EP0788662A1 - Semiconductor x ray detector - Google Patents

Abstract

The invention provides an X ray detector comprising at least one high resistivity, type II-VI semiconductor material on which are arranged two or more electrical contacts, at least one of which is selected from the group of blocking contacts.

Description

 SEMICONDUCTOR-BASED X-RAY DETECTION DEVICE

DESCRIPTION

Technical area

The present invention relates to an X-ray detection device based on semiconductors.

State of the art

Many types of detectors have been devised for the detection of X or γ radiation. If the nature of the detector medium is very varied, solid, liquid or gaseous, the principles of detection, for their part, are generally based on the same processes of ionization or excitation of the detector medium by the passage of charged particles.

However, the number of charged particles created in the detector as well as the means of measuring the resulting signal are very different depending on whether it is to detect X-ray or γ-radiation. (see reference [1]).

The use of solid-state detectors based on semiconductors has been the main contribution of the last thirty years to techniques for detecting X or γ radiation, which mainly use gas or scintillation detectors.

Semiconductor-based detectors directly convert energy into X or γ radiation in matter without going through intermediate steps such as emission of photons visible in the case of scintillators. This eliminates the problems of coupling synonymous with loss of yield. The energy required to create an electron-hole pair in a semiconductor is much lower than in a gas or in a scintillator (around 4eV in semiconductors versus 30eV in gases and 300eV in a scintillator photomultiplier system). Consequently, the number of free charges created per photon detected is greater, which makes it possible to obtain high gains with low noise. In addition, their high atomic number and density makes it possible to use significantly lower detection volumes than those of gases or scintillators, while retaining the same quantum detection efficiency (see reference [2]).

All of these advantages have made it possible to use detectors based on semiconductors in the following three fields of application, which are given in chronological order of their study: - nuclear detection, the objective of which is to measure the energy deposited by a γ photon from a source of nuclear radiation,

- scientific instrumentation, since it is necessary to detect brief X-ray pulses and measure their temporal evolution and intensity,

- X-ray detection whose objective is to produce the radiological image of an object irradiated by an X-ray generator. This last field of application of X-ray detection by detectors based on semiconductors is very recent and therefore much less studied than that of the detection of γ radiation, which began in the 1960s. Among the semiconductor materials, Cadmium telluride (CdTe) is the best choice having regard to its electronic properties (see reference [3]). However, other detectors, in particular based on type IV (Si, Ge, ...), II-VI (ZnS, ...), III-V (GaAs, InP ...) semiconductors, or II-VII (Hgl2, ...), can be used both in the X domain and in the γ ray domain. The use of these semiconductor materials as X-ray detectors involves the deposition of two electrical contacts on the surface of the material, at the terminals of which a bias voltage is applied. The charge carriers, that is to say the electron-hole pairs created by the interaction of the photon X with the material, separate under the action of the electric field, the electrons migrating towards the positive electrode and the holes towards the negative electrode. The ability of these charge carriers to migrate towards the electrodes without being trapped by the faults present in the semiconductor material conditions the value of the measured signal. This aptitude also called "transport property" of the charge carriers is all the higher as the electric field applied over the entire thickness of the detector is strong, because it limits their transit time in the detector.

These transport properties of the charge carriers as well as the resistivity of the material, which impose the so-called dark current (detector current in the absence of radiation), and the useful detection area depend on the purity of the material. ie the presence of active faults in the prohibited band. These active defects appear systematically during the crystallogenesis of the material and this, whatever the pulling method used (THM ("Traveling Heater Method" or mobile radiator method), HPBM ("High Pressure Brigman Method" or Brigman high pressure) or BM ("Brigman Method" or Brigman method ) for CdTe). The bibliography concerning the study of these defects using the pulling method is abundant and the latest developments show that it has not been possible to remove them (see reference [4]).

The choice of the nature of the metallic contact deposited on the semiconductor material is dictated by the need to limit the dark current, to limit the contact resistance and to impose the electric field on the entire thickness of the detector, this which ensures a high useful detection area. Here again, the literature on the various possible detection structures such as ohmic structures (metallic deposit), junctions (implantations, diffusions), diode structures (heterojunctions, etc.) is abundant (see reference [5]).

But in addition to the choice of the detector material according to its purity, and the nature of the contact in order to make the detector's performances optimal, the detection structure (contact - semiconductor - contact) thus formed must meet a common specification. to the detection of X and γ radiation, namely obtaining a high signal with a minimum of noise which is constant during the time of its acquisition.

However, if these optimal detection structures of the junction, diode type make it possible to obtain a high signal with a minimum of noise, they unfortunately have a polarization effect which consists in an evolution over time of the spatial distribution of the electric field applied between the two electrodes (see reference [6]).

Again many publications relate to the functioning of these ideal structures, but these are used exclusively for the detection of γ radiation. These publications show that the polarization effect is linked to the presence of active defects in the semiconductor material (for example CdTe: Cl), which the optimal structures (type junctions, diodes) demonstrated. Today, only the use of a specific detection structure, a so-called "electroless" contact deposited on a previously chemically polished, so-called "ohmic" surface, makes it possible to make the measured signal constant over time. In return, the dark current is high (high noise), which limits the applied electric field and therefore the transport properties of the charge carriers (low efficiency) (see reference [7]).

Such a non-optimal detection structure is the only one used for the detection of γ radiation. The interpretation of the polarization effect (evolution over time of the spatial distribution of the applied electric field) has led new users of X-ray detection systems to use such structures.

The present invention relates to an X-ray detection device which overcomes these various drawbacks. Statement of the invention

The present invention relates to an X-ray detector made of a high resistivity semiconductor material of type II-VI on which are arranged at least two electrical contacts, at least one of these being taken from the family of blocking contacts. .

The nature and principle of X-ray detection, different from that of γ radiation detection, thus allow the use of optimal detection structures (Pin diodes, pn junction .... based on CdTe for X detection , while they do not work in γ detection. The polarization effect responsible for abandoning optimal structures in γ detection can be suppressed under certain conditions in X detection.

Such a structure allows the application of a strong electric field while limiting the dark current by a factor of 3 to 10 and eliminating the polarization effect specific to the CdTe material.

Such a structure, for which there is no stacking and no drag, makes it suitable for carrying out X-ray imaging. Advantageously, a detection structure of the head-to-tail diode type blocking / CdTe / blocking can be deposited on any CdTe material.

The detection of X-rays at room temperature using optimal structures based on CdTe therefore makes it possible to:

- limit the current of darkness,

- allow a high electric field and therefore a high signal and a weak memory effect;

- obtain a constant signal over time. As regards blocking contacts (such as aluminum, indium, silver), the present invention goes against what was done in the prior art. Indeed, such blocking contacts which are stable over time for X-ray radiation, make it possible to greatly improve the quality of X-ray detection. In addition, these blocking contacts (such as aluminum, silver, indium) were quickly abandoned in Gamma detection because they were not stable over time. Since X detection has developed on the basis of Gamma detection, those skilled in the art therefore use ohmic contacts.

Brief description of the drawings

FIGS. IA, IB and 1C illustrate a detection device γ, FIGS. 2A and 2B illustrate a detection device X according to the invention,

FIGS. 3A, 3B and 3C illustrate current-voltage characteristics for different structures according to the invention,

FIGS. 4A, 4B and 4C illustrate curves of detection of γ radiation,

FIGS. 5A, 5B and 5C illustrate X-ray detection curves,

FIGS. 6A, 6B and 6C illustrate a characterization by time of flight of the device of the invention with a source of γ radiation,

FIGS. 7A to 7C illustrate a characterization by time of flight of the device of the invention with a source of X-ray radiation. Detailed description of embodiments

To better understand the invention, it is necessary to explain the differences between X and γ radiation, the differences in the principle of detection of X and γ radiation as well as the different physical criteria required for X and γ detectors.

Both X and γ radiation are made up of photons whose energies are roughly of the same order of magnitude. The differences lie in the sources of emissions and their control.

The γ radiation comes from radioactive sources whose emission of photons is random, therefore not controllable. The energy of each photon is quantified, because the photon comes from disintegrations of the atomic nucleus. The activity (number of disintegrations per second) is variable, but generally low. X-ray radiation comes from a generator whose emission of photons is controllable. We obtain an energy spectrum of photons which we can control the maximum energy (by the high voltage of the tube) and the number of photons per unit of time (by the intensity of the tube). The photon flow is generally quite high. The emission of X photons can be continuous or chopped in the form of repetitive pulses with the use of a chopper.

Γ radiation is mainly used in nuclear medicine. The objective is to perform γ spectrometry of photons from tracers that have been injected into the patient. This γ spectrometry consists in detecting all the photons emitted and measuring their energy. X-ray is mainly used in radiography. The objective is to produce the image of an object by subjecting it to a spectrum of photons, by measuring the signal from the transmitted photons which have not interacted with the object during the acquisition time.

Unlike γ spectrometry, we do not measure the energy of each photon, but the signal resulting from the interaction of photons transmitted during the acquisition time in the detector.

The measurement of the energy of each photon produced by γ spectrometry is very different and more restrictive than that of the signal from a set of photons interacting in the detector produced in X-ray radiography.

FIG. 1A illustrates a detection device γ, with a source of rays γ 10. FIGS. 1B and 1C respectively represent curves of the integrated current Q as a function of time and of the number of strokes as a function of the measured value Q _mes .

FIG. 2A illustrates a detection device X, with an X-ray generator 11. FIG. 2B illustrates the measured current I as a function of time, with integrated current values Q.

The detection of γ radiation by a CdTe semiconductor began in the 1960s. Numerous studies have been carried out in order to optimize the detection structure for better spectrometry. Today, only the ohmic structure with deposit of two electroless contacts (Gold or Platinum) on a surface previously chemically etched allows to obtain an acceptable (but not optimal) spectrometry with a constant signal over time. The other structures of detection (junction type, diodes ...) display much better performance in terms of energy resolution, but this only in the first few seconds or minutes of acquisition at the end of which no more signal is observed. This so-called polarization effect is due to the presence of defects in the material which progress in crystallogenesis has not yet made it possible to remove.

The detection of X-rays by a CdTe semiconductor began in the 1990s and therefore much later than that of γ radiation. This explains the reasons which have pushed the many users of CdTe in X detection to use the only structure operating in γ detection. The polarization effect which causes an evolution of the spatial distribution of the electric field in the detector until its extension in a few seconds after its application, is highlighted in γ detection and should in the same way be observed in X detection .

The object of the present invention is to demonstrate that a certain optimal detection structure works in X detection, while it does not work in γ detection.

The detection device of the invention consists of a semiconductor material of high resistivity of type II-VI: CdTe to Cl, CdTei- _x Se _x , Cd _ι _ _x Zn _x Te: Cl, CdTeι_ _x Se _x : Cl, GaAs, Hgln on which is deposited a blocking contact by displacement of cations in solution thus conferring on the Metal / Semiconductor contact properties remarkable electrics. The blocking contact can be placed on one side, but better still on both sides. Such a structure with two blocking contacts deposited on the opposite faces of a CdTe detector has a resistivity 3 to 10 times greater than that of this same material provided with gold or platinum electrode contacts (so-called ohmic structure). Consequently, this blocking / CdTe / blocking structure is the seat of a dark current approximately 3 to 10 times weaker for the same polarization. It behaves like a head-to-tail diode structure.

Figures 3A, 3B and 3C illustrate the current-voltage characteristics respectively:

- for an ohmic / CdTe / ohmic structure

- for a blocking / CdTe / blocking structure - for a blocking / CdTe / ohmic structure.

The contacts are, in fact, divided into two families: blocking contacts (such as aluminum, indium, silver) and ohmic contacts (such as gold or platinum).

For the first structure (FIG. 3A), a material resistivity of 10 ⁹ Ω-cm is obtained; for the second (Figure 3B), we obtain an apparent resistivity of 10 ¹¹ Ω-cm. In addition, the contact resistance turns out to be very low compared to that of CdTe, which allows the electric field to be applied over the entire volume of the CdTe detector and not to be consumed under the blocking electrodes. Such structures have already been studied in γ detection (reference [8]), but nothing has been published on their use in X detection, because the reasons for their abandonment in γ detection do not apply in X detection. Detection structures based on CdTe of high resistivity used in γ detection are commonly called "ohmic structures". FIGS. 4A to 4C present results of detection of γ radiation obtained with the CdTe semiconductor of high resistivity provided with ohmic contacts and diode contacts. The source of γ radiation is a radioactive source of cobalt 57 for which the emitted photons have the following energies: 14 keV (9.1% of cases), 122 keV (85.7% of cases), 136 keV (10.7 % of cases).

FIG. 4A presents the ideal theoretical spectrum incident to the CdTe detector.

. The Au / CdTe / Au ohmic structure (with a 3x3x3 mm detector, a 150-volt polarization, a dark current. -10 ^" A) makes it possible to obtain γ spectrometry with average performance as shown in FIG. 4B, because the resolution in measured energy (between 5 and 8%) is far from the theoretical resolution (2%) The ohmic structure does not allow the application of a strong electric field which would certainly allow the charge carriers created in the volume of the CdTe detector to migrate towards the electrodes without being trapped by the active defects of the material, but which would generate a too high dark current. Thus high electric field and weak dark current are incompatible with an ohmic structure. The ohmic contacts mean that the dark current is not limited, but imposed by the resistivity of the material. Thanks to this dark current, the ohmic detection structure does not polarize, i.e. the spectrum measured remains stable during the time of its acquisition (a few minutes). The blocking diode blocking / CdTe / blocking structure (with a 3x3x3 mm ³ detector, 300Volts polarization, 10 ^"9 A dark current) does not allow to obtain a γ spectrometry, as shown in FIG. 4C, no signal not being detected. The dark current being 3 to 10 times lower than that of the previous structure for the same bias voltage, a higher bias voltage can be applied. However, the absence of spectrum shows that the field electric is not applied to the entire volume of the detector and that, subjected to a DC bias voltage, the blocking / CdTe / blocking detector polarizes.

No detection structure combining the application of a strong electric field, a very low dark current and a constant signal over time has been proposed for the detection of CdTe-based γ radiation at room temperature.

X-ray radiation is most often made up of a train of pulses of a few milliseconds at the frequency of a few tens of Hertz. The high voltage of generator X varies between 20 and 160 kV, the intensity between 2 and 40 mA. In FIG. 5A, there is a train of pulses of duration 2 ms, of frequency 50 Hz, with a voltage 120kV / 20mA.

The Au / CdTe / Au ohmic structure (with a 10 × 10 × 10 mm detector, 50Volts polarization, 10 ^" A dark current) displays good sensitivity, but the presence of a trail 20 which appears at the end of each pulse X- , as illustrated in FIG. 5B, causes a stacking of the measured signal. This drag is linked to the trapping of the charge carriers which have been trapped during pulse X due to the presence of CdTe faults and the weak electric field applied.

The blocking / CdTe / blocking diode structure (with a 10 x 10 × 1mm detector, 150Volts polarization, 10 ^"9 A dark current) displays a sensitivity equivalent to the Au / CdTe / Au ohmic structure without presenting the polarization effect (see Figure 5C). This unexpected finding is remarkable, because it opens the way to the use of a structure allowing the application of a strong electric field for a weak dark current. strong electric field makes it possible to limit the trapping / trapping of the charge carriers and thus to limit the drag and consequently to eliminate the stacking. These blocking / Cd / Te / blocking diode structures seem to perfectly follow the theoretical temporal evolution of the train d pulse X with a dynamic attenuation at the cutoff of the radiation close to four decades.

To better understand these phenomena, we study these structures using a characterization manipulation called "time of flight" (or "Time of flight" in English). It allows by the use of a very fast (500ps pulses) and repetitive (30Hz max.) Ultraviolet laser, as shown in FIGS. 6A and 7A., To observe the temporal evolution of the spatial distribution of the electric field ( see Figure 6B). In γ detection, as shown in Figure 6A with a Cobalt 57 radioactive source, everything happens as if the detector was constantly in the dark since the γ photon incident to the detector creates only very few charge carriers and in a space infinitely smaller than the volume of the detector. The flight time experiment confirms the presence of a constant electric field over time, higher on the cathode side for the Au / CdTe / Au structure (with a 10 × ¹⁰ × ¹⁰ mm detector, 54V polarization, 10 ^" dark current ⁶ A) (see FIG. 6C, curves 30 and 31 corresponding to use with and without filter. It also confirms the absence of an electric field for the blocking / Cd / Te / blocking structure (with a detector 10 X 10 X 1mm , 90V polarization, 10 ^" dark current A). The 100ms signal disappears after switching on.

In X detection (see Figure 7A) with a 120kV, 20mA generator, one side of the detector is irradiated by the ultra-violet laser, the other side is irradiated by X photons from the generator; This time, the detection structures are subjected to a much higher flux of photons than in γ detection, the X photons are absorbed throughout the volume and numerous charge carriers are created. The results of the time of flight show that the electric field of the Au / CdTe / Au structure (with a 10x10x1mm ³ detector, 90V polarization, 10 ^"6 A dark current) is little modified unless the rate of incident photons is too high, in which case the electric field becomes higher towards the opposite electrode to that which is irradiated by generator X (see figure 7B). The results concerning the blocking / CdTe / blocking structure

(lOxlOxlmm detector, 72V polarization, dark current 10 ^" A) show the presence of an electric field which, under the presence of many created charge carriers, has regenerated when it should be absent (see figure 7C The presence of these numerous charge carriers, by their trapping, seems ability to compensate for the effect of faults responsible for the polarization effect.

REFERENCES

[1] "Cadmium telluride and related materials as X and gamma-ray detectors. A review of recent progress" by P. Siffert (SPIE, volume 2305 pages 98 to 109, 1994)

[2] "CdTe in photoconductive applications. Fast detector for metrology and X-ray imaging" by Marc Cuzin (Nuclear Instruments and methods in physic research, section A, pages 341 to 351, 1992)

[3] "CdTe detectors responses to pulsed X-rays;

Comparison of different materials "by L. Verger, M. Cuzin, F. Glasser, J. Lajzerowicz, F. Mathy and J.

Rustic (Materials Research Society Symp. Proc, volume 302, 1993, pages 169 to 181)

[4] "Structural defects in high resistivity cadmium telluride" by M. Samimi, B. Biglari, M. Hage-Ali, JM Koebel and P. Siffert (Nuclear Instruments and Methods in Physics Research A283 (1989) p. 243-248 )

[5] "A review of ohmic and rectifying contacts on cadmium telluride" by JP Ponpon (Solid State Electronics vol. 28, number 7, pages 689 to ^' 706, 1987)

[6] Polarization in cadmium telluride nuclear radiation detectors "by P. Siffert, J. Berger, C. Scharager,

A. Cornet, R. Stuck, R.O. Bell, H.B. Serreze, F.V.

Wald (IEE transactions on nuclear science, volume

NS-23, number 1, February 1976, pages 159 to 170) [7] "Polarization-free semi-insulating chlorine doped cadmium telluride" by M. Hage-Ali, C. Scharager, JM Koebel and P. Siffert (Nuclear Instruments and Methods, 176, 1980, pages 499 to 502)

[8] G.-A-511 410.

Claims

1. An X-ray detector based ^{'semiconductor} characterized in that it consists of a semiconductor material of high resistivity on which are arranged at least two electrical contacts of which at least one is taken in the family of blocking contacts. 2. Device according to claim 1, characterized in that said semiconductor material is of type II-VI. 3. Device according to claim 1, characterized in that the material is chosen from the following materials: CdTe: Cl, Cd _ι _ _x Zn _x Te, CdTei_ _x Se _x , Cd _! _ _x Zn _x Te: Cl, CdTe _ι _ _x Se _x : Cl, GaAs, Hgln. 4. Device according to claim 1, characterized in that it comprises two metal contacts positioned on two opposite sides of the detection means.