An electrical charge amplifier including a filter circuit is AC coupled to a semiconductor (CdTe) detector and generates a voltage pulse in response to an electrical charge generated in the detector by an incident pulse of radiation. The filter allows only frequencies within a predetermined range to contribute to the voltage pulse. The selected range of frequencies is determined in accordance with the duration of the incident radiation pulse such that the voltage pulse faithfully represents the magnitude of incident radiation in spite of undersirable detector characteristics which would otherwise introduce distortions. Exemplary charge amplifier and detector structures are also described.
Some of the earlier CT scanners obtain the requisite multiplicity of absorption measurements by synchronously scanning the body section with a highly collimated X-ray beam and an aligned single detector. The beam and detector are together translated to scan the body section and generate a set of measurements along parallel paths. This assembly is then rotated with respect to the body section and the translation-rotation operation is repeated to provide sets of measurements at different angular dispositions.
It is desirable, however, to minimize the overall scanning time. Accordingly, more recent CT scanners generally utilize a plurality of individual radiation detectors in co-operation with a fan-shaped X-ray beam wide enough to irradiate the entire body section. In one example of such systems, a fan-beam source rotates about the body section, irradiates a stationary detector array forming the outside perimeter of the scanning frame. As many as 600 scintillation crystals and associated photomultipliers tubes (PMT) are typically used in such systems. Thus, the cost of the scintillation crystal-PMT detectors, and particularly the photomultiplier tubes, presently represent a substantial portion of the cost of such CT scanners. Accordingly, it is desirable to directly convert the radiation to an electrical signal and eliminate the need for large numbers of photomultiplier tubes and associated circuits.
Various direct conversion detectors such as Xenon gas and high purity germanium (HPGe) semiconductors have been utilized in the past. However, such detectors are disadvantageous for other reasons.
The properties of high purity germanium are well known. A high density position sensitive array has been described in "Two Detector 512-Element High Purity Germanium Camera Prototype" Kaufman et al, IEEE Transactions Nuclear Science, NS-25, February, 1978. (Presently in press) However, the production of high purity germanium is a relatively complex and costly procedure, and further high purity germanium detectors require special cooling apparatus, (e.g. liquid nitrogen). Additionally, it is desirable in CT scanning applications to use a thinner detector than generally can be made with high purity germanium.
Room temperature semiconductor detectors would thus appear to be a natural choice for application in CT scanners. Since photon energy from the radiation is directly converted into an electrical charge, compact arrays could be assembled without the requirement of more bulky photomultiplier tubes. Such higher detection element density is especially advantageous for CT scanning applications. These high density semiconductor detector arrays could also be mass produced by batch processing.
However, semiconductor detector materials now available in reasonable quantities have been considered by others in the past and discounted as unsuitable for CT scanning applications. For example, cadmium telluride (CdTe) has specifically been studied with respect to use as a detector in nuclear medicine, and has been found not compatible with the needed accuracies. Specific reference is made to Allemand et al, "Present Limitations of CdTe Detectors in Nuclear Medicine", Revue de Physique Appliquee, 12: 365-367, February, 1977. Other semiconductor detectors are discussed in Armantrout et al, "What Can Be Expected from High-Z Semiconductor Detectors", IEEE Transactions on Nuclear Science, Vol. NS-24, No. 1, Feb. 1977. Of the semiconductor materials that are described, various ones have been tested and show unsuitable characteristics similar to CdTe. Computer studies project that similar properties are present in other semiconductor detectors--not presently tested.
More specifically, CdTe detectors have been found to exhibit adverse "leakage current", "Tailing", "polarization" and "memory" properties, which are generally considered to render CdTe unsuitable for CT scanning usages. Other detectors, such as H.sub.g I.sub.2, have been found to exhibit similar properties.
The high "leakage current" property varies with temperature, and further, varies with time once bias has been applied to the detector. The "tailing" phenomenon causes the detector response to portray a monochromatic (single energy) source as having energies ranging all the way down to noise levels. "Tailing" increases as a function of the elapsed time after bias is applied to the detector. The "polarization" property is manifested by the most frequently encountered detector response (hereinafter termed the photo-peak) to a constant monochromatic source, shifting towards the low energy end of the spectrum. The "memory" property is manifested by a continued current flow after a pulse of incident radiation has ceased, sometimes with a time constant of hours.
Various techniques have been proposed for dealing with such adverse properties. However, the proposed techniques have tended to merely add to the complexity of the detector circuitry and, in fact, to introduce sources of errors that have made application of CdTe detectors in CT scanners impractical. For example, the high current leakage could be compensated by sampling the detector output just before the X-ray beam is activated to produce each pulse of incident radiation. However, such a scheme will compromise the signal-to-noise ratio of the system when absorption is high, as happens when the beam goes through bone. Pre and post measurement calibration techniques for memory and polarization compensation, similarly tend to add signal-to-noise problems. Further, polarization compensation techniques wherein the bias to the detectors is turned off between measurements, are inadequate due to hysteresis effects.
SUMMARY OF THE INVENTION
In accordance with the present invention, it has been found that room temperature semiconductor detectors, and in particular CdTe detectors, can be made compatible with CT scanning operations. The detector is used in cooperation with a pulsed radiation source and frequency filtering techniques are utilized in measuring the resulting pulse of electrical charge developed by the detector. Only selected frequencies within a predetermined range, chosen in accordance with the duration of incident radiation pulse, contribute to the output pulse. More specifically, such range of frequencies is preferably a narrow band of frequencies centered about a frequency approximately equal to the inverse of twice the radiation pulse duration.
It has also been found that plural individual detectors can be formed as a unitary array on a single CdTe or H.sub.g I.sub.2 wafer by merely providing plural electrically separated electrodes. Thus, CdTe or H.sub.g I.sub.2 detectors are particularly advantageous in that entire high density arrays can easily and quickly be formed by batch processing. Further, two dimensional arrays can be formed by providing a grid work of such electrodes. Moreover, a stacked array for taking measurements at a plurality of differing energy levels can be provided by disposing a multi-electrode detector such that the radiation is incident on a wafer surface transverse to the electrodes.
BRIEF DESCRIPTION OF THE DRAWINGS
A preferred exemplary embodiment of the present invention will hereinafter be discussed with reference to the following drawings wherein like designations denote like elements and:
FIGS. 1A and 1B are spectral responses of CdTe detectors when used as in the prior art illustrating "tailing";
FIGS. 2A-2D are spectral response of CdTe detectors when used as in the prior art illustrating "polarization";
FIGS. 3A-3B are graphs illustrating "memory" of CdTe detectors when used as in the prior art;
FIG. 4 is a block diagram of a radiation detecting system in accordance with the present invention;
FIG. 5 represents the shape of a typical X-ray pulse;
FIG. 6 is a schematic diagram of a suitable charge amplifier and filter means for use in the system of FIG. 4;
FIG. 7 is the spectral response of a CdTe detector used in accordance with the present invention;
FIG. 8 illustrates the non-uniform response of a CdTe detector as used in the prior art;
FIGS. 9A and 9B illustrate the uniform response of the CdTe detector of FIG. 8 when used in accordance with the present invention;
FIGS. 10A and 10B are schematic diagrams of alternative charge amplifier and filter means;
FIGS. 11A-11D are pictorial schematics of CdTe detectors in accordance with another aspect of the present invention;
FIG. 12 illustrates the localized charge phenomenon of CdTe, and lack of crosstalk between elements in accordance with this invention;
FIGS. 13 and 14 illustrate the charge response of CdTe detectors to bias voltage;
FIG. 15 compares the radiation stopping power of various detector materials;
FIGS. 16 and 17 illustrate dual-energy detectors; and
FIG. 18 is a block diagram of a CT scanner utilizing detectors in accordance with the present invention.
DETAILED DESCRIPTION OF AN EXEMPLARY EMBODIMENT
The present inventor has examined the properties of CdTe detectors by observing the spectral response the single photons from a monochromatic source (AM-241). FIGS. 1A and 1B illustrate the spectral response of two different CdTe detectors to a monochromatic source (AM-241) when used as in the prior art. Specifically, FIGS. 1A and 1B show the count of photons registered by the detector at various energy levels during a preset time interval, the energy level being indicated by the pulse height of the detector output.
Ideally, since the response was to a monochromatic source, all of the counts should have registered at a single energy providing a single impulse-like spike response. However, as illustrated in FIG. 1A, due to the tailing effect, the spectral response is broadened. Photons which should be depicted as counts at the photo-peak 10 (predominant energy level) are indicated as having energies ranging down to the noise level of the detector, and a plateau region 12 is manifested in the spectral response.
It is apparent from a comparison of FIGS. 1A and 1B, that the tailing effect is not uniform between individual detectors. The response shown in FIG. 1A is that of a relatively good (by prior art standards) CdTe detector, whereas the response shown in FIG. 1B of another CdTe detector to the same source evidences a worse peak-to-plateau ratio and poor energy resolution.
The polarization phenomenon shifts the photo-peak response toward the lower energy levels and the peak-to-plateau ratio decreases (a measure of the tailing effect) as a function of the time from the instant the detector is placed under bias is illustrated in FIGS. 2A--2D. FIGS. 2A and 2C show the spectral response of two different CdTe detectors to monochromatic AM-241 immediately after bias is applied. FIGS. 2B and 2D illustrate the spectral response of the respective detectors 15 minutes after bias is applied. It should be noted that the photo-peak 10 shifts towards the lower energy level, and the response widens as tailing increases. Such polarization has generally been attributed (apparently in error, as will be explained) to the loss of depletion depth with time. If such is the case, it follows that X-ray detection efficiency would be reduced as a function of time and since tailing worsens with time, the charge collection efficiency of the detector would also degrade with time. In any event, irrespective of the physics involved in a uncompensated detector, a steadily decreasing response to a steady X-ray source would be manifested.
FIGS. 3A and 3B illustrate the memory property exhibited by CdTe detectors with current continuing to flow after an applied radiation pulse has terminated. More specifically, the current output of two different CdTe detectors are shown respectively in FIGS. 3A and 3B as a function of time, following exposure to a short X-ray pulse. It should be noted that leakage current does not decrease below 1% of maximum until approximately 25 .mu.sec. and 800 .mu.sec. after the end of the X-ray pulse for the respective detectors associated with FIGS. 3A and 3B. DC relaxation times were found to be approximately 0.5 minutes and 1.5 minutes respectively. Such memory phenomenon, absent compensation, would be unusable with CT scanners at present scanning rates.
Data is available which tends to indicate that the tailing effect is the result of the CdTe detector producing pulses of varying rise times. Reference in this regard is made to Jones, "the Use of CdTe .gamma. Spectrometers in Monitoring Activity Deposited in Nuclear Power Stations", Revue de Physique Appliquee 12: 379-384, 1977. When processed with conventional amplifiers having Gaussian-shaped frequency responses, the varying rise times are manifested as pulses of different amplitudes. The above-noted Jones article appears to demonstrate that the fastest pulses correspond to photons of the highest energy level registered (the photo-peak) and that various bands in rise time correspond to energy bands in the tail.
While the polarization phenomenon is generally considered to be the result of a decreasing depletion depth, as noted above, the present inventor speculates that the polarization effect is the result of further degradation of rise time distributions during the period after bias is applied. Accordingly, polarization and tailing effects could effectively be eliminated if the detector is presented with a pulse of photons having a duration which is long compared to the longest detector rise times (approximately 1 .mu.sec.). Thus, the pulse of radiation in effect drives the rise time response, and spurious signals not in accordance with the rise time of the pulse can be filtered from the detector output to provide an undistorted representation of the incident radiation pulse. Experimental results appear to verify this theory. In any event, irrespective of the physics of the phenomenon, the experimental results have proven that semiconductor detectors can be made compatible with the requisites of CT scanning when practiced in accordance with the present invention.
With reference now to FIG. 4, a semiconductor detector 14 such as CdTe, suitably biased by circuitry 16, is AC coupled via means 18, (e.g., a capacitor), to an amplifier/filter 20. Amplifier/filter 20 is coupled to suitable storage and/or data processing means 22.
Detector 14 cooperates with a pulsed radiation source 24. Source 24 can be a conventional pulse mode X-ray generator, or can be a constant source such as, for example, a Phillips MG160 X-ray generator, used in conjunction with a wheel chopper. Where wheel choppers are used, the rotational frequency of the wheel and ripple in the X-ray tube output signal tend to produce beats in pulse amplitude. Such beat frequencies can be compensated for by averaging a number of pulses. A typical X-ray pulse shape is shown in FIG. 5.
The effective time constant of biasing circuitry 16 and AC coupling 18 is chosen in accordance with the pulse repetition rate of source 24; the time constant being chosen to allow dectector 14 to recharge (recover) between radiation pulses.
Amplifier/filter 20 comprises means for generating a signal indicative of the charge developed by detector 14, and has a frequency response such that frequency components not within a predetermined range (band) of frequencies are effectively filtered from that signal. The predetermined frequency band is determined in accordance with radiation pulse duration and is preferably a narrow band of frequencies centered about a frequency approximately equal to the inverse of the radiation pulse duration or the inverse of twice the radiation pulse duration. For example, amplifier/filter 20 suitably comprises a charge sensitive first stage having a RC time constant TD, and a second stage having an effective RC time constant TS. It was found that by chosing the time constants such that TS is approximately equal to the duration of the X-ray pulse (TX) and much larger than TD, a strongly differentiated detector output pulse results, which provides faithful representation of the actual X-ray pulse shape. Accordingly, the integration of such a differentiated "X-ray pulse" yields an accurate indication of the total incident X-ray energy. Such an amplifier/filter will hereinafter be described in more detail in conjunction with FIG. 6.
In the alternative, the charge sensitive first stage can be made to have an RC time constant much larger than the duration TX of the X-ray pulse, and the second stage effectively shaping the first stage output by differentiation with time constant TD and integration with a time constant TS considerably shorter than TD but comparable to TX. Such a filtering amplifier (20B) will be more fully described in conjunction with FIGS. 10A and 10B. While such an arrangement successfully eliminates the effects of the detector characteristics from the output pulse it is often difficult to obtain adequate pole-zero compensation and tends to exhibit undesirable overshoot characteristics.
With reference now to FIG. 6, amplifier/filter 20A will now be described. Exemplary component values are given for use in conjunction with X-ray pulse durations on the order of 100 .mu.sec.
A CdTe detector 14 is biased by a biasing circuit 16, and coupled through AC coupling means 18 to amplifier/filter 20A. AC coupling means 18 may be a 0.005 .mu.f. capacitor 26. Capacitor 26 is coupled to the gate terminal of a field effect transistor 28, suitably of the type TIS 75, having grounded source and a drain suitably biased by a resistor-capacitor biasing network 30. Biasing network 30 may comprise resistors 32 (422.OMEGA.) and 34 (215.OMEGA.) serially connected between the source of FET 28 and a 24 volt source. Bypass capacitors 36 (0.1 .mu.f.) and 38 (212 .mu.f.), are respectively coupled from the voltage source to ground and from the juncture between resistors 32 and 34 to ground.
The drain of FET 28 is also connected to the emitter of an PNP transistor 40 (e.g., of the type MPS 6523). The base of transistor 40 is coupled through a resistor 42 to a positive 4.7 volt source, and is AC coupled to ground through a capacitor 44 (0.01 .mu.f.). The collector of transistor 40 is coupled through two serially connected resistors 46 and 48 to a negative 12 volt source. An AC bypass capacitor 50 is provided for the negative 12 volt source.
The collector of transistor 40 is also coupled to the base of an NPN transistor 52, (e.g., of the type MPS 6521). The collector of transistor 52 is connected to a positive 12 volt supply, the 12 volt supply being coupled to ground through each of two capacitors 54 (0.01 .mu.f.) and 56 (0.1 .mu.f.), respectively. The emitter of transistor 52 is coupled to the negative 12 volt source through a resistor 58 (5.1 K.OMEGA.). The emitter of transistor 52 is also connected through a capacitor 60 (1 .mu.f.) to the juncture between resistors 46 and 48, (connected between transistor collector 40 and the negative 12 volt supply).
Additionally, a fast differentiator 62 is connected in a feedback loop between the emitter of transistor 52 and the gate of FET 28. Differentiator 62 may comprise a resistor 64, (1.5 M.OMEGA.), in parallel with a capacitor 66, (1 pf.).
The emitter of transistor 52 is also coupled to a second stage 67, comprising an integrator 68 preceded by a voltage amplifier 69. Voltage amplifier 69 suitably comprises a conventional operational amplifier (op amp) 70, for example, of the type MC1712. The emitter of transistor 52 coupled to the non-inverting input terminal of op amp 70 through a precision resistor 71 (100.OMEGA.). The inverting input terminal of op amp 70 is connected to ground potential through a precision resistor 72 (10K.OMEGA.), and feedback between the output terminal and inverting input terminal of op amp 70 is provided by a further precision resistor 73 (56K.OMEGA.). A connection is also made from the output terminal of op amp 70 to a negative 6 volt source through a resistor 74 (2K) to provide for generation of bipolar signals by amplifier 69. The output terminal of voltage amplifier 67 is connected to integrator 68, suitably comprising a resistor 75 (690.OMEGA.) and a capacitor 76 (0.1 .mu.f.).
The circuit of FIG. 6 can be adapted for radiation pulse durations other than on the order of 100 .mu.s, by adjusting the time constants of differentiator 62 and integrator 68. For example, the circuit can be made compatible with a 800 .mu.s radiation pulse duration by adjusting the values of resistor 75 and capacitor 76 to increase the time constant of integrator 68 by a factor of 8 to 10.
A number of different detectors were tested in conjuction with the circuit of FIG. 6. The structure of such detectors will be discussed below. For comparison, the uncompensated signal photon characteristics of the detectors (as used in the prior art) were first noted. The results of the testing are shown in tabular form in TABLES I-V. Table I is a brief description of the individual detectors tested. The detector structures, per se, are conventional devices, except that in some instances the detectors are adapted, as will be explained, for operation with plural electrodes to form an array. Tables II and III show the uncompensated characteristics of the detectors, Table II shows the detector characteristics when bias voltage is first applied at the detector, and Table III shows the measurements taken 15 minutes after bias is applied.
TABLE I
__________________________________________________________________________
DETECTOR SPECIFICATIONS
Detector
Detector
Code Model No. Description Electrodes
__________________________________________________________________________
A1 Radiation Rectangular Platinum
Monitoring
2mm .times. 7mm .times. 2mm
(Pt)/Pt
Device (R.M.D.)
NRD 0201
A2 M272-2 Round (single electrode)
Pt/Pt
8mm Dia. .times. 2mm
A3 M141-4 Round (single electrode)
Pt/aluminum-
10mm Dia. .times. 2mm
oxide O-Al
(Pt. side neg.)
(Purposely
fabricated for
poor results)
AA4 NDO40-01 Round (split electrode)
Pt/Pt
10mm Dia. .times. 2mm
AA5 NDO40-02 Round (split electrode)
Pt/Pt
10mm Dia. .times. 2mm
A6 M353-4 Rectangular 10mm .times. 15mm
Pt/Pt
.times. 2mm (multi-electrode)
B7 Hughes 10 .times. 10 .times. 2mm.sup.3
Aquadhag/
CTD-UCSF-1 Aquadhag
__________________________________________________________________________
TABLE II
______________________________________
DETECTOR CHARACTERISTICS
(Single Photon Mode)
Energy Resolution
Full-width-at-half
Peak
Detector
Bias Leakage maximum-height
to
Code Voltage Current FWHM (at 60 Kev)
Plateau
______________________________________
A1 60V 35 nA 23.1% 2.68
A2 60V 500 nA 42.3% 2.75
A3 50V 14 nA 19.3% 7.28
AA4 Side 1
80V 300 nA -- 2.19
Side 2 60V 150 nA -- 2.05
AA5 Side 1
60V 90 nA -- 1.69
Side 2 60V 77 nA -- 1.66
A6 50V 120 nA -- --
per element
B7 350V 38 nA 24.8% 5.38
______________________________________
TABLE III
______________________________________
DETECTOR STABILITY
(Single Photon Mode)
Leakage* Photopeak* Peak to Plateau*
Detector
Current Shift Change
Code Change (at 60 keV) (at 60 keV)
______________________________________
A1 35-50 nA -0.56 keV 2.68 .fwdarw. 2.67
A2 no change no change no change
A3 no change -10.5 keV 7.28 .fwdarw. 4.22
AA4 1 breakdown
2 220 .fwdarw. 260 nA
no change 2.05 .fwdarw. 1.98
AA5 1 90 .fwdarw. 110 nA
-1.6 keV 1.69 .fwdarw. 1.61
2 77 .fwdarw. 89 nA
-0.8 keV 1.66 .fwdarw. 1.58
A6 120 .fwdarw. 170 nA
not measured
not measured
B7 38 .fwdarw. 40 nA
no change 5.38 .fwdarw. 4.92
______________________________________
*Change over 15 minutes after the detector is under bias
As shown by Tables I-III the detectors were found to have initial leakage currents ranging from 14nA to 500nA, energy resolutions ranging from 19.3% to 42.3% and peak-to-plateau ratios (see FIG. 1) ranging from 7.28 to 1.66. Subsequent measurements showed that 15 minutes after the detectors were placed under bias, changes in leakage currents ranging from 0 to 50nA were evidenced, shifts in photo-peak ranged from 0 to -10 keV and the peak-to-plateau ratio evidenced changes ranging from 0 to almost 3.
The detectors were then tested in pulsed X-ray mode,in conjunction wth amplifier/filter 20A (FIG. 6), using X-ray pulse durations of approximately 100 .mu.sec. and 800 .mu.sec. Each of the detectors when used in cooperation with the amplifier 20A (FIG. 6) provided output signals with spectral responses such as that shown in FIG. 7, approaching the ideal impulse spectral response with no appreciable memory or tailing effects manifested.
The memory effects were tested by measuring the amplitude of a small pulse that was interlaced with a large pulse (approximately four times the amplitude of the smaller pulse). A wheel chopper was utilized to produce a pulsed beam with pulses occurring at intervals of approximately 20 msec., equivalent to the rate necessary for obtaining 1.degree. increments of measurement in rotary CT scanner operating at a speed such that a 360.degree. rotation is completed in 7.2 seconds. The large pulse was blocked with a synchronous second chopper to provide for observation of the change in measured amplitude of the smaller pulse. If the large memory effect observed in the uncompensated detector existed, a large shift would be manifested in the representation of the amplitude of the smaller pulse. However, no measurable shift was observed within a 1% measurement limit. Measurements were taken at 3.5 minute intervals over a period of 35 minutes. The results are summarized in Tables IV and V. The results demonstrate that after a short turn-on period, stable operation was attained by utilizing the pulsed source and circuit shown in FIG. 6, for measurement periods over approximately half an hour.
TABLE IV
______________________________________
DETECTOR STABILITY
(PULSED MODE)
De- Time to % Shift Pulsed No. of Pulse
tector
Stable From Time Width Pulses %
Code Response Zero Used Per Event
FWHM
______________________________________
A2 3.5 min.*
4.6% .about.100.mu.s
20 2.5%
3.5 min. 2.9% .about.800.mu.s
20 2.8%
A3 10.5 min.
6.2% .about.100.mu.s
20 4.0%
10.5 min.
6.7% .about.800.mu.s
20 3.8%
AA5 7.0 min. 1.9% .about.100.mu.s
20 2.8%
A6 3.5 min.*
0.5% .about.800.mu.s
10 4.6%
B7 3.5 min.*
0.7% .about.100.mu.s
20 3.6%
______________________________________
*NOTE:
Measurements made at 3.5 min. intervals
TABLE V
______________________________________
DETECTOR STABILITY
(PULSED MODE)
OUTPUT (.+-..010)
TIME A2 A3 AA5 A6 B7
(min.)
* + * + * + *
______________________________________
0 1 1 1 1 1 1 1
3.5 .954 .971 .967 .973 .997 .993 .995
7 .960 .970 .945 .933 .982 .993 .999
10.5 .969. .974 .938 .917 .988 .992 .978
14 .970 .972 .937 .896 .987 .991 .979
17.5 .975 .968 .937 .876 .985 .995 .990
21 .984 .938 .855 .977 .981 .997
24.5 .982 .849 .980 .971 .997
28 .985 .857 .986 .979 .992
31.5 .982 .981 .980 .973
35 .984 .990 .984
______________________________________
* 100 .mu.sec. pulse width
+ 800 .mu.sec. pulse width
One of the problems with prior art usage of semiconductor detectors in general, the CdTe in particular, is the non-uniformity of the response exhibited between respective detectors, and, in addition, non-uniformity of response over the surface of a respective detector. The uniformity of response of the detectors when used in conjunction with a pulsed source and amplifier/filter 20 was checked, by collimating the X-ray source to a small area of the detector.
Referring now to FIGS. 8 and 9, localized measurements were taken with the collimated radiation directed to various positions A-I on the detector surface. Such positions are illustrated in FIG. 8, along with spectral responses to a monochromatic source (AM-241) at various of those positions for the case of the most non-uniform detector (B7 in TABLES I-V) when used in accordance with the prior art. FIG. 9B shows the spectral response found at each of the positions A-I when the detector B7 was used in conjuction with a pulsed X-ray source and filtering amplifier 20. FIG. 9A more particularly indicates the normalized photo-peak response of the CdTe detector to collimated radiation at positions A-I using this invention.
It should be noted that while the CdTe detector has extremely non-uniform response when used in accordance with the prior art techniques, the deviations in response are of the order of .+-.2% when used in accordance with the present invention. More significantly, it was found that even using individual detectors of different geometry, (and in fact, made by different manufacturers) when the radiation source was collimated to provide equal irradiated areas, these detectors provided readings within7% of each other when used in accordance with the present invention.
As noted above the amplifier/ filter 20B depicted in FIG. 10A also successfully compensates for the characteristics of detector 14. The exemplary component values here given are again suitable for operating with X-ray pulse durations on the order of 100 .mu.sec. In the circuit of FIG. 10, FET 28 and transistors 40 and 52 are interconnected in the same manner as in FIG. 6, with the exception that fast differentiator 62 is replaced by long time constant circuit 80 comprising a parallel combination of a resistor 82 (120 M.OMEGA.), and capacitor 84 (10 pf.).
The emitter of transistor 52 is coupled to an active second stage 85. Active stage 85 comprises a differentiator portion 86 comprising an operational amplifier (op amp) 88, (e.g., of the type MC1712) and an integrator portion 100. The emitter of transistor 52 is coupled through a precision resistor 90, (100.OMEGA.) to a noninverting input of op amp 88. The inverting input of op amp 88 is coupled through a serially connected precision resistor 92 (100.OMEGA.) and capacitor 94 (0.01 .mu.f.) to ground. Feedback from the output terminal of amplifier 88 to its inverting input is provided through a further precision resistor 96, (6.19K.OMEGA.). A connection from the output terminal of op amp 88 is made through a resistor 98 (2K.OMEGA.) to a -6 volt source. The output terminal of op amp 88 is also connected to an integrator 100 comprising a resistor 102 (690.OMEGA.) and a capacitor 104 (1000 pf.). Amplifier/filter 20B can be adapted for other radiation durations by changing the time constant of differentiator 86, integrator 100, and, where necessary, circuit 80. Similarly, for long pulse durations it may be necessary to adjust the effective time constant of biasing circuit 16 and coupling circuit 18.
FIG. 10B depicts an amplifier/filter 20C similar to amplifier/filter 20B, but including provisions for pole zero compensation. The exemplary values given are again associated with a radiation pulse duration of on the order of 100 .mu.s.
The long time constant circuit 80 of amplifier/filter 20B is replaced in amplifier/filter 20C by a long time constant circuit 80a of slightly different time constant, formed of a parallel combination of resistor 82a (1000M.OMEGA.) and capacitor 84a (18 pf). The emitter of transistor 52 is coupled through two serially connected capacitors 105 (1 .mu.f) and 106 (1 .mu.f) to the input of a voltage amplifier 107, suitably identical to voltage amplifier 67 in FIG. 6. A potentiometer 108 (5K) is connected between the juncture of capacitors 105 and 106 to ground potential. The wiper of potentiometer 108 is connected through a precision resistor 109 (10K) to the input of voltage amplifier 107. Capacitor 106 and resistor 109 cooperate to form an effective differentiator having time constant of lms.
Resistor 109 and capacitor 105 also cooperate with capacitor 105 and potentiometer 108 as a pole zero (p/z) compensation circuit.
The output of voltage amplifier 107 is connected to an integrator 100a, comprising a resistor 102 (690.OMEGA.) and capacitor 104a (0.1 [f). The time constant of integrator 100a is thus approximately 70 .mu.s.
It should be appreciated that any of amplifier/filters 20A, 20B or 20C can be made compatible with faster pulse repetition frequencies by selectively discharging the capacitor (76, 104, 104a) of the integrator (68, 100, 100a) after the peak charge developed by detector 14 in response to a given radiation pulse has been attained. This can easily accomplished by shunting the conduction path of a transistor across the capacitor (76, 104, 104a) to ground, and controlling conduction through the transistor in accordance with the output signals of a conventional peak detector responsive to the output signal of the integrator.
As noted above, a conventional semiconductor detector such as CdTe or HgI2 can be made compatible with the requisites of CT scannning in accordance with the above described aspect of the present invention. As illustrated in the FIG. 11A, conventional semiconductor detectors comprise a wafer of the semiconductor material, (e.g. CdTe 110) having deposited on either side thereof conductive electrodes 112 and 114. The electrodes are typically formed of platinum or aquadhag.
Such unitary detector arrays are shown in FIGS. 11B and 11C. Assuming the incident radiation to scan across the detector in the longitudinal direction, an array of adjacent detectors can readily be provided by merely removing strips of electrode material in the transverse direction at longitudinally spaced locations to provide a plurality of electrically isolated electrodes FIG. 11B. Similarly, by additionally removing a strip of electrode material in the longitudinal direction (FIG. 11C) a two-dimensional array for simultaneously detecting multiple cross-sections in a CT scanner is provided.
In FIG. 11D a suitable biasing configuration is shown for such a multi-electrode detector. The detector is disposed such that it is irradiated by the X-ray pulses on its cathode (114) side, and individual anode electrodes 112a-112f are disposed on the far side of CdTe wafer 110. Cathode 114 is connected to ground potential and each individual anode connected to a high voltage source through a respective resistor, suitably of the value 1 M.OMEGA.. The high voltage source is on the order of 50 volts. Each anode is AC coupled through a respective capacitor 26 to an associated amplifier/filter 20.
It should be appreciated that radiation can also be directed to the anode side of the detector. Further, a unitary anode and plural cathodes could be utilized to form the unitary array. This is equivalent to connecting electrode 114 to the high voltage source and grounding each of the electrodes 112a-112f.
Experiments show that crosstalk between the respective elements in the multi-element array of FIGS. 9B-9D are not appreciable. FIG. 12 illustrates the effect of scanning a narrow (1.5 mm) beam over one of the individual electrodes of such a detector, having one electrode per 2 mm in the direction of scanning. It appears that CdTe exhibits a localized charge property in response to incident radiation. When taking beam width into account, scattering effects within the detector contribute to less than 1% crosstalk at 1 mm. Such crosstalk produces an effect analogous to smoothing the data over the 1 mm range. If desired, a simple digital de-convolution procedure can be utilized, in effect, to unfold the data, and thus eliminate the inter-element crosstalk effect.
Measurements of the electrical pulse output of the detectors as a function of detector bias was performed to insure that all stopped photons contributed to the charge collected by the detector. It was found, as expected, that the response essentially flattens at high bias voltages where depletion is complete. Reference is made to FIGS. 13 and 14. It was also found that detectors with electrodes smaller than the detector area, in instances where the X-ray beam reaches the electrode edges, continue to yield larger pulses as the bias voltage is increased. This is apparently attributable to improved charge collection at the electrode edge. It was also found that where the incident radiation was collimated and only the central region of the electrode was irradiated, the detector output/detector bias response was the same as for the full coverage electrode detector.
It has also been determined that CdTe wafers only 2 mm thick provide adequate stopping power for CT scanning operations. Table VI shows the stopping power of a 2 mm CdTe wafer in cooperation with various pre-hardening radiation filters.
TABLE VI
______________________________________
STOPPING POWER OF 2mm CdTe
150 kV X-RAY BEAM
Prehardening Stopping
Filter Power
______________________________________
6mm Al 96%
6mm Al, 1.3mm Cu 93%
6mm Al, 2.6mm Cu 90%
6mm Al, 3.9mm Cu 89%
______________________________________
FIG. 15 shows a comparison of the stopping powers of 2 mm thicknesses of various detector materials used in CT scanning (including scintillator crystals). It should be noted that while bismuth germanate has a high stopping power, it is a scintillator as opposed to being a direct electrical conversion material.
It is sometimes desirable to measure radiation occuring at more than one effective energy level. Measurements at different energies can be taken by disposing two detectors in the path of the radiation, such that radiation passing through the first detector is detected by the second, etc. as illustrated in FIG. 16. Such measurements can also be accomplished in accordance with one aspect of the present invention by disposing a multiple electrode detector, as shown in FIG. 17 such that the radiation is incident on a surface of the detector transverse to the electrodes. Here, the first electrode develops a charge in accordance with the number of photons passing through a given distance of CdTe and the second electrode develops a signal in accordance with the charge produced by the photons not blocked by the first portion of CdTe.
FIG. 18 illustrates the usage of this invention in a CT scanner. A plurality of multi-element detector arrays 14 (FIGS. 11B-11D) are disposed in a stationary circular array. A pulsed radiation source 24 provides a pulsed fan-shaped beam of radiation, which irradiates a body section 120 located in a patient circle 122. Source 24, is made to rotate about body section 120 to effect a 180.degree.-360.degree. scan. The radiation passing through body section 120 is detected by detectors 14. Each element of detector 14 is AC coupled, through associated means 18, to an associated amplifier/filter 20. The respective amplifier/filters are multiplexed by a conventional multiplexer 124 through an analog-to-digital converter 12 to suitable data processing circuitry 130. Data processing circuitry 130 may be any suitably programmed data processor or computer for reconstructing a point-by-point electron density profile of the scanned body section from the individual measurements. It should be appreciated that since only one group of detectors is irradiated at a given time, a number of detectors (each from a different mutually exclusive group), can be interconnected to share a given amplifier/filter 20. The scanning of the X-ray source itself effectively multiplexes respective interconnected detectors to a common amplifier/ filter 20.
In summary, it has been found that when used in accordance with the present invention, room temperature semiconductor detectors such as CdTe, can be advantageously used in CT scanning. This involves use of a pulsed radiation source and an amplifier/filter which provides a processed output pulse from a predetermined range of frequency components present in the law detector poutput. The range of frequencies is chosen in accordance with the duration of the X-ray pulse. Such apparatus effectively removes the adverse properties of the detector as a factor in the overall performance. In effect, the pulse of incident radiation drives the rise-time response of the detector so that any adverse variatons in detector rise-times are not apparent in the final output signal from the detector electronics interface.
Further, it appears that CdTe detectors in particular, are especially advantageous. Electrodes appear to form only ohmic contacts with the CdTe wafer, without any appreciable boundary layer. The localized charge collction phenomenon also permits densely packed arrays of detectors to be formed on a single wafer by simple batch fabrication techniques. Densities of one detector per 2 millimeters have been defined on the face of a detector slab merely by scratching out sections of electrode material.
CdTe detectors are urged, easily handled and, when used in accordance with the present invention, very stable in electrical response. Since light pipes and PMT's are not needed, there is much more flexibility in designing detector configurations. For example, for dual energy detection, a thin CdTe detector (less than 0.5 mm thick) could be backed by a thicker, (2 mm) detector as shown in FIG. 16. Alternatively, a unitary configuration could be utilized such as shown in FIG. 17.
The same considerations that apply to CdTe also apply to HgI2. It is also thought that the following semiconductor materials among others, exhibit properties similar to those of CdTe and may be advantageously utilized in accordance with the present invention; gallium arsenide (GaAs), aluminum antimonide (AlSb), indium phosphate (InP), zinc telluride (ZnTe), tungsten selenide, (WSe.sub.2), cadmium selenide (CdSe), bismuth iodide (BI.sub.3) and cesium antimonide (Cs.sub.3 Sb). Similarly, germanium (Ge) of low quality typically exhibits properties not unlike those of CdTe.
While various conductors shown interconnecting the elements of some of the figures of the drawing are shown in single line form, they are not shown in a limiting sense and may comprise plural connections as is understood in the art. Nor should exemplary component values be construed in a limiting sense. The exemplary circuits can also be optimized for particular applications.
It should be further understood that the above description is of illustrative embodiments of the present invention, and that the invention is not limited to the specific forms shown. For example, while the present invention is described primarily in the environment of CT scanners, it should be appreciated that it may find application in other types of apparatus requiring the de
