Comparison of CMOS and CCD for Solid Image Sensor
At present, the market application of solid image sensor mainly CCD and CMOS two. In this paper, from the perspective of technical performance, the internal and external structure of the device, the principle, the application, the manufacturing process and the equipment and so on the comparison between the two, from the present, both have the pros and cons; from the development, the CMOS image sensor will Replace the CCD and get more widely used than the CCD.
Solid image sensor (also known as solid photoelectric imaging device) has two kinds of CCD and CMOS. CCD is "charge-coupled device" (ChargeCoupledDevice) short, and CMOS is "complementary metal oxide semiconductor" (ComplementaryMetalOxideSemiconductor) short. The CCD was invented by W · B · Boyle and G · E · Smith of the Bell Labs in 1970, which opened the prelude to the charge transport device. Since then, people use this technology to create a camera and digital camera, the image processing industry to advance to a new field. CCD is a photographic semiconductor chip used to capture images, widely used in scanners, copiers, camcorders and filmless cameras and other equipment. As a camera, similar to the principle of the film, the optical image (ie, the actual scene) is projected onto the CCD through the lens. Unlike the film, the CCD does not have the "exposure" capability, nor the ability to record and store image data, but instead sends the image data into an A / D converter, a signal processor, and a storage device, but Repeated shooting and real-time adjustment, the image can be unlimited copy without reducing the quality, but also easy to save forever.
CMOS was originally an important chip in the computer system, it can save the system to guide the required large amounts of information. In the early 1970s, it was found that CMOS will be introduced into the semiconductor photodiode can also be used as a sensor, but in the resolution, noise, power consumption and image quality than the current CCD difference, and thus not developed The With the development of CMOS process technology, the use of standard CMOS technology to produce high-quality, low-cost CMOS imaging devices. This device is easy to mass production, its low power consumption and low cost characteristics are businesses are dreaming. Today, CCD and CMOS coexist, CCD temporarily or "mainstream", but CMOS will replace the CCD and become the mainstream of image sensors. The following from the technical performance, the device's internal and external structure, principles, applications, manufacturing technology and equipment, etc. will be a comparison between the two, and finally summarized a comparison table, and look forward to its development prospects.
Basic structure and working principle comparison
CCD is based on the MOS transistor developed, the basic structure is MOS (metal - oxide - semiconductor) capacitor structure. It is in the semiconductor P-type silicon (si) as the substrate on the surface of the oxidation method to generate a layer of about 1500 ~ 1500? SiO2, and then in the SiO2 surface of a layer of metal (such as aluminum), in the substrate And a metal electrode with a bias voltage (called gate voltage), it constitutes a MOS capacitor. Therefore, the CCD is composed of a row of rows of MOS capacitors arranged closely on the silicon substrate.
Current CCD devices use photodiodes instead of past MOS capacitors, ie, diffuse an N + region on a P-type Si substrate to form a P-N junction diode. Through the reverse bias of the polysilicon relative to the diode, so in the diode to produce a directional charge area (called the depletion area). In the directional charge region, the photo-generated electrons are separated from the holes, and the photo-generated electrons are collected in the space charge region. The space charge region is a region with a particularly low potential for negatively charged electrons, and is therefore commonly referred to as potential wells. The photogenerated charge generated by the projected light is stored in this potential well. The maximum charge amount that the potential well can store is also known as the potential well capacity. The potential well capacity is approximately proportional to the applied gate voltage. Compared with the MOS capacitor, the photodiode has the characteristics of high sensitivity, wide spectral response, good blue light response and dark current. If a series of MOS capacitors or photodiodes are arranged and the corresponding electrodes are connected together in a two-phase, three-phase, or four-phase mode of operation, and a constant timing drive pulse is applied to each set of electrodes, The basic functions of the CCD.
In general, the most basic CMOS image sensor is a low impurity concentration of P-type silicon substrate as a substrate, with the diffusion method in the surface of the production of two highly doped N + -type area as the electrode, the source of the source And the surface of the silicon with high temperature oxidation method of covering a layer of silica (SiO2) insulation layer, and the source and the drain between the insulating layer of a layer of metal aluminum deposition, as a The gate of the field effect transistor. Finally, a photodiode is placed above the metal aluminum, which constitutes the most basic CMOS image sensor.
In order to make CMOS image sensor work, must be in the P-type silicon substrate and the source connected to the negative power supply, drain connected to the power supply cathode. There is no current between the source and the drain when no image optical signal is irradiated onto the photodiode, so that no signal is output; when the image optical signal is irradiated onto the photodiode, the valence band of the photosensitive member obtains an energy excited transition To the conduction band to form an image photoelectron, thereby forming a current path between the source and the drain and outputting an image electrical signal. The stronger the optical signal of the incident image, the more conductive particles (electrons and holes) excited in the photosensitive material, so that the greater the current between the source and the drain, the larger the output signal. Therefore, the size of the output signal directly reflects the intensity of the incident light signal.
In CMOS camera devices, the electrical signal is read directly from the CMOS transistor switch array, without the need to read the line as a CCD.
From the basic structure and the principle we can see from the imaging device itself, the internal and external structure, the two are different.
Internal structure contrast
The imaging points of the array CCDs are arranged in the X-Y vertical and horizontal matrix, and each imaging point consists of a photodiode and an adjacent charge storage area (temporary zone) controlled by its transfer. As the arrangement and composition of different ways, area array CCD frame transfer type, inter-row transfer type, inter-frame transfer type, line transfer type and virtual type. When the photodiode converts the optical image into a charge image (the amount of charge is proportional to the light intensity) is stored in the potential well, it is quickly transferred to the buffer register in the buffer and charge transfer directions by transfer control, and then through two or three Or four-phase clock drive pulse to the output side of a bit shift, the output circuit charge / voltage conversion and amplifier output video signal. This configuration produces images with low noise and high performance, but requires two or three or four phase clock drives, gate bias, transfer control, and reset pulses, so the entire structure is complex and increases power consumption. Increased cost.
The CMOS imager is constructed like a memory that integrates digital logic circuits, clocks, and A / D conversions in the same machining program. Each imaging point contains a photodiode, a charge / voltage conversion, a reset and select tube with an amplifier. The entire imager covers the metal interconnector for timing applications and read signals as well as the longitudinally arranged output signal interconnects. The signal readout is read directly from the switch amplifier array by simple XY addressing technology, And convenient.
From the external structure of the imager in the product application
CCD imager requires a peripheral drive circuit to work, it can only output analog signal, this signal to the subsequent address decoder, analog-to-digital converter, image signal processor processing, integration is very low. Such as by the array CCD digital camera usually has six chips, some up to eight, at least three, so that the volume can not be reduced, the production cost is higher.
The CMOS imager does not require peripheral drive circuit, it is the photodiode, image signal amplifier, signal read circuit, analog-to-digital converter, image signal processor and controller integrated into a chip, and manufacturing only need to use Semiconductor manufacturers can produce integrated circuit processes. If the composition of digital cameras, digital cameras can be all the components are integrated into this chip, that is, "single-chip camera." Therefore, the use of CMOS chip photoelectric image conversion system, not only can reduce the overall cost of the system and assembly time required, but also greatly reduce the system size and complexity.
Comparison of CMOS and CCD in Manufacturing Process and Equipment
Since the signal of the CCD imager needs to be read one line at a time, the charge must be read out a number of times from the output amplifier. Therefore, in the production of the CCD imaging device, it is necessary to make each pixel Perfect (that is, intact), otherwise a column of the signal can not be completely read out. In addition, the transmission of each charge must also be very clean, that is, the charge transfer efficiency is very high, otherwise it will be due to the accumulation of charge caused by image stains. Therefore, although the CCD material is silicon chip, but must have a special manufacturing process technology and equipment, and also accumulated considerable manufacturing experience, in order to ensure high pixel quality and high charge transfer efficiency. So the production of CCD easily monopolized by a small number of manufacturers.
The CMOS imaging device is relatively simple and easy to manufacture, because it only needs to use the general semiconductor manufacturing process and equipment, do not have to bring their own expensive semiconductor manufacturing process equipment. CMOS circuit design only, please some of the semiconductor factory on behalf of the processing can be. As a result, CMOS imaging devices are very competitive in development and manufacturing costs.
Comparison of CMOS and CCD in Technical Performance
Comparison of information reading
CCD photoelectric imaging device stored in the charge information, the need for two-phase or three-phase or four-phase clock drive pulse control, one by one to implement the transfer after the progressive order to read.
And the optical image information of the CMOS photoelectric imaging device is photoelectrically converted to produce a current or voltage signal. This electrical signal does not need to read it as a CCD, but is read directly from the CMOS transistor switch array. flexibility. The CCD no such function.
By the knowledge, CCD imaging devices need to be in the two, three, four-phase clock drive pulse control, the behavior of a unit to output information, so slower.
The CMOS imaging device in the collection of photoelectric image signal can be taken out at the same time the electrical signal, which can simultaneously deal with the unit image information, so the speed is much faster than the CCD imaging device. As the rows of CMOS imaging devices, column electrodes can be driven at high speed, coupled with the same chip to do A / D conversion, the image signal can be quickly removed, so it can operate at a very high frame rate. Such as some designed to do machine vision CMOS, claiming to be up to 1000 frames per second frame rate.
Power and power consumption of the comparison
Since the pixels of the CCD are composed of MOS capacitors, it is necessary to use a two-phase or three-phase or four-phase sequential pulse signal with a relatively large voltage (at least 12V) to read the charge signal. So the CCD image acquisition system in addition to a number of power, the peripheral circuit will consume a considerable power. Some CCD image acquisition system to consume 2? 5W power.
The CMOS photoelectric imaging device only need to use a single power supply 5V or 3V, power consumption is very small, only the CCD 1/8 1/10, and some CMOS image acquisition system consumes only 20 ~ 50mW power.
Comparison of image quality
CCD imaging device fabrication technology started early, mature technology, using PN junction or silicon dioxide (sio2) isolation layer isolation noise, so low noise, good image quality.
Compared with the CCD, CMOS main drawback is the high noise and low sensitivity, because the CMOS imaging device integration is high, the photoelectric components, the distance between the circuit is very close to each other between the light, electricity, magnetic interference, noise on the image The quality of a great impact, began a long time can not enter the practical. Later, the noise problem was reduced with active pixels (ActivePixel) design and noise correction lines. In recent years, with the CMOS circuit noise reduction technology continues to progress, for the production of high-density high-quality CMOS imaging devices provide a good condition. Manufacturers have claimed that the technology developed, the image quality is no worse than the CCD.
CMOS imaging device sensitivity is low, because the pixel part of the area is used to make amplifiers and other lines. In the fixed chip area, unless the use of more sophisticated manufacturing process, or in order to maintain a considerable level of sensitivity, imaging device resolution can not be done too high (in turn, fixed resolution sensor, chip size can not do too small). But at present, the use of 0.18μm manufacturing technology has developed a 4096 × 4096 ultra-high resolution CMOS image sensor.
Comparison of CMOS and CCD in Application
CCD development has been 32 years of history, can be said to be quite mature products. The current technology development is mainly how to reduce the sensor area, reduce production costs and improve commercial performance. CCD mainstream applications are gradually evolving from analog cameras, security surveillance cameras and other digital multimedia applications, such as digital video cameras (DigitalVideoCamcorder), digital camera DSC (DigitalStillCamera) and so on. As the portable digital camera requires light, thin, short, megapixel CCD has done 1/3 ", while the low resolution of the CCD has been developed to 1/4" or 1/5 ", the future toward 1 / 8 "and even 1/10" development. The CCD of the analog camcorder is equipped with a PAL / NTSC TV system, the pixel aspect ratio is also 4: 3 and the interlaced read signal is used, while the dedicated digital camera emphasizes Founder Pixels (which can reduce the time required for signal processing) and progressive scan (can enhance the speed of the image).
Because CCD has good image quality, high sensitivity and high resolution, so the current high, mid-range video applications are mainly CCD, and for some emerging products, such as PC cameras, mobile phones, etc., can not meet the use of CCD power consumption , Volume and other aspects of the requirements.
CMOS imaging devices are currently focused on CIF and VGA grade products. CIF-level CMOS imaging devices have been shifted from 1/3 ", 1/4" to 1/4 "to 1/5", or even 1/7 ". Current 1/7" CIF-level CMOS imaging devices consume power The amount has been reduced to 30mw below, so more used in mobile phones and other portable products. VGA grade CMOS imaging device from the past 1/3 "to 1/4", go to 1/4 "to 1/5". This level of CMOS imaging device power consumption from the past 300mw down to 100mw or so, it is mainly used in PC cameras, followed by the DSC, PHS and TOY also have applications.
As the CMOS imaging device is small, low power consumption, but also for security concealed surveillance camera and interactive toys for the "eyes." In addition, can also open up the medical market applications, such as CMOS after the use of lost camera (like the size of the pill), so that patients swallow, it will go through the stomach, small intestine and large intestine, with feces excreted. When the camera crosses the body, it illuminates itself, takes two shots per second, and transmits the radio waves to the recorder that is hanging from the patient's belt. The doctor then sends the data to the computer. The camera battery can be used for 8 hours, enough to shoot the entire small intestine (due to the small intestine check has been "blind spot"). Obviously, this is CCD can not do.
CMOS imaging devices in the VGA level above the performance of the resolution, has been better than the CCD is excellent, because the high pixel CMOS imaging device must reduce the noise and improve the sensitivity to face the CCD competition. But the CMOS imager can also take advantage of its combination of photographic and image processing and identification capabilities, making it a high value-added intelligent image sensor, and applied in some of the more special purposes, such as machine vision, the human body Face and fingerprint recognition, dynamic detection and so on.
CMOS image sensor and CCD compared to a number of prominent features, such as small size, low power consumption, low cost, single-chip integrated system, random access, lossless reading, anti-corona image without tail, high frame rate , High dynamic range and so on. Therefore, the CMOS image sensor has irresistible broad market temptation and good prospects for development.
In order to expand the market share of CMOS imaging devices in mobile video communications equipment (such as notebook computers, PDAs, PHSs, etc.), and with the development of portable products market, integrated, small size, low cost, low power consumption is still the future development the key of. VGA level, power consumption has been developed below 100mw; CIF level to 40mw below; QCIF level to 25mw following break. In a single power supply low operating voltage, there are already manufacturers to develop 2.8V.
Another development is the combination of integrated photographic and image processing and identification functions, making CMOS a high value-added intelligent image sensor and applied in more specific applications (such as machine vision, dynamic detection and automatic identification, etc.).
CMOS Imaging Sensors To replace CCDs in digital cameras and digital camera markets over megapixels, it is necessary to break down from reducing noise and increasing sensitivity in terms of resolution.
Now, CMOS image sensor is moving towards high resolution, high dynamic range, high sensitivity, high frame rate, integrated, digital, intelligent direction. Now, two step-by-step reset gate voltage technology can increase the dynamic range of active pixel sensor (APS) in CMOS to more than 90dB, using linear-logarithmic output mode, and CMOS-DPS technology, but also the dynamic range Increased to 120dB; and the general CCD is only about 66dB. There are indications that the field of image sensors is facing a major turning point, although from the current situation, CMOS and CCD image sensor application market still has a demarcation, but the boundaries seem increasingly blurred. With more than 3 million pixels of CMOS image sensor listed, the image sensor is about to enter the "CMOS era." And CMOS in making high pixels also has a certain advantage.
With the video surveillance technology to the development of intelligent, high-definition monitoring is a high profile, high demand to enter our world. At present, high-definition network camera has become an inevitable trend. Many manufacturers are scrambling to launch their own high-definition network camera. Although the widely used sensor is still CCD, but the CCD response speed is low, does not apply to high-definition surveillance cameras used in high-resolution progressive scan mode, so high-definition surveillance cameras need to generally use CMOS as a photosensitive element. And now committed to the CMOS research vendors have developed 720P and 1080P dedicated CMOS devices, its wide dynamic, strong light suppression, sensitivity and other properties have been greatly improved.
As the CMOS image sensor has a small size, low power consumption, high integration, wide dynamic range, a USB computer interface, can be made DPS, automatic timekeeping, random access, lossless read, anti-halation and radiation resistance , And irresistible broad market temptation and more extensive application prospects than CCD
Copyright © Jaste Solar Technology Co.,Ltd All Rights Reserved.