Mercuric Iodide and Lead Iodide X-Ray Detectors for Radiographic and Fluoroscopic Medical Imaging

G. Zentai, L. Partain, R. Pavlyuchkova, C. Proano and G. Virshup Ginzton Technology Center of Varian Medical Systems, Mountain View, CA 94043 USA L. Melekhov, A. Zuck, B.N. Breen, O. Dagan and H. Gilboa Real Time Radiography, Jerusalem Technology Park, Jerusalem 91487 Israel P. Bennet, K. Shah and Y. Dmitriev Radiation Monitoring Devices, Watertown, MA 02172 USA J. Thomas Uniformed Services University of the Health Sciences, Bethesda, MD 20814 USA M. Yaffe and D. Hunter University of Toronto and Women’s Health Centre, Sunnybrook, Toronto, Ontario M4N3M5 Canada

ABSTRACT For the purpose of digital x-ray imaging, mercuric iodide (HgI2) and lead iodide (PbI2) have been under

development for several years as direct converter layers. Previous reports have covered the basic electrical and physical characteristics of these and several other materials. As we previously reported on 5cm x 5cm and 10cm x 10cm size imagers, direct digital radiography X-ray detectors, based on photoconductive polycrystalline mercuric iodide deposited on a flat panel thin film transistor (TFT) array, have a great potential for radiographic and fluoroscopic applications in medical imaging as well as in NDT, and security applications. This paper, for the first time, presents results and comparison of both lead iodide and mercuric iodide imagers scaled up to 20cm x25cm sizes.

Both the mercuric iodide and lead iodide are vacuum deposited by Physical Vapor Deposition (PVD). We have now successfully scaled up this coating technology to the 20cm x 25cm size required in common medical imaging applications. A TFT array with a pixel pitch of 127 microns was used for this imager.

In addition to the increase in detector size, more sophisticated, non-TFT based small area detectors were developed in order to improve analysis methods of the mercuric and lead iodide photoconductors themselves. These small area detectors were evaluated in radiographic mode, continuous fluoroscopic mode and pulsed fluoroscopic mode. Mercuric iodide coating thickness between 140 microns and 300 microns and lead iodide between 100 microns and 180 microns were tested with beam energy between 40 kVp and 100 kVp utilizing exposure ranges typical for both fluoroscopic and radiographic imaging.

Diagnostic quality radiographic and fluoroscopic images at up to 15 pulses per second were demonstrated. Image lag characteristics of mercuric iodide appear adequate for fluoroscopic rates but currently the longer image lag characteristics of lead iodide make it suitable for radiographic mode only. For both material the MTF is determined primarily by the aperture and pitch of the TFT array (Nyquist frequency of ~3.93 mm-1 (127 micron pixel pitch).

Keywords: imaging, X-ray radiology, polycrystalline, lead iodide, mercuric iodide, imaging detectors, Flat-Panel imaging arrays.

1. INTRODUCTION Direct detector materials must exhibit several attributes including high x-ray absorption, high charge collection, low dark current and good uniformity. These are difficult to achieve in a single material. Nevertheless because blurring due to spreading of light is eliminated, higher resolutions are possible with direct detectors than with detectors utilizing phosphor coatings.

Films composed of the polycrystalline semiconductors, PbI2 and HgI2, directly convert X-rays into electrical signals and show promise as X-ray detectors for digital radiography1-14. Table 1 summarises some of the physical properties of amorphous Se, polycrystalline HgI2 and polycrystalline PbI2 films, which are the materials considered for direct imaging. This paper describes measurements performed on high-resolution image sensors using HgI2 and PbI2 photoconducting layers in the direct detection mode of operation. The high Z values of HgI2 and PbI2 indicate that they are efficient materials for absorbing X-rays for clinically used X-ray exposure energies. In addition, the X-ray energy required to generate an Electron-Hole Pair both in the mercuric iodide and lead iodide photoconductors, as designated by the parameter W, is relatively low. The lower the W, the higher the number of charges liberated by interacting X-ray, and the higher the X-ray sensitivity. As Table 1 indicates, the parameter W is much smaller both for HgI2 and PbI2 compared to Se. The larger the mobility-lifetime (µτ-product), the greater the distance the electrical charges can move in the detector. Greater distances result in higher sensitivity due to better charge collection. The high Z number, low W and high µτ-product result in very high signal which helps overcome noise sources in fluoroscopic modes of operation. The ability to operate at low voltages allows low-voltage electronic design.

Poly-HgI2 Poly-PbI2 a-Se Comments Atomic Number (Z) 80, 53 82,53 34 Absorption increases with Z

Energy Band Gap (Eg) eV 2.1 2.3 2.2 Wide gap reduces dark current Effective Charge Pair Formation

Energy (W), eV ~5 ~5.5 ~42 Lower W increases the gain

Mobility Life-time Product (µτ) cm2/V

1.5x10-5 (hole) 1.8x10-6

(electron) 7x10-8

10-6 – 10-5 Higherµτ increases the charge collection

Operational Electric Field (E) V/micron

0.2-1 0.2-1 10 Lower E reduces electrical breakdown

Table 1. Comparison of Amorphous Se, poly HgI2 and poly PbI2

2. SAMPLE PREPARATION

2.1 Photoconductor deposition

The lead iodide coatings are produced via physical vapor deposition, or more specifically, thermal evaporation. A charge of solid material is heated to incur an evaporative flux that impinges upon an appropriately positioned substrate and condenses into a polycrystalline film. Sublimation, at less than 408 °C source temperatures (below the melting point of PbI2), can also provide substantial growth rates (> 10 µm/hr). Typical conditions for better quality films range from 1 to 10 µm/hr and are obtained at a substrate temperature of 200 to 230 °C. Higher temperatures lead to substantial re-“evaporation” and present risk to a-Si circuitry. Starting material is commercially available, 99.999% pure PbI2,

subsequently treated to multiple passes of zone purification. The evaporation is carried out in a 10-6 Torr vacuum, and followed by a deposition of several hundred angstroms of palladium (Pd) to form a top electrode. Films are generally not encapsulated, and remain stable while stored in air. Most films are deposited onto either conductive (ITO coated) glass plates or a-Si TFT arrays, and little difference is seen between the two in the resulting films’ grain size or adhesion. Generally, grains can be described as hexagonal platelets with the longest dimensions being 10 µm or less. The platelets show preference to growing perpendicular to the substrate, thus producing films less dense (3 to 5 g/cc) than bulk crystalline material (6.2 g/cc). The HgI2 material deposition has been published earlier1. For the first time, a full-scale 20x25cm imager has been manufactured using mercuric iodide photoconductor. Manufacture of such imagers required scale-up of the small reactors that have been used until now for production of 5cm x 5cm and 10cm x 10cm imagers. The substrates of the larger imagers (for both HgI2 and PbI2) are TFT arrays with 127µm x 127µm pixel size, 10cm x 10cm (768x768 pixels) and 20cm x 25cm (1536x1920 pixels) active area. The reactor is based upon a 25” (63cm) diameter stainless steel vacuum vessel. Highly purified mercuric iodide powder is loaded into evaporators in the base of the reactor. The TFT array is suspended over the evaporators. By proper choice of evaporator temperature and array temperature, a highly oriented (c-axis) and dense layer of polycrystalline mercuric iodide is deposited on the TFT array. This reactor has been used to deposit polycrystalline mercuric iodide layers in a thickness range of 40µm to >300µm. After deposition of the mercuric iodide photoconductor, a bias electrode is deposited on top followed by a polymer encapsulation layer.

3. EXPERIMENTAL RESULTS ON SMALL AREA (NON-TFT BASED) PHOTOCONDUCTOR

SAMPLES

3.1 Characterization of Mercuric Iodide Photoconductor as a Function of X-Ray Energy on Small Area Samples Measurements were carried out on two basic parameters of the mercuric iodide photoconductor: Sensitivity as a function of pulse rate for pulsed fluoro mode and response linearity as a function of exposure rate. Data were obtained under various conditions of X-ray generator voltage (kVp), generator current and added filtration. Figures 1 and 2 demonstrate the linear response of a 290µm thick polycrystalline mercuric iodide photoconductor as a function of X-ray exposure rate for various conditions of X-ray voltages (70 kVp and 100kVp) and added 2mm aluminum filtration, for both radiographic mode and continuous fluoro mode. The linearity is quite reasonable for both modes but the slope is slightly lower for continuous fluoro mode than for radiographic mode. Fig 3 shows that the sensitivity is very high, close to 16µC/cm2/R and it is practically independent of the pulse rate when tested in pulsed fluoro mode.

Fig. 1. Current density versus X-ray exposure rate at 70kVp for both radiographic and continuous fluoro modes (HgI2)

Fig2: Charge collection as a function of X-ray exposure rate at 100 kVp for both radiographic and continuous fluoro modes (HgI2)

Fig 3: Sensitivity as a function of pulse rate for pulsed fluoro mode

3.2 Characterization of Lead Iodide Photoconductor on Small Area Samples

Collected Charge Fluence vs. Dose Rate in Continuous Fluoro Mode @ 70 kVp, 2.25 mm Al filter

y = 1.3257x + 0.0412R2 = 0.9999

y = 4.4299x - 0.0008R2 = 0.9985

0.00

0.05

0.10

0.15

0.20

0.00 0.02 0.04 0.06 0.08 0.10 0.12Exposure Rate, R/s

Cha

rge

Flue

nce

(µC

/cm

²/s)

Second sample

First sample

Fig. 4. Current density versus X-ray exposure rate at 70kVp for continuous fluoro mode (PbI2)

We also evaluated how the charge fluence depends on the exposure rate in lead iodide at low fluoro rates. It shows an interesting behavior. When we increase the exposure rate over 0.015R/s, then the slope of the curve changes. The slope (sensitivity) at very low exposure rate is more than 3 times higher than above 0.015R/s (slope drops from 4.43 to about 1.33 – see the trendlines of Fig. 4, First sample). This phenomenon exists in different samples.

3.3 Electron-Hole Pair Generation Energy (W) of PbI2 Photoconductor

One of the most important parameters of a photoconductive type x-ray detector is the x-ray energy is needed to generate one detectable electron-hole pair in the photoconductor. This energy is given in eV and lower values give better sensitivity, i.e. a larger number of electron-hole pairs are generated by the same x-ray photon energy and collected by the applied electrical field. This means that the effective W value depends both on the number of electron-hole pairs generated by a given x-ray photon also on the charge collection efficiency. Different PbI2 films were evaluated for W value. The irradiation was done by a mammography system set for 100mR exposures over 0.5 seconds. The system consisted of a tungsten anode with molybdenum filter, and a small amount of added Al filtering. W values were calculated by integrating current versus time plots. The absorptions were experimentally measured for the same beam. Fig. 5 shows W values and dark current of a typical sample with 232 µm thick PbI2 layer but we need to note that some other samples had higher W values. For the test we used three different integration periods: 1) integrating over the same 0.5 second period as the irradiation (indicated by squares in Fig. 5), 2) integrating from X-ray ON to 0.5 seconds beyond X-ray OFF (shown by diamonds), and 3) integrating from X-ray ON to 1.3 seconds beyond X-ray OFF (triangles). At negative bias, short integration time and low electrical field the effective W value is high. This occurs because some of the generated charges are trapped and not released before the end of the integration and, therefore, are not collected. Higher electrical fields provide shorter trapping times yielding lower effective W value because of better charge collection. At longer integration time and high electric field we can obtain W value of 6eV/ehp, close to the 5eV/ehp theoretical limit for PbI2. Surprisingly, with positive bias we get lower W values than the theoretical limit. This can only be explained by charge injection of the electrodes, which results in photoconductive gain. This theory is supported by the fact that the dark current (see the CIRCLES) also drastically increases at increasing positive bias.

0

2

4

6

8

1 0

1 2

1 4

1 6

-4 -3 -2 -1 0 1 2 3 4

E -fie ld (V /µ m )

W (e

V/eh

p)

0

5 0 0 0

1 0 0 0 0

1 5 0 0 0

2 0 0 0 0

2 5 0 0 0

3 0 0 0 0

Id (pA/mm

2)

X -ra y O N+ .5 s

+ 1 .3 s

F S 7 7 5 , c o n ta c t 1 2 3 2 µm > > 9 4 % a b s o rp tio n

Fig. 5. W values and dark current as a function of bias polarity and electric field. Three different integration periods are plotted on the graph: 1) SQUARE = integrating over the same 0.5 second period as the irradiation, 2) DIAMOND = integrating from x-ray ON to 0.5 seconds beyond x-ray OFF, and 3) TRIANGLE: integrating from x-ray ON to 1.3

seconds beyond x-ray OFF. CIRCLES: denote leakage currents.

The high dependence of the W value on the negative bias (electrical field) and on the integration time can help also in the understanding of the bias dependence of the sensitivity and the image lag in PbI2.

4. CHARACTERIZATION OF PbI2 AND HgI2 IMAGERS

We measured and evaluated the most important parameters of PbI2 and HgI2 X-ray imagers. The parameters included are the dark current, sensitivity at a given X-ray energy, image lag, MTF. In addition, radiographs were produced for comparison.

4.1 HgI2 imagers

4.1.1. Dark current

The dark current of HgI2 imagers is generally increases superlinearly with the applied bias voltage. Typical dark current values for HgI2 imagers are given in Fig. 6. The dark current at 1V/µm electrical field is about 600pA/mm2 for the large imager (HgI2-#22). This value is about twice as high as was measured, about a year ago, on imager #10 (10cm x 10cm size). One of the most recent 10cm x 10cm size imagers (HgI2-#21) has much lower dark current, ~130pA/mm2 at the same 1V/µm relative bias, which indicates that this value is also obtainable for the larger imager by improving the deposition method in the new large PVD system. We have been working on modified process parameters to improve this value further and have some other deposition technique we have observed less than 10pA/mm2 dark current at 0.5V/µm relative bias on small special samples14.

Dark current

0.00

100.00

200.00

300.00

400.00

500.00

600.00

700.00

0.000 0.500 1.000 1.500 2.000 2.500Relative bias (V/µm)

Dar

k cu

rren

t (pA

/mm

2)

HgI2-#21 (10cm x 10cm),230um thick, 15fr/s

HgI2-#22 (20cm x 25cm),300um thick, NO binning,7.5fr/sHgI2-#10 (10cm x 10cm),165um thick, 15fr/s

Fig. 6. Dark current of typical PVD-HgI2 10 cm x 10 cm size imagers and a new 20cm x 25 cm size imager.

4.1.2. Sensitivity

Most of the sensitivity measurements were taken at 60kVp x-ray energy with 0.1mm Cu filtering. The imagers were tested in continuous fluoroscopic mode, mostly at 15fr/s frame rate except for the 20 cm x 25 cm size imager in non-binning mode where the frame rate was 7.5fr/s. (We can run this imager up to 30 fr/s rate when we bin the pixels 2x2). The sensitivity data presented here are absolute values and are not corrected for the absorption of the imagers. The

sensitivity of some of the imagers is close to 10µC/cm2/R at 1V/µm electrical field (Fig. 7). This value is about three times higher than the best value measured on a 600µm thick CsI2 indirect type imager of 3.3 µC/R/cm2 at the same energy (Varian internal data). It is also very promising that the sensitivity of the new 20 cm x 25 cm size new imager is high and it is also in the same range as the sensitivity of some smaller (e.g. HgI2 #10, 10cm x 10cm size) imagers.

Sensitivity

0.00

2.00

4.00

6.00

8.00

10.00

12.00

0.000 0.500 1.000 1.500 2.000 2.500Relative bias (V/µm)

Sens

itivi

ty (u

C/R

/cm

2)

HgI2-#21 (10cm x 10cm), 230um, 15fr/s,60kVp 0.1mmCu filter, 425uR/frame

HgI2-#22 (20cm x 25cm), 300um, 7.5fr/s,60kVp 0.1mm Cu filter, 164uR/frame

HgI2-#22 (20cm x 25cm), 300um, Binned2x2,15fr/s, 60kVp 0.1mm Cu filter,41uR/frameHgI2-#10 (10cm x 10cm), 165um, 15fr/s,60kVp, 0.1mm Cu filter, 425uR/frame

Fig. 7. Sensitivity of some HgI2 imagers including the new 20 cm x 25 cm size imager

4.1.3. Image Lag

Image lag of HgI2 #8 at different bias voltages after fluoro exposure

100.00

9.964.02 2.97 2.51 2.15 1.87 1.69 1.51 1.33 1.28 1.14 1.050.00

10.00

20.0030.00

40.00

50.00

60.00

70.0080.00

90.00

100.00

0 5 10

Frames

Rel

ativ

e si

gnal

Lag at 200V bias(0.8V/um)

Lag at 100V bias(0.4V/um)

Fig. 8. Image lag as a function of bias voltage (HgI2)

The image lag was checked in pulsed fluoro mode by using 3ms x-ray pulses at 15fr/s rate (Fig. 8). A continuous series of x-ray pulses was applied for a minimum of 30s time length. The average exposure was 650µR/frame. Signal decay was measured immediately following the exposure series. The first frame image lag is less than 10% of the signal at 0.4V/µm bias and decreases further to less than 8% by increasing the bias to 0.8V/µm. After the first frame the residual signal drops gradually reaching about 1% at the 15th frame (1s).

4.1.4. Resolution

The MTF of the 20cm x 25 cm HgI2 imager with 300µm thick HgI2 layer was measured with a slit of 50µm width and corrected for the slit width (Fig. 9). The MTF is very reasonable; it is only slightly below the theoretical sinc function up to the Nyquist frequency, which is 3.93lp/mm for our 127µm pitch pixels .

MTF: HgI2 imager #22 (20cm x 25cm), 300µm

0

0.2

0.4

0.6

0.8

1

1.2

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5cycles/mm

MTF

MTFsinc

Fig. 9. MTF of the large area (20cm x 25cm) size HgI2 imager compared to the theoretical sinc function

4.1 PbI2 imagers

4.1.1. Dark current

The dark current increases with the applied bias voltage but unlike the situation with HgI2 the dependence on with bias is sublinear rather than supralinear. Typical dark current values for PbI2 imagers are given in Fig. 10. The dark current value for the 20cm x 25cm imager (FS819) is over 200pA/mm2 at 1V/µm bias. This is lower than that of the 10cm x 10cm size imagers (FS 818), which had over 200 pA/mm2 dark current already at 0.4V/µm bias. The sensitivity curves (Fig. 11) show that we already get nearly maximum gain values at 0.2V/µm field where the dark current is about 100pA/mm2. However even this dark current value is still too high for long exposure time applications.

Dark current

0.00

50.00

100.00

150.00

200.00

250.00

300.00

0.000 0.200 0.400 0.600 0.800 1.000 1.200

Relative bias (V/µm)

Cur

rent

(pA

/mm

2)

PbI2-FS818 (10cm x 10cm),120um thick, 15fr/s

PbI2 FS 819 (20cm x25cm),180um thick, 7.5fr/s

Fig. 10. Dark current of a typical PbI2 10 cm x 10 cm size imager and a new 20cm x 25 cm size imager

4.1.2. Sensitivity

The sensitivity measurements were taken at 60kVp X-ray energy with 0.1mm Cu filtering. The imagers were tested in continuous fluoroscopic mode, mostly at 15fr/s frame rate except for the 20cm x 25cm size imager in non-binning mode when the frame rate was 7.5fr/s. The signal was read out after 30 sec continuous X-ray illumination. The sensitivity data presented here (Fig. 11) are absolute values and are not corrected for the absorption of the PbI2 layer. The sensitivity of the large imager (FS819) is about 5µC/R/cm2 even at a field of 0.2V/µm and it does not change significantly above that. The sensitivity of the smaller area imager (FS818) is just slightly higher.

Sensitivity

0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

0.000 0.200 0.400 0.600 0.800 1.000 1.200Relative bias (V/µm)

Sens

itivi

ty ( µ

C/R

/cm

2 )

PbI2-FS818 (10cm x 10cm), 120um thick, 15fr/s,60kVp, 0.1mm Cu filter, 425uR/frame

PbI2-FS819 (20cm x 25cm), 180um thick, 7.5fr/s,60kVp, 0.1mm Cu filter, 850uR/frame

Fig. 11. Relative sensitivity of a 10 cm x 10cm imager and a new 20 cm x 25 cm size PbI2 imager

4.1.3. Image Lag

PbI2 FS761 image lag at 60kVp (negative bias)

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140Frames

Rel

ativ

e si

gnal

leve

l Fluoro decay after longexposureRad shot decay

Fig. 12. Fluoro and radiographic lag of a PbI2 imager. Dose per frame of the fluoro was the same (55µR) as the integrated dose of the single radiographic shot

The image lag (Fig. 12) was checked in pulsed fluoro mode similar to the method described for the HgI2 lag measurements. Generally, lead iodide has a much longer image lag decay time than mercuric iodide. This fact already showed up as bias and integration time dependent W value, at negative biases. The lag also depends on the “prehistory” of the material. If we irradiate the imager for a longer period of time in fluoro mode, then the first frame lag value is

MTF of PbI2 compared to CsI2

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 1 2 3 4 5

lp/mm

MTF

MTF of PbI2 FS818120um thicksinc

20x25 CsI2 imager

Fig. 13. MTF of a PbI2 imager

much higher and the decay is much longer than for applying only a single radiographic shot. The decay is also faster in rad than in fluoro mode. But even in rad mode the lag is still significantly higher than that of the HgI2 imager. We get more than 10% lag after 8 seconds by reading out continuously at 15fr/s. This fact practically excludes the fluoroscopic application of PbI2 imagers (at least by using negative bias). The dependence of W on the bias voltage and the image lag results of PbI2 are in good agreement to the results of Street at al10 published earlier on smaller arrays.

4.1.4. Resolution

We measured the MTF of the FS818 (10cm x 10cm) imager with 120 µm thick PbI2 layer (Fig.13). The MTF is also reasonable although it is slightly lower than that of the HgI2 imager above. The MTF at the Nyquist frequency is about 35%, which still gives a much better resolution than a 600um thick CsI2 imager (see CsI2 curve for comparison).

5. IMAGING WITH PbI2 AND HgI2 ARRAYS Here we are presenting some sample images taken by both PbI2 and HgI2 imagers. We used a foot phantom to briefly check the medical imaging capability of the imagers and a resolution phantom to see how the MTF correlates with the visible spatial resolution.

Fig. 14. Foot phantom image taken by the HgI2 #22 Fig. 15. Resolution pattern image taken by the same

20cm x 25 cm imager HgI2 imager (enlarged)

Fig. 14 is a foot phantom and Fig. 15 is a resolution pattern taken by a new 20cm x 25 cm size imager. It is a pity that the 20cm x 25cm size TFT array used for this imager had some wide line defects that degrades the image quality. However in the good parts of the imager the image is sharp and we can see line resolution up to 3.93 lp/mm, which is the Nyquist frequency of this 127µm pixel pitch TFT plate. We need to add that we had much better quality smaller

(10cm x10cm) TFT plates with HgI2 coating but we wanted to demonstrate the first 20cm x 25 cm size HgI2 imager here.

Fig. 16. Foot phantom image taken by the PbI2 FS819 Fig. 17. Resolution pattern image taken by the same

20cm x 25 cm imager PbI2 imager (enlarged)

The TFT array for the PbI2 based 20 cm x 25cm size array was somewhat better quality, which shows up as much lower number of line defects. This imager also gives reasonable medical quality image of the foot phantom and a good resolution is shown in Fig. 17.

These 20cm x 25cm size images demonstrate that both HgI2 and PbI2 can be used for photoconductor based X-ray imagers and they are able to generate excellent quality radiographic medical images.

Finally we also give one frame of a fluoroscopic image series of a running alarm clock taken by a 10cm x 10cm size HgI2 imager. This image was shot by a pulsed fluoro X-ray source at 70kVp with 200mA 3ms pulses at 15 fr/s. No image lag of the fast running balance wheel is visible but we can see details of the fine spring. This image also proves that line defects, which are visible on Fig. 14 and Fig.15, are not characteristics for the HgI2 material.

Fig. 18. Running alarm clock. One frame of a fluoroscopic image series shot by 70kVp 200mA 3ms x-ray pulse running at 15 fr/s.

6. SUMMARY Images obtained with 10cm x 10cm and 20cm x 25cm imaging pixilated arrays using PVD-HgI2 and PbI2 films are shown. Data are also presented of measurements performed on mercuric iodide and lead iodide photoconductor detectors based upon striped electrodes (non-TFT). Both HgI2 and PbI2 arrays demonstrate high sensitivity to x-rays and excellent spatial resolution. High resolution and high absorption were also shown. However, the very long image lag of PbI2, at least with negative bias applied, excludes the fluoroscopic application of this detector. We need to test the PbI2 imagers with positive bias very soon to see if the image lag properties are better with that bias polarity. The low image lag property of the HgI2 detectors opens new applications for this material, especially those requiring extreme sensitivity such as fluoroscopy, which is beyond the capabilities of the existing thin film detectors. Further reduction in dark current are necessary to apply any of these materials for long exposure applications (radiography, mammography); however the high spatial resolution, which is comparable to that of a-Se, and the higher sensitivity and temperature stability than that of a-Se make these materials good candidates for those fields.

7. ACKNOWLEDGMENTS The authors are grateful to the members of the RTR Testing group, I. Baydjanov and M. Kaminsky for carrying out the X-ray response linearity testing and dark current dependence on temperature and Chris Webb from Varian for helping in the MTF measurements and evaluation.

8. REFERENCES

1. “Polycrystalline mercuric iodide detectors”, M. Schieber, H. Hermon, A. Zuck, A. Vilensky, L. Melekhov, R. Shatunovsky, and R. Turchetta, Proc. of the SPIE, Vol. 3770 (1999) 146.

2. “Mercuric Iodide Thick Films for Radiological X-ray Detectors”. M. Schieber, H. Hermon, R.A. Street, S.E. Ready, A. Zuck, A. Vilensky, L. Melekhov, R. Shatunovsky, M. Lukach, E. Meerson, Y. Saado and E. Pinhasy, Proc. of the SPIE Vol. 4142 (2000) 197.

3. “Radiological X-ray Response of Polycrystalline Mercuric Iodide Detectors”. M. Schieber, H. Hermon, R. Street, S. Ready, A. Zuck, A. Vilensky, L. Melekhov, R. Shatunovsky, E. Meerson, Y. Saado. In Proc. of the SPIE on Medical Imaging 2000 San Diego, Vol. 3977 (2000) 48.

4. “Theoretical and experimental sensitivity to X-rays of single and polycrystalline HgI2 compared with different single crystal detectors”. M. Schieber, H. Hermon, A. Zuck, A. Vilensky, L. Melekhov, R. Shatunovsky, E. Meerson, H. Saado, NIMA Vol. 458 (2001) p 41.

5. “Characterization of CZT Detectors Grown From Horizontal and Vertical Bridgman”. H. Hermon, M. Schieber, M. Goorsky, T. Lam, E. Meerson, H. Yao, J. Erickson, and R.B. James, in Proc. of the SPIE on Hard X-ray and Gamma-ray Radiation, Vol. 4141 (2000) 186.

6. “Comparison of Cadmium Zinc Telluride Crystals Grown by Horizontal and Vertical Bridgman and From the Vapor Phase”. M. Schieber, R.B. James, H. Hermon, A. Vilensky, I. Baydjanov, M. Goorsky, T. Lam, E. Meerson, H.W. Yao, J. Erickson, E. Cross, A. Burger, J.O. Ndap, G. Wright, and M. Fiederle, Accepted for publication in JCG (2001).

7. “Thick Films of X-Ray Polycrystalline Mercuric Iodide Detectors”. M. Schieber, H. Hermon, A. Zuck, A. Vilensky, L. Melekhov, R. Shatunovsky, E. Meerson, Y. Saado, M. Lukach, E. Pinhasy, S.E. Ready, and R.A. Street, Journal of Crystal Growth, 225 (2-4) (2001) pp. 118-123.

8. “Deposition of Thick Films of Polycrystalline Mercuric Iodide X-Ray Detectors”. H. Hermon, M. Schieber, A. Zuck, A. Vilensky, L. Melekhov, E. Shtekel, A. Green, O. Dagan, S.E. Ready, R.A. Street, G. Zentai, and L. Partain, Proc. of the SPIE, MI 2001 – Vol. 4320 (2001) pp. 133-139.

9. “Comparative Study of PbI2 and HgI2 as Direct Detector Materials for High Resolution X-ray Image Sensors”. R.A. Street, M. Mulato, S.E. Ready, R. Lau, J. Ho, K. Van Schuylengergh, M. Schieber, H. Hermon, A. Zuck, A. Vilensky, K. Shah, P. Bennett and Y. Dmitryev. Proc. of the SPIE MI 2001 - Vol. 4320 (2001) pp. 1-12.

10. “Non Destructive Imaging with Mercuric Iodide Thick Film x ray detectors”. M. Schieber, H. Hermon, A. Zuck, A. Vilensky, L. Melekhov, M. Lukach, E. Meerson, Y. Saado, E. Shtekel, B. Reisman, G. Zentai, E. Seppi, R. Pavlyuchkova, G. Virshup, L. Partain, R. Street, S.E. Ready and R. James. Published in Proc. of the SPIE NDT 2001 - Vol. 4335 (2001) pp. 43-51.

11. “Comparison of PbI2 and HgI2 for Direct Detection Active Matrix X –Ray Image Sensor” R.A. Street, S.E. Ready, K.V. Schuylenbergh, J. Ho, J.B. Boyce, P. Nylen, K. Shah, L. Melekhov and H. Hermon J. App. Phys. Vol 91, No 5 P 3345 (2002).

12. “Approaching the Theoretical X-ray Sensitivity with HgI2 Direct Detection Image Sensors”. R.A. Street, S.E. Ready, L. Melekhov, J. Ho, A. Zuck and B. Breen. To be published in Proc. of the SPIE MI 2002.

13. “Large Area Mercuric Iodide X-Ray Imager”. G. Zentai, L. Partain, R. Pavlyuchkova, G. Virshup, A. Zuck, L. Melekhov, O. Dagan, A. Vilensky and H. Gilboa. To be published in Proc. of the SPIE MI 2002.

14. “Large area mercuric iodide thick film X-ray detectors for fluoroscopic (on-line) imaging”. G. Zentai, L. Partain, R. Pavlyuchkova, C. Proano, G. Virshup, B.N. Breen, A. Zuck, B. Reisman, A. Taieb and M. Schieber. To be published in Proc. of the SPIE NDT 2002.

15. “Technology and Applications of Amorphous Silicon” R. A. Street, ed., Springer, 1999.JM Boone, J A Seibert, Med. Phys. 24 (1997), 1661-1670.

16. http://physics.nist.gov/PhysRefData/XrayMassCoef/cover.html, JH Hubbell, SM Seltzer, (NIST public domain).

17. “Handbook of Medical Imaging” Vol 1, JM Boone in (J Beutel, HL Kundel, RL Van Metter, eds.), SPIE Press 2000.

18. “X-Ray Imaging Using Amorphous Selenium: Feasibility of a Flat Panel Self- Scanned Detector for Digital Radiology” W Zhao, JA Rowlands, Med. Phys. 22 (1995), 1595-1604.