WO2022219515A1

WO2022219515A1 - A diode radiation sensor

Info

Publication number: WO2022219515A1
Application number: PCT/IB2022/053407
Authority: WO
Inventors: Alberto Gola; Giacomo BORGHI; Maurizio Boscardin; Giovanni PATERNOSTER
Original assignee: Fondazione Bruno Kessler
Priority date: 2021-04-14
Filing date: 2022-04-12
Publication date: 2022-10-20
Also published as: EP4324032A1; JP2024515298A

Abstract

A diode radiation sensor having one or more charge multiplication diodes (2) and comprising: a substrate (3; 303) having a front surface (4) and a rear surface (5); a first layer of semiconductor material (8) doped with a doping of a first type and made on said front surface (4); a second layer of semiconductor material (9; 109; 209; 309) doped with a doping of a second type of electrically opposite sign to the first type and made to a first depth so as to make between the first layer (8) and the second layer (9; 109; 209; 309) a high electric field region (10); an isolation region (15; 315) made peripherally to the substrate (3; 303) and extending in depth to an intermediate area between the front surface (4) and the rear surface (5); a locking element (18; 118; 218; 318; 418) made in depth in the substrate (3; 303) to electrically isolate the first layer (8) and to hinder charge collection paths between the substrate (3; 303) and the front surface (4) extending laterally to the second layer (9; 109; 209; 309).

Description

A DIODE RADIATION SENSOR

DESCRIPTION

Field of application

The present invention can be applied to the field of diode sensors and, in particular, to the field of radiation sensors.

More in detail, the present invention relates to a diode radiation sensor having one or more diodes with charge multiplication structure powered so as to work in linear multiplication regime.

Background art

Radiation sensors are used in a wide variety of applications, ranging from industrial to scientific. In many cases, the detector is constructed on a single body of semiconductor material, for example silicon, divided into several microcells or channels (also called pixels), each typically consisting of an individually accessible diode.

A typical example of an ionizing radiation detector consists of a silicon microstrip with a typical thickness of a few hundred pm. Such a device is used for detecting ionizing radiation (such as charged particles or X-rays) in scientific experiments and for industrial applications.

The active area of the detector is divided into several parallel strips, with a width usually between 25 pm and a few hundred micrometres.

As a first approximation, the above sensors do not have an internal gain and thus have a limit if the amount of charge created by the ionizing radiation is too low to be measured accurately.

To improve performance, it was thus decided to introduce a charge multiplication structure inside the diode which allows the diodes to operate, when appropriately polarized, in linear charge multiplication regime, which means that the charge collected at the output of each channel is proportional to the charge generated by the radiation which interacts with the sensor.

In some cases, such as in the case of measuring the interaction time of charged particles at the minimum ionization in high-energy physics experiments, the charge amplification provided by the diode must be sufficient to allow the detection of the radiation and to obtain an optimal working point considering the ratio between the signal and the noise, but not excessive to avoid impairing the accuracy of the measurement due to the worsening of the signal/noise ratio due to the excess noise determined by the charge multiplication process. For this reason, the diodes used are usually powered so that the gain is particularly limited and with values in the range of 10 to 20. In such a case LGADs are spoken of, i.e., Low-Gain Avalanche Diodes.

Given the type of function which LGADs have and the type of radiation they must measure, the sensors implementing them have active thicknesses which typically range from a few tens to a few hundreds of pm.

Generally, as shown in fig. 1 , a microcell M of an LGAD is made in a substrate S of semiconductor material as said of a thickness from a few tens to a few hundreds of pm. In such a substrate S, a first layer P1 of doped semiconductor material is identified with doping of a first type (which can be of type n or of type p indifferently, in any case of the sign opposite the doping of the substrate S). Such a first layer P1 is made on the front surface of the substrate S.

Then there is a second layer P2 of doped semiconductor material of opposite sign to the first layer P1 and made in depth in the substrate S. The latter is generally doped with the same sign of the second layer P2, but in smaller amounts. There is then a bottom layer PF doped with the same sign as the second layer P2 and the substrate S, but typically in higher amounts of the latter. The diode supply typically, but not necessarily, is between the first layer P1 and the bottom layer PF and, in accordance with the supply, between the first P1 and the second layer P2 or between the first P1 and the bottom layer PF an emptying region is created. In any case, between the first layer P1 and the second layer P2, in the presence of an appropriate polarization, a region with a high electric field E is created for generating the multiplier effect of the diode charge.

Each microcell M must be electrically isolated from its neighbours in order to function properly as envisaged. In this sense, there is usually a third layer P3 called ‘p-stop’ or ‘n-stop’, i.e., an area with doping sign opposite that of the first layer P1 and which has a depth in the substrate S comparable to that of the first layer P1, although typically greater. By design setting, the magnitude of the electric field is constant between the first P1 and the second layer P2. Moreover, it is important that such an electric field uniformly decreases at the edges, from its peak value at the centre to a significantly lower value below the third layer P3, without any high electric field peaks outside the volume comprised between the first layer P1 and the second layer P2. The main reasons for this requirement are to avoid a dielectric breakage (also called electrical breakdown) of the edge between the first layer P1 and the third layer P3 under operating conditions. Such a phenomenon would comprise the device functionality.

To achieve this goal, among others, the technique of terminating the edges of the first layer P1 with what is typically referred to as a Junction Termination Extension (JTE) is known, which consists of a fourth layer P4 having doping of the same type as the first layer P1 and deeper in the substrate S with respect to the first layer P1. Typically, the fourth layer P4 has a lower doping dose with respect to the first layer P1. This allows the electrical potential to be shaped at the edges, decreasing the electric field in a controlled manner. Furthermore, the width of the high electric field region E is optimized, typically reduced with respect to the first layer P1 so as to further reduce the electric field at the edge. Such a structure is usually referred to as a ‘virtual guard ring’ VGR.

As can be seen in the figure, however, it is necessary that between the first layer P1, or the fourth layer P4, and the third layer P3 there is a sufficient distance to avoid a high electric field between the two.

Although effective, the expedients described above have some drawbacks.

A first drawback is that the minimum distance between the second layer P2 and the fourth layer P4 as well as the minimum distance between the latter and the third layer P3, in fact form a dead edge, i.e., a peripheral area to the nominal active area at full gain of the diode (defined by the second layer P2) in which the electrical charges generated by the radiation are collected according to charge collection paths which do not pass through the high electric field region of the avalanche diode, since they are not in fact subject to linear multiplication. This forms a loss of efficiency of the radiation sensor.

A second drawback consists of the fact that a part of the charges generated under the second layer P2 moves according to charge collection paths which lead to the front surface of the substrate S without passing through the high electric field region of the diode, i.e., without multiplication, effectively increasing the dead edge.

In other words, the size of the actual dead edge of each microcell is thus determined both by the design rules (the minimum distances above) and by the additional effect of charge collection paths which do not pass through the high electric field region of the diode. This introduces sensitivity losses of the LGAD radiation sensor (i.e., only a part of the radiation sensor area is actually sensitive) and, for example in the case of a generation of X-ray incidence charges (affecting substantially point-like areas), impaired spectroscopic performance due to an uneven gain in the sensor area.

It is clear that the larger the dead edge, the greater the negative effects just mentioned.

Presentation of the invention

An object of the present invention is to at least partially overcome the drawbacks noted above, providing a radiation sensor having improved performance with respect to the equivalent sensors with regard to the dead edge effects.

In particular, an object of the present invention is to provide a radiation sensor whose microcells have, in their active area, as uniform a gain as possible.

Another object of the present invention is to provide a radiation sensor whose microcells have areas with less or no charge multiplication of smaller size, if not null, with respect to the equivalent radiation sensors of the prior art.

In particular, therefore, an object of the present invention is to provide a radiation sensor whose microcells have smaller, if not null, dead edges with respect to the equivalent known sensors.

Such objects, as well as others which will become clearer below, are achieved by a diode radiation sensor in accordance with the following claims, which are to be considered as an integral part of the present disclosure.

In particular, it includes one or more typically polarized charge multiplication diodes for working in the linear multiplication area. In such a sense, the sensor comprises a substrate made of semiconductor material (whose depth is typically at least 20 pm) and having a front surface and a rear surface opposite the front surface.

At least near the front surface of the substrate, at least a first layer of doped semiconductor material is made with a doping of a first type so as to cover at least a first central area of said front surface.

There is then a second layer of semiconductor material doped with a doping of a second type of electrically opposite sign to the first type and made at a first depth in the substrate. The second layer extends substantially parallel to the first layer so as to affect a second area such as to identify a high electric field region between the two layers. In other words, the two layers create the charge multiplication structure of a diode of such a type whose working area, and therefore whose charge multiplication level, will be determined by the power supply of the same diode.

According to an aspect of the invention, each charge multiplication diode comprises at least one isolation region made peripherally to the substrate and extending in depth from the front surface to an intermediate area between it and the rear surface. Thereby it is arranged laterally at least at the first and second layers.

It follows, advantageously, that there is an electrical, and typically also optical, shielding, at least partial between the diodes forming the sensor. In this sense, still advantageously, the doped semiconductor layer called p-stop (or n- stop) and the termination semiconductor layer present in the known equivalent radiation sensors are no longer necessary.

As a result, advantageously, there are no more problems of distances between the different layers which forced limiting the extensions of the first and second layers.

Advantageously, therefore, the dead edge of the single microcell is strongly decreased, resulting in an increase in sensor efficiency. Moreover, this makes allows to substantially make the electric field between the first and the second layer uniform, eliminating distortion factors of the amplification provided by the charge multiplication diode.

According to another aspect of the invention, the substrate also comprises at least one locking element made in depth in the substrate near the surface of the isolation region. Such a locking element is a region in contact with the isolation region which is not emptied during the normal operation of the device, i.e., which under operating conditions contains a sufficiently high concentration of majority carriers of the same type as those of the substrate. Typically, but not necessarily, this locking element is made of semiconductor material doped with a doping of the second type. Equivalently, according to different embodiment variants, it can be obtained naturally in accordance with special components inserted in the isolation region and with the type of substrate doping, for example by inserting a dielectric in the isolation area having a fixed charge such as to induce on the surface of the isolation region a free charge accumulation of the type of the majority carriers which are present in the substrate. In any case, the locking element has the function of electrically isolating the first layer, interrupting the conductive paths between layers of the first type which can be created along the surface of the isolation region. Furthermore, if suitably shaped, the locking element also serves to obstruct the charge collection paths between the substrate and the front surface which extend laterally to the second layer and to the high electric field region, avoiding to cross it.

In other words, a locking element is also inserted in the substrate which isolates the first layer and focuses the charge collection paths in the direction of the high electric field region. In other words, the locking element performs a substantially funnel function for the charges generated in the substrate which are thus directed towards the high electric field region.

Advantageously, therefore, the charges generated in the substrate have a higher probability of crossing such a high electric field region, thus obtaining the desired charge multiplication effect.

Still advantageously, the radiation response of the sensor will thus be more uniform with respect to the equivalent known sensors since the edge area with null or reduced gain will be greatly reduced.

Brief description of the drawings

Further features and advantages of the invention will become more evident in light of the detailed description of a preferred but non-exclusive embodiment of a radiation sensor according to the invention, illustrated by way of non-limiting example with the aid of the accompanying drawings, in which:

FIG. 1 depicts a radiation sensor according to the state of the art in schematic view;

FIG. 2 depicts a radiation sensor according to the invention in schematic view;

FIGS. 3 to 6 depict embodiment variants of the sensor of FIG. 2.

Detailed description of an exemplary preferred embodiment

With reference to the above-mentioned figures, and in particular fig. 2, a diode radiation sensor 1 is described having one or more charge multiplication diodes 2 which will be polarized so as to work in a linear multiplication area. For ease of description, the sensor 1 depicted in the figures comprises a single diode 2, but it is evident that such an aspect must not be considered limiting for the present invention.

Therefore, the sensor 1 comprises a substrate 3 made of semiconductor material and having two surfaces, a front surface 4 and a rear surface 5 opposite the front surface 4. Such a substrate, given the use within the aforementioned LGAD, has a typically high depth and in the order of a few hundred pm or, typically, of at least 20 pm.

On the front surface 4 there is a first layer 8 of semiconductor material doped with a first type of doping. In the figures, such a doping is of type n, but also this aspect must not be considered limiting for the present invention. In fact, the reversal of the types of doping cited in the present description does not make any difference for the purposes of the present patent.

The thickness of the first layer 8 can also be any in accordance with the design parameters of the sensor 1. In general, it is specified that regardless of what can be deduced from the figures, the thicknesses of all the layers indicated in the present patent will be in accordance with the design parameters of the radiation sensor without any limitations for the invention.

The position of the first layer 8 on the front surface of the substrate 3 is also a feature to be considered non-limiting for the invention, since there are embodiment variants not depicted here where the same first layer is made deep in the substrate (although near the front surface) and connected with the front surface by an electrical contact. The first layer 8 is made to cover a first central area of the front surface 4 of the substrate.

There is also a second layer 9 of semiconductor material doped with a doping of a second type of electrically opposite sign to the first type. In such a sense, the doping is of type p, but as mentioned, this aspect must not be considered limiting for the present invention, the inversion of the types of doping does not involve any difference for the purposes of the present patent.

The second layer 9 is made at a first depth in the substrate 3 and extends substantially parallel to the first layer 8 so as to affect a second area. Furthermore, between the second layer 9 and the first layer 8, in the presence of an appropriate polarization of the device, a high electric field region 10 is identified for generating the charge multiplication effect.

Typically, also the substrate 2 is doped with a doping of the second type, but with a lower doping level than that of the second layer 9. Similarly typically, on the rear surface 5 of the substrate there is a further doped layer with the second type of doping, generally at higher doping with respect to that of the substrate. Although it is observed in the figure that such a further doped layer covers the entire rear surface 5, such an aspect must not be considered limiting for different embodiments of the invention where the further doped layer covers only a portion of the rear surface of the substrate.

According to an aspect of the invention, the radiation sensor 1 comprises an isolation region 15 made peripherally to the diode 2 and extending deep in the substrate 3 from the front surface 4 to an intermediate area between it and the rear surface 5. In particular, the figure shows that said first isolation region 15 appears to be arranged laterally at least at the first 8 and the second layer 9. However, such an aspect must not be considered limiting for different embodiments of the invention where, for example, the isolation region is deeper.

Typically, the isolation region 15 is made by etching the substrate 3 and inserting one or more materials into the groove thus obtained, at least one of which is isolating (typically oxide of the semiconductor material of which the substrate 3 itself is formed), but also this aspect must not be considered limiting for the present invention. Typically, but not necessarily, the semiconductor material is silicon and the oxide is thus silicon.

In any case, advantageously, the function of the isolation region 15 is of electrical, and typically also optical, shielding between the charge multiplication diodes 2 forming the sensor 1. In this sense, advantageously, the layer called p- stop (or n-stop in the case of doping inversion) and the termination layer present in the known equivalent radiation sensors are no longer necessary.

Advantageously, therefore, there are no longer problems of distances between layers and doped portions which forced the extensions of the first 8 and the second layer 9 to be limited.

Advantageously, therefore, the dead edge of the single microcell is strongly decreased, resulting in an increase in the efficiency of the sensor 1. Moreover, this allows to improve the behavioural uniformity of the LGAD device by reducing distortion factors due to dead edges.

Moreover, still advantageously, the isolation region 15, which in fact forms a trench, does not affect both surfaces, but only the front surface 4. This is particularly advantageous in the case of radiation sensors 1 which are intended to be illuminated on the rear surface 5 since the latter is continuous and not affected and thus there are no elements which can influence the correct incidence of the radiation.

According to another aspect of the invention, the substrate 3 further comprises a locking element 18. In the figure it is observed that it is made of a semiconductor material doped with a doping of the second type. However, according to different embodiment variants it can be obtained naturally in accordance with special components inserted in the isolation region and with the type of substrate doping.

In any case, the locking element is positioned deep in the substrate 3 to electrically isolate the first layer 8, interrupting the conductive paths between doped layers of the first type which can be created along the surface of the isolation region. Furthermore, if appropriately shaped, the locking element also serves to obstruct charge collection paths between said substrate 3 and the front surface 4 extending laterally to the second layer 8 avoiding to cross it (and therefore avoiding to cross the high electric field region 10).

In other words, a locking element 18 is also inserted in the substrate 3 which isolates the first layer 8 and focuses the conductive paths in the direction of the high electric field region 10. Advantageously, therefore, the charges generated in the substrate 3 are more likely to cross the high electric field region 10, obtaining the desired charge multiplication effect.

Still advantageously, the radiation response of the sensor 1 will thus be more uniform with respect to the equivalent known sensors since the edge area with null or reduced gain will be greatly reduced.

In the figures it can be observed that the locking element 18 consists of two technologically distinct details. Firstly, the locking element 18 comprises the second layer 9 which, for this purpose, extends over the entire width of the substrate 3 so as to be in contact with the isolation region 15 along the entire perimeter of the substrate 3 itself. Advantageously, therefore, the focusing effect of the charges is inevitable.

However, to increase the conveying to the central portion of the high electric field region 10, the locking element 18 also comprises a third layer 20 of semiconductor material doped with a doping of a second type and made peripherally to the substrate 3 as well as below and in contact with the isolation region 15.

In substance, advantageously, the third layer 20 substantially forms a frame for the charge multiplication diode 2 and its doping obtains a device for focusing the charges towards the charge multiplication area of the diode 2 itself.

However, it is evident that such an embodiment of the invention must not be considered limiting for different embodiment variants which still fall within the scope of protection of the present patent.

In particular, according to a possible embodiment variant shown in fig. 3, in the sensor 100 the locking element 118 consists only of the second layer 109.

According to a further embodiment variant shown in fig. 4, the locking element 218 in the sensor 200 consists only of the third layer 220 which conveys the charges generated towards the central area of the second layer

209.

In all the embodiments described so far it is observed that the isolation region 15 has a first end 25 at the front surface 4 of the substrate 3 and a second end 26, opposite the first end 25, located deep in the substrate 3. Such a second end 26 can assume a polarization especially at the edges.

According to another embodiment variant of the sensor 300, depicted in fig. 5, accordingly, the locking element 318 comprises: the second layer 309; the third layer 320; a fourth layer 328 of semiconductor material doped with a doping of the second type and interposed between the isolation region 315 and the substrate 303 for a stretch near the second end 326.

The fourth layer 328 advantageously allows to passivate at least an end stretch of the isolation region 315 contributing to focusing the charges towards the charge multiplication area of the diode 302 and avoiding the formation of parasitic electric fields.

Such a conformation can also be used with the embodiment variant of fig. 4, giving rise to a further embodiment variant depicted in fig. 6 where the locking element 418 in the sensor 400 only comprises the third layer 420 and the fourth layer 428.

For what has been said so far, it is clear that it is possible to provide further embodiment variants aimed at creating the locking element in different manners.

In any case, observing the embodiment variants highlighted above, it is observed that they have a further common feature. In fact, in all the variants described so far, at least one of the first layer of semiconductor material and the second layer of semiconductor material is spaced from the isolation regions in order to identify specific substrate areas in which the structure indicated above can be defined with the name ‘virtual guard ring’.

Returning to the main embodiment, it is observed that the diode 2 also comprises a fifth layer 30 of semiconductor material doped with a doping of the first type and made on the front surface 4 of the substrate 3 above the first layer 8.

In particular, the doping of the fifth layer 30 is greater than the doping of the first layer 8. Thereby, substantially, the first layer 8 is patterned, i.e., it has a doping graduality which advantageously allows to model the electric fields which involve it especially at the edges with this, moreover, increasing the isolating effect of the virtual guard ring. Potentially, therefore, such a virtual guard ring could be reduced in extension.

However, such a detail should not be considered a limiting detail for the invention. In fact, in all cases in which the electrical power supply of the diode does not have high values or in which the isolation region is sufficiently wide, the risk of electric arcs is substantially cancelled, allowing to obtain embodiment variants in which the first layer extends over a first area substantially covering the entire front surface of the substrate until contact with the isolation region. Two examples of such embodiment variants are depicted in figs. 4 and 6. Obviously, such an arrangement is independent of the embodiment of the locking element.

The position of the fifth layer 30 on the front surface of the substrate 3 is also a feature to be considered non-limiting for the invention, since there are embodiment variants not depicted here where the same fifth layer is made deep in the substrate (although near the front surface and in any case at least partially interposed between the front surface and the first layer) and connected with the front surface by an electrical contact. In other embodiment variants, however, the same fifth layer is shaped and comprises the aforesaid electrical contact.

In light of the foregoing, it is understood that the radiation sensor of the invention achieves all the preset objects.

In particular, it has improved performance with respect to equivalent sensors in terms of dead edge effects. In fact, not only do the microcells have, in their active area, a uniform gain, but in fact they do not comprise, if not in a very limited manner, areas with little or no charge multiplication.

The invention might be subject to many changes and variants, which are all included in the appended claims. Moreover, all the details may furthermore be replaced by other technically equivalent elements, and the materials may be different depending on the needs, without departing from the protection scope of the invention defined by the appended claims.

Claims

1. A diode radiation sensor having one or more charge multiplication diodes (2), said radiation sensor (1; 100; 200; 300; 400) comprising: a substrate (3; 303) made of semiconductor material, having a front surface (4) and a rear surface (5) opposite said front surface (4); at least one first layer of semiconductor material (8) doped with a doping of a first type and made at least near said front surface (4) of said substrate (3; 303) so as to cover at least a first central area of said front surface (4) of said substrate (3; 303); at least one second layer of semiconductor material (9; 109; 209; 309) doped with a doping of a second type of electrically opposite sign to said first type and made to a first depth in said substrate (3; 303), said second layer (9; 109; 209; 309) being substantially parallel to said first layer (8) so as to affect a second area and so as to make between said first layer (8) and said second layer (9; 109; 209; 309), with a polarization of said sensor (1; 100; 200; 300; 400), a high electric field region (10); at least one isolation region (15; 315) made peripherally to said substrate (3; 303) and extending deep in said substrate (3; 303) from said front surface (4) to an intermediate area between said front surface (4) and said rear surface (5) so as to be arranged laterally at least to said first (8) and second layer (9; 109; 209; 309); at least one locking element (18; 118; 218; 318; 418) made deep in said substrate (3; 303) to electrically isolate said first layer (8) and to hinder charge collection paths between said substrate (3; 303) and said front surface (4) extending laterally to said second layer (9; 109; 209; 309) avoiding crossing said high electric field region (10).

2. Radiation sensor according to claim 1 , wherein said substrate (3; 303) is at least 20miti deep.

3. Radiation sensor according to claim 1 or 2, wherein said locking element (18; 118; 218; 318; 418) is made of semiconductor material doped with a doping of said second type

4. Radiation sensor according to any of the preceding claims, wherein said locking element (18; 118;318) comprises said second layer (9; 109; 209; 309) extending along the entire width of said substrate (3; 303) so as to be in contact with said isolation region (15; 315) along the entire perimeter of said substrate.

5. Radiation sensor according to one or more of the preceding claims, wherein said locking element (18; 218; 318; 418) comprises at least one third layer of semiconductor material (20; 220; 320; 420) doped with a doping of said second type and made peripherally to said substrate (3; 303) as well as below and in contact with said isolation region (15).

6. Radiation sensor according to one or more of the preceding claims, wherein said isolation region has a first end at said front surface of said substrate (303) and a second end opposite said first end and located deep in said substrate (303), said locking element (318; 418) comprising at least one fourth layer of semiconductor material (328; 428) doped with a doping of said second type and interposed between said isolation region and said substrate (303) at least for a stretch near said second end.

7. Radiation sensor according to one or more of the preceding claims, comprising at least one fifth layer of semiconductor material (30) doped with a doping of said first type and made at least near said front surface (4) of said substrate (3; 303) above said first layer (8), said doping of said fifth layer (30) being greater than said doping of said first layer (8) so as to obtain a conductivity of said fifth layer (30) greater than the conductivity of said first layer (8).

8. Radiation sensor according to one or more of the preceding claims, wherein said first area covers substantially all of said front surface of said substrate until contact with said isolation region.

9. Radiation sensor according to one or more of the preceding claims, wherein said doping of said first type is an n-type doping, said doping of said second type being a p-type doping.

10. Radiation sensor according to one or more of claims 1 to 8, wherein said doping of said first type is a p-type doping, said doping of said second type being an n-type doping.