EP2726905A2

EP2726905A2 - Radiation detector comprising a circuit for injecting a calibrated quantity of counter-charges

Info

Publication number: EP2726905A2
Application number: EP12730909.4A
Authority: EP
Inventors: Marc Arques
Original assignee: Commissariat a lEnergie Atomique CEA; Trixell SAS; Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Current assignee: Trixell SAS; Commissariat a lEnergie Atomique et aux Energies Alternatives CEA
Priority date: 2011-06-30
Filing date: 2012-06-25
Publication date: 2014-05-07
Also published as: WO2013000849A3; JP2014526161A; FR2977413B1; US9234969B2; CN103874933A; FR2977413A1; WO2013000849A2; US20150034832A1

Abstract

The invention relates to an electronic circuit especially comprising: a comparator (12) receiving a threshold potential (Vcomp) and the potential of an integration node (N), said node being able to store electrical charges generated by a photosensitive element (11); a meter (13) connected at the outlet of the comparator (12); and a circuit (41) for injecting counter-charges. Said injection circuit comprises: a capacitor (143) storing counter-charges; a transfer transistor (M2) that can be controlled in the conducting state for transferring counter-charges from a terminal (A) of the capacitor (143) to the integration node (N) on every oscillation of the comparator (12), the transfer of the counter-charges triggering a variation of a potential (Va) at said terminal (A) of the capacitor (143); and a regulation circuit (42) for controlling the transfer transistor (M2), said regulation circuit comprising means (425) for controlling the transfer transistor (M2) in the conducting state when the potential at the terminal (A) of the capacitor (143) is between two pre-determined potentials (Vp2, Vcomp2) that are independent of the transfer transistor (M2).

Description

RADIATION DETECTOR COMPRISING A COUNTER-LOAD INJECTION CIRCUIT IN QUANTITY CALIBRATED

The invention relates to an electronic circuit for a radiation detector capable of quantifying photon radiation received by a counter-charge injection circuit. In particular, it relates to matrix radiation detectors intended for X-ray or gamma-ray radiological imaging, comprising a pixel matrix of CMOS technology associated with a structure for converting X-rays or gamma into electrical charges.

A matrix radiation detector comprises a matrix of pixels and an electronic circuit forming reading means. Each pixel comprises a photosensitive element generating electric charges in proportion to the quantity of photons received. These electric charges, also called photocharges, are processed by the reading means in order to provide information representative of the quantity of photons received by each photosensitive element. The use of CMOS technology has made it possible to integrate the reading means at the level of each pixel. Thus, the electrical charges can be converted into digital signals within the pixels themselves to facilitate the transfer of the result of the detection to the outside of the array. A common solution for realizing the reading means is to use a circuit operating by integration of the electric charges. This integration circuit comprises an integration capacitance receiving the charges from the photosensitive element, a threshold comparator, a counter and a counter-charge injection circuit. During an exposure phase, the arrival of electrical charges on the integration capacitance causes a decrease in the voltage at its terminals. During a phase of reading of the electric charges, the threshold comparator switches a certain number of times, as long as the voltage at the terminals of the integration capacitance is lower than a threshold voltage. Each switching of the comparator increments the counter by one unit and controls the injection circuit the injection of a counter-charge packet whose quantity - Q0 is calibrated. The minus sign is used arbitrarily to indicate that the injected counter charges have a polarity opposite to that of the charges received from the photosensitive member. The counter is thus incremented by number of charge packets necessary to bring a voltage higher than the threshold voltage across the integration capacitance. In practice, the injection of counter-charges is generally carried out as photocharges are collected, a counter determining the number of latches of the comparator, in order to estimate the total amount of charges injected. The reading then corresponds to reading the content of the counters. The number of increments of the counter provides a numerical value representative of the amount of photons received by the photosensitive member.

The counter-charge injection circuit is a critical element of the integration circuit. Indeed, the accuracy of the measurement is based on the calibration of the quantity - Q0 of counter-charges. On the one hand, the quantity - Q0 of counter-charges must be relatively small since it corresponds to the step of the quantization of the charges; on the other hand, this quantity must be identical for each packet of counter-charges since it quantifies the charges received by the integration capacity.

However, in the current counter-charge injection circuits, the quantity of charges injected during each switching of the comparator can fluctuate. Indeed, these circuits include field effect transistors whose channels are affected by a random noise called rts "Random Telegraph Signal". The random nature of this noise affects the amount of counter-charges injected: some injections are affected by this noise, but not others. Also, when it is desired to estimate the total charge injected by a number of injections, injections that have or have not been affected are not known.

For example, a counter-charge injection circuit frequently comprises two series-connected field effect transistors (FETs) and a capacitor connected between the point of connection of the transistors and a fixed voltage, for example ground. A first transistor makes it possible to charge the capacitor at a first voltage value, called the charging voltage, controlled by the gate voltage of this transistor. The second transistor makes it possible to discharge the capacitor to a second voltage value, called the discharge voltage, controlled by the gate voltage of this transistor. The quantity - Q0 of counter-charges injected from the capacitor to the integration capacity of the integration circuit is a function of the value of the capacity of the capacitor and the difference between the charging and discharging voltages. However, the charge and discharge voltages are not directly deductible from the gate voltages of the transistors. The charge and discharge voltages correspond to the internal potentials of the transistors, which are not precisely known because of the noise due to the trapping of charges in the channel of each transistor. This noise rts is all the more significant as the components have reduced dimensions in order to generate relatively low quantities - Q0 of counter-charges. In practice, this noise rts modifies the value of the quantity - Q0 by a few percent. This modification has a direct impact on the evaluation of the quantity of photons received, and thus on the quality of the image obtained. However, such an error is generally unacceptable in the detectors, in particular in the field of medical imaging.

An object of the invention is in particular to remedy all or part of the aforementioned drawbacks by precisely determining the amount of counter-charges injected for the evaluation of the amount of charges generated by a photosensitive element. For this purpose, the subject of the invention is an electronic circuit for a radiation detector comprising:

^■ a comparator having a first input receiving a predetermined threshold potential and a second input is adapted to be connected to an integration node capable of storing electric charges generated by a photosensitive member upon receipt of a photon of radiation, charges causing variation of a detection potential on the integration node,

^■ a counter connected to the output of the comparator, so as to recognize the crossing of the threshold potential by the detection potential, and

A counter-charge injection circuit for counterbalancing the charges, said circuit comprising:

a capacitor storing electric counter-charges,

a transfer transistor that can be controlled in the on state to transfer counter-charges from one terminal of the capacitor to the integration node each time the comparator is switched, said terminal of the capacitor forming a node of the injection circuit, the transfer of the counter-charges causing a variation of a potential at said node of the injection circuit, and

a control circuit for controlling the transfer transistor, said circuit comprising means for controlling the transfer transistor in the on state when the potential at the node of the injection circuit is between two predetermined and independent potentials of the transfer transistor . With such a circuit, the quantity of charges injected at each switchover of the comparator is controlled. The total amount of charge collected by a detector is then improved, which increases the accuracy of the measurement. According to a preferred embodiment, the regulation circuit further comprises means for generating a reference potential at a point whose variation is representative of a variation of the potential at the node of the injection circuit, the means for controlling the transfer transistor controlling it in the on state when the reference potential is between two predetermined and independent potentials of the transfer transistor.

The invention also relates to a radiation detector comprising a photosensitive element generating electrical charges on the integration node upon reception of photon radiation, and an electronic circuit as described above, the second input of the comparator being connected to the integration node.

The invention has the particular advantage of allowing the use of components of reduced dimensions while avoiding the noise rts. The quantity - Q0 of counter-charges injected can be very small, thus leading to a precise quantification of the charges generated by each photosensitive element. In addition, the use of components of reduced dimensions makes it possible to limit the area occupied by the counter-charge injection circuit in each pixel. The invention then allows the use of these compact components, without degrading the accuracy of the measurement. By compact components, one understands components whose greatest length is of the order of the micron, or less. The invention will be better understood and other advantages will appear on reading the description which follows, made with reference to the attached drawings in which:

- Figure 1 shows the electrical diagram of a pixel in a radiation detector according to the state of the art;

FIGS. 2A to 2E illustrate the operating principle of a counter-charge injection circuit in the pixel of FIG. 1;

FIGS. 3A to 3E illustrate, by a representation similar to FIGS. 2A to 2E, the impact of noise rts on the operation of the counter-charge injection circuit of FIG. 1;

FIG. 4 represents the electrical diagram of a first pixel example according to the invention;

FIG. 5 represents the electrical diagram of an alternative embodiment of a regulation circuit in the pixel of FIG. 4;

FIG. 6 represents the electrical diagram of a second pixel example according to the invention;

FIGS. 7A to 7E illustrate, by a representation similar to FIGS. 2A to 2E, the operating principle of the counter-charge injection circuit in the pixel of FIG. 6;

FIG. 8 represents, in chronograms, the operating principle of the counter-charge injection circuit in the pixel of FIG. 6.

FIG. 1 represents the electric diagram of a pixel 10 in a matrix radiation detector according to the state of the art. Each pixel 10 forms a photosensitive dot of the matrix detector. It comprises a photodiode 1 1, a threshold comparator 12, a counter 13 and a counter-charge injection circuit 14. The threshold comparator 12, the counter 13 and the injection circuit 14 form an electronic circuit allowing a reading the photodiode 1 1. The photodiode 1 1 could be replaced by a phototransistor or, more generally, by any photosensitive element generating electric charges in proportion to the amount of photons it receives. The photons considered have, for example, a wavelength in the visible range or in the X-ray range. In the latter case, either the photosensitive element directly produces electric charges under the effect of X-radiation, or is sensitive to visible radiation, a scintillator then being interposed between the X-ray source and the photosensitive element.

The photodiode 11 has a parasitic capacitance used as an integrating capacitance for storing electric charges generated during an exposure phase. The parasitic capacitance of the photodiode is generally sufficient. Nevertheless, a capacitor could be connected in parallel with the photodiode to increase the integration capacity. The anode of the photodiode 1 1 receives a fixed voltage. It is for example connected to the electrical ground. The threshold comparator 12 receives on a positive input a threshold potential Vcomp. The negative input is connected to the cathode of the photodiode 1 1. An output of comparator 12 is connected to an input of counter 13.

The counter-charge injection circuit 14 comprises two field effect transistors (FETs): a first transistor M1 and a second transistor M2, two voltage sources 141 and 142 and a capacitor 143 of capacitor C. The drain of the first transistor M1, the source of the second transistor M2 and a terminal of the capacitor 143 are connected at a point A, called the node of the counter-charge injection circuit.

The first transistor M1 is connected by its source to the voltage source 141 and by its drain to the source of the second transistor M2. The first transistor M1 makes it possible to constitute a load at the node of the charge injection circuit. It will be called a precharge transistor. The connection point between the transistors M1 and M2 corresponds to the point A previously defined.

The drain of the second transistor M2 is connected to the cathode of the photodiode 1 1, so as to inject counter charges. Also, the second transistor M2 may be called a transfer transistor. It should be noted that the cathode of the photodiode 11 also corresponds to the point where charges generated by the interactions of the radiation in the detector accumulate. This point can be called integration node N of the pixel. In other words, the node N is a point of connection between the photodiode 1 1 and its electronic reading circuit. It can receive, on the one hand, electrical charges of the photodiode 1 1 when exposed and, on the other hand, electric counter-charges of the counter-charge injection circuit 14.

The collection of electric charges and the injection of counter-charges on the N node cause a variation of its potential. By way of example, it is considered that, during the reception of photons, the photodiode 11 generates negative charges (electrons) stored on its cathode. These negative charges cause a reduction of the potential at the node N. When the potential becomes lower than the threshold potential Vcomp, the comparator 12 switches. Each switchover is counted by the counter 13. The gate of the transistor M1 is biased to a fixed potential Vg1. The gate of the transistor M2 is biased by the voltage source 142. The capacitor 143 is connected between the point A and a fixed voltage source, for example ground. The voltage sources 141 and 142 respectively deliver potentials Phi 1 and Phi 2. They are controlled by a control circuit 15 receiving as input a switching information of the threshold comparator 12, and outputting a first signal Phil c controlling the voltage source 141, and a second signal Phi2c controlling the voltage source 142. The counter-charge injection circuit 14 may not include a capacitor 143, the capacitor C being in this case provided by the parasitic capacitances of the transistors M1 and M2.

FIGS. 2A to 2E illustrate the operating principle of the counter-charge injection circuit 14 according to a hydraulic model, which is conventional in the field of charge-coupled circuits (CCD). In these figures, it is considered that the transistors M1 and M2 are strongly inverted. Transistors M1 and M2 could work just as well in low inversion. The quantitative values would be modified, but the qualitative explanations given below remain valid. In each of these figures, the columns represent, from left to right, the potential Phi1, the internal potential Vg1 s of the transistor M1, the potential Va at the point A, the internal potential Phi2s of the transistor M2, and the drain potential Vd2 of the transistor M2. The attention of the reader is drawn to the fact that it is the internal potentials of the transistors M1 and M2 which are considered, that is to say the potentials of the channels, and not the potentials applied to the grids. The first order approximation of the internal potential of the gate of a transistor is Vg - VT, with Vg the potential applied to the gate of the transistor, and VT the threshold voltage of the transistor. The injection of counter-charges to the photodiode 11 requires the succession of a precharge step, a skimming step, and a transfer step. The injection of counter-charges is for example triggered by the tilting of the comparator 12. FIG. 2A represents the counter-charge injection circuit 14 during the precharging step. In this step, the potential Phi1 is at a high level Phi1_h. The Phi2 potential is at a high Phi2_h level. The internal potential Phi2s is therefore at a high level Phi2s_h. The potentials Phi1_h, Vg1 and Phi2_h are determined in such a way that the potential Phi1_h is greater than the potential Vg1 s and lower than the potential Phi2s_h. The potential Va can thus stabilize at the potential Phi1_h. FIG. 2B represents the injection circuit 14 during the skimming step. During this step, the potential Phi1 is at a low level Phi1_b, lower than the potential Vg1 s. The Phi2s potential is maintained at its high Phi2s_h level. Transistor M1 conducts and discharges excess charges from capacitor 143 to voltage source 141, resulting in a decrease in potential Va. FIG. 2C shows the injection circuit 14 at the end of the skimming step. In this figure, it appears that the potential Va has stabilized at the potential Vg1 s. During the transfer step, represented by FIG. 2D, the potential Phi2 is kept at a low level Phi2_b. The internal potential Phi2s is therefore at a low level Phi2s_b. The potential Phi2_b is determined in such a way that the potential Phi2s_b is lower than the potential Vg1 s. The transistor M2 conducts and discharges the excess charges of the capacitor 143 to the photodiode 1 1. The transfer of the charges causes a decrease in the potential Va and an increase of the potential on the node N. FIG. 2E represents the injection circuit 14 at the end of this transfer step. Potential Va has stabilized at Phi2s_b potential. The step of transferring the excess charges (countercharges) to the photodiode 1 1 thus made it possible to reduce the potential Va from Vg1 s to Phi2s_b. The quantity - Q0 of counter-charges injected on the photodiode 1 1 is therefore C x (Vg1 s - Phi2s_b). For the next injection of counter-charges, the precharge, skimming and transfer steps are repeated. Injections are performed until the potential on node N reaches the threshold potential Vcomp. It should be noted that the capacitance of the photodiode 11 may be sufficiently small for the injection of a single counter charge to bring the potential of the node N to the threshold potential Vcomp, which causes the comparator 12 to switch over.

In order to inject counter-charges in a limited quantity, the transistors M1 and M2 must be controlled by low potentials and the capacitance C of the capacitor 143 must be minimal. As a result, the dimensions of the transistors must be relatively small. The FET transistors then see a significant noise appear, particularly annoying by the fact that it is random. This noise rts is due to the trapping of one or more charges in one or more traps of the transistor channel. The duration during which charges are trapped is typically of the order of one second. Throughout this period, the operation of the FET transistor is changed. This modification can be seen as a conduction variation of the transistor, or as a variation of the transistor channel potential for the same potential on the gate. FIGS. 3A to 3E illustrate the impact of noise rts on the operation of the counter-charge injection circuit 14 of FIG. 1. FIGS. 3A to 3E are respectively equivalent to FIGS. 2A to 2E, in the case where a hole is trapped in the channel of transistor M1. This positive trapped charge increases the potential Vg1 s in the channel. It is represented in Figures 3A to 3E by a bump, by analogy with a pebble placed on the bottom of a stream. This hump disturbs very little the precharging step, shown in FIG. 3A. The potential Va also stabilizes at the potential Phi1_h. On the other hand, the trapped charge slows down the skimming step, shown in FIG. 3B, and above all, it modifies the level at which the potential Va stabilizes. At the end of the skimming step, shown in FIG. 3C, the potential Va is slightly greater than the potential Vg1 s. During the transfer step, shown in FIG. 3D, a larger quantity of charges is therefore injected onto the photodiode 1 1. It is noted that even though the trapped charge does not alter the potential of the channel over its entire width, it necessarily slows down the skimming stage. However, the durations attributed to each stage are in practice always limited. In other words, the trapping of a load necessarily results in a change in the potential Va. Similarly, a hole could be trapped in the channel of the transistor M2, changing the stabilization level of the potential Va at the end of the transfer step. In summary, the noise rts implies a variation of the quantity - Q0 of counter-charges injected. The variation of this quantity - Q0 is typically of the order of a few percent, which may be unacceptable in certain imaging fields, especially medical imaging.

Thus, in a general manner, the so-called rts noise randomly affects the potential in the channels of the transistors of the counter-charge injection circuit. Consequently, during each injection of counter-charges, the quantity of charges injected can fluctuate uncontrollably.

The invention aims to overcome the noise rts in the FET transistors and to allow the injection of a constant amount of counter-charges. For this purpose, the counter-charge injection circuit according to the invention comprises means for controlling the transfer transistor M2 so that, during the transfer step, the voltage variation at the node A of the circuit injection is equal to a variation between two predetermined independent potentials of the transistor M2, that is to say independent of the potential of its channel. In addition, the voltage variation at the node A of the injection circuit is independent of the potential of the channel of the precharge transistor M1.

FIG. 4 represents the electrical diagram of a first exemplary pixel 40 according to the invention. The electronic reading circuit of the pixel 40 differs only from the electronic reading circuit of the pixel 10 of FIG. 1 by the counter-charge injection circuit. Said circuit 41 also comprises two series-connected FET transistors M1 and M2, a voltage source 141 and a capacitor 143 of capacitance C, typically from a few fF to a few tens of fF. This capacity C can be called the counter-charge capacity, because it participates in the creation of a counter-load at the node A of the counter-charge injection circuit 41. Said circuit 41 furthermore comprises a regulation circuit 42. The transistor M1 is connected by its source to the voltage source 141 and by its drain at the point A. The transistor M2 is connected by its source to the point A and by its drain to the node N, that is to say at the cathode of the photodiode 1 1. The capacitor 143 is connected between the point A and a fixed voltage source, here the mass. The gate of the transistor M1 is always biased to a fixed potential Vg1. On the other hand, the gate of the transistor M2 is biased by a control potential Phi2 generated by the regulation circuit 42. The regulation circuit 42 receives the potential Va from the input point A and controls the control potential Phi2 as a function of the variation of this potential. It comprises a follower 421, a capacitor 422 of capacitance C2, a controlled switch 423, a voltage source 424 driving the controlled switch 423, a threshold comparator 425, a switch 426, and voltage sources delivering a fixed potential Vp2 , a fixed potential Vcomp2 and a fixed potential Phi2_h. An input of the follower 421 is connected to the point A. An output of the follower 421 is connected to a first armature of the capacitor 422, a second armature being connected to a point B, whose potential forms a reference potential Vb, potential that the the image potential of the potential at the point A (node of the injection circuit) can be qualified. The follower 421 and the capacitor 422 form means for generating a reference potential Vb whose variation is representative of the variation of the potential Va. The point B is also connected to a negative input of the comparator 425 and to the switch 423. The voltage source 424 controls the switch 423 by control pulses Phi_Vp2, so as to apply a potential Vp2 to the point B. positive input of the comparator 425 receives a fixed potential Vcomp2, lower than the potential Vp2. An output of the comparator 425 delivers the control potential Phi3. This potential can take two values Phi3_h and Phi3_b, according to the result of the comparison between the potential Vb at the point B and the potential Vcomp2. A switch 426 placed downstream of the comparator 425 makes it possible to connect the output Phi3 of the comparator to the control Phi2 of the transistor M2 during the transfer phase, this phase being triggered at a predetermined time after the comparator 12 has switched. Naturally, Phi3_h and Phi3_b correspond respectively to Phi2_h and Phi2_b. Thus, during the transfer phase, the output of the comparator 425 forms a means for controlling the transistor M2. Means for adjusting the values Phi2_h and Phi2_b, not shown, are provided in order to adjust the bias of the transistor M2. These adjustment means can be integrated in the comparator 425, or interposed between the comparator 425 and the gate of the transistor M2. The voltage sources 141 and 424, and the switch 426 may be controlled by a control circuit, not shown, receiving as input a switching information of the threshold comparator 12.

Outside the transfer phase, the switch 426 makes it possible to connect the command Phi2 to its high level Phi2_h, independently of the output Phi3.

The counter-charge injection circuit 41 operates in the following manner. During the precharging step, for example triggered by the control circuit following the switching of the comparator 12, at the time t ₀ , the potential Phi1 is at the high level P hi 1 _h and the internal potential Phi2s is at the high level Phi2s_h, with Phi1_h greater than Vg1 s and lower than Phi2s_h. During this step, the potential Phi_Vp2 is raised to its high level, so as to make the switch 423 passing. This makes it possible to set the potential of B independently of the potential of A. The output Phi3 of the comparator 425 is at its low level Phi3_b. The gate potential Phi2, controlling the transistor M2, is at its high level Phi2_h. This step lasts a predetermined duration, between t ₀ (or t ₀ + ε) and t i. The potential of A is then Phi1_h.

The skimming step consists of an adjustment of the charge generated at node A of the charge injection circuit. This step is triggered at a time t-ι, predetermined, after the switching of the comparator 12. The potential Phi1 is brought to the low level Phi1_b, lower than the potential Vg1 s. At the end of the skimming step, the potential Va is therefore at the level Vg1 s plus a possible variation due to the noise rts in the transistor M1. During this step, the potential Phi_Vp2 is maintained at its high level, which makes the switch 423 passing, so as to set the potential of B independently of the potential of A, and therefore the noise rts. The output of the comparator Phi3 is at its low level Phi3_b. The potential Phi2, controlling the transistor M2, is at its high level Phi2_h. This step lasts a predetermined time, between ti (or ti + ε) and t ₂ .

The end of skimming takes place at a time t ₂ , the time difference between t ₂ and t ₀ being predetermined. At this moment, the potential Phi_Vp2 switches to its low level, which opens the switch 423. The third step is the load transfer. This step takes place from a time t ₃ , the time difference between t ₀ and t ₃ being predetermined. During this step, the potential Phi_Vp2 is kept at its low level, which opens the switch 423 so that the potential of B follows the evolution of the potential of the node A of the injection circuit 41 via the follower 421 and the associated capacitor C2. In addition, the gate voltage Phi2 of the transfer transistor M2 is made equal to the output voltage Phi3. The output Phi3 of the comparator 425 is then at its low level Phi3_b. The gate potential Phi2, controlling the transistor M2, is at its low level Phi2_b.

The application of Phi2_b on the gate of the transistor M2 triggers the transfer of counter-charges from point A to the node N. As the counter-charges are injected on the node N, the potential Va decreases. This reduction is transmitted, via the follower 421 and the capacitor 422, to the point B. The point B and its potential Vb can thus be qualified, vis-à-vis the point A, image point of the point A and potential potential voltage potential Va, respectively, because when the switch 423 is open, which is the case during the charge transfer, the potential Vb decreases following the evolution of the potential of the point A. The potential Vb decreases until reach the potential Vcomp2. The comparator 425 then switches, and its output Phi3 switches to its high level Phi3_h. Thus, Phi2 switches to its high level Phi2_h which stops the transfer of counter-charges by the transistor M2. During this transfer, the potential Va goes from Vp2 to Vcomp2.

The transfer step therefore involves the transfer of a quantity - Q0 of counter-charges equal to C x (Vp2 - Vcomp2) Thus, during the injection of counter-charges, the variation of the potential of the charging point A is bounded by fixed and controlled potentials With respect to the counter-charge injection circuit 14 of FIG. 1, the injection circuit 41 no longer makes the quantity Q0 of the channel potentials of the transistors M1 and M2 dependent. , but of potentials external to the transistors M1 and M2 More precisely, the quantity - Q0 no longer depends on the variation of the potential Va between the internal potentials Vg1 s and Phi2s_b, but on the variation of the potential Vb (identical to the variation of Va ) between potentials Vp2 and Vcomp2, which can be precisely determined, for example in applying these potentials by voltage sources. Of course, for the quantity of counter-charges injected to depend neither on the internal potential Vg1 s nor on the internal potential Phi2s_b, it is necessary to choose the potentials Vp2, Vcomp2, Phi2_b and Vg1 such that (Vp2 - Vcomp2) <(Vg1 s - Phi2s_b). In order to limit the number of necessary potentials, it is possible to choose Vp2 = Phi2s_b and Vcomp2 = Vg1 s if the comparator 425 is slightly asymmetrical, that is to say if the switching threshold is slightly greater than Vcomp2. In the state of the art (FIG. 3), the counter-charge depends on the difference of the internal potentials of M1 and M2. It is therefore important that these two transistors are as matched as possible (proximity, same dimensions, same operating regime skimming) to eliminate as much as possible thermal or technological variations.

But in the invention, the voltage of the precharge end of A is stored during the blocking of the switch 423, and then only the variation of this voltage is dealt with. There is therefore more freedom to preload A. Particularly, the precharge transistor M1 can be realized by a relatively large MOS transistor, its capacity not influencing the precharging step.

To allow the manufacture of counter-charge packets whose quantity - Q0 is relatively low, the follower 421 is preferably weakly capacitive on its input, that is to say a few femtofarads. It is therefore subject to noise rts, which results in a variation of its offset (potential difference between the input and the output of the follower). However, the offset is eliminated by the capacitor 422, the variation of the potential Va remaining always identical to the variation of the potential Vb. On the other hand, the variation of the offset causes a difference between the variations of Va and Vb. However, this difference only remains during the generation of the counter-charge packet during which the variation of the offset takes place. In other words, if the offset of the follower 421 changes once in the course of one of N injections of counter-charges, only one packet of counter-charges sees its quantity - Q0 altered. It can also be noticed that the comparator 425 is controlled by a follower via a capacitor, that is to say by a Together with relatively low AC output impedance, it does not need to be very weakly capacitive on its input. It can thus be realized by FET transistors whose dimensions are large enough so that it does not present rts noise. In order to transmit the variation of the potential Va to the potential Vb with the minimum of losses, the capacitor 422 preferably has a large capacitance C2 relative to the parasitic capacitances existing at the point B. Typically, the capacitor 422 has a capacitance C2 of a few tens from fF to a few hundred fF. Finally, it can be noticed that the follower 421 could be replaced by a linear amplifier, the potentials Vp2 and Vcomp2 then having to be adapted as a function of the gain of amplification.

FIG. 5 represents the electrical diagram of an alternative embodiment of the regulation circuit 42 shown in FIG. 4. This variant allows a pixel-by-pixel adjustment of the quantity - Q 0 of injected countercharts. Such an adjustment may be necessary when the amount of counter-charges is very small, for example a few hundred holes. The capacitor 143 of each injection circuit must then have a very low capacitance C, of the order of the femtofarad. This capacitance C is for example obtained by the parasitic capacitances of the transistors M1 and M2. In any case, the capacitance C can undergo large variations from one pixel to another, and it is therefore desirable to have a means for adjusting the quantity -Q0 of counter-charges from one pixel to the next. 'other. According to the embodiment variant of FIG. 5, this adjustment is performed indirectly by means for modifying the transmission gain between the variation of the potential Va and the variation of the potential Vb. In other words, it is sought to modulate, for each pixel, the value of the potential of the charging point A before the injection of counter-charges, that is to say the value of one of the previously defined limits. One could also adjust the value Vcomp pixel by pixel. The control circuit 51 of FIG. 5 comprises, in addition to the control circuit 42 of FIG. 4, a set of capacitors 521, 522, 523 and 524, of respective capacitance Cg1, Cg2, Cg3 and Cg4, a controlled switch 531. , 532, 533 or 534 for each capacitor, and control means 54 of the switches. In Figure 5, consider a set of four capacitors. Nevertheless, he can be considered any number of capacitors, depending on the degree of precision desired. Each capacitor 52i, where i here takes integer values between 1 and 4, is connected between a fixed voltage source, for example ground, and the associated switch 53i. Each switch 53i is also connected to point B. The control means 54 comprise for example a memory in which is stored a memory weight for each switch. Each memory weight indicates whether the switch 53i must be controlled on or off. When the switches 53i are controlled in the on state, the capacitors 52i are connected in parallel and their capacitors Cgi add up. The sum of the capacities Cgi connected to the point B is noted Cadd. It is noted that Cadd may also have parasitic capacitances. The transmission gain between the variation of the potential Va and that of the potential Vb is then equal to the capacitive ratio C2 / (C2 + Cadd). The values of the Cgi abilities may differ from each other. For example, it is possible to choose them in a progression in power of two (1, 2, 4, 8, etc.), so as to be able to choose all the multiples of the weakest capacity Cgi by the intermediate of a binary code stored in the control means 54.

Thus, during each injection, the potential of the charging point A varies between two controlled terminals, the value of one of these terminals can be adjusted pixel by pixel. Thus, the amount of counter-charges injected at each injection can be modulated from one pixel to another, as a function of the capacitor C. FIG. 6 represents the electrical diagram of a second exemplary pixel 60 according to the invention. The pixel 60 also comprises a photodiode 11, a threshold comparator 12, a counter 13 and a counter-charge injection circuit 61 connected to each other as for the pixel 40 of FIG. 4. Structurally, the injection circuit counter-charge 61 differs from the injection circuit 41 essentially by its control circuit 62. It also differs in that the gate of the transistor M1 is no longer biased to a fixed potential, but by a control potential Phi3 generated by the control circuit 62. Said circuit 62 always receives the potential Va from point A at the input and outputs the control potential Phi2 and the control potential Phi3. It can include a follower 621, a threshold comparator 622 and a sequential logic block 623. The follower 621 receives the potential Va on an input. This follower 621 is optional. It makes it possible to maintain a small capacity at point A while supplying comparator 622, which is useful when the input capacitance of the latter is high. An output of the follower 621, forming the point B, image of the point A, of potential Vb, is connected to a negative input of the comparator 622. An output of the comparator 622, forming a reference point C of potential Vc, is connected to a input of the logic block 623. The logic block 623 outputs the control potential Phi2 and the control potential Phi3, respectively biasing the gate of the transistor M2 and the gate of the transistor M3. The control potentials Phi2 and Phi3 can take a high value, respectively Phi2_h and Phi3h, and a low value, respectively Phi2_b and Phi3_b. The logic block 623 also delivers a comparison potential Phi4 injected on a positive input of the comparator 622. The potential Phi4 can take a high value Phi4_h and a low value Phi4_b. In this counter-charge injection circuit 61, the control circuit, not shown, receiving as input the switching information of the threshold comparator 12, can control the voltage source 141 and the logic block 623.

FIGS. 7A to 7E illustrate, in a similar manner to FIGS. 2A to 2E, the operating principle of the counter-charge injection circuit 61. The columns respectively represent, from left to right, the potential Phi1, the internal potential Phi3s of the transistor M1, the potential Vb at the point B, the internal potential Phi2s of the transistor M2, and the potential at the drain of the transistor M2. FIG. 8 represents this same operation in the form of chronograms. A first timing diagram 81 represents the potential Vb at point B. For the sake of simplicity, it is assumed that the follower 621 is perfect, that is to say that the potential Vb at the point B is exactly equal to the potential Va at point A. The counter-charge injection circuit 61 can nevertheless operate with the same precision if the follower introduces an offset voltage, or if it is replaced by a voltage amplifier operating in linear mode. Timing diagrams 82, 83, 84, 85 and 86 respectively represent the potential Vc at the point C, the potential Phi1, the potential Phi4, the potential Phi2 and the potential Phi3. Figure 7A represents the counter-charge injection circuit 61 at the end of the precharging step. This step is for example performed at the end of the previous injection. The potential Phi1 is at the high level P hi 1 _h, the potential Phi2 is at the high level Phi2_h (and thus Phi2s is at the high level Phi2s_h), the potential Phi3 is at the low level Phi3_b (and therefore Phi3s is at the low level Phi3s_b), and the potential Phi4 is at the high level Phi4_h. The potentials P hi 1 _h, Phi2_h and Phi3_b are determined in such a way that Phi1_h is greater than Phi3s_b and less than Phi2s_h. At the end of the precharging step, the potential Va (and therefore the potential Vb) is thus stabilized at the potential Phi1_h. FIG. 7B shows the injection circuit 61 during the skimming step. This step is for example triggered by the control circuit following the switchover of the comparator 12. It starts at a time t0, where the voltage source 141 makes the potential Phi1 go low Phi1_b. Since the potential P hi 1 _b is lower than the potential Phi3s_b, the transistor M1 drives and discharges the excess charges from the capacitor 143 to the voltage source 141, resulting in a decrease of the potentials Va and Vb towards the asymptotic value Phi3s_b. The potential Phi4_h is chosen so that the potential Vb reaches this value Phi4_h before the potential Va reaches the potential Phi3s_b. In the present case, where the potentials Va and Vb are equal, it suffices to choose Phi4_h higher than Phi3s_b. When the potential Vb reaches the value Phi4_h, at a time t1, the comparator 622 switches. The potential Vc for example goes from a low state "0" to a high state "1". The logic block 623 then makes the potential Phi3 go high Phi3_h, which blocks skimming, as shown in Figure 7C. Immediately after the instant t1, at a time t2, the logic block 623 shifts the potential Phi4 to the low level Phi4_b. Since the potential Phi4_b is lower than the potential Vb, substantially equal to the potential Phi4_h at the instant t2, the comparator 622 switches again at a time t3, the potential Vc returning to the low state. At a time t4, after time t3, the logic block 623 switches the potential Phi2 to the low level Phi2_b. The potential Phi2_b is determined in such a way that the potential Phi2s_b is lower than the potential Phi4_h. The transistor M2 then conducts the charges of the capacitor 143 to the photodiode 1 1. This transfer step is represented by FIG. 7D. During this step, the potentials Va and Vb tend towards the asymptotic value Phi2s_b. The potential Phi4_b is chosen so that the potential Vb reaches this value Phi4_b before the potential Va reaches the potential Phi2s_b. In the present case, where the potentials Va and Vb are equal, it suffices to choose Phi4b greater than Phi2s_b. When the potential Vb reaches the value Phi4_b, at a time t5, the comparator 622 switches and shifts the potential Vc to the high state. The logic block 623 then passes the potential Phi2 high phi2_h, which blocks the transfer, as shown in Figure 7E. At the end of the transfer step, a precharging step can be performed for the next injection. In this step, the logic block 623 shifts the potential Phi3 to the low level Phi3_b at a time t6; it shifts the potential Phi4 to the high level Phi4_h at a time t7; and the voltage source 141 shifts the potential Phi1 to the high level P hi 1 _h at a time t8. As a result, the potential Vb stabilizes at the level Phi1_h. This potential P hi 1 _h being greater than the potential Phi4_h, the comparator 622 switches again at a time t9, the potential Vc returning to the low state. It is noted that the various operations of the precharging step can be performed in any order. The instants t6, t7 and t8 are for example concomitant. Moreover, the precharging step can be performed either at the beginning or at the end of each injection. In the first case, the precharging step is useful for the current injection. In the second case, it is useful for the next injection. Each transfer involves the injection of a quantity - Q0 of counter-charges equal to C x (Phi4_h - Phi4b). The quantity - Q0 is therefore independent of the internal potentials of the transistors M1 and M2. It is also independent of the offsets of follower 621 and comparator 622, provided that these offsets are stable during skimming and transfer steps. In case of offset variation during these steps, only the injection during which this variation occurs is altered, and not the following injections.

In the control circuit 62 of FIG. 6, the follower 621 could be replaced by a gain amplifier G. The quantity - Q0 then becomes 1 / G x C x (Phi4_h - Phi4b). The Phi4_h and Phi4_b potentials can be easily adjusted accordingly. Thus, during the injection of counter-charges, the potential variation of the charging point A is limited by fixed and controlled potentials.

Claims

1. An electronic circuit for a radiation detector comprising:

^■ a comparator (12) having a first input receiving a predetermined threshold voltage (Vcomp) of which a second input is adapted to be connected to an integration node (N) capable of storing electric charges generated by a photosensitive member (1 1) at the reception of a photon radiation, the charges causing a variation of a detection potential on the integration node (N),

^■ a counter (13) connected to the output of the comparator (12), so as to recognize the crossing of the threshold voltage (Vcomp) by the detection potential, and

^■ an injection circuit (41, 61) against load-to counterbalance loads, said circuit comprising:

a capacitor (143) storing electric counter-charges; a transfer transistor (M2) that can be controlled in the on state to transfer counter-charges from a terminal of the capacitor (143) to the integration node ( N) at each switching of the comparator (12), said terminal of the capacitor (143) forming a node (A) of the injection circuit (41, 61), the transfer of the counter-charges causing a variation of a potential (Va ) at said node (A) of the injection circuit (41, 61), and

a control circuit (42, 62) for controlling the transfer transistor (M2), said circuit comprising means (425, 622, 623) for controlling the transfer transistor (M2) in the on state when the potential at the node (A) of the injection circuit (41, 61) lies between two predetermined potentials (Vp2, Vcomp2, Phi4_h, Phi4_b) and independent of the transfer transistor (M2).

An electronic circuit according to claim 1, wherein the control circuit (42, 62) further comprises means (421, 422, 621) for generating a reference potential (Vb) at a point (B) whose the variation is representative of a variation of the potential (Va) at the node (A) of the injection circuit (41, 61), the means (425, 622, 623) for controlling the transfer transistor (M2) the commander at the on state when the reference potential (Vb) is between two predetermined potentials (Vp2, Vcomp2, Phi4_h, Phi4_b) and independent of the transfer transistor (M2).

An electronic circuit according to claim 2, wherein the control circuit (42) further comprises means (423, 424) for forcing the reference potential (Vb) to the first predetermined potential (Vp2), the means for controlling the transistor (M2) having a comparator (425) whose first input receives the reference potential (Vb), a second input of which receives the second predetermined potential (Vcomp2), and an output of which delivers a control signal (Phi3) according to the result of the comparison between the reference potential (Vb) and the second predetermined potential (Vcomp2), the control signal (Phi3) for controlling the transistor (M2) in the on state.

An electronic circuit according to claim 3, wherein the means for forcing the reference potential (Vb) to the first predetermined potential (Vp2) comprises a controlled switch (423) connected between the point (B) whose potential serves as a potential of reference (Vb) and a voltage source providing the first predetermined potential (Vp2).

An electronic circuit according to one of claims 2 to 4, wherein the means for generating a reference potential (Vb) comprises a linear amplifier (421) having an input connected to the capacitor (143), and a second capacitor ( 422) connected between an output of the linear amplifier (421) and the point (B) whose potential serves as reference potential (Vb).

An electronic circuit according to claim 5, wherein the control circuit (51) further comprises a set of capacitors (521 -524), a controlled switch (531 -534) for each capacitor, and control means (54) switches, each capacitor (521-524) being connected between a fixed voltage source and the associated switch, each switch (531 -534) being further connected to the point (B) whose potential serves as a potential of reference (Vb).

7. An electronic circuit according to one of the preceding claims, wherein the counter-charge injection circuit further comprises a field effect transistor precharge transistor (M1) having a gate biased to a potential. fixed (Vg1), said transistor for transferring counter-charges to the capacitor (143) and outside the capacitor (143), the transfer of the counter-charges causing a voltage variation (Va) across the capacitor (143) .

An electronic circuit according to claim 2, wherein the counter-charge injection circuit (61) further comprises an on-state controllable precharge transistor (M1) for transferring counter charges to the capacitor (143) and outside the capacitor (143), the transfer of the counter-charges causing a voltage variation (Va) across the capacitor (143), the control circuit (62) further comprising:

^■ means (622, 623) for controlling the precharging transistor (M1) in the on state so as to bring the reference potential (Vb) to the first predetermined potential (Phi4_h).

An electronic circuit according to claim 8, wherein the means for controlling the transistors (M1, M2) comprises a comparator (622) and a logic block (623), the comparator (622) receiving the reference potential (Vb) on a first input and a comparison potential (Phi4) on a second input, the comparison potential being able to take either the value of the first predetermined potential (Phi4_h) or the value of the second predetermined potential (Phi4_b), the logic block (623) receiving on an input an output of the comparator (622) and outputting the comparison potential (Phi4), a first control signal (Phi2) controlling the transfer transistor (M2) and a second control signal (Phi3) controlling the transistor of precharge (M1).

An electronic circuit according to claim 9, wherein the means for generating a reference potential (Vb) comprises a linear amplifier (621) having an input connected to the capacitor (143) and an output of which provides the reference potential (Vb ).

1 1. Radiation detector comprising a photosensitive element (1 1) generating electric charges on the integration node (N) upon reception of a photon radiation, and an electronic circuit according to one of the preceding claims, the second input of the comparator (12) being connected to said integration node (N).