Image Lag Research Articles

Comparison of PbI2 and HgI2 for direct detection active matrix x-ray image sensors

The factors determining the x-ray sensitivity of HgI2 and PbI2 as direct detector materials for large area matrix addressed x-ray image sensors are described, along with a model to explain their different properties. The imaging studies are made on test arrays with 512×512 pixels of size 100 μm. The x-ray sensitivity and spatial resolution are reported, along with measurements of the various mechanisms that influence the sensitivity, such as charge collection, x-ray absorption, fill factor, and image lag. The spatial resolution of PbI2 decreases with increasing film thickness, but this effect is not observed in HgI2. The x-ray response data are used to compare the sensitivity to the theoretical values for the ionization energy and to identify the various loss mechanisms. We find that the sensitivity of HgI2 can be explained by a few small and well characterized loss factors. This material exhibits good spatial resolution, high fill factor, and high charge collection. PbI2 films exhibit lower sensitivity, principally attributable to a very large image lag. We propose that the x-ray response of the two materials is distinguished by their different depletion layer properties, and present a model that accounts for the sensitivity, image lag, and spatial resolution of PbI2.

Measurement and correction of the effects of lag on contrast–detail test results in fluoroscopy

Persistence of the video signal between TV frames, an effect also known as image lag, can lead to anomalously good contrast–detail test results for fluoroscopy systems. In this practical paper, a simple method is described which quantifies lag in fluoroscopy systems and corrects for its effect on threshold contrast. A digital framestore was used to acquire temporally contiguous fluoroscopy images. Correlation of the variance between an initial base TV frame and successive later frames was then measured via the correlation coefficient. Plotted against time, this function defines a time constant which characterizes the rate at which the initial variance pattern is replaced by incoming quantum noise. A survey of seven fluoroscopy units incorporating vacuum TV camera tubes found a mean time constant of 0.06 s.The relative change in contrast–detail performance was then measured as a function of applied digital frame averaging for two separate fluoroscopy units. A time constant was found for each frame averaging mode using the correlation of variance between frames. These measurements were used to derive a function which corrects contrast–detail results obtained for a unit with a measured nominal time constant to the typical vacuum camera tube time constant of 0.06 s. The correction is shown to significantly reduce the spread of contrast–detail results obtained over a range of temporal filtration settings.

A performance comparison of direct- and indirect-detection flat-panel imagers

A comparison of the performance of a direct- and an indirect-detection amorphous silicon flat-panel X-ray imager is presented for a 6 MV beam. Experimental measurements of the noise characteristics, image lag, spectral response, spatial resolution and quantum efficiency are described, compared and discussed. The two systems are comprised of 512×512 pixel, 400 μm pitch, arrays of a-Si:H p–i–n photodiodes and thin-film transistors. In the direct-detection system, X-rays interact to produce electron/hole pairs directly in the silicon photodiodes. For the indirect-detection system, a phosphor screen converts energy from the incident X-rays into visible light, which is then detected by the photodiodes. Both systems are shown to be quantum noise limited, with the total electronic noise in the detector 10–15 times smaller than the Poisson noise level in detected signal. The measured lag for both systems is 1.0±0.1% or less in the first frame with subsequent signals decaying exponentially with frame read-out, with a half-life of between 3.3 and 3.8 frames. Both systems are demonstrated to have a pronounced sensitivity to low-energy multiply scattered photons, although this is shown to be effectively filtered out using a 2 mm copper build-up plate. The direct-detection system, with the 2 mm Cu build-up, shows greater sensitivity to scattered radiation than the indirect system. The spatial resolutions of both systems were effectively equal with an f 50 of 0.25 mm −1 when pixels are binned 2×2, although a slight contribution from optical scattering in the phosphor screen is seen for the indirect-detection system. The quantum efficiency of the direct-detection system is a factor of 0.45 lower than that of the indirect-detection system. The application of these detectors to megavoltage CT is discussed, with the conclusion that the indirect-detection system is to be preferred.

Two-color amorphous silicon image sensor

A large-area two-color image sensor array made with amorphous silicon (a-Si:H) technology is described. Mesa-isolated double-junction p-i-n-i-p a-Si:H sensors discriminate the two spectral bands—blue/green and red—according to the polarity of the applied voltage bias. The 512×512 element active-matrix array with 75 μm pixel pitch is addressed using a-Si:H thin-film switching transistors. Under steady state illumination, the array exhibits a linear response, good color separation, and good spatial response as measured by the line-spread function, so that images obtained with the array are clear and sharp. The response to transient illumination exhibits image lag and a strong dependence of the signal on integration time, both of which depend on the bias polarity. Switching the bias voltage also induces strong transient properties. These effects are attributed to the back-to-back diodes, which act as capacitative dividers in the generation and readout of the signal. The transient effects compromise the practical application of the color sensors for video color imaging.

Analysis of temporal noise in CMOS photodiode active pixel sensor

Temporal noise sets the fundamental limit on image sensor performance, especially under low illumination and in video applications. In a CCD image sensor, temporal noise is primarily due to the photodetector shot noise and the output amplifier thermal and 1/f noise. CMOS image sensors suffer from higher noise than CCDs due to the additional pixel and column amplifier transistor thermal and 1/f noise. Noise analysis is further complicated by the time-varying circuit models, the fact that the reset transistor operates in subthreshold during reset, and the nonlinearity of the charge to voltage conversion, which is becoming more pronounced as CMOS technology scales. The paper presents a detailed and rigorous analysis of temporal noise due to thermal and shot noise sources in CMOS active pixel sensor (APS) that takes into consideration these complicating factors. Performing time-domain analysis, instead of the more traditional frequency-domain analysis, we find that the reset noise power due to thermal noise is at most half of its commonly quoted kT/C value. This result is corroborated by several published experimental data including data presented in this paper. The lower reset noise, however, comes at the expense of image lag. We find that alternative reset methods such as overdriving the reset transistor gate or using a pMOS transistor can alleviate lag, but at the expense of doubling the reset noise power. We propose a new reset method that alleviates lag without increasing reset noise.

Two-Dimensional Amorphous Silicon Color Sensor Array

ABSTRACTThis paper reports on the first full realization and characterization of a two-dimensional array of amorphous silicon (a-Si:H) color sensors, addressed by integrated amorphous silicon-based thin-film transistors (TFTs). The array includes 512 × 512 pixels with 75-µm pitch, or about 340 dpi. Each pixel features a color sensor realized by a p-i-n-i-p stack of doped and undoped a-Si:H layers, and the TFT. The color sensors are made of two back-to-back p-i-n diodes, which selectively sense the illumination according to the polarity of the applied bias voltage. The sensor layers are grown on top of the TFTs to improve the array fill factor. The p-in-i-p sensor stack is mesa-isolated into single sensors to reduce cross-talk.Images are acquired using two bias voltages and yield the red and blue/green components of the original with a good color separation. A color image is reconstructed using the information from the two images acquired. Aside from a color bias, which is expected for a two-color reconstruction, the imaging system works well. In particular, the array shows very low leakage currents, which enable a very large dynamic range and sensitivity. In the response of the array to a light pulse, the bottom thick diode ensures a fast drop in the signal after the flash, while the top thin diode exhibits some residual image lag.

イメージセンシング技術 ＣＭＯＳイメージセンサにおける低電圧駆動埋込みＰＤの解析 埋込みＰＤ構造による低リーク電流の検討

A CMOS image sensor with a very-low-leakage photodiode and a low-voltage buried photodiode has been developed that has sensitivity comparable to that of conventional CCD imagers. Photodiode leakage current caused by stress originates in the surface damage and in the dislocations and defects near the local oxidation of silicon. Use of a dislocation free stress-release process and of a buried photodiode reduced the damage and eliminates the stress. A 1/4-inch 330 K pixel CMOS image sensor using this buried photodiode structure and a sensor-specified process had a low dark current of 0.1 nA/cm2 in 5.6 μm pixels. The buried photodiode can operate in complete charge-transfer mode at voltage as low as 3.3 V, thereby suppressing the image lag and FPN of the photodiode.

Effects of buried insulator-sensor interface on the lateral conduction of high fill factor aSi:H imagers

ABSTRACTThe pixel cross-talk is investigated in two-dimensional amorphous silicon (a-Si:H) imager arrays based on the new high fill factor design. In this configuration a continuous a-Si:H sensor extends over the whole surface of the imager, and a buried insulator material with low dielectric constant is used to separate the sensor from the underlying active matrix readout circuit. We find that the lateral conduction between neighboring pixels is mainly determined by the quality of the buried insulator-sensor interface, rather than the specific buried material itself. Minimum cross-talk values below 1% are obtained for different insulator materials including silicon oxynitride and thicker polymer based resins. The quality of this interface also affects trapping and recombination of the photogenerated carriers, influencing important imager properties such as sensitivity and image lag.

Cone-beam computed tomography with a flat-panel imager: effects of image lag.

A system for cone-beam computed tomography (CBCT) has been developed based upon the technology of active matrix flat-panel imagers (FPIs), and the system has demonstrated the potential for fully three-dimensional volumetric imaging with high spatial and contrast resolution. This paper investigates the effects of image lag (arising from charge trapping and release in the FPI pixels) upon CBCT reconstructions. Hypotheses were derived based upon a simple, geometrical/physical model, suggesting that image lag in the projection data results primarily in two artifacts: a spatial blurring artifact in the direction opposite to the direction of rotation (called a "comet") and a line artifact along the direction of the first few projections (called a "streak"). The hypotheses were tested by means of computer simulations and experimental measurements that yielded CBCT images of a simple cylindrical water phantom containing an attenuating rod of varying size and composition. The computer simulations generated projection images based upon analysis of the system geometry and a simple model of the FPI that allowed free adjustment of the image lag. Experimental measurements involved CBCT scans of the phantom under various conditions and modes of acquisition followed by examination of the resulting CBCT axial slices for lag artifacts. Measurements were performed as a function of exposure level, position and contrast of the rod, and for three modes of acquisition designed to isolate and/or minimize the two hypothesized artifacts. The results clearly illustrate the comet and streak artifacts, particularly in relation to high-contrast objects imaged at high exposure levels. The significance of such artifacts under clinical conditions is expected to be small, considering the magnitude of the effect relative to the morphology and composition of typical anatomy. The artifacts may become appreciable, however, in the presence of high-contrast objects, such as marker BBs, dental fillings, and metal prosthetics. A procedural method of reducing lag artifacts is demonstrated.

A ghost story: spatio-temporal response characteristics of an indirect-detection flat-panel imager.

Spatial and temporal imaging characteristics of an amorphous silicon flat-panel imager (FPI) were investigated in terms relevant to the application of such devices in cone-beam computed tomography (CBCT) and other x-ray imaging modalities, including general radiography, fluoroscopy, mammography, radiotherapy portal imaging, and nondestructive testing. Specifically, issues of image lag (including the magnitude, spatial uniformity, temporal-frequency characteristics, and dependence upon exposure and frame time) and long-term image persistence ("ghosts") were investigated. As part of the basic characterization of the FPI, pixel dark signal and noise (magnitude, temporal stability, and spatial uniformity) as well as radiation response (signal size, linearity, gain, and reciprocity) were also measured. Image lag was analyzed as a function of frame time and incident exposure. First-frame lag (i.e., the relative residual signal in the first frame following readout of an exposure) was approximately 2-10%, depending upon incident exposure and was spatially nonuniform to a slight degree across the FPI; second-, third-, and fourth-frame lag were approximately 0.7%, 0.4%, and 0.3%, respectively (at 25% sensor saturation). Image lag was also analyzed in terms of the temporal-frequency-dependent transfer function derived from the radiation response, allowing a quantitative description of system components contributing to lag. Finally, the contrast of objects as a function of time following an exposure was measured in order to examine long-term image persistence ("ghosts"). Ghosts were found to persist up to 30 min or longer, depending upon the exposure and frame time. Two means of reducing the apparent contrast of ghost images were tested: (i) rapid scanning of the FPI at maximum frame rate, and (ii) flood-field exposure of the FPI; neither was entirely satisfactory. These results pose important considerations for application of FPIs in CBCT as well as other x-ray imaging modalities. For example in CBCT, the magnitude of image lag is such that significant artifacts in tomographic reconstructions may result if strategies are not adopted either to reduce or correct the lag between successive projections (e.g., rapid scanning between projections or iterative correction algorithms, respectively). Similarly, long-term image persistence may necessitate frequent recalibration of offset corrections.

Dependence of Image Transient Behavior on Operating Parameters of Amorphous Silicon Image Sensors

AbstractThe growing interests for real-time applications of our amorphous silicon X-ray image sensors requires a better understanding of transient behavior in imaging characteristics. In our pixel structure, which consists of a p-i-n photodiode and an addressing thin film transistor (TFT), the image lag can stem from a number of mechanisms, including charge retention on the photodiode, TFT charge transfer efficiency, and data-line charge sharing. The exact prevailing mechanism depends on the array design and the operating conditions of the sensors. This paper describes the results of our systematic studies on image lag, and its dependence on various operating conditions of the image sensors, such as applied bias, light intensity, frame time, and applied gate voltages. We find that the charge retention in the intrinsic region of the photodiode is an important source of the image lag, and it is correlated to the forward transient characteristics of the diode. This correlation can be explained by the common underlying mechanisms leading to depletion region formation and charge transport through the diode. Furthermore, we notice that the time dependent decay of the image lag, originated mainly from charge retention in the diode, presents power-law time dependence, which is consistent with the release of trapped charges from a continuous distribution of defect states. We also find only a small variation in the image lag, <%3, for a wide range of gate on-periods, 10μs-80μs, and gate on-voltages, 15V- 30V, which indicates an efficient charge transfer by the switching TFT. These results demonstrate that in our devices, the image lag is controlled by the intrinsic defect levels in the a-Si:H photodiode.

Device design with automatic simulation system for basic CCD characteristics

A simulation system has been developed to automatically analyze basic electrical characteristics of a charge-coupled device (CCD) image sensor from a process simulation result. This system shortened the simulation period to approximately 1/10 by getting rid of complicated repetitious procedures. A high-performance new cell technology has been developed successfully with improving impurity distribution in shorter development time by using this system. This technology has been realized as a CCD cell pixel with CCD charge quantity of 1.8 times, effective transfer efficiency of over 99%, no image lag for driving read-out pulse voltage in comparison with conventional technology. A 1/4-in 330 K square pixel progressive-scan CCD was fabricated with this technology. These results are described to demonstrate the effectiveness of the automatic simulation system.

CMOS active pixel image sensor with simple floating gate pixels

A new pixel structure using a simple floating gate (SFG) has been proposed. The pixel consists of a coupling capacitor, a photogate, a barrier gate and a MOS transistor. It features complete reset that results in no kTC noise and no image lag, high blooming overload protection, nondestructive readout (NDRO), and CMOS compatibility. Its basic operation has been confirmed with a 32(H)/spl times/27(V) pixel area array. Since the pixel structure is relatively simple, small pixel size is feasible. >

ＣＭＤ撮像素子のＦＰＮ抑圧駆動法

This paper describes a new method of FPN suppression that does not use memories on CMD imagers. This method does not require any changes in the conventional device structure. As part of this suppression method, we have employed a new drive method, which we also developed. In this drive method, the pixels are reset sequentially one by one. FPN can be reduced by removing dark signal, which is read out just after resetting, from output signal. We have shown that this method causes no deterioration in terms of photoelectric conversion and image lag, as compared with the conventional method. By this method, We achieved a FPN suppression ratio of 15.8 dB and suppressed FPN to-40.2 dBp-p in comparison with the saturation signal.

A 2/3-in 2 million pixel STACK-CCD HDTV imager

A 2/3-in optical format 2 M pixel STACK-CCD imager has been developed. The STACK-CCD imager, overlaid with an amorphous silicon photoconversion layer, realizes a large signal charge handling capability of 1.0/spl times/10/sup 5/ electrons, a -140 dB smear noise and a 90 dB dynamic range with a newly introduced device configuration and its unique operation. The 5.0 /spl mu/m(H)/spl times/5.2 /spl mu/m(V) unit pixel imager with sufficiently low image lag has been realized by a novel bias charge injection scheme These device performances are the best ever achieved by a CCD HDTV imager. This is the first such imager satisfying the device performances required for a practical use 2/3-in 2 M pixel HDTV imager. >

120 × 100 pixel antiblooming array based on optically addressed ferroelectric liquid-crystal cells

We describe the principle and the operation of a 120 x 100 pixel antiblooming array. Each cell in the array is made up of a hydrogenated amorphous silicon photoconductor layer driving a ferroelectric liquid-crystal shutter. The device is designed to be coupled with a CCD camera to prevent high-blooming overload and image lag.

A high-packing density pixel with punchthrough read-out method for an HDTV interline CCD

A new pixel structure for a high-packing-density interline CCD is proposed, in which signal charges are read out from the photodiodes to the vertical CCD by a punchthrough mechanism. This read-out method makes it possible to reduce the depth of the VCCD channel and the second p-well by implanting these two layers after diffusion of the photodiode n layer. Spreading resistance measurements on dummy wafers show that the depths of these layers are 0.28 /spl mu/m and 0.6 /spl mu/m, respectively. Moreover, the photodiode n-layer is covered with a surface p/sup +/-layer, even at the transfer region. We describe the results of simulations and experiments on a test image sensor with pixel dimensions of 7.3 /spl mu/m (H)/spl times/7.6 /spl mu/m (V). From the experimental data, we estimate the characteristics of an image sensor with pixel dimension 5.0 /spl mu/m (H)/spl times/5.2 /spl mu/m (V). Such a device should have a maximum charge handling capability of 1.4/spl times/10/sup 5/ electrons, a smear level of -88 dB, a sensitivity of 1.5/spl times/10/sup 3/ electrons/Ix with a 30% fill factor, no image lag, and a low photodiode dark signal of less than 14 electrons at 60/spl deg/C. These results indicate that an IL-CCD with a punchthrough readout structure is suitable for image sensors with a high pixel density such as 2/3 inch 2 million pixel image sensors for high-definition TV applications. >

埋込電極をもった増幅型撮像素子の提案

The proposed structure allows opaque driving electrode to be removed from the cell surface, increasing high quantum efficiency and resulting in higher photo-sensitivity. Mode selection (such as reset, charge storage and readout) can be performed by driving electrodes buried in silicon substrate. Readout signal amplitude, which is the output of a source follower, is determined by the number of signal holes stored in each photodiode. Simulation of a device with 1000×1000 of these cells/cm2 indicated that the sensor could be operated without image lag, blooming and low smear. The simulated photo-sensitivty, cell charge capacity and charge-to-voltage coversion factor were high enough for videocameras.

2D Image Sensing Arrays with NIP Diodes

2D image sensor arrays made with a-Si devices on glass over large area are of considerable interest as document scanners and in medical applications. We have made a test array containing a-Si NIP diodes for both the sensors and the active matrix switching devices. The issues of vertical crosstalk and image lag are discussed in relation to the device performance of the switching diode. The vertical crosstalk is controlled by the diode capacitance and the image lag by the high transient current in the device. We speculate that the transient current is a trap filling current in the deep states of the switching diode.

１インチ２００万画素ＣＭＤイメージセンサ

This paper presents the experimental results of a high resolution CMD (Charge Modulation Device) image sensor. This sensor features ultra-low power dissipation and a low smear level.We designed, fabricated and studied a 2M pixel CMD image sensor. A pixel size of 7.3 μm (H) ×7.6μm (V) was used to obtain a 1-in optical format HDTV image sensor. We achieved a die size of 16.6 mm (H) ×10.7mm (V), even though the clock drivers for the scanners were fabricated in the sensor chip.The smear level was-123 dB and the power dissipation was 206 mW. The saturation signal current per pixel was 28 μA at an exposure of 76 lx (Tc=5100 K, 1/60s integration). The signal had good linearity of photoconversion characteristics. Furthermore, the image lag was lower than the detectable limit (0.1% of the saturation signal).