Semiconductor-based particle detector with electrical properties dedicated to particle differentiation.

Abstract

The invention relates to a particle detector comprising: - a semiconductor substrate including a sensitive volume capable of generating charges upon the passage of a particle, at least one front contact island formed under a front face of the semiconductor substrate, and at least one rear contact island formed under a rear face of the semiconductor substrate; - at least one front electrode formed on the front face of the semiconductor substrate, in contact with the front contact island; - at least one rear electrode formed on all or part of the rear face of the semiconductor substrate, in contact with the rear contact island; - an acquisition chain capable of detecting transient signals originating from said at least one front electrode. Said at least one front electrode is a voltage application and transient signal detection electrode.Said at least one front electrode is made on only a portion of the front face of the semiconductor substrate, without guard electrodes, so that a non-homogeneous electric field is created in the sensitive volume when said at least one electrode is energized. Figure for the abbreviation: Fig. 12.

Description

Title of the invention: Semiconductor-based particle detector with electrical properties dedicated to particle differentiation. technical field

[0001] The present invention relates to a particle detector.

[0002] Such a detector is designed to measure energy deposits generated by the passage of ionizing or non-ionizing radiation. Such detectors are used in many technical fields where knowledge of the ambient radiation environment (flux, nature of particles, etc.) is desired, such as radiation monitors onboard satellites dedicated to the study of radiation belts. However, the invention has a broader scope, as it also applies to the calibration of particle beams in radiation protection, or to the study of other types of radiation besides electrons or protons, in space or on the ground. Prior art

[0003] Generally, commonly used ionizing radiation detectors comprise a sensitive volume, usually corresponding to the substrate of a semiconductor SSD (Solid State Detector), typically silicon, and an amplification chain. When only the deposited energy is sampled, the amplification chain acts as an integrator to converge to a value directly comparable to the quantities of charge generated and thus to the energy deposited by the incident particle.

[0004]

[0005] It sometimes happens that the data measured by radiation monitors are erroneous. Indeed, in a mixed environment such as the radiation belts, radiation monitors can encounter significant difficulties in discriminating between electrons and protons, particularly when they deposit the same amount of energy in the detector. This is due to the commonly used measurement method, where the only information gathered during the detection of a particle is the total energy deposited in the diode. This misidentification of the nature of the particles leads to contamination of the monitor data. Data contamination occurs when measurements attributed to protons are actually due to electrons, and vice versa; thus, the wrong type of particle is not being measured. This results in a lack of reliability in the measurement of particle fluxes in the radiation belts at a given time.

[0007] It is known that observing transient signals allows information to be extracted regarding the nature and energy of the incident ionizing particle. These detectors include an amplification chain designed to preserve the integrity of the dynamics of the collection of charges generated by the incident particle. Therefore, the function is no longer to act as an integrator but solely to amplify the signal as it is at the output of the SSD.

[0008] The generation of transient signals in a semiconductor detector is based on the Shockley-Ramo theorem. This theorem allows the current measured on a given electrode to be calculated as a function of time z, according to the following expression:

[0009] VM “ / KE t(O - -qv(t) -Tv

[0010]

[0011] With i(0) the current induced on the electrode, q the electric charge of the carriers moving in the electric field of the sensitive volume with a velocity, and the virtual electric field or "Weighting Field" normalized by a potential of 1 Volt. This virtual field is calculated individually for each measuring electrode of the detector, by removing all charges, setting the potential of the electrode in question to 1 V, and the potential of all other electrodes to 0 V. Each electrode thus has its own "weighting field," defining the current generated on it independently of the electric field actually applied in the detector.

[0012] The electric field in the sensitive volume and the displacement of the carriers in this field are calculated from the solution of the Poisson drift-diffusion equations:

[0013] îyj rit 0 J n

[0014] T

[0015] VEqE^V = - {p - n + N^- NA)

[0016] JqnpnE(x, y, t)+ qD,^n

[0017] Jp = qniij^(x, y,t} +qDp^p

[0018] With n and p the electron and hole densities, Jn and JP the electron and hole current densities, V the electrostatic potential, A'p and NA the donor and acceptor site density, Dn and Dp the diffusion coefficients for electrons and holes, and and ^p the electron and hole mobilities.

[0019] This type of detector is used in accelerators, particularly to identify the isotopes of an incident ion. However, identifying the nature of the particle often relies on the signal generated when it is immobilized in the detector, which does not allow discrimination of particles passing through the detector.

[0020] In the case of the detectors described above, the first makes it completely impossible to discriminate between particles depositing the same amount of energy, and the second depends on the post-processing applied to the captured transients. Currently, scientists' efforts have focused solely on heavy particles and in controlled environments. There is no solution for measurements in uncontrolled environments (such as in the radiation belts) and for distinguishing lighter particles.

[0021]

[0022] Figure 1 shows the LET (Linear Energy Transfer) of 200 keV electrons and 500 MeV protons in a silicon volume. The LET is the amount of energy per unit length transferred by the incident particle to the target material. In Figure 1, it can be seen that the electrons and protons in this example deposit the same amount of energy in a volume up to several tens of micrometers thick. Thus, if these two particles were to pass through a detector that only considered the amount of energy deposited, this detector would be unable to differentiate between the electron and the proton. However, if one focuses on the details of the particle's path, and not just the amount of energy deposited, it is possible to identify trends specific to each type of particle.Indeed, low-energy electrons will mostly interact elastically with the matter they traverse, while protons will very often interact inelastically. Consequently, electrons will tend to have a very diffuse path, while protons will traverse the volume in a very straight line. These differences lead to different tracks, and therefore different ways of depositing the same amount of energy, as illustrated in [Fig. 2].

[0023]

[0024] The design of SSDs (Solid State Detectors) is primarily oriented towards meeting "static" criteria (typically leakage current, which is to be as low as possible). Figure 3 shows the classic structure of an SSD: an n- or p-doped substrate, and (n+ or p+) junctions for the contacts.

[0025] Such SSDs, combined with transient signal processing during particle passage through the detector, have already been used for heavy ion discrimination in controlled environments. Signals are generated in the device when the particle passes through the semiconductor diode(s) composing the detector and deposits charges in the diode's sensitive volume. These signals are measured using a detection track integrated into each diode composing the detector. Each measurement track is connected to an acquisition and processing chain. The signal is capable of sampling the shape of the transient signal generated on the detection track. The shape of the transient signals generated on the track can then be used to analyze their characteristics. Through a method of analyzing and classifying the collected signals and their characteristics, these devices can identify the incident particle and its initial energy. However, this identification is only possible when the particle has not completely passed through the detector.

[0026]

[0027] Thus, current transient measures do not allow for effective discrimination of particle types in uncontrolled environments and / or for light particles.

[0028]

[0029] CN103954988B is known to describe a detector for discriminating particles. This detector comprises several SSD diodes in parallel. One of the three diodes has an input window for the radiation to be measured. The radiation enters through the first and second diodes and is then blocked by the third. The recovered signals allow the particles and their energies to be discriminated.

[0030] We know that document CN104656117B exploits the shape of the signal generated by alpha particles and protons (heavy particles) to differentiate them.

[0031] Document EP3821278B1 also describes a detector based on guard rings for particle detection. The guard rings serve to ensure a homogeneous magnetic field within the volume of the SSD diode.

[0032]

[0033] The present invention aims at a new detector capable of efficiently discriminating between heavy and light particles.

[0034] Another object of the invention is a new detector capable of evolving in a controlled or uncontrolled environment.

[0035] Description of the invention

[0036] At least one of the aforementioned objectives is achieved with a particle detector comprising:

[0037] - a semiconductor substrate comprising a sensitive volume capable of producing charges during the passage of a particle, at least one front contact island made under a front face of the semiconductor substrate, and at least one rear contact island made under a rear face of the semiconductor substrate,

[0038] - at least one front electrode made on the front face of the semiconductor substrate, in contact with the front contact island;

[0039] - at least one rear electrode made on all or part of the rear face of the semiconductor substrate, in contact with the rear contact island;

[0040] - an acquisition chain capable of detecting transient signals from said at least one electrode before.

[0041] Said at least one front electrode is a voltage application and transient signal detection electrode, and said at least one front electrode is made on only a part of the front face of the semiconductor substrate, without guard electrodes, so that a non-homogeneous electric field is created in the sensitive volume when said at least one electrode is voltage supplied.

[0042]

[0043] Transient signals are obtained for example by connecting at least one front electrode and at least one rear electrode to the acquisition chain.

[0044]

[0045] With the detector according to the invention, an inhomogeneous electric field is generated in the sensitive volume of the detector in order to enhance the dynamics of charge collection, these charges being generated by the passage of an incident particle through the sensitive volume of the detector. The geometry of the electrode is defined so as to make the electric field non-homogeneous.

[0046] With enhanced dynamics, it is thus possible to discriminate between different types of particles using signal processing techniques familiar to those skilled in the art. Indeed, a non-homogeneous electric field within the detector's sensitive volume causes singular behaviors in the measured signals, resulting in easier discrimination. Each particle generates a highly distinctive signal shape measured via the electrode.

[0047] The electrode in the invention plays a dual role of disturbing the field to make it inhomogeneous and detecting the measured signal.

[0048]

[0049] The principle of the invention is to allow manipulation of the two terms governing the current collected by the detector during the passage of a particle, which are given by the Shockley-Ramo equation. By modulating the electric field, the transport of charges created within the sensitive volume is modified. This allows, among other things, the charge carriers to be directed towards the measuring electrode(s), and limits the loss of information. This modulation depends on the voltage applied to the measuring electrode, as well as on the geometry of the detector.

[0050]

[0051] According to an advantageous feature, said at least one front electrode has a geometric shape having at least an inter-electrode spacing suitable for creating a non-homogeneous electric field in the sensitive volume.

[0052] The inter-electrode spacing means that the electrode comprises at least two parts separated by a zone leaving the surface of the semiconductor substrate uncovered. This zone can be linear, circular, or any other shape. This allows for the creation of a non-homogeneous distribution on the surface of the semiconductor substrate and, when an electric field is created by applying a voltage, an inhomogeneous electric field within the sensitive volume.

[0053] According to one embodiment, said at least one front electrode is a spiral covering at least a part of the surface of the sensitive volume.

[0054] When the spiral covers the surface of the sensitive volume, there is the inter-electrode spacing which ensures inhomogeneity.

[0055]

[0056] According to an advantageous feature, the semiconductor substrate is circular, square or rectangular in shape and:

[0057] - said at least one front electrode consists of a plurality of front electrodes annular electrodes distributed concentrically around a central front electrode, with an inter-electrode spacing present between two front electrodes and

[0058] - said at least one rear electrode consists respectively of a plurality annular rear electrodes distributed concentrically around a central rear electrode; each front-rear electrode pair being capable of receiving a bias voltage different from that intended to be applied to the neighboring front-rear electrode pair, an inter-electrode spacing being present between two rear electrodes.

[0059] Using multiple front or rear electrodes further allows the virtual collection field to be decomposed. This makes it possible to obtain individual signals across the entire width of the detector and to reconstruct the shape of the particle's track. Each front-rear electrode pair constitutes an individual measurement track, allowing the measurement to be discretized.

[0060] Ideally, for each front-rear electrode pair, the voltage applied to the front electrode is different from the voltage applied to the rear electrode. Preferably, one electrode is grounded and the other is held at a given potential.

[0061]

[0062] The semiconductor substrate may be circular, square or rectangular in shape, and:

[0063] - said at least one front electrode comprises in a first front electrode a central electrode of circular, square, or rectangular shape and at least one second front electrode, peripheral or not, of circular, square, or rectangular shape; an inter-electrode spacing being present between the two front electrodes; and

[0064] - said at least one rear electrode consists of a single rear electrode distributed across the entire back face of the semiconductor substrate.

[0065] Such an implementation has the advantage of limiting the number of measurement tracks and yet ensuring active in-homogenization.

[0066]

[0067] According to one embodiment, where the semiconductor substrate can be a diode structure, said at least one front electrode is an anode or a cathode; said at least one rear electrode being respectively a cathode or an anode. Those skilled in the art will understand that changing from an anode+cathode assembly to a cathode+anode assembly requires an adjustment of the voltage direction.

[0068]

[0069] Advantageously, the semiconductor substrate can have a thickness between 30 µm and 5 mm, ideally 300 µm. This range of values ​​has been shown to be optimum. Too thin a thickness presents a risk that the detector will not detect charges. Too thick a thickness presents a risk of detector breakdown due to an excessively high applied voltage.

[0070]

[0071] The semiconductor substrate can also have a width between 30 pm and 5 cm, ideally 1.5 mm. A width that is too small risks resulting in an insufficient particle flux. A width that is too large risks an excessively large detector size, leading to integration difficulties.

[0072]

[0073] According to one embodiment, the inter-electrode spacing may have a width between 30 pm and 1 mm, ideally 200 pm.

[0074] An inter-electrode spacing that is too small would result in an excessively high electric field, with a risk of detector breakdown. A spacing that is too large would lead to a loss of efficiency, with a risk of charge loss in the sensitive volume area below the spacing.

[0075]

[0076] Advantageously, when this detector includes several front electrodes, the acquisition chain can be configured to apply voltages respectively to the front electrodes with a voltage difference between two consecutive front electrodes of between 0 and 20 V, ideally 0 V.

[0077] We can have zero voltage at the center and 100V at an nth ring.

[0078]

[0079] According to a preferred embodiment, the sensitive volume has a doping level between 10¹⁰ and 10¹⁴ cm³, ideally 10¹¹ cm³. These values ​​correspond to relatively low doping levels, allowing for a completely depleted. This ensures a full collection of the charges created in the sensitive volume.

[0080]

[0081] Furthermore, each contact island can have a doping level between 10¹⁰ and 10¹⁸ cm³, ideally 10¹⁶ cm³. Preferably, the doping level of the islands is different from the doping level of the sensitive volume. This can help to obtain an inhomogeneous electric field in the sensitive volume.

[0082]

[0083] The detector can include as many front electrodes as there are front contact islands.

[0084]

[0085] According to another aspect of the invention, a system for detecting and discriminating particles is proposed. This system comprises:

[0086] - a particle detector as described above,

[0087] - a processing unit configured to compare one or more parameters of a transient signal acquired with one or more predetermined parameters in order to identify the nature of the particle at the origin of this acquired transient signal.

[0088]

[0089] A method for detecting and discriminating particles is also provided. Advantageously, the method comprises the following steps implemented by a processing unit of the above system:

[0090] - application of one or more voltages to one or more front electrodes of the particle detector such that a non-homogeneous electric field is created within the sensitive volume,

[0091] - acquisition of a transient signal on said one or more electrodes before, following the passage of a particle through the sensitive volume,

[0092] - comparison of one or more parameters of the transient signal to one or more predetermined parameters in order to identify the nature of the particle at the origin of this acquired transient signal.

[0093]

[0094] The parameters can be characteristics of the measured signals: slope, amplitude, derivative, mean, area, ...

[0095] Description of figures and embodiments.

[0096] Other advantages and features of the invention will become apparent upon reading the detailed description of implementations and embodiments, which are by no means limiting, and the following accompanying drawings:

[0097] Figure 1 is a graph illustrating the linear energy transferred by 200keV electrons and 500MeV protons according to the prior art,

[0098] Figure 2 includes a map of the charge densities generated by 500 MeV protons (left) and a map of charge densities generated by 200 keV electrons (right),

[0099] Figure 3 is a schematic view of a typical SSD structure,

[0100] Figure 4 includes three examples of calculated electric field maps for a classic geometry (a) and for multi-electrode geometries (b and c),

[0101] Figure 5 includes four examples of virtual potential maps calculated for (a) a simple electrode made over the entire front surface of a particle detection diode, (b) a central electrode covering half the front surface of the diode, (c) a central electrode covering one-third of the front surface of the diode, and (d) for an annular electrode surrounding the electrode of example (c),

[0102] Figure 6 is a graphical representation illustrating the curves of the effect of the virtual electric field on the current collected by a central junction of variable width: 100% (a); 50% (b); and 30% (c) of the front surface of the diode,

[0103] Figure 7 is a schematic view of a system for detecting and discriminating particles,

[0104] Figure 8 is a schematic view of a semiconductor substrate according to the invention with front and rear measuring electrodes,

[0105] [Fig. 9] comprises a) a schematic top view of a semi- a) a conductor with a serpentine-shaped front electrode, and b) a schematic cross-sectional view of the same semiconductor substrate with the front electrode and a rear electrode,

[0106] Fig. 10 comprises a) a schematic top view of a semiconductor substrate with front electrodes in the shape of concentric circles, b) a schematic bottom view of the same semiconductor substrate with rear electrodes in the shape of concentric circles, and c) a schematic cross-sectional view of the same semiconductor substrate,

[0107] Fig. 11 comprises a) a schematic bottom view of a semiconductor substrate with two rear electrodes of square and annular shape respectively, and b) a schematic cross-sectional view of the same semiconductor substrate with an anode on the front surface and the rear electrodes on the rear surface,

[0108] Fig. 12 comprises a) a schematic top view of a semiconductor substrate with two front electrodes of circular and annular shapes respectively, and b) a schematic cross-sectional view of the same semiconductor substrate with the front electrodes and a rear cathode, and

[0109] The [Fig. 13] is a graphical representation illustrating the curves of the transient signals generated by the passage of a proton and an electron depositing the same amount of energy.

[0110] The embodiments described below are not in any way limiting; in particular, variants of the invention may be implemented comprising only a selection of the features described below, isolated from the other features described, if this selection of features is sufficient to confer a technical advantage or to differentiate the invention from the prior art. This selection includes at least one preferably functional feature without structural details, or with only a portion of the structural details if this portion alone is sufficient to confer a technical advantage or to differentiate the invention from the prior art.

[0111] The principle of the invention is to allow manipulation of the two terms governing the current collected by the detector during the passage of a particle, which are given by the Shockley-Ramo relation. To this end, the invention provides a detector comprising a semiconductor substrate and at least one measuring electrode formed on a surface of the substrate. The geometry of the electrode on the substrate is such that the electric field created in the sensitive volume of the substrate when the substrate is powered via the electrode is inhomogeneous. The electric field in the sensitive volume is modulated according to the shape of the electrode formed on the surface of the substrate.

[0112] By modulating the electric field within the sensitive volume of a particle detector, the transport of charges created within the sensitive volume is consequently modified. This allows, among other things, the charge carriers to be directed towards the measuring electrodes, and limits the loss of information. This modulation depends on the geometry of the detector and / or the voltage applied to the measuring electrodes.

[0113] Figure 4 shows examples of electric field maps calculated for different detector geometries. Figure 4(a) shows an example of the electric field in a conventional SSD structure such as that of Figure 3. Figures 4(b) and (c) show the electric field obtained from multi-electrode geometries, but with different dimensions of the measuring electrode and different voltage values ​​applied to the measuring electrode. It is therefore possible to manipulate the different geometric parameters to change the electric field and charge transport, and thus the shapes of the transient signals.

[0114]

[0115] By changing, in particular, the number of measuring electrodes and their dimensions, it is possible to influence the virtual field associated with each measuring electrode and to modify their individual response. Figure 5 shows some examples of virtual potential maps calculated for a single electrode spanning the entire width of the diode (a), a central electrode spanning half (b) or one-third (c) of the diode surface width, and for an annular electrode (d).

[0116] First, a significant difference can be observed between the uniform virtual field of the classical case (a), and the distorted fields in the case of the other tracks (b), (c), and (d), which will lead to very different signal shapes measured on the measuring electrodes according to the Shockley-Ramo relation. Indeed, the collected signal is directly proportional to the virtual electric field.

[0117] Figure 6 shows the effect of the virtual field on the signal generated by a proton of 1.56 MeV with an angle of incidence of 19° and injected into the center of the sensitive volume of a detector, in the case of the geometries in Figures 5a), 5b), and 5c), corresponding respectively to a simple junction representing 100%, 50%, or 30% of the total diode width (300 µm in this example). A junction is the contact area of ​​an electrode on the surface of the semiconductor substrate. It can be seen that the smaller the junction width, the weaker the generated signal. Furthermore, the amount of energy collected by the junction also decreases. With a full junction, 100%, [Fig. 5a], all the energy is collected (1.56 MeV). A 50% junction, [Fig. 5b], collects only 20% of the total charge, and a 30% junction [Fig. 5c] collects only 5% of the total charge.This is due to the fact that the angle of incidence of the proton is such that it passes by the partial junctions (not totally covering the surface of the diode), which reduces the proportion of charge recovered by the electrode.

[0118]

[0119] It is thus possible to see the effect of the detector's geometric parameters by comparing the signals of the geometries in Figures 5b) and 5c). In [Fig. 5c), the central measuring electrode is narrower than that in [Fig. 5b), which tends to reduce the proportion of charge collected by the electrode, as seen in the case of a proton with an angle of incidence. However, this central measuring electrode can be combined with an annular measuring electrode (see [Fig. 6d)) arranged around the central measuring electrode. The annular measuring electrode makes it possible to measure the signal induced on the sides of the diode. Measurement capability is found over almost the entire surface of the diode, but with disturbances in the electric field according to the invention.

[0120] In view of the significant difference between the potential maps 5 c) and 5 d), the response of the annular track of [Fig.5] d) will also be very different from that of the central electrode 5 c).

[0121] By manipulating the two parameters, the real electric field and the virtual electric field, it is thus possible to modify the shapes of the signals measured by the electrodes, which is then used to optimize electron / proton discrimination. Unlike prior art designs using guard rings to control the electric field (to keep it uniform in the sensitive volume) and separate electrodes for measuring the deposited charge, the electrodes according to the invention have a dual function: deformation of the electric field and measurement of the signal generated in the diode.

[0122]

[0123] Figure 7 shows an example of an embodiment of a system according to the invention. A semiconductor substrate, such as a diode 4, is shown, adapted to receive ionizing particles to be measured. Each ionizing particle interacts within the semiconductor substrate by transferring energy into the sensitive volume of the substrate. This energy creates electron-hole pairs, which are charge carriers. When an electric field is applied to the sensitive volume of the semiconductor substrate via the electrodes, the electrons and holes move in opposite directions. This movement of charge carriers results in the generation of an electrical signal across the electrodes.

[0124] This electrical signal is collected by an acquisition chain 5 which includes hardware and software components capable of processing this signal, in particular by digitizing it and recording it over time so as to preserve its dynamics.

[0125] A processing unit 6 retrieves the measured signal and applies a classification, correlation, comparison or deep learning technique to identify, based on the shape and dynamics of the signal, whether the particle at the origin of the signal is an electron 7 or a proton 8. The processing unit 6 is also capable of determining the NRJ energy generated by the ionizing particle at the origin of the measured signal.

[0126]

[0127] Figure 8 shows in greater detail an assembly according to the invention comprising a semiconductor substrate 9 on which are arranged front 10 and rear 11 measuring electrodes. Under each measuring electrode is an island 12, 13, that is to say, an area of ​​the semiconductor substrate doped inversely to the doping of the sensitive volume 14 forming the interior of the semiconductor substrate. Each island can have a width of a few tens to a few hundred micrometers. The doping of the islands can range from 1E+11 to 1E+14 # / cm³.

[0128]

[0129] At least one measuring electrode can be provided on one side of the semiconductor substrate to apply an electric field creation voltage and to measure the signal generated by an ionizing particle detected by the semiconductor substrate. The other side can be equipped with a grounded electrode, also a measuring electrode, or a floating electrode.

[0130] Different compatible configurations are possible provided that the electric field is made inhomogeneous in the semiconductor substrate. For example, when the semiconductor substrate is designed as a diode, one or more measuring electrodes can be made on the anode (front side) and / or on the cathode (rear side).

[0131] With, for example, a single measuring electrode 10 on [Fig. 8], the electric field is sufficiently perturbed in the sensitive volume to allow efficient discrimination of particles according to the measured signals. To further perturbed the electric field, several configurations are possible.

[0132] Figure 9 describes a very schematic embodiment in which a parallelepiped-shaped diode 15 with a rectangular upper and lower face is distinguished. The diode comprises a sensitive volume 16 in which particles passing through the diode deposit charges. The electric field in the sensitive volume is created by applying a voltage V to a measuring electrode 17. An island (not shown) with a shape substantially symmetrical to that of electrode 17 is formed within the sensitive volume 16, in contact with electrode 17. Electrode 17 is a spiral-shaped detection anode as illustrated in the top view of Figure 9 on the left. The spiral covers a large part of the front surface of the diode 15, thus enabling the collection of almost all the charges generated during the passage of a particle. The spiral shape has an inter-electrode spacing 19.The bias voltage V applied to electrode 17 generates, thanks to the inter-electrode spacing, an inhomogeneous electric field in the sensitive volume 16.

[0133] Unlike the design with guard rings, whose role was to uniformize the electric field, the use of a spiral shape here allows the electric field within the diode to be distorted. This has the effect of modifying the propagation of charge carriers and, according to the Shockley-Ramo relation, of modifying the shape of the measured signal. It may be advantageous, however, to use several measurement anodes or cathodes to also decompose the virtual collection field. Figure 10 describes such an implementation.

[0134] Figure 10 shows a circular diode 20, composed of a sensitive volume 21. Seven annular detection anodes 22 are distributed concentrically around a central anode 23. Seven annular detection cathodes 24 are distributed concentrically around a central cathode 25. A gap or inter-electrode spacing 26 is present between the two consecutive electrodes.

[0135] Each anode 22 can receive a bias voltage Vn. Each cathode can receive a bias voltage V'n, different from Vn. The bias voltages Vn and V'n are determined so as to create an inhomogeneous electric field in the sensitive volume 21. The voltage difference between two bias voltages applied to two adjacent rings can be between zero and 20 volts. This allows modification of their charge collection properties and control of the deformation of the electric field in the sensitive volume.

[0136] The use of multiple electrodes makes it possible to act on both terms of the Shockley-Ramo relation. Indeed, in addition to the distortion of the real electric field, each electrode possesses a specific virtual collection field and therefore its own signal. Compared to the previous spiral design, the total signal can thus be decomposed with each electrode, which makes it possible to obtain individual signals across the entire width of the detector and to reconstruct the shape of the particle's track.

[0137] To improve upon the example in [Fig. 10], a smaller number of measuring electrodes can be used. This would reduce the complexity of manufacturing and operation, particularly since each anode and cathode would need its own signal acquisition chain. In the present example in [Fig. 10], sixteen acquisition chains would therefore be required.

[0138]

[0139] The device of [Fig. 11] comprises a square diode 30, consisting of a sensitive volume 31 and a single anode 32. The cathode is divided into two square measuring electrodes: a first central electrode 33 and a second electrode 34 following the perimeter of the diode around the first electrode. Each electrode 33 or 34 can receive an individual bias voltage (respectively Vi or V2) in order to modify its charge-collecting properties. A gap or inter-electrode spacing 35 is present between electrodes 33 and 34. Electrode 34 can be located at the periphery of the diode or simply around electrode 33 without being at the edge of the diode surface.

[0140] The use of two measuring electrodes, unlike the sixteen in the example in [Fig. 10], simplifies the use of the diode and reduces the complexity of the electronics required for its operation. Since the device has several electrodes, it is always possible both to distort the electric field in the sensitive volume and to decompose the virtual collection field. Here, decomposing the cathode into two electrodes allows for modulation of the hole collection. However, since holes are less mobile than electrons, it is more difficult to make them converge towards a given measuring electrode, and there is a risk of charge loss in the band gap. This loss could be reduced by decomposing the anode instead of the cathode. In addition, the square geometry of the diode creates peak areas where the local electric field can be very strong.Therefore, an alternative geometry can be considered to limit the risk of breakdown.

[0141]

[0142] The device in [Fig. 12] comprises a circular diode 40, consisting of a sensitive volume 41 in which particles passing through the diode deposit charges. A central sensing anode 42 and a second sensing track 43, forming a ring around the first, are included at the end of the diode. A gap or inter-electrode spacing 44 is present between electrodes 42 and 43. The second sensing track is a measuring electrode located at the periphery (edge) or not of the diode surface. Each electrode 42 or 43 can receive an individual bias voltage (V1 or V2, respectively) to modify its charge-collecting properties.

[0143]

[0144] In order to optimize the device according to the different embodiments, the following properties can be used: - A diode thickness between 30 µm and 5 mm - A diode width between 30 µm and 5 cm, - An inter-electrode gap or spacing (between two electrodes or between two parts of the same electrode on the same surface), the width of which is between 30 pm and 1 mm, - A voltage difference of 0 to 20 V between two electrodes on the same surface, - A doping level of the semiconductor substrate between 1E+10 and 1E+14 # / cm3, - A doping level of the islets associated with the electrodes between 1E+10 and 1E+18 # / cm3.

[0145] An example of optimal values ​​for the geometry of [Fig. 12] can be given below: - A diode thickness of 300 µm, - A diode width of 1.5 mm, - A gap or inter-electrode spacing of 200 pm between the two electrodes 42 and 43, - A voltage difference of 0 V between the two electrodes 42 and 43, - A doping level of the semiconductor substrate of 1E+11 # / cm3, - A doping level of the islets associated with the electrodes of 1E+16 # / cm3.

[0146]

[0147] In the example of [Fig. 12], the leakage current can be kept around nA, ten times lower than when the optimization according to the invention is not applied. The leakage current can be on the order of 1–10 nA. Although higher, this current remains on the order of 0.1% to 1% of the amplitude of the signal generated by a particle passing through the detector (pA). Thus, the properties of the detector are not significantly degraded by the leakage current. The intensity of the leakage current depends on the size of the gap, as well as the potential difference between the two electrodes. If the difference The potential difference between the electrodes is too high, and if the gap is sufficiently narrow, the resulting electric field in this area can be strong enough to disrupt the formation of the depletion region between the anodes and the cathode and generate a significant leakage current.

[0148]

[0149] The advantage of the optimized geometry according to the invention compared to the conventional geometry of SSDs is evident when comparing their particle identification performance. The confusion matrices calculated in both cases (without and with optimization) are presented below, for different energy ranges. The values ​​are given for incident protons between 100 and 250 MeV and incident electrons between 0.1 and 7 MeV, which correspond to the ranges where the total energy deposited by the electrons and protons is very close. The prediction results are classified by total energy deposited in the detector by the particles according to three ranges: Table 1: 0-150 keV, Table 2: 150-300 keV, and Table 3: 300-450 keV.

[0150] Confusion matrices allow the accuracy of the prediction of the particle's nature to be quantified. They show the proportion of electrons and protons correctly identified.

[0151] In the example of the first matrix (a) for the 0-150 keV range (Table 1), the method used in this case correctly identifies 47% of the protons, while 53% of the protons are identified as electrons. In the second row of the matrix, it can be seen that 32% of the electrons are identified as protons, and 68% of the electrons are correctly identified.

[0152] [Tables 1] 0-150 keV (a) Protons (prediction) Electrons (prediction) 0-150 keV (b) Protons Electrons Protons (True) 47 53 Protons 85 15 Electrons (True) 32 68 Electrons 22 77

[0153]

[0154] The confusion matrices for the conventional single-track detection design (a) are obtained solely from the total amount of energy deposited by the particles, which is the only accessible information in this case. The confusion matrices allow comparison of the percentages of correctly identified particles (columns) with the true nature of the particles (rows). These matrices demonstrate a significant improvement in predicting the nature of the particle by analyzing the shape of transient signals.

[0155] The confusion matrices for the optimized design according to the invention with multiple detection tracks are obtained using statistical processing of the shapes of the transient signals induced by the particles. In this processing, the characteristic quantities of the signals (maximum amplitude, rise time, full width at half maximum, integral, proportion of the total charge deposited on the ring, etc.) are first studied by comparing their distributions for electrons and protons. By calculating the overlap of these distributions, the sorting criteria for which the electron-proton separation is most pronounced can be selected, and the probability that a signal is associated with one or the other particle can be calculated. For example, from the distributions of the maximum amplitudes of the electron and proton signals, ranges of values ​​where protons are more prevalent than electrons, and vice versa, can be identified.Next, when a signal from an unidentified particle is measured, its maximum amplitude can be extracted and placed back into the previously established distribution, giving us the probability of having measured a proton. There are, of course, ranges of values ​​where the maximum amplitudes of the electron and proton signals are identical, and where it is impossible to determine the identity of the measured signal. In such cases, the discrimination must be refined using other criteria. It is possible to further improve the efficiency of the discrimination by using machine learning methods, which automate the logical path to follow to discriminate a measured signal.

[0156] [Tables2] 150-300 keV (a) Protons Electrons 150-300 keV (b) Protons Electrons Protons 61 39 Protons 78 22 Electrons 26 74 Electrons 20 80 [Tables 3] 300-450 keV (a) Protons Electrons 300-450 keV (b) Protons Electrons Protons 57 43 Protons 91 9 Electrons 35 65 Electrons 23 77

[0157] It is also possible to demonstrate the contribution of the present invention by examining more closely the different forms of transient signals generated by particles depositing the same amount of energy in the sensitive volume of the diode. Figure 13 shows examples of transient signals calculated for the passage of a 240 MeV proton and a 1 MeV electron in a sensitive volume 300 pm thick, having deposited the same amount of energy (200 keV). Current conventional detectors only allow measurement of the total amount of energy deposited in the sensitive volume, making differentiation impossible with this type of detector in the example presented here. In [Fig. 13], it is possible to compare the transient signal that would be obtained with a conventional geometry ([Fig. 3]) with the multiple signals obtained using the multi-track geometry of the present example of [Fig. 12].

[0158]

[0159] The solid curves show the transient signal obtained with the geometry of a conventional detector shown in [Fig. 3], namely a single measurement track spanning the entire width of the diode. The shapes of these signals remain similar, and it is difficult to distinguish the proton signal (thin line) from the electron signal (thick line). This shows that, even using a conventional diode geometry connected to an acquisition chain capable of obtaining transient signal shapes (instead of an acquisition chain that only returns the integral of this signal), electron / proton separation remains complex due to the similarity of the transient signals.

[0160] The advantage of the optimized geometry of the invention is that it allows the total signal to be decomposed, and the shape of the signals generated on each track to be obtained. In the case of [Fig. 13], the annular track has been divided in two to provide three measurement tracks (left, center, and right), which makes it possible to determine the direction from which the particle passes. If this information is not required, the left and right tracks can be combined into a single annular electrode. This has the advantage of reducing the number of acquisition chains required (one per electrode).

[0161] The shapes of the signals generated on each individual track are very different between the proton and the electron. The signal generated at the center by the proton is wider than that of the electron, while the signal generated on the left by the electron is not symmetrical with the signal generated on the right, unlike the case of the proton. This is due to the fact that the path of electrons is much more chaotic than the straight path of protons. The present example shows that this difference in behavior between electrons and protons has a more pronounced effect in signals measured with an optimized geometry than with a standard geometry, which would make it easier to identify trends specific to each type of particle. In this particular example, it is easy to differentiate the electron from the proton using only the central measuring electrode, since the proton passed straight through the center of the diode while the electron was deflected.In more realistic situations, where particles arrive at an angle of incidence and without . because it necessarily enters through the center of the diode, the electron-proton distinction is no longer as easy and it becomes necessary to cross-reference the information from several electrodes.

[0162]

[0163] With the invention, it is possible to influence the shapes of transient signals using optimized geometry by modifying the electric field in the detector ([Fig. 4]) or the virtual field of each measurement track ([Fig. 6]). This makes it possible to optimize the detector dimensions (sensitive volume, track size, etc.) as well as the applied voltages, in order to distort the signals and further highlight the transport differences between electrons and protons, as seen, for example, in [Fig. 8].

[0164] In the case of the geometry used in the invention, the optimal design is the one that allows for the most marked separation possible between the shapes of the signals generated by electrons and protons, while maintaining a leakage current low enough not to degrade the performance of the detector.

[0165]

[0166] Of course, the invention is not limited to the examples just described. Many modifications can be made to these examples without departing from the scope of the present invention as described.

Claims

Demands

1. 1. Particle detector comprising: - a semiconductor substrate including a sensitive volume capable of producing charges during the passage of a particle, at least one front contact island made under a front face of the semiconductor substrate, and at least one rear contact island made under a rear face of the semiconductor substrate, - at least one front electrode made on the front face of the semiconductor substrate, in contact with the front contact island; - at least one rear electrode made on all or part of the rear face of the semiconductor substrate, in contact with the rear contact island;- an acquisition chain capable of detecting transient signals from said at least one front electrode, characterized in that said at least one front electrode is a voltage application and transient signal detection electrode, and in that said at least one front electrode is made on only a part of the front face of the semiconductor substrate, without guard electrodes, so that a non-homogeneous electric field is created in the sensitive volume when said at least one front electrode is supplied with voltage.;

2. 2. Detector according to claim 1, characterized in that said at least one front electrode has a geometric shape having at least one inter-electrode spacing suitable for creating a non-homogeneous electric field in the sensitive volume.

3. 3. Detector according to claim 2, characterized in that said at least one front electrode is a spiral covering at least a part of the surface of the sensitive volume.

4. 4. Detector according to claim 1 or 2, characterized in that the semiconductor substrate is circular, square, or rectangular in shape and: - said at least one front electrode consists of a plurality of annular front electrodes distributed concentrically around a central front electrode, an inter-electrode spacing being present between two front electrodes, and - said at least one rear electrode consists respectively of a plurality of annular rear electrodes distributed concentrically concentric around a central rear electrode; each front-rear electrode pair being capable of receiving a bias voltage different from that intended to be applied to the neighboring front-rear electrode pair, an inter-electrode spacing being present between two rear electrodes.

5. 5. Detector according to claim 1 or 2, characterized in that the semiconductor substrate is circular, square or rectangular in shape, and: - said at least one front electrode comprises a first central front electrode of circular, square or rectangular shape and at least one second front electrode, peripheral or not, of circular, square or rectangular shape; an inter-electrode spacing being present between the two front electrodes; and - said at least one rear electrode consists of a single rear electrode distributed over the entire rear face of the semiconductor substrate.

6. 6. Detector according to any one of the preceding claims, characterized in that, where the semiconductor substrate is a diode structure, said at least one front electrode is an anode or a cathode; said at least one rear electrode being respectively a cathode or an anode.

7. 7. Detector according to any one of the preceding claims, characterized in that the semiconductor substrate has a thickness of between 30 pm and 5 mm, ideally 300 pm.

8. 8. Detector according to any one of the preceding claims, characterized in that the semiconductor substrate has a width between 30 pm and 5 cm, ideally 1.5 mm.

9. 9. Detector according to any one of claims 2 to 8, characterized in that the inter-electrode spacing has a width between 30 pm and 1 mm, ideally 200 pm.

10. 10. Detector according to any one of the preceding claims, characterized in that when this detector comprises several front electrodes, the acquisition chain is configured to apply voltages respectively to the front electrodes with a voltage difference between two consecutive front electrodes of between 0 and 20 V, ideally 0 V.

11. 11. Detector according to any one of the preceding claims, characterized in that the sensitive volume has a doping level between 1010 and 1014 cm3, ideally 1011 cm3.

12. 12. Detector according to any one of the preceding claims, characterized in that each contact island has a doping level between 1010 and 1018 cm3, ideally 1016 cm3.

13. 13. Detector according to any one of the preceding claims, characterized in that it comprises as many front electrodes as front contact islands.

14. 14. System for detecting and discriminating particles, characterized in that it comprises: - a particle detector according to any one of the preceding claims, - a processing unit configured to compare one or more parameters of an acquired transient signal with one or more predetermined parameters so as to identify the nature of the particle at the origin of this acquired transient signal.

15. 15. A method for detecting and discriminating particles, characterized in that it comprises the following steps implemented by a processing unit of a system according to claim 14: - application of one or more voltages on one or more front electrodes of the particle detector so that a non-homogeneous electric field is created in the sensitive volume, - acquisition of a transient signal on said one or said several front electrodes following the passage of a particle through the sensitive volume, - comparison of one or more parameters of the transient signal to one or more predetermined parameters so as to identify the nature of the particle at the origin of this acquired transient signal.