Yong Ki Chi

A Pixelated CsI (Tl) Scintillator for CMOS-based X-ray Image Sensor

Direct and indirect imaging detectors for digital radiography have been developed during the past a few years. Among scintillators, columnar CsI(Tl) screens are used for indirect digital X-ray imaging for industrial and medical radiography. By the help of recent developments CMOS (Complementary Metal Oxide Semiconductor) is replacing the CCD (Charge Coupled Devices) in X-ray image sensor because of low operating power, standard CMOS manufacturing process, and low cost. An X-ray imaging detector consisting of pixelated CsI(Tl) scintillator and CMOS imaging sensor for application in dental radiography was developed. The manufactured CMOS imaging sensor has 128 by 128 pixels with the pitch size of 50 mum and gap width of 5 mum. A 0.5 mum CMOS process was used for sensor fabrication. Pixel structured CsI(Tl) scintillator was directly deposited on the patterned CMOS imaging sensor by thermal evaporation and photo-lithography process. The fabricated monolithic X-ray imaging sensor showed dramatic improvement in spatial resolution due to enhancement of the scintillation light guiding effect and reduction of the light cross-talk between scintillator pixels.

A computer code for the simulation of x-ray imaging systems

A computer code was developed to simulate an X-ray imaging system, which is composed in X-ray source, object and detection parts. The physical principles are the X-ray attenuation law and the detecting probability based on deterministic methods. In the part of an X-ray source, it uses a source spectrum and prepares corresponding attenuation coefficients for object materials. A computer-aided drawing (CAD)-formatted file is used as the object geometry to make it convenient to describe complex 3-D object components. The number of pixel and the size of detector are variable in the part of a detector. Transmission images can be simulated with various X-ray energies, including spectrum and various system geometries. The simulation code has its graphic user interface (GUI), which makes it easy to perform selection of simulation parameters.

The application of 3-D X-ray microtomography with FEM analysis for trabecular bone/cement interface

Recently, microfinite element techniques based on high-resolution images have been introduced that allow modeling of trabecular bone structures since 3-D cone beam tomographic system was developed. Although microtomographic system has been paid much attraction for microstructure study, the interface between each other objects is not completely. Many published studies with FEM analysis suggested approximately assumed models of the interface for the application of biomechanical study. Therefore, mechanical failure was dependent on simple interface models assumed by the experiment and was predicted with FEM analysis. For FEM analysis with the real structure and interface, we obtained 3D cone beam microtomography of trabecular bone/cement of which the failure is considered to be an important cause of aseptic loosening of total hip replacements in orthopedic implants. The sample was manufactured by implanting cement in a pig hip bone and the system was composed of microfocusing X-ray tube, CMOS image sensor and microprecision motor. In case of microtomographic imaging trabecular bone/cement, it has white-out appearance due to low contrast nature with low energy X-ray. Therefore the multicriterion method for region and volume was applied to the segmentation of white-out appearance. The segmented data was used for FEM analysis as determining node of hexahedron meshing from it. The results of this paper show that the interface of trabecular bone/cement correctly can be distinguished by multicriterion method and the finite element modeling using microtomography can be used for predicting the failure mechanism of the interface with FEM analysis.

Noise response in a CMOS active pixel sensor due to the radiation effects

In a viewpoint of electric noise, the main provider of the CMOS APS is the pixel among other components and generally the radiation response to it makes additional noise source leading to change of leakage current. For the change of leakage current, the main mechanism of total ionization dose (TID) effect on CMOS APS had been investigated in other research activities. However, it is difficult to distinguish dominant noise among several other noise sources, affecting the performance of active pixel since that is composed of photodiode and three MOS transistors, reset, source follower, and selection transistor. In this paper, we have performed characterization of dominant noise source after measuring total noise and dominant source in each operation step. To derive each source we designed and fabricated the combination set of discrete components as follows: (1) only photodiode with two styles n-well/p-sub and n+/p-sub, (2) only MOS transistor, (3) photodiode and reset transistor, (4) photodiode, reset transistor, and source follower, finally (5) photodiode, reset transistor, source follower, and select transistor. The size of the designed pixels is 100 mum and the fill factor is approximately over 73%. The test chip was fabricated using ETRI 0.8 mum (2P/2M) CMOS standard process. These test samples were irradiated using DC X-ray source to the 4 krad dose. It was found that reset noise and readout noise is not changed, but only number of dark charge is changed to that exposure dose.

An experimental study on the variation of MTF and NPS caused by x-ray beam conditions for three indirect digital radiographic imagers

Digital radiographic X-ray imagers are increasingly used in many clinical areas. So it is important to evaluate the performance of imagers. There is growing consensus in the scientific world that the detective quantum efficiency (DQE) is the best-suited parameter for describing the imaging performance of an X-ray imaging device. DQE is calculated with modulation transfer function (MTF) and noise power spectrum (NPS). As a preliminary study, MTF and NPS of three indirect digital radiographic imagers were measured and analyzed for comparison of imager performance for three imagers. The imagers were evaluated using two standard X-ray beam qualities, RQA5 and RQA9 recommended by an International Electrotechnical Commission (IEC) standard. For MTF and NPS measurement, edge method and two-dimensional Fourier analysis method were adopted each other. In result, the highest MTF is observed in the imager with smallest pixel pitch and preprocessing of raw data among three imagers. There is an obvious MTF difference with preprocessing steps and a little difference between RQA5 and RQA9. NNPS is also affected by preprocessing steps. There are the significant increase of noise at low frequency because of the increase of the multiplicative noise component as exposure increases except CDXI-40G (Canon). The low NPS is relatively caused by high exposure level because of noise conversion process. In comparison, the lowest NPS imager was found. The evaluation of imaging performance using DQE concept will be done as our further study.

Calculation of detective quantum efficiency for CMOS image sensor coupled with micro-channel plate in low energy X-ray imaging

Micro-channel plate (MCP) detector has a combination of unique properties such as high gain, high spatial resolution and high temporal resolution. By these advantages, it was shown that the possibility of MCP detector is very promising in important fields such as UV, soft X-ray and hard X-ray detection. Recently an X-ray detection model of MCP has been developed according to several energy ranges and X-ray imaging techniques based on MCP have been tested for new radiation detection systems. In this paper, we proposed a radiation detection system using CMOS active pixel sensor (APS) coupled with MCP. The theoretical detection efficiency of this system will be used for showing its usefulness in low energy X-ray. Applying our detection system to linear cascade system, we consider the MCP model for the X-ray induced secondary election emitted from MCP bulk. A multiple stochastic spread, fluence gain and additive noise were described as cascade stages caused by system components such as the MCP, phosphor screen and CMOS APS. Finally, the analytical investigation and numerical calculation of the quantum accounting diagram (QAD) and the detection quantum efficiency (DQE) were carried out. Through these results, we will discuss the performance of proposed system which is used for a high resolution image detector in low energy X-ray imaging.

Development of a 3-D X-Ray Micro-tomography System and its Application to Trabecular Bone/Cement Interface

In recent years,the interface analysis of micro-structure based objects is an important research in osteoporosis,vascular imaging since a 3-D X-ray micro-tomography system was developed. However,the micro-tomographic image shows the white-out appearance in case of imaging of similar density objects with low energy X-ray.Therefore these images must be analyzed about the interface between micro-sucture based objects for its application to biomechanical study. Many published studies suggested approximately assumed model of interface and predicted mechanical failure by means of Finite Element Method (FEM) but these FEM analysis has not used for modeling the real structure and interface between objects such as roughness,voids and pores of objects. We developed micro-tomography system and suggest the application of micro-tomographic image for predicting mechanical failure at the interface.The micro-tomography system consists of a 5µm micro-focus X-ray tube,a CMOS-based image sensor and a rotating sample holder controlled by a precision motor. CMOS image sensor has 62×62 mm2 sensing area and uses opticallenses system for increasing resolution.The sample which was manufactured by implanting cement in a pig hip bone was used and its fracture is considered to be an important cause of loosening of hip joint replacement in orthopedic implants. A Feldkamp’s cone-beam reconstruction algorithm on the equispatial detector case was used for bone/cement 3D volume data and the analysis of a trabecular bone/cement interface containing white-out appearance was performed by using multiple criterion segmentation of region and volume. Finally,the segmented data can be used for fracture prediction of FEM by determining node ofhexahedron meshing. In this paper,we present development of a 3-D cone beam micro-tomographic system with CMOS image sensor and its application to a complex structure of a trabecular bone and implanted cement for predicting the failure mechanism of orthopedic implants due to stress and pressure. We wi1l also show that how the segmented data is used as geometric input data of I-DEAS for FEM simulation.