METHOD FOR CALIBRATING AN IONISING RADIATION DETECTOR AND ASSOCIATED DEVICE

Abstract

A method for calibrating an ionising radiation detector, with the aim of determining a correction factor in order to establish an amplitude-energy correspondence The invention first relates to a method for calibrating a device for detecting ionising radiation, the detector comprising a semiconductor or scintillator detection material capable of generating a signal S of amplitude A upon interaction between ionising radiation and the detection material, the method including the determination of a weighting factor of amplitude A.

Claims

1. Method for calibrating a device (1) for detecting ionising radiation, the device (1) comprising a detector (10) capable of generating electrical charges under the effect of an interaction between ionising radiation in said detector, the method including the following steps: a) irradiation of the detector (10) using a known reference source (Sref) so as to generate interactions in the detector, b) selection of a plurality of correction factors (j), and determination of reference spectra (Sp.sub.refj), each reference spectrum being associated with a correction factor (j), each reference spectrum representing a distribution of the amplitude (A) of the signal (S) detected when the detector is irradiated by said reference source (Sref), the amplitude (A) of the detected signal being modulated by said correction factor (j); c) detection of a plurality of interactions in the detector during said irradiation and, for each interaction, measurement of the amplitude A of an electronic signal S generated by said detector; d) for each detected interaction, determination of a calibration function F such that: each calibration function F is defined on a plurality of correction factors j, each term F(j) of said calibration function represents a probability associated with said correction factor j each term F(j) being updated at each detected interaction, by combining the amplitude A measured during this interaction with each reference spectrum Sp.sub.refj, e) determination of a correction factor j.sub.corr by identifying a remarkable point of said calibration function F.

2. Calibration method according to claim 1, in which each step d) comprises: i initialisation of said calibration function F; ii at each detected interaction, update the calibration function F, by adding the value of the corrected reference spectrum Spref (A) to each term F (j) in this function, at said amplitude A, iii repeat step ii until a predefined stop criterion is reached.

3. Calibration method according to claim 1, in which step e) includes determination of the correction factor j.sub.corr by identifying the value of j that maximises the value of the function F, in which case: j.sub.corr=argmax.sub.i[F(j)].

4. Calibration method according to claim 1, in which: steps c), d) and e) are repeated iteratively, each step e) leading to the determination of a partial correction factor, where q denotes the iteration rank, and in which, following the first iteration, step d) comprises the update of each term F(j) by weighting said amplitude A by the product of partial correction factors j.sup.1. . . j.sup.q, these factors having been determined during the previous q iterations. step e) in the last iteration of the method then consisting of determining the correction factor j.sub.corr as a function of the product of partial correction factors j.sup.1 . . . j.sup.Q determined during said successive iterations, where Q denotes the number of iterations.

5. Calibration method according to claim 1, in which step b) comprises: i load an initial reference spectrum Sp.sub.ref previously saved in a memory, corresponding to said reference source S.sub.ref, this spectrum Sp.sub.ref representing the distribution of the amplitude of the detected signal when the detector is irradiated by said reference source S.sub.ref; ii starting from the initial reference spectrum Sp.sub.ref, determine a plurality of reference spectra Sp.sub.refj, each reference spectrum Sp.sub.refj having been obtained from the initial reference spectrum Sp.sub.ref, each value Sp.sub.refj(i) of the reference spectrum being obtained from the value Spref (i/j).

6. Calibration method according to claim 1, the detector (10) comprising a scintillator detection material or a semiconducting detection material (11).

7. Device for detection of ionising radiation comprising: a) a detector (10), the detector being capable of generating charge carriers under the effect of an interaction with said ionising radiation b) an electronic amplification circuit (22, 23) capable of forming a signal S from charge carriers collected by an electrode during an interaction and determining its amplitude A; an electronic calibration circuit (40) configured to implement steps b) to e) in the method according to claim 1.

Description

DESCRIPTION OF THE FIGURES

[0053] FIG. 1A represents a diagram of the detection device;

[0054] FIG. 1B represents a diagram of another embodiment of the detection device;

[0055] FIG. 2A represents the different steps of the calibration method, according to a first embodiment of the invention;

[0056] FIG. 2B represents the different steps of the calibration method, according to a variant of a first embodiment of the invention;

[0057] FIG. 3 represents an example of a reference spectrum Spref,

[0058] FIG. 4 represents a plurality of corrected reference spectra Spref.sub.j;

[0059] FIG. 5 represents a 3-dimensional view of a plurality of corrected reference spectra Spref.sub.j;

[0060] FIG. 6 represents a calibration function F, from which a correction factor j.sub.corr can be extracted.

DETAILED PRESENTATION OF PARTICULAR EMBODIMENTS

[0061] FIG. 1 represents a device 1 for the detection of ionising radiation according to the invention. It comprises a detector 10 made from a semiconducting material 11, that can be polarised by a cathode 12 and an anode 13. The semiconducting material is CdTe in this example. It could be another semiconducting material frequently used for the detection of ionising radiation (Si, CdZnTe . . . ). The thickness of the material is a few millimetres, for example 5 mm.

[0062] Electrons have better charge mobility properties in a CdTe type material. Also, signals giving more precise information about the energy deposited by an interaction in the detector are signals output from the anode 13.

[0063] The device also comprises an electronic amplification circuit 23, associated with the anode, so as to produce an anode signal with an amplitude proportional to the number of charge carriers collected by the anode after an interaction. Such a circuit is known and will not be described in detail.

[0064] The electronic amplification circuit can also be connected to the cathode instead of the anode. According to one variant, the anode and the cathode each have their own amplification circuit.

[0065] Thus in general, each interaction in the detector leads to the collection of charge carriers by at least one electrode E (anode A or cathode K) so as to generate an electronic signal S with an amplitude A.

[0066] The term amplitude denotes the maximum height of a signal, when the signal is in the form of a pulse. It can also be the integral of this signal, or any other function of the maximum height or the integral. In general, the amplitude represents the quantity of charges collected by the E during an interaction.

[0067] The detection device 1 also comprises an electronic calibration circuit 40 capable of implementing the detector calibration method. This circuit is an important element of the invention and will be described below.

[0068] The detection device 1 also comprises an electronic correction 50, capable of applying a correction factor j.sub.corr to an electrode signal S. In general, this correction factor is assigned to the amplitude A of the signal S from the anode A (or possibly to the signal from the cathode K).

[0069] The detection device is capable of detecting ionising radiation. The term ionising radiation denotes radiation capable of generating charge carriers in the detector 11. In particular, it can be X-radiation, γ radiation, charged particles (particularly α or β particles) or neutrons.

[0070] The amplification circuit 23 is known and will not be described in detail.

[0071] According to another embodiment, the anode 13 of the detector is divided into a plurality of elementary anodes 13.sub.1. . . 13.sub.x, each elementary anode being connected to an elementary amplification circuit 23.sub.1 . . . 23.sub.x. A localisation circuit 30 is capable of determining which elementary anode collected the signal output from the interaction and applying the calibration method to this anode. The cathode 12 can also include an electronic amplification circuit 22.

[0072] The calibration method used by the electronic calibration circuit 40 will now be described with reference to FIG. 2. Its function is to determine a correction factor j.sub.corr, this correction factor subsequently being applied to the amplitude of a signal S collected by an electrode, and more particularly by the anode. This calibration circuit is implemented when the detection device 1 is in a calibration mode.

[0073] The basic concept of calibration is to irradiate the detection device 1 with a known source of ionising radiation, called the reference source S.sub.ref, for which the theoretical response of the detection device is known. The objective then is to analyse signals generated in response to this irradiation, and to determine an appropriate correction based on the theoretical response of the detector facing this reference source, this theoretical response possibly being qualified as a reference response.

[0074] FIG. 2 represents the steps in a first embodiment of the calibration method. In a first step 100, a reference source S.sub.ref is placed facing the detection device 1, knowing that the response Sp.sub.ref of the device in this configuration is known. In this example, the response of the detector is an energy spectrum, produced by the detector in response to the source.

[0075] An energy spectrum means the amplitude distribution of signals S output from an electrode E during exposure of the detector to a radiation source. Such a spectrum is in the form of a histogram Sp in which each term Sp(i) represents the number of signals detected by the electrode considered, with an amplitude equal to i. Amplitude equal to i means an amplitude within the amplitude range i containing the value i and for example centred around the value i. Each channel with amplitude i corresponds to a known energy.

[0076] This response can also be determined by modelling the detector. It can also be determined experimentally, by adopting a sufficiently long acquisition time to be statistically representative.

[0077] In the example considered, the reference response Sp.sub.ref is an energy spectrum corresponding to the spectrum “theoretically” produced by the detector 11 in response to the reference source Sref.

[0078] FIG. 3 represents such a reference spectrum Sp.sub.ref. It comprises several peaks P1, P2, P3, P4 standing out from the background noise B. It is preferable if the initial reference spectrum Spref comprises at least one and preferably a plurality of peaks, preferably between 1 and 20 peaks. Preferably, the width of a peak at its mid-height is less than a tenth or even a twentieth or thirtieth of the range of the spectrum, which is the range of amplitudes over which the spectrum is recorded. Preferably, each peak has a height equal to at least twice, and preferably at least five times, the noise level on each side of the peak. The example in FIG. 3 represents the spectrum for a .sup.57Co source. The photoelectric peak corresponding to the emission of gamma photons with an energy of 122 keV, and the X-fluorescence peaks of the material from which the collimator of the source is made, in this case Tungsten, can be seen.

[0079] Such a spectrum .sub.Spref can be treated like a probability distribution of measuring a given amplitude knowing the reference source irradiating the detector.

[0080] The first step 100 in the calibration method is to load a reference spectrum Spref, corresponding to a reference source Sref. This reference spectrum is designated the initial reference spectrum.

[0081] The purpose of the second step 200 is to start from the initial reference spectrum Sp.sub.ref and produce a plurality of reference spectra Sp.sub.refj, each being associated with a correction factor j. Each spectrum Sp.sub.refj corresponds to the initial reference spectrum Sp.sub.ref taking account of a correction factor equal to j. A correction factor j is a positive coefficient between a minimum value j.sub.min and a maximum value j.sub.max, j.sub.min and j.sub.max are defined arbitrarily. In this example, 0.8≦j≦2. Preferably, the interval j.sub.min<1 and j.sub.max>1. The values of j are selected and predetermined.

[0082] More precisely, if Sp.sub.ref (i) designates the value of the spectrum Sp.sub.ref corresponding to the channel with amplitude i, the reference spectrum associated with the correction coefficient j can be written Spref.sub.j(i)=Sp.sub.ref(i/j). Each channel of a spectrum Sp.sub.ref; represents the number of signals collected with an amplitude within an interval

[00001]$\frac{\Delta .Math..Math.i}{j}$

centred around the value

[00002]$\frac{i}{j}.$

[0083] The above equation can also include a normalisation term K.sub.norm(j), that can be equal to 1 or equal to 1/j, in which case

[00003]$\begin{array}{cc}\mathrm{Sprefj}\left(i\right)=\frac{\mathrm{Spref}\left(i/j\right)}{j}.& \left(1\right)\end{array}$

Using this normalisation term then makes it possible to keep the integral of the spectrum constant.

[0084] When the correction factor j is more than 1, the reference spectrum Spref.sub.j corresponds to compression of the initial reference spectrum Spref, to the extent that the information in the spectrum is concentrated on a number of channels smaller than the number of channels making up the spectrum Spref.

[0085] Conversely, when the correction factor j is less than 1, the reference spectrum Spref, corresponds to expansion of the initial reference spectrum Spref, the information in the spectrum being expanded to a number of channels larger than the number of channels making up the initial reference spectrum Spref.

[0086] FIG. 4 shows several reference spectra Spref.sub.j associated with a correction coefficient j between 0.8 and 2. More precisely, each column in FIG. 4 corresponding to an abscissa j represents a reference spectrum Spref.sub.j. Thus, a weight P.sub.ji is assigned to each point (j,i) indicated by the grey levels scale, this weight representing the value of the spectrum Sp.sub.refj at amplitude i, such that P.sub.ji=Sprefj(i). The abscissas scale in FIG. 4 is not linear.

[0087] Thus, each column in FIG. 4 represents a probability distribution of the amplitude measured by the detector, knowing the correction factor j, when the detector is irradiated by the reference source Sref, the correction factor j being the abscissa of the column. Furthermore, each line in FIG. 4 corresponding to an amplitude i represents a probability distribution of the correction coefficient j, when this amplitude is measured, the detector being irradiated by the reference source Sref.

[0088] In other words, when the detector is irradiated by the reference source Sref, if a signal S with amplitude A is measured, the distribution formed by the set of values Sprefj(A), where j is between j.sub.min and j.sub.max, corresponds to a probability distribution of the correction factor j to be applied, knowing the measurement of this amplitude A.

[0089] By constructing a plurality of reference spectra Spref.sub.j, corresponding to a column with abscissa j on FIG. 4, we can determine a plurality of probability distributions, a correction factor j, knowing a measured amplitude A. These distributions will be used in the next step.

[0090] FIG. 5 is a 3-dimensional view of FIG. 4. Each profile, corresponding to a fixed value of the correction factor j, represents a reference spectrum Sp.sub.ref.

[0091] The third step 300 consists of irradiating the detector using a reference radiation source Sref, corresponding to the reference spectrum Spref. During each interaction, a signal S is collected at the terminals of an electrode E (anode 13, elementary anode 13x or cathode 12), this signal then being processed by the amplification circuit 3 attached to the electrode E.

[0092] The fourth step 400 consists of creating a calibration function F. More precisely, during this step, the calibration function F is updated as a function of the signal S collected during step 300, and particularly as a function of its amplitude.

[0093] Initially, each calibration function F is initialised to a predetermined value, for example a null value: regardless of the value of j, F(j)=0. This fourth step in the method consists of updating each function F such that every time that an interaction is detected, the value of the corrected reference spectrum Spref.sub.j corresponding to the amplitude A of the signal measured at each interaction is added to each term F(j), for j between j.sub.min and j.sub.max.

[0094] In other words, the calibration function F is updated for each detected interaction, using the following iterative equation:

For all values of j between j.sub.min and j.sub.max, F.sup.P(j)=F.sup.P−1(j)+Spref.sub.j(A)

[0095] in which:

[0096] F.sup.P−1(j) is the value of the term F(j) before iteration i,

[0097] F.sup.P−1(j) is the value of the term F(j) after iteration i,

[0098] A is the amplitude of the signal S measured by an electrode, for example the anode, following each interaction.

[0099] This iterative update is made for a plurality of detected interactions, until a stop criterion is reached, for example a predetermined number of iterations or a maximum value reached by one or several values of the function F.

[0100] FIG. 6 represents a calibration function obtained after 1000 updates, this number usually being between a few hundred and a few thousand. The abscissas axis represents correction factors j between j.sub.min and j.sub.max. The ordinates axis represents values of the calibration function F(j).

[0101] Each term F(j) is proportional to the probability that the correction factor j.sub.corr, to be applied to each interaction, is equal to j.

[0102] The fifth step 500 is determination of the correction factor j.sub.corr as a function of a remarkable point of the calibration function F, and for example as a function of the correction factor j for which the value of the function F is maximum.

[0103] In this example, j.sub.corrargmax [Fk(j)]. Returning to the example in FIG. 6, the maximum value of this calibration function occurs at j=1.03. This, in this example, j.sub.k=1.03. This means that the correction factor to be applied to the signal S from the anode, and particularly its amplitude, is equal to j.sub.corr=1.03.

[0104] According to one variant of this embodiment shown on FIG. 2B, steps c) to e) can be performed iteratively, the correction factor determined during step e) of each iteration q being qualified as a partial correction factor and denoted j.sup.q. After the first iteration, during the update of the calibration function F, the amplitude A of the signal S is weighted by the product of the “partial” correction factors j.sup.1 . . . j.sup.q−1, determined during the previous iteration or during the q−1 previous iterations, where q denotes the rank of the iteration. This product should be considered as an intermediate correction factor.

[0105] For example, step d) comprises an iterative update of the calibration function F such that

[0106] For any value of j between j.sub.rmin and j.sub.max; in which: F.sup.p,q(j)=F.sup.p−1,q(j)+Sprefj(Π.sub.l=1.sup.q−1j.sup.l×A) [0107] F.sup.p−1,q(j) is the value of the term F (j) before the p.sup.th iteration (iteration of the update to the calibration function), during the rank q iteration (iteration in which a partial correction factor is determined) [0108] F.sup.p,q(j) is the value of the term F.sub.k(j) after the p.sup.th iteration (iteration of the update to the calibration function), during the rank q series (iteration in which a partial correction factor is determined) [0109] A is the amplitude of the signal S measured by an electrode following an interaction. In this example, the signal S.sub.A0 measured by the winning anode A0 is considered, therefore A=A.sub.A0. [0110] p is the iteration index corresponding to the rank of the update to the calibration function. During the first update, p=1. [0111] j.sup.l is the partial correction factor, in which 1≦l≦q−1

[0112] Steps c) to e) are then repeated until a stop criterion is reached; this may be a predetermined number of iterations, or a small difference between two successive partial correction factors j.sup.q and j.sup.q+1. Typically, according to one variant, the number Q of iterations is between 2 and 10.

[0113] It is found that during the iterations, the values of the partial correction factors j.sup.q tend towards 1.

[0114] During the final iteration, step e) includes determination of the correction factor j.sub.corr by calculating the product of partial correction factors j.sup.1 . . . j.sup.Q successively acquired during each series.

[0115] Thus, during step e) of the last iteration, we can determine j.sub.corr such that:

[00004]$j_{\mathrm{corr}}=\underset{q=1}{\overset{q=Q}{.Math.}}.Math..Math.j^q$

[0116] During a step 600, the correction factor j.sub.corr is stored in a memory 50.

[0117] When the detector 1 comprises a plurality of anodes as shown on FIG. 1B, the calibration is made for each elementary anode, each being assigned a correction factor j.sub.corr specific to it. Each correction factor corresponding to an elementary anode, is then stored in the memory 50.

[0118] We will now describe an example operation of a detection device as shown diagrammatically on FIG. 1A.

[0119] When an interaction occurs in the detection material 11, it generates a signal on the cathode 12 and on the anode 13. The correction circuit 60 applies the correction parameter j.sub.corr, stored in the memory 50, to the signal emitted by the anode, and corrects the measured amplitude A such that A*=j.sub.corr×A, where A* designates the corrected amplitude. The calibration circuit 40 is not used during this operating mode.

[0120] In the embodiment shown in FIG. 1B, when an interaction occurs in the detection material 11, it generates a signal on the cathode 12 and on the elementary anodes 13.sub.1 . . . 13.sub.x. The localisation circuit 30 determines the elementary anode that collected the maximum signal. The correction circuit 60 applies the correction parameter j.sub.corr associated with the selected anode, stored in the memory 50, to the signal emitted by the previously selected anode, and corrects the measured amplitude A such that A*=j.sub.corr×A, where A* designates the corrected amplitude. The calibration circuit 40 is not used during this operating mode.

[0121] In the examples described above, the detection material 11 is a semiconductor. The invention is applicable when the detector 10 comprises a scintillator material coupled to a photodetector, the photodetector comprising an electrode capable of generating a signal for which the amplitude depends on the energy lost in the detection material during each interaction.