WO2012109729A1

WO2012109729A1 - X-ray imaging heterostructure comprising two different photoconductor/semiconductor materials

Info

Publication number: WO2012109729A1
Application number: PCT/CA2012/000119
Authority: WO
Inventors: David M. HUNTER; Martin Yaffe
Original assignee: Hunter David M; Martin Yaffe
Priority date: 2011-02-14
Filing date: 2012-02-14
Publication date: 2012-08-23

Abstract

A digital x-ray imaging device comprising: a backthinned charge-coupled device comprising, polysilicon electrodes; an insulating layer; a n channel layer; and a low doped p-layer; and a high-Z photoconductor deposited on a backthinned face of the charge- coupled device comprising, a layer of photoconductor material; and a thin bias electrode. The polarity of the device may be inverted by using a charge-coupled device based on an n- type substrate (with a p-type channel) in which the signal is comprised of holes and applying a positive bias to the electrode.

Description

X-RAY IMAGING HETEROSTRUCTURE COMPRISING TWO DIFFERENT

PHOTOCONDUCTOR SEMICONDUCTOR MATERIALS

FIELD OF THE INVENTION

This invention relates to a digital x-ray imaging device, with high spatial resolution (>= 20 cycles mm-¹ ), low readout noise (200 q), high quantum efficiency (QE), high detective quantum efficiency (DQE), low lag, and which is applicable for breast imaging, breast-tissue sample characterization, and other medical and non-medical applications. The applicable range of x-ray energies is 2-200 keV.

BACKGROUND TO THE INVENTION

The TFT (thin film transistor) "passive" amorphous silicon based switching arrays currently in widespread use for "full field" digital radiography are limited in spatial resolution ( 7 cycles mm ¹) and their readout noise floor is about 1000 q. "Full field" refers to the fact that the x-ray image is achieved at one time using a detector which is at least as large as the object being imaged. The term "passive" means that no signal amplification occurs at the pixel/ del. The spatial resolution of full field TFT imagers is limited in part by the practicality of making large arrays comprised of very small (< 50 pm) pixels or detector elements (dels). Additionally, for passively switched TFT arrays, the noise floor cannot be reduced much below existing attained values, because of the associated capacitance of the long control lines addressing the individual pixel TFTs. In an attempt to reduce readout noise, active TFT arrays have been proposed which include an amplification element at each pixel/ del. However this introduces new complexities, plus cost, and in itself may not lead to much improved performance due to the poor characteristics of amplifying transistors based upon amorphous or polysilicon semiconductors. Furthermore, the poor operating properties of the amplifying transistors becomes progressively worse due to 1/f noise as the device is made smaller (< 100 pm ). This is because the irreducible 1/f noise of the devices becomes larger due to an increase in the surface to volume ratio of smaller transistors (the 1/f noise is associated with surfaces). On the other hand it is well known that CCDs are intrinsically capable of very low noise readout (< 200 q, even as small as 2 q for low readout frequencies) (see Janesick, J. R., [Scientific Charge-Coupled Devices], SPIE Press (2001)) and they can achieve higher spatial resolution ( > 20 cycles mm-¹ ). The low noise readout of CCDs is due, in part, to the direct deterministic management of the image signal as a discrete charge rather than the signal being based on an intermediate voltage. The image charge is transferred very efficiently across the device, with low losses, and is only converted to a voltage signal at the periphery of the device by an amplification output stage. The highest resolution capability (< 5 pm pixel pitch or equivalently a Nyquist sampling frequency of 100 cycles mm-¹) is due to the ease in which small pixels or dels can be made. This is because the simple stacked planar structure of a CCD is easy to scale to small sizes. Despite the possible advantage of x-ray detectors based upon CCDs, x-ray imaging devices based upon them has to date been limited to 10 cycles mm-¹. This is almost entirely due to the phosphor methods used to convert the x-ray fluence into a light signal which is then detected optically by the CCD.

Another attribute of full field TFT based x-ray detectors which prevents them from achieving high spatial resolution is the point source of x-rays used to illuminate them. This causes blurring due to the large angle of incidence with which the x-rays enter the detector over large parts of the image field which are away from the direct line of sight from the point source to the detector. Whereas the thickness of the x-ray absorbing material (Csl or a-Se) is typically 200-500 pm or greater, the absorbed energy is thus distributed laterally during penetration through the thickness of the material, causing blurring. This problem can be overcome in a scanning system in which the x-rays are kept normal to the detector entrance plane.

Another type of digital x-ray system exists wherein light from an x-ray absorbing phosphor screen is optically coupled to one or more optical imaging CCDs. However this type of device also suffers from the intrinsic spatial resolution limits inherent of phosphors. Furthermore, this type of device suffers from a poor detective quantum efficiency (DQE) due to the poor optical coupling of the light from the screen to the CCD(s). In general, this type of device performs more poorly than TFT array methods.

It should be noted that there are two generic classes of digital x-ray imaging systems. The first is called "indirect" because the x-ray energy is first converted into an optical signal (light) before being imaged and finally converted into an electronic signal by an optical device such as a CCD or the photodiodes at each pixel/ del location of a TFT array. Because phosphors are intrinsic to the design, spatial resolution is limited by the light diffusion in the phosphors rather than by a further reduction in size of the pixel/ del.

The second generic class of digital x-ray imagers is called "direct" because the x-ray energy is converted into an electronic signal or charge directly by the x-rays as they are absorbed by a photoconductor, usually selenium in the amorphous form (a-Se). In full field TFT arrays the signal charge is individually stored at a capacitor at each pixel/ del location until it is switched onto a readout line and converted to a voltage by an amplifier at the periphery of the TFT array. TFT full field imagers based upon a-Se have the potential to achieve very high spatial resolution because a strong vertical electric field exists in the selenium which prevents lateral blurring of the image signal. However, in practice this type of device is limited by the pixel/ del size of the TFT array and the oblique angle of incidence of the x-rays.

Two digital x-ray scanning systems have been introduced which are noteworthy. One uses a CCD detection system operating in the TDI (Time Delay Integration) mode which was developed by the Fischer corporation in collaboration with Sunnybrook Health Science Centre. However the absorbed x-ray energy is turned into light by a phosphor based absorption layer which introduces spatial blurring. The image light is optically sensed by the CCD(s) but is intrinsically limited in performance by the blurring nature of the phosphors. It should be noted, that like most x-ray imaging devices, this system develops an integrated energy x-ray signal.

Work was carried out at the Sunnybrook Health Sciences Centre (James Mainprize et al.) in which a direct method of detecting x-rays was combined with CCD readout methods in a scanned readout method. The method used CdZnTe as the photoconductor. This approach showed promise but was limited in part by the feasibilty of obtaining good CdZnTe crystals.

A second type of x-ray scanning system (the Sectra system) uses thick layers of silicon in its detector design. The silicon must be thick (3.6 mm) because silicon is a poor absorber of x- rays. However silicon has a small W value (the mean energy required to create an electron/hole pair) and thus the signal resultant from the absorption of a single x-ray photon is sufficiently large to be discriminated as a single event. This type of digital x-ray imaging device operates in the "counting mode", rather than the energy integration mode, to create x-ray images. In theory this method is nearly an ideal device but it suffers from the complexity of the supporting counting electronics which becomes ever more burdensome and expensive as the del/ pixel size is reduced to achieve higher spatial resolution.

It should also be noted that scanning x-ray imaging systems have an intrinsic advantage over full field imaging methods because scanning systems do not need an x-ray grid to reduce detrimental x-ray scatter. They do not need a grid because the volume of the object imaged at any time is small and little x-ray scatter is produced. Because x-ray grids unavoidably block some of the unscattered x-rays, the QE or DQE of full field systems is poorer than that of scanned systems by about a factor of two.

Another class of digital x-ray imaging devices can be based upon CMOS (complimentary metal oxide semiconductor) readout circuitry. The device could in principle be either of the indirect (phosphor) or direct type (photoconductor). CMOS devices, like CCD devices are low noise readout, although not as low as CCDs. Also, like CCDS, they are made from crystalline silicon and for this reason it is not as practical to make large area detectors as with TFT devices. In principle a TDI device could be constructed from CMOS, but the readout noise would exceed the low values possible with CCDs.

Summarising, in principle, a scanning system based upon a direct x-ray converter readout by a low noise readout CCD, operating in a energy-integrating, scanned TDI mode, could provide a very high resolution, high DQE, and simple x-ray imaging system which would surpass the capabilty of existing systems. Such a system would use a photoconductor, such as a- Se, which is easily deposited to the CCD readout detector and the relatively high x-ray absorption coefficient of the photoconductor (a-Se, CdZnTe, Pbl) would provide a high QE and DQE.

Applications for such a device would be breast imaging, breast-tissue sample characterization, and other medical and non-medical applications that would be apparent to a person skilled in the art. SUMMARY OF THE INVENTION

In a first embodiment, the present invention is directed to a digital x-ray imaging device comprising: a backthinned charge-coupled device comprising, polysilicon electrodes; an insulating layer; a n channel layer; and a low doped p-layer; and

a high-Z photoconductor deposited on a backthinned face of the charge-coupled device comprising,

a layer of photoconductor material; and a thin bias electrode.

The polysilicon electrodes of the backthinned CCD may be protected with a passivation material.

In a preferred embodiment, the insulating layer of the backthinned CCD comprises silicon dioxide.

The layer of photoconductor material may comprise a primary constituent selected from any one of selenium, lead, tellurium, cadmium, zinc, or other photoconductive materials or compounds having an atomic number larger than that of silicon. In a preferred embodiment, the layer of photoconductor material comprises a layer of amorphous selenium.

The thin bias electrode may be selected from Pt, Cr or Au, or other suitable electrode materials.

In yet a further embodiment, the high-Z photoconductor of the digital x-ray imaging device may further comprise an electron trapping layer between the layer of photoconductor material and thin bias electrode. In yet a further embodiment, the high-Z photoconductor may also comprise a layer of amorphous selenium with arsenic doping, and a layer of crystalline or polycrystalline selenium.

Such a digital x-ray imaging device in accordance with the foregoing is capable of being used in low dose, high resolution digital x-ray mammography, high resolution x-ray tissue biopsy, high resolution x-ray semiconductor inspection systems, high resolution x-ray materials inspection systems, and micro-computed tomography.

The polarity of the device may be inverted by using a charge-coupled device based on an n-type substrate (with a p-type channel) in which the signal is comprised of holes and applying a positive bias to the electrode. As such, in a second embodiment, the present invention is also directed to a digital x-ray imaging device comprising: a backthinned charge-coupled device comprising,

polysilicon electrodes;

an insulating layer; a p channel layer; and

a low doped n-layer; and

a high-Z photoconductor deposited on a backthinned face of the charge-coupled structure comprising, a layer of photoconductor material; and

a thin bias electrode.

In this embodiment, the polysilicon electrodes may once again be protected with a passivation material.

Also, in a preferred embodiment, the insulating layer of the backthinned CCD once again comprises silicon dioxide.

The layer of photoconductor material may again comprise a primary constituent selected from any one of selenium, lead, tellurium, cadmium, zinc, or other photoconductive materials or compounds having an atomic number larger than that of silicon. In a preferred embodiment, the layer of photoconductor material is a layer of amorphous selenium.

The thin bias electrode may once again be selected from Pt, Cr or Au, or other suitable electrode materials.

In a further embodiment, the high-Z photoconductor of the digital x-ray imaging device further comprises a hole trapping layer between the layer of photoconductor material and thin bias electrode. In yet a further embodiment, the high-Z photoconductor may further comprise a layer of amorphous selenium with arsenic doping, and a layer of crystalline or polycrystalline selenium.

A digital x-ray imaging device in accordance with the foregoing is capable of being used in low dose, high resolution digital x-ray mammography, high resolution x-ray tissue biopsy, high resolution x-ray semiconductor inspection systems, high resolution x-ray materials inspection systems, and micro-computed tomography.

BRIEF DESCRIPTION OF THE DRAWINGS

In order that the invention may be better understood, preferred embodiments will now be described by way of example only with reference to the accompanying drawings wherein:

FIGURE 1 represents an embodiment of the invention in a basic form. The various elements are described starting at the bottom of the figure and working upwards. The bottom elements are the polysilicon electrodes of the backthinned CCD. They may be protected with a passivation material not shown. The next layer is the CCD silicon dioxide insulating layer. Next is the CCD n channel layer (silicon) followed by the low doped p- layer (silicon). This completes the description of the CCD structure. Evaporated on top of the surface of the p- layer of the backthinned CCD is a thick layer of selenium in the amorphous form (a-Se). The a-Se thickness may range from a few tens of microns to one millimeter in thickness. The top x-ray facing side of the a-Se is provided with a thin bias electrode made of Pt, Cr, or other suitable electrode materials. It is possible to replace the a-Se with another photoconductor material (Cd, Zn, Te, etc.) which also has a bandgap larger than c-Si. It is also possible to invert the polarity of the device by inverting the doping arrangement of the c-Si (i.e. n->p, p->n) for the two silicon doped layers. Also in such a case the a-Se would then be biased positively.

FIGURE 2 represents a slightly different embodiment to that shown in Figure 1, in which an electron trapping layer has been added just under the a-Se bias electrode. The purpose of this is to lower the injection rate of electrons into the a-Se. If the polarity of operation of the device were to be reversed, the layer would be hole blocking layer.

FIGURE 3 represents yet another slightly different embodiment to that shown in Figures 1 or 2, in which an electron trapping layer has been added just under the a-Se bias electrode. The purpose of this is to lower the injection rate of electrons into the a-Se. If the polarity of operation of the device were to be reversed, the layer would be hole blocking layer, and the semiconductor types would be reversed (i.e. p -> n, n->p).

FIGURE 4 displays a band diagram of a-Se/ c-Si interface with the a-Se negatively biased. The c- Si is lightly p-doped, moving the intrinsic c-Si Fermi level £, to E/ , nearer to the c-Si valence band E_v. The a-Se bandgap E_ss_e, 2.2 eV, is twice the c-Si bandgap, E_gsi. As the diagram is drawn, electrons would be free to enter the c-Si from the a-Se to the c-Si, whereas the freely available holes from the p-doped c-Si would be blocked from entering the a-Se and therefore accumulate at the a-Se/ C-Si interface. X-ray photons are absorbed in the a-Se and the energy deposited creates free electron-hole pairs.

FIGURE 5 displays non-imaging sample measurement equipment and arrangements for three different sample current test conditions: (1) dark signal only, (2) time of flight (TOF) using pulsed (several

blue semiconductor laser and (3) x-ray exposures (-100 ms). The a-Se bias electrode for the three different configurations was set by the programmable HV supply. The leftmost arrangement used an electrometer for the dark current measurement. The TOF signal in the center arrangement was measured directly through the 50 Ohm input of a digital scope. The x-ray current signal was measured by a custom built current amplifier. For the x-ray measurements the time response of the x-ray pulse was independently measured by a photodiode exposed to the x-ray beam.

FIGURE 6 displays a cross section of a backthinned CCD coated with a-Se.

FIGURE 7 displays a linear plot of a typical I-V curve of a non-imaging a-Se/c-Si heterostructure demonstrating the rectifying property of the sample. The c-Si substrate was p- type and the a-Se bias electrode was platinum. When the bias electrode is negative very little current flows in the structure.

FIGURE 8 displays a log scale plot of current with negative field applied to the a-Se. The absolute value of the current is plotted and the absolute value of the electric field is greatest at the left of the plot. The dark current remains small (-100 nA cnr² at 14 Volt μητ¹) at the field strengths (~ lOVpnrr¹) required to operate a-Se with a good x-ray sensitivity.

FIGURE 9 displays optical time of flight measurement of electron transport across a heterostructure sample. The initial rising slope of the current is due to the 5 ps length of the laser pulse. The duration of the laser pulse is illustrated in the upper curve. FIGURE 10 displays temporal x-ray response of a a-Se/c-Si heterostructure (upper curve). The length of the x-ray exposure is 100 ms and the total exposure was 96 mR. Plotted in the lower curve is the response of a small photodiode also placed in the x-ray beam. It can be seen that the a-Se/c-Si structure has a small increase in the signal during the exposure (ideally it should be constant) and that there is a small residual lag signal at the termination of the exposure.

FIGURE 11 displays x-ray signal as a function of the applied electric field. As expected the sensitivity of the a-Se increases sublinearly as the field is increased.

FIGURE 12 displays an a-Se CCD x-ray image of a lead bar pattern at an angle to the CCD scan direction. Most of the visible response is due to the signal from x-ray absorption in the c-Si, but there are spots with much higher sensitivity which is attributed to the a-Se. Due to window and leveling limitations the spots may not be readily visible in this image. The vertical lines are due to component defects.

FIGURE 13 displays x-ray response of the CCD for the regions attributed to silicon and selenium. The saturation of the a-Se response at exposure times greater than ~ 300 ms is due to the CCD wells becoming full.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a low noise readout (<200 q) scanning (slot) or staring mode (single picture) x-ray imaging detector capable of high spatial resolution (> 20 cycles mnv ^a) with a high QE and/ or high DQE in the x-ray energy range of 2 to 200 keV. The device has low image lag due to the direct transfer of the x-ray generated signal charge from the x-ray photoconductor layer to the image storage layer of a backthinned CCD. A backthinned CCD is prepared from any ordinary CCD by removing the majority of the silicon substrate (-0.5 mm thick) so that access (normally optical) to the active region of the charge storage c-Silicon (normally an epitaxial layer 10-20 thick) is possible rather than access through the "top" polysilicon electrode face. The surface from which the substrate has been removed is herein referred to as the backthinned face. The backthinned CCD may be mechanically supported by a passive silicon substrate (0.5 mm or thicker) placed on the polysilicon electrode face. In this invention, a high-Z photoconductor, nominally a-Selenium, is deposited on the backthinned face of the CCD. The device operates by having the x-ray image field expose the photoconductor from the backthinned side, rather than through the polysilicon electrode face. The transfer of the charge is partially assisted by a large electric field which is impressed across the photoconductor/CCD structure. Because the need for discrete charge transfer switching elements has been obviated, the pixel/ del size can be made very small and is principally limited by the minimum pixel/ del size of the CCD. Furthermore there is no need to pixellate the photoconductor and align a separate pixellated photoconductor structure to a pixellated CCD structure via indium bump bonding or other similar methods. In the case of a-Se, the structure is assembled by vacuum deposition processes. Because the photoconductor and the CCD act as a single continuous united element, the low noise TDI imaging performance of CCD is retained and comparable image resolution is obtained in both the slot and scan directions.

The invention employs two different photoconductor/ semiconductor materials of different band-gap energies in a non-pixellelated heterostructure layer/ interface. The heterostructure interface is constructed in a manner to minimize interface induced mid band- gap states so that recombination of charge carriers at the interface is minimized, and such that a continuous vertical electric field exists across the interface of the two materials.

The device has a first layer material comprising a primary x-ray absorbing photoconductor material (which may be amorphous, polycrystalline, or crystalline), the primary constituent of which may be selected from selenium, lead, tellurium, cadmium, zinc, or other photoconductive materials or compounds having an atomic number larger than that of silicon. A preferred embodiment of the primary x-ray absorbing photoconductor material comprises amorphous selenium (a-Se). This is shown in Fig. 1. The primary photoconductor constituent may also be alloyed with other elements that would be apparent to a person skilled in the art, such as arsenic, chlorine, iodide, or tellurium, amongst others.

The device has a second layer material comprising a semiconductor layer made of crystalline silicon (c-Si) employed in an image readout structure and is comprised of a back- thinned CCD. The x-ray absorbing photoconductor material is formed on the back-thinned surface of the CCD. The amorphous selenium (a-Se) has a relatively high atomic number (Z) and high x-ray attenuation coefficient, while the crystalline silicon (c-Si) has a relatively low atomic number and x-ray attenuation coefficient.

A thin bias electrode (Pt, Cr, Au, or other suitable materials), which absorbs only a small fraction of the incident x-rays, is formed on the free surface of the x-ray absorbing photoconductor which faces the mcoming x-rays. In this manner, the device is generally held under a high voltage bias with the x-ray facing side of the a-Se layer held negatively (mamtaining a very low dark current as required for an x-ray imaging device), and the c-Si layer near ground potential. The natural "semiconducting'' properties of the two different materials are conducive to facilitating the transfer of the x-ray generated signal electrons in the a-Se layer to the p-type CCD c-Si layer, whilst simultaneously preventing the "backflow" of a dark signal from the c-Si to the a-Se layer. It is possible to invert the polarity by using a CCD based on an n-type substrate (with a p-type channel) in which the signal is comprised of holes and applying a positive bias to the a-Se electrode. The spatial binning of the x-ray signal charges occurs through the vertical (no lateral spreading) transfer of the a-Se signal charge to the c-Si interface and subsequent sequestration into separate positive potential wells inherent in the CCD readout structure. The larger band-gap of the a-Se (>2.0 eV) compared to the band- gap of the c-Si (1.1 eV at room temperature) is an essential feature of this interface because the image electrical signal, which may be either electrons or holes, is generated in the x-ray absorbing, higher atomic number photoconductive layer with the large band-gap material and transferred directly to the lower band-gap c-Si material. This transfer of the electrical charge between the two different materials is not based upon the existence of discrete pixellized switching devices, but is rather transferred naturally and in a spatially contiguous manner across the interface of the two different materials. Only once the signal is captured within the CCD, where it encounters spatially separated potential wells, does the signal become "pixellated".

The high spatial resolution of the device is attributable to properties of the a-Se or high- Z photoconductor under the high bias, the thin layer of the active portion (epitaxial layer) of the c-Si device which minimises lateral diffusion of signal charge, and the ability to make high pitch potential wells in the c-Si device. Because the signal charges may undergo a diffusion process once injected into the CCD, it is generally preferred that the thickness of the backthinned layer be limited to under 15 pm.

In another preferred embodiment, additional layers may be introduced to enhance function. One such enhancement is the introduction of an electron trapping layer at the x-ray bias electrode as shown in Fig. 2. This aids in the suppression of injection of dark electron signal.

In yet another embodiment, it may be preferred to enhance the natural properties of the a-Se/c-Si interface by the generation of tertiary interface materials (e.g. deliberate creation of a thin crystalline layer of selenium (such as the trigonal form) at the c-Si interface of the CCD) and stabilised (impervious to crystallization) a-Se. The purpose of this layer is to facilitate transfer of the x-ray signal layer in the a-Se layer above it to the c-Se and from the c- Se to the c-Si. A layer of a-Se doped with arsenic may be positioned just above the c-Se in order to suppress spontaneous crystallization of the pure a-Se layer above it. These arrangements are illustrated in Fig. 3. Other arrangements are possible and would be readily understood by persons skilled in the art in view of the foregoing.

It may also be preferred, either prior to or during the operation of the device, to uniformly flood the backthinned face (the x-ray facing side) with a weak light in order to fill charge traps which might exist at the c-Si/ photoconductor interface.

Additional Interface Preparation Considerations

Prior to evaporating selenium or other high-Z photoconductors onto the backthinned CCD silicon surface, the silicon surface may need special preparation such as removal of native S1O2 by a weak solution of HF and deionized water. It may be preferable to prepare the free surface of the Si with a (mono)layer of Se by MBE (molecular beam epitaxy). In general, it may be preferable to prevent formation of S1O2 (prior to the x-ray photoconductor deposition) or in the contrary it may be preferable to form, in a controlled way, a thin S1O2 layer with good Si passivation properties not obtainable from the native S1O2 form.

In general the interface of the c-Si and x-ray photoconductor should be charge neutral and as free as possible from defects which might act as traps or recombination centres.

It may be preferable to alter the "built in" electric field profile at the c-Si interface by delta doping the c-Si surface prior to forming the x-ray photoconducting layer (nominally selenium).

The present invention has application in low dose, high resolution digital x-ray mammography, high resolution x-ray tissue biopsy, high resolution x-ray semiconductor inspection systems, high resolution x-ray materials inspection systems, and micro-computed tomography, among others, that would be known to persons skilled in the art to which this invention pertains.

The present invention provides a high spatial resolution (20 mm-¹) low noise digital radiography method for breast imaging, breast-tissue sample characterization and other possible medical applications. In particular, the present invention provides a low noise readout (<200 q) scanning (slot) or staring mode x-ray imaging detector capable of high spatial resolution (> 20 mm⁴) with a high QE (quantum efficiency) and/ or high DQE (detective quantum efficiency) in the x-ray energy range of 2 to 200 keV. The device has low image lag, thereby permitting imaging performance comparable to the orthogonal slot direction when in the TDI (time delay integration) scanning mode.

Various aspects of the device and method of the present invention may be demonstrated by reference to the following experiment discussed below.

Experiment

We investigated the creation of an x-ray imaging method based on a planar heterostructure of a-Se and c-Si. In such a device, the a-Se layer absorbs x-rays efficiently and retains high spatial frequency content due to the vertical electric fields in the a-Se. The thin c-Si layer (10-20 pm), part of a back-thinned CCD, receives the electron charge signal directly from the a-Se layer. The creation of large area x-ray images with such a device is achieved by operating it in a scanned, time delay integration (TDI) mode. However, for such a heterostructure device to work, it was necessary that it meet certain requirements, namely 1) it had to be possible to apply a large potential to the structure which would support a large electric field in the a-Se (10 Vpnv¹) but not disturb the operation of the CCD; 2) electrons generated in the a-Se had to be able to transfer to the c-Si CCD; 3) holes from the p-doped c-Si had to be blocked from entering the a-Se otherwise an excessive dark signal would occur and the electric field in the a-Se would not be maintained; and 4) at the interface of the a-Se and c-Si, recombination of carriers had to be minimized by controlling the density of recombination centres and an electric field of the correct polarity should be present to push the electrons into the bulk of the c-Si.

It was important to note that the band-gap/mobility-gap of a-Se (2.2 eV) is twice the band-gap of c-Si (1.1 eV). This can promote the transmission of electrons from the a-Se to p- type c-Si and simultaneously block the transmission of holes from p-type c-Si into the a-Se, when the a-Se is negatively biased. Theory

Although a-Se is often described as a semiconductor, a more accurate description would be that it is a photo-insulator with a mobility/ bandgap of E_gs_e - 2.2 eV and capable of supporting non-dispersive transport of excess electron and hole carriers created by the absorption of light or x-ray energy. Shown in Fig. 4 is a simplified band-gap diagram of the preferred a-Se/c-Si structure made with p-type c-Si (~5xl0¹⁴ cm ³ ) and with the a-Se biased negatively. The structure is very similar to a metal oxide semiconductor (MOS) capacitor. For p- type c-Si, when the device is biased negatively, the a-Se/ c-Si interface is in accumulation mode with an accumulation of holes at the interface. When the device is biased positively, the a-Se/c- Si interface can go into inversion mode at a bias threshold voltage which is dependent upon the thickness of the a-Se layer.

The work function differences between the metal and the a-Se were ignored (see Johanson, R. E., Kasap, S., Rowlands, J., and Polischuk, B., "Metallic electrical contacts to stabilized amorphous selenium for use in X-ray image detectors," Journal of Non-Crystalline Solids 227-230, 1359-1362 (May 1998)), but the exact interface physics of the a-Se metal bias electrode was of secondary interest to the physics of the a-Se/c-Si interface. The band-gap of the c-Si (1.1 eV) is ~ 1/2 that of the a-Se. It was assumed that the Fermi level of the a-Se is mid-gap, pinned by trap states. With the alignment of the Fermi levels between the a-Se and c-Si occurring based on this assumption, under a negative a-Se bias the accumulated holes at the a- Se/c-Si interface from the p type c-Si experienced a very large blocking barrier of ~ 1 eV. Ignoring trapping states, a-Se has very few thermally generated free carriers and can be considered as an insulator, thus almost all of the bias potential appears across the a-Se layer maintaining a large electric field in the a-Se layer. Electron-hole-pairs (EHP) are created in the a- Se when x-ray energy is absorbed and are collected by the a-Se electric field. As in existing a-Se imaging detectors which are biased negatively, the holes created in the a-Se bulk are collected at the bias electrode and recombine with bias metal conduction electrons. The signal electrons move under the action of the a-Se field to the a-Se/c-Si interface. As can be seen from Fig. 4 at the conduction band alignment of the a-Se/c-Si interface there should be a modest electric field pushing the a-Se electrons into the c-Si.

E E

In Fig. 4 the c-Si is lightly p-doped, moving the intrinsic c-Si Fermi level ' to ' nearer to the c-Si valence band F ^v . The a-Se bandgap F

s& , 2.2 eV, is twice the c-Si bandgap, gSi . As the diagram is drawn, electrons would be free to enter the c-Si from the a-Se to the c- Si, whereas the freely available holes from the p-doped c-Si would be blocked from entering the a-Se and therefore accumulate at the a-Se/C-Si interface. X-ray photons are absorbed in the a-Se and the energy deposited creates free electron-hole pairs.

Summarizing, the configuration as shown in Fig. 4 meets the goals of creating a a-Se/ c-Si imaging device: under negative bias there was little dark current because of hole-blocking a- Se/c-Si interface, while electrons created in the bulk a-Se from x-radiation were able to travel to the c-Si and be pushed into the c-Si layer.

Materials and Methods

We conducted three different types of non-imaging experiments on samples using c-Si substrates with an a-Se layer deposited on the c-Si substrate, as in Fig. 5.

Figure 5 shows non-imaging sample measurement equipment and arrangements for the three different sample current test conditions: (1) dark signal only, (2) time of flight (TOF) using pulsed (several β) blue semiconductor laser and (3) x-ray exposures (~100 ms) . The a-Se bias electrode for the three different configurations was set by the programmable HV supply. The leftmost arrangement used an electrometer for the dark current measurement. The TOF signal in the centre arrangement was measured directly through the 50 Ohm input of a digital scope. The x-ray current signal was measured by a custom built current amplifier. For the x-ray measurements the time response of the x-ray pulse was independently measured by a photodiode exposed to the x-ray beam.

We fabricated a back thinned CCD imaging device with an a-Se x-ray detection layer deposited directly on the thinned side of the CCD. For the non-imaging samples the following four parameters were assessed: 1) the dark current was measured as a function of applied bias (both the polarity and magnitude of the bias were varied); 2) optical time of flight (TOF) measurements of electrons/ holes transiting through the sample were measured (the charges were photo-injected from the top semi-transparent bias electrode on the a-Se (see Fig. 5); 3) the temporal signal response of the samples to a pulsed x-ray source were measured; and 4) the x- ray signal response was quantitatively measured and the effective ionization energy (W) of the a-Se was calculated.

Sample Preparation

The non-imaging samples were prepared by vacuum deposition of selenium onto the c- Si samples, which were p-doped. In some samples the c-Si substrate was obtained from the same wafers used to make certain commercial CCDs. These wafers had a p-minus epitaxial layer of approximately 20 pm. The a-Se was deposited with the c-Si native S1O2 left intact or in some cases the native oxide was removed with a weak HF (1% hydrofluoric acid) solution immediately prior to the vacuum selenium deposition. The non-imaging samples were made with pure selenium or selenium stabilized with 0.5 % As. The thickness of the a-Se was typically 50 pm, with the a-Se thickness of each sample determined by an evaporation rate meter which was part of the selenium vacuum deposition system. A thin circular bias electrode (Pt) with an area of 1 cm² was added to the a-Se surface of each sample. It is important to note that no special procedures were undertaken to produce blocking layers at either the a-Se/ C-Si interface or the a-Se/bias electrode interface. In the literature these are sometimes referred to as "P-I-N" structures, where "P" indicates a region filled with electron traps, "I" refers to a region with a relatively few number of traps, and "N" refers to a region filled with hole traps. The creation of electron or hole traps can be facilitated by the use of impurities or by reducing the temperature of the substrate at the time of the selenium deposition (see Kasap, S. O., Koughia, K. V., Fogal, B., Belev, G., and Johanson, R. E., "The influence of deposition conditions and alloying on the electronic properties of amorphous selenium," Semiconductors 37, 789-794 (July 2003) and Belev, G. and Kasap, S., "Reduction of the dark current in stabilized a-Se based X-ray detectors," Journal of Non-Crystalline Solids 352, 1616-1620 (June 2006)). During the selenium evaporation the substrates were kept at 50 deg C, the nominal glass transition temperature a of selenium. Dark I-V measurement apparatus

The dark currents were measured with a Keithley 617 programmable electrometer with a IEEE-488 bus interface. A IEEE-488 programmable Keithley 248 High Voltage (HV) power supply was used to apply the bias to the structure. Measurements were computer automated using a C++ program which controlled the equipment via a USB to IEEE-488 bus interface. The samples were put inside a light-tight, electrically-shielded box, with the appropriate connections made to the HV power supply and electrometer as shown in the left of Fig. 5. All measurements were performed at room temperature (22 deg C). At the beginning of a measurement sequence the bias applied to the sample was zero. The bias was incremented in approximately 10 V increments. When the bias is incremented, the dark current jumps, but rapidly declines to a lower value in several minutes. Therefore after each bias increment the current was measured after a consistent wait period of several minutes. The nature and cause of this transitory current behaviour is complex and has been studied elsewhere (see Kasap, S. O. and Belev, G., "Progress in the science and technology of direct conversion X-ray image detectors : The development of a double layer a-Se based detector," Journal of Optoelectronics and Advanced Materials 9(1), 1 - 10 (2007)). The I-V curves were created by the monotonic increase of the absolute value of the bias.

Optical Time of Flight measurement apparatus

The optical TOF measurements were performed using a pulsed blue solid-state laser. The laser light was guided to the surface of the sample by a fibre optic cable. The sample substrate was grounded through the 50 ohm input of a 500 MHz digital oscilloscope as indicated in Fig. 5. The current response was directly measured by the oscilloscope. The mobility of electrons and holes in a-Se depends on sample preparation methods. The mobility for holes may range between 0.1-0.16 cm² V⁴ s-¹ while for electrons the range is 0.002-0.007 cm² V"¹ s ¹ . Because the electron mobility of electrons is ~ 10 x less than that of holes, and due to the limited brightness of the laser, the laser excitation pulse for the electron TOF measurements was about 10 x longer than for the hole TOF measurements. The amount of space-charge injected by the laser light was maintained small compared to the stored charge on the biased sample. X-ray signal measurement apparatus

The x-ray source was a mammographic x-ray tube with a tungsten target and molybdenum filtration (30 pm). The tube was operated at a constant potential of 45 kV. It was pulsed with a precisely timed logic signal. It is estimated that the rise and fall time of the x-ray pulse was of the order of 1-2 ms at the x-ray tube current employed. The finite rise and fall time of the x-ray pulse was due primarily to the capacitance of the HV cables and the x-ray tube. The x-ray signal was independently monitored by a small photodiode. The x-ray exposure was measured using a Radcal model 2026C dosimeter. A transresistance current-to-voltage amplifier employing an instrumentation amplifier (INA128P) was built to measure the x-ray response.

The measured transresistance ' / of the circuit was 2.5xl0⁶ Ohms, where V is the output voltage of the amplifier and / is the input current. A fourth order Butterworth low-pass filter with bandwidth 5 kHz was incorporated in the design to reduce pickup noise from the high voltage x-ray power supply operation.

Backthinned CCD

The CCD, part number STA0510A, was obtained from the University of Arizona (UA), Imaging Technology Laboratory (ITL), Tucson, Arizona. The device is normally used for astronomical optical applications. The CCD is a three phase device with a 15 μπτ pixel/DEL pitch in a bifurcated (1200x400)x2 array with two output channels. The CCD was backthinned and packaged by the UA ITL. The selenium deposition onto the backthinned free surface was done at the University of Saskatchewan. The CCD data acquisition electronics and software were created at the Sunnybrook Health Sciences Centre. Only one of the CCD output channels was used and the CCD was clocked in a manner to achieve a single 1200 x 800 image frame.

The CCD was backthinned so that approximately 17 μιη of the c-Si wafer was remaining. The backthinning was completed in the normal way for optical CCDs, which includes the final deliberate careful formation of a Si0₂ layer (see Lesser, M. P., "Improving CCD quantum efficiency," in [Proc. SPIE 2198, 782], (1994)) of the backthinned surface. The purpose for this is to passivate the c-Si surface and reduce surface states. Because an oxide layer seems opposed to our intention of transferring charge into the CCD, the decision to retain this process was not taken lightly, but it was thought the best choice for our initial experiments. The backthinned CCD itself is supported on a secondary support wafer piece as shown in Fig. 6. Selenium containing arsenic (0.2% wt), for stabilization against crystallization, was vacuum deposited onto the CCD in a manner similar to that for the non-imaging samples. A thin platinum HV bias electrode was deposited on the x-ray facing side of the a-Se layer.

Results

I-V dark current

In general the a-Se/c-Si structures were rectifying as can be seen from the data of Fig. 7. Figure 7 shows a linear plot of a typical I-V curve of a non-imaging a-Se/c-Si heterostructure demonstrating the rectifying property of the sample. The c-Si substrate was p-type and the a-Se bias electrode was platinum.

When the bias electrode is negative very little current flows in the structure. When the a- Se is negatively biased the dark current is much smaller than when positively biased. The very small level of dark current when the device is biased negatively is shown more clearly in Fig. 8. Figure 8 shows a log scale plot of current with negative field applied to the a-Se. The absolute value of the current is plotted and the absolute value of the electric field is greatest at the left of the plot. The dark current remains small (-100 nA cm-² at 14 Volt μητ¹ ) at the field strengths (~10V μην¹) required to operate a-Se with a good x-ray sensitivity. The lowest dark currents with a negative bias were obtained from the samples made from the CCD wafers with the p-doped epitaxial layer and which had been etched with a weak HF solution to remove the native S1O2 immediately prior to the selenium evaporation.

Optical TOF results

Shown in Fig. 9 is the electron time of flight response of a a-Se/ c-Si structure to a pulse of optical excitation. The initial rising slope of the current is due to the 5 μ s length of the laser pulse. The duration of the laser pulse is illustrated in the upper curve. If it is assumed that the bias drops entirely across the a-Se the electron mobility calculated is in agreement with the range of a-Se mobilities measured by other research groups. There is a small residual current signal at times greater than expected for the simple non-dispersive transit of electrons across the a-Se layer. TOF measurements for holes seemed to show a smaller residual current after the transit time of the last charge carrier front should have been completed. X-ray signal results

The x-ray signal temporal response is shown in Fig. 10. The length of the x-ray exposure is 100 ms and the total exposure was 96 mR. Also plotted in Fig. 10 is the response of a small diode which was also placed in the x-ray beam. The plot of the diode response is vertically shifted downwards from the a-Se/c-Si data plot for clarity. There is a slight increase in the a-Se signal current during the 100 ms duration x-ray exposure and a small lag signal at the end of the exposure. These effects may be due to trapping / detrapping in shallow trap states of the a-Se layer. In the paper by Stone et al. (See Stone, M. F., Zhao, W., Jacak, B. V., OConnor, P., Yu, B., and Rehak, P., "The x-ray sensitivity of amorphous selenium for mammography," Medical Physics 29(3), 319 (2002)) the x-ray response of their samples shows a large increase in the x-ray signal over time, larger than in our samples, but their exposure was longer (~1 s), and the total exposure is not directly stated.

Shown plotted in Fig. 11 is the magnitude of the a-Se/ c-Si response as a function of the applied electric field. The x-ray tube potential was 45 kV and assuming a 30 μιη Mo filter the mean energy E of the x-rays was estimated to be 19.8 keV, and the photon fluence rate ^ψ was estimated to be 5.21xl0⁶ photons mR-¹ cm-² . The estimates were determined using the semiempirical x-ray spectra generation methods of Tucker et al. (See Tucker, D. M., Barnes, G. T., and Charkraborty, D. P., "Semiempirical model for generating tungsten target x-ray spectra,"

Med. Phys. 18(2), 211-218 (1991)). Using the simulated spectrum, quantum efficiency Ί of the 45 μη thick a-Se layer was estimated at 0.51, although this value has not explicitly considered

^ ^~~ fluorescence escape mechanisms. With a 450 Volt bias applied to the sample (Fig. 11) the selenium electric field £¾ ⁼ Ιθνμπν¹ and the measured signal was 435 mV, which referred to the input of the amplifier, is a current / of 174 nA or l.lxlO¹² q s-¹ , where q is the unit of elementary charge. The measured exposure rate X was 960 mRs ¹ . The following equation

ψΧΑηΕ [ εΥ \ (1)

in which A is the area of the detector (1 cm² ), can be used to calculate W± and the value of W± evaluates to about 47 eV. This value is consistent with values reported in the literature (See Stone, M. F., Zhao, W., Jacak, B. V., OConnor, P., Yu, B., and Rehak, P., "The x-ray sensitivity of amorphous selenium for mammography," Medical Physics 29(3), 319 (2002); and Blevis, I. M, Hunt, D. C, and Rowlands, J. a., "Measurement of x-ray photogeneration in amorphous selenium," Journal of Applied Physics 85(11), 7958-7963 (1999)) where it has also been shown that there is a weak energy dependence of W± on the value of E, the x-ray energy. The value of W± for selenium decreases (sublinearly) as Cs_e increases or equivalently the expected signal response, ^s , should increase sublinearly as Cs_e increases. Shown plotted in Fig. 11 is a curve of the form S=k G^Se (k a constant) which has been fitted to the data. The value of the sublinear parameter γ is approximately 0.64 which is slightly smaller than 0.70 as reported elsewhere (See Stone, M. F., Zhao, W., Jacak, B. V., OConnor, P., Yu, B., and Rehak, P., "The x-ray sensitivity of amorphous selenium for mammography," Medical Physics 29(3), 319 (2002)).

CCD Results

Prior to the backthinned CCD being coated with a-Se, the CCD and associated readout electronics were tested optically. The electronics were adjusted, and ultimately the CCD was found to be operating properly, although the dark current was quite high. The CCD, without selenium, would saturate with dark signal after about 2.0 s, at room temperature. The clocking readout time of one frame was -640 ms, and CCD clocking permitted arbitrary length frame integration times to be applied by "freezing" the clocks.

A mask was prepared for the selenium deposition in order to limit evaporation to the active imaging area of the CCD and not to contaminate the wire-bonding at the periphery of the CCD. The selenium was evaporated, using a substrate temperature of 50 deg C, and finally a thin Pt bias was sputtered onto the top of the a-Se. The thickness of the a-Se selenium layer was 200 μπτ.

Tests were then conducted with the backthinned CCD coated with selenium. Dark current images were obtained with and without the selenium bias applied. With the selenium deposition on the CCD, the dark current was found to be reduced by about a factor of two so that the dark signal saturation did not occur until 5 s at room temperature. The maximum bias applied to the bias electrode was -1000V. The CCD continued to function when the bias was applied, and there was only a small increase in the dark signal under bias, so that most of the dark signal in the device did not originate from the a-Se. X-ray images were obtained with and without the selenium bias applied. There appeared at first to be no difference with the bias on or off. It was then realized that a weak x-ray image was visible from the direct absorption of x-rays, which had passed through the a-Se, into the c-Si layer of the CCD. Furthermore it was also noticed that there were a few spots where the x-ray signal was much larger. It was determined that the spots showing increased x-ray sensitivity only did so when the a-Se bias was applied. The spots did not show any significant increase in dark current.

Figure 12 shows an a-Se CCD x-ray image of a lead bar pattern positioned at an angle to the CCD scan direction. Most of the image content is due to the signal from x-ray

absorption in the c-Si, but there are spots with much higher sensitivity which is attributed to the a-Se. The working spot regions had all of the correct attributes that the invention requires such as being dependent upon the selenium bias, low dark current, and spatially localized with respect to the x-ray image information. This was tested by movement of the bar pattern to verify that the response only occurred when illuminated by the x-rays. The response of these regions was quantified and compared to theoretical expectations and found to agree with theory. This is shown in detail in Fig. 13. Furthermore, annealing the whole device, by bringing the device near the glass transition of a-Se increased the number of working regions significantly, providing evidence that an interfacial layer of crystalline selenium may under certain circumstances be needed in a preferred operation of the device. Due to window and levelling limitations the spots may not be readily visible in this image. The vertical lines are due to component defects.

The dark and x-ray signal were measured for those regions showing increased x-ray sensitivity and for those not; the hypothesis to test was that the two different responses were due to signal generation by the silicon layer and the selenium layer. The x-ray signal was increased by increasing the exposure time. The x-ray signal for both regions was corrected by subtracting the dark current. Shown in Fig. 13 is plotted the relative response of the two different regions. The smaller response Si has a relative slope (AS / X) of 2.5 whereas the greater response Se has a slope (ASJAX) of 15.9. Taking into consideration the thickness of the a-Se, the c-Si thickness, and the x-ray spectrum, it was estimated that the fraction of the incident x-ray energy absorbed by the c-Si, ESi, was 2.8x10³ and for the a-Se, * ' it was 0.72. It follows then that the relative signal gain of the two regions is:

where the W± values for Si and a-Se have been taken to be 3 and 100 eV respectively.

The calculated ratio Q is 7.7 whereas the measured ratio of the two responses S_< S„ is 6.4, which is close to the calculated expected ratio.

Figure 13 shows the X-ray response of the CCD for the regions attributed to silicon and selenium (see Fig. 12). The saturation of the a-Se response at exposure times greater than ~ 300 ms is due to the CCD wells becoming full, this provides further direct evidence that the invention is working properly in the regions showing the much greater x-ray response. The calculated relative responses of the image parts (Fig. 12) attributed to silicon and a-Se were carried as described in the main body of the text.

We believed the reason why small regions permitted the x-ray signal developed in the a- Se layer to be transmitted to the c-Si of the CCD was that trigonal selenium had formed at those locations. If this were true, annealing the whole device near the T s of selenium would be expected to promote the formation of more regions. This hypothesis was tested by heating the device to temperatures near T s for increasing lengths of time (1/2, 1, 2 hours) and checking if the detector developed more regions of increased x-ray sensitivity after each annealing. The x- ray sensitive regions increased in number over the entire device with the annealings.

Conclusions

It has been shown experimentally that an a-Se/c-Si heterostructure exhibits rectifying properties which are consistent with a simple band diagram of the interface which would indicate electrons can be transferred to the c-Si from the a-Se while holes are blocked from leaking into the a-Se when the a-Se is negatively biased. It is recognized that this asymmetry may also be due in part to the details of selenium bond and trap formation at the interfaces. The dark currents measured from the a-Se/c-Si heterostructures when the a-Se is reversed biased are remarkably low and exceed the requirements necessary for an x-ray imaging device using a-Se and a CCD readout method. TOF measurements made on the non-imaging heterostructures showed normal electron and hole transit times showing that the a-Se was properly formed and that the a-Se had the electric field magnitude one would expect if the bias applied to the heterostructure dropped almost entirely across the a-Se layer, as expected. The x-ray response of the non-imaging samples was measured and the response was as expected. The W ± value determined for an a-Se layer 45 μπι thick, at C

and a mean x-ray energy of 20 keV, was found to be 47 eV.

A backthirmed CCD has been coated with a layer of a-Se and a bias electrode on the x- ray facing side of the a-Se. The CCD device is fully functional as an imaging device when a large bias (1000 V) is applied to the a-Se layer. However only small regions of the a-Se/ c-Si interface permit proper transfer of the x-ray generated electron signal in the a-Se layer into the CCD. For the sample tested, the regions where the electron signal is properly transferred is attributed to the presence of selenium crystallization.

Claims

We claim:

1. A digital x-ray imaging device comprising: a backthinned charge-coupled device comprising, polysilicon electrodes; an insulating layer; a n channel layer; and a low doped p-layer; and a high-Z photoconductor deposited on a backthinned face of the charge-coupled device comprising, a layer of photoconductor material; and a thin bias electrode.

2. The device of claim 1, wherein the polysilicon electrodes are protected with a passivation material.

3. The device of claim 1, wherein the insulating layer comprises silicon dioxide.

4. The device of claim 1, wherein the layer of photoconductor material is a layer of amorphous selenium.

5. The device of claim 1, wherein the layer of photoconductor material comprises a primary constituent selected from any one of selenium, lead, tellurium, cadmium, zinc, or other photoconductive materials or compounds having an atomic number larger than that of silicon.

6. The device of claim 1, wherein the thin bias electrode is Pt, Cr or Au.

7. The device of claim 1, wherein the high-Z photoconductor further comprises an electron trapping layer between the layer of photoconductor material and thin bias electrode.

8. The device of claim 7, wherein the high-Z photoconductor further comprises a layer of amorphous selenium with arsenic doping, and a layer of crystalline or polycrystalline selenium.

9. A digital x-ray imaging device as claimed in any one of claims 1 to 8 that is capable of being used in low dose, high resolution digital x-ray mammography, high resolution x-ray tissue biopsy, high resolution x-ray semiconductor inspection systems, high resolution x-ray materials inspection systems, and micro-computed tomography.

10. A digital x-ray imaging device comprising: a backthinned charge-coupled device comprising, polysilicon electrodes; an insulating layer; a p channel layer; and a low doped n-layer; and a high-Z photoconductor deposited on a backthinned face of the charge-coupled structure comprising, a layer of photoconductor material; and a thin bias electrode.

11. The device of claim 10, wherein the polysilicon electrodes are protected with a passivation material.

12. The device of claim 10, wherein the insulating layer comprises silicon dioxide.

13. The device of claim 10, wherein the layer of photoconductor material is a layer of amorphous selenium.

14. The device of claim 10, wherein the layer of photoconductor material comprises a primary constituent selected from any one of selenium, lead, tellurium, cadmium, zinc, or other photoconductive materials or compounds having an atomic number larger than that of silicon.

15. The device of claim 10, wherein the thin bias electrode is Pt, Cr or Au.

16. The device of claim 10, wherein the high-Z photoconductor further comprises a hole trapping layer between the layer of photoconductor material and thin bias electrode.

17. The device of claim 16, wherein the high-Z photoconductor further comprises a layer of amorphous selenium with arsenic doping, and a layer of crystalline or polycrystalline selenium.

18. A digital x-ray imaging device as claimed in any one of claims 10 to 17 that is capable of being used in low dose, high resolution digital x-ray mammography, high resolution x-ray tissue biopsy, high resolution x-ray semiconductor inspection systems, high resolution x-ray materials inspection systems, and micro-computed tomography.