumption under the premise of product quality. But noise has become the biggest obstacle to CMOS success, which will lead to a decline in image quality. This is one of the reasons why noise problems must be solved. The LBCAST JFET has many advantages, in addition to the JFET related to the amplifier, but also related to its internal structure and working characteristics.
2.1 LBCAST JFET reading method
At present, the common reading methods of CCD and CMOS are sequential charge transfer mode and XY address and transmission mode. Fig. 1(a) shows the sequential charge transfer mode usually used by the conventional Interline CCD image sensor. The electrical signal converted from the optical signal is first transmitted to the column transfer register and finally output to the image processing unit, so the speed is limited. In addition, in theory, since the sequential charge transfer method requires a continuous, high-speed drive conversion register, this requires more electric power. Figure 1 (b) shows the XY addressing and transmission method commonly used in CMOS image sensors. In this way, each pixel has its own amplifier, which transmits signals through column scan and line scan, and outputs them to the image processing unit. . It has a separate data transmission line, so it can achieve high speed, but if you carefully observe the output image, you can find that image distortion is easy to occur on separate lines.
   The JFET image sensor also uses XY addressing and transmission. The data is read out in different colors through two signal lines, which makes reading images faster and has the advantage of being able to extract high-density pixel data at will. The data distribution line of the JFET image sensor replaces the area by the color method (green, blue, and red), which improves the image speed while improving the operation speed, and solves the problem that the output image is easily reproduced on separate lines. A problem.
The JFET image sensor separates the source based on color; all green signals are output through one line, and all blue and red signals are output through another line, which allows the image to be unaffected by fluctuations in the output amplifier, ensuring image quality. Since the human eye is particularly sensitive to green, the green signal line only processes the green signal, and it is especially important in sharpening the image and setting the contrast of the image in green. When reading the first line of data, apply the data line to the left of the column (the G signal is output above, the B signal is output below), and then read the second line of data. At this time, the data line to the right of the column is applied (the G signal is output above, and the output R is below) The signal), which continues in turn (Fig. 2), shows that the R:G:B ratio is 1:2:1, and the ratio of the color filter is usually designed based on this principle.
2.2 JFET function
The transistor for extracting pixel data in the LBCAST JFET is a JFET, and each pixel includes a pair of charge accumulation portions (ie, photosensitive elements) and JFET transistors for detecting amplification, which can realize photoelectric conversion, storage, and amplification. The amplifier in the CMOS image sensor is a MOSFET amplifier. When the camera shutter is closed, the light reception ends, the gate of the MOSFET for transfer is turned on, and all stored charges are transferred to the JFET gate. In addition, the JFET gate is equivalent to a measuring cup, and it is possible to read out how much light load is transferred into this "cup" by the JFET. The JFET gate voltage rises with the charge transferred from the photodiode. At this time, the JFET causes the signal voltage to rise accordingly, and outputs it as data read by the column signal line. After the image signal is read, the JFET gate sends a charge to the MOSFET to reset it, which controls the JFET gate on and off. In other words, the function of a JFET is like a pixel switch, which can be closed when a signal needs to be read. Compared to CMOS image sensors, the JFET's path is much simplified, which results in greatly improved speed, increased reliability, and reduced defect rate.
2.3 Analysis of internal structure characteristics
In the LBCAST JFET, since the charge accumulation portion is laterally embedded, the JFET becomes a channel structure sandwiched between Gates, making it an ideal amplifying amplifying element with higher sensitivity and lower noise than CMOS. .
First, for a given signal, LBCAST uses a smaller measuring cup, but provides a larger voltage increment and higher resolution. Second, in CMOS image sensors, the signal passes through the channel to the silicon surface; in LBCAST, the signal is transmitted through the internal channel, so the large noise is reduced to almost 1/3 of the previous level, while the dark current is extremely small. Can effectively suppress dark noise. On the other hand, the pixel signal is extracted simultaneously in two channels, enabling high-speed processing. In terms of structure, the wiring structure of the LBCAST imaging pixel is one metal layer smaller than that of the CMOS, and the wiring density is also relatively low, and the interlayer connection holes are also relatively small. This achieves the goals of simple structure, less manufacturing failure, and high yield.
   The JFET sensor pixel select switch consists of three transistors: transfer, JFET, and reset. The CMOS sensor consists of four transistors, and the fourth transistor is used for pixel selection. Therefore, LBCAST has a simpler structure than CMOS, higher efficiency, and because the photodiode per unit area can be increased, its functionality is improved. The internal wiring (including the non-transparent layer) structure is also simple, suitable for one polysilicon layer and two material layers. The design of CMOS is composed of four layers. The smaller the number of layers required, the shorter the distance between the photodiode and the microlens. Through the BPD (Burie d PhotoDiode), internal FPN (Fixed Pattern Noise) and other technologies, the LBCCAST sensor can effectively reduce the image noise captured under dark light. It is similar to the CMOS dual-channel read mode, which improves the data readout rate. The microlens that is very close to the photosensitive unit improves the light efficiency and improves the consistency of the center and corner of the picture. The LBCAST JFET pixel structure is shown in Figure 3. The LBCAST JFET pixel profile is shown in Figure 4.
3 Conclusion
The main goal of the newly developed LBCAST JFET sensor is to focus on "speed." Its total pixels are not very high, only 4 million. Therefore, Nikon will target the D2H camera used in the LBCAST JFET image sensor for news reports and sports photography.
Compared with CMOS, LBCAST has higher sensitivity and low noise, and has a simple structure, fewer manufacturing failures, and high yield. Because of its simple structure, it can adopt the same manufacturing process as CMOS. It is expected that the manufacturing cost can be greatly reduced in the future, and the application prospect is very promising.
Electronic Wire Harness Co., Ltd. , http://www.nsusbcables.com