WO2015028148A1

WO2015028148A1 - Position-sensitive detector having digital evaluating electronics for detecting photon distributions or particle distributions

Info

Publication number: WO2015028148A1
Application number: PCT/EP2014/002330
Authority: WO
Inventors: Peter Fischer
Original assignee: Ruprecht-Karls-Universität Heidelberg
Priority date: 2013-08-27
Filing date: 2014-08-26
Publication date: 2015-03-05
Also published as: DE102013014360A1

Abstract

The invention relates to a position-sensitive detector for detecting photon distributions or particle distributions, wherein the detector receiving surface (1) is formed by a plurality of detector cells (2) having individual detector elements (7). A read-out device for reading out the detector elements (7) associates at least one read-out channel (5) with each detector cell (2) used for the detection in accordance with a specified association rule. The read-out device has one or more counting devices, which count detection events occurring at the detector elements (7) separately for different groups of detector elements (7), which groups are formed by means of the association rule, and a counting result for each group can be temporarily stored in a memory and/or can be output or read out via the read-out channels (5) in accordance with the association rule. The association rule is selected in such a way that a position of a photon distribution or particle distribution hitting the detector receiving surface (1) can be locally determined from the signals of the read-out channels (5). The detector can be economically realized and permits high positional resolution with a low number of read-out channels.

Description

Location-sensitive detector with digital

Evaluation electronics for the detection of photon or

particle distributions

Technical application

The present invention relates to a

A location sensitive detector for detecting photons or particle distributions, having a detector receiving plane formed by a plurality of detector cells with individual detector elements and a number N of readout channels for the detector cells that is less than the number of detector cells, each detector cell having at least one the readout channels is assigned. Site-sensitive photodetectors are

used for example for the detection of gamma quanta in positron emission tomography (PET). The gamma quanta to be detected are in

Scintillation crystals that produce several thousand optical photons due to the interaction with the gamma quantum. This light must be detected with a spatial resolution of 0.5 to 3 mm. The scintillation crystals are usually subdivided into columns of 0.5 to 3 mm in width, owing to their high thickness of several cm required for the absorption of the gamma quanta, in order to obtain the required spatial resolution. The photons emitted by the individual columns must then be detected with a photodetector which also achieves this spatial resolution.

An embodiment of the photodetector with one of

However, the number of channels corresponding to a high number of channels is complicated and cost-intensive. A

The possibility of reducing the cost is the light emerging from the scintillation crystals However, this results in an optical coupling ("light spreader") to several larger detector elements, from whose signals the respective exit location of the light is interpolated

Edge effects and greatly restricts the possible mechanical structure.

State of the art

DE 102005055656 B3 describes an apparatus for processing detector signals that are less

Has readout channels as detector elements. In this device, each detector element is with each

Readout channel connected. A position determination of incident photons is achieved by a suitable weighting of the detector signals with a binary code.

US 2011/0001053 A1 discloses a detection device comprising a plurality of detector cells, in which individual channels are combined in signal processing units for the detector signals. The number of readout channels connecting the detector cells to the signal processing units thereby becomes

but not reduced.

From DE 102011111432 AI is a location-sensitive detector for detecting photon or

Particle distributions are known in which each detection cell used for detection associated with at least one of the readout channels and is connected thereto. The assignment of the detector cells to the readout channels is chosen such that signals from the

Readout channels, the position of a center of gravity of a incident on the detector receiving surface photon or particle distribution can be determined. The object of the present invention is to provide a location-sensitive detector for the detection of photon or particle distributions, which achieves a high spatial resolution with a small number of read-out channels, can be produced inexpensively and flexibly adapted to different requirements.

Presentation of the invention

The task becomes with the location sensitive

Detector according to claim 1 solved. Advantageous embodiments of the detector are the subject of the dependent claims or can be taken from the following description and the exemplary embodiment.

The proposed detector has a detector receiving surface which is formed by a plurality of detector cells with individual detector elements. The detector receiving surface is thus in the individual

Segmented detector cells that can be read via readout channels of the detector. The detector has for this purpose a number N of readout channels for the

Detector cells, which is much smaller than the number of detector cells. The detector preferably has a number of N = 3 or N = 4 evaluation channels. The number of detector cells is preferably at least 30x30 detector cells. The detector furthermore has a read-out device for reading out the detector elements, by means of which each detector cell used for the detection has at least one of the predetermined order codes which are preferably programmable in the detector

Readout channels is assigned, preferably in each case exactly one readout channel. The read-out device has one or more counting devices which detect the detection signals occurring at the detector elements. Separate events separately for different groups of detector elements, which are formed by the assignment rule. A count result for each group is buffered in a memory and / or output according to the ZuOrdungsvorschrift on the readout channels or can be read out. The ZuOrdungsvorschrift is in the proposed

Detector selected such that from the signals of the Auslese- channels the position of the impinging on the detector receiving surface photon or particle distribution

can be determined, for example, the position of the center of gravity of this photon or particle distribution.

Preferably, this assignment is selected such that this position can be calculated from the signals of the readout channels via a focus formation. In a preferred embodiment, the detector is as

Photodetector formed with photodiodes as detector elements. By the proposed realization of

Detector with a read-out device, which has one or more digital counting devices, the detector can be produced inexpensively, for example. In a CMOS technology-based chip with integrated single photon avalanche diodes (SPADs). At a

Realization as a particle detector can also be appropriate particle sensors such as. MAPS

(monolithic active pixel sensors). The corresponding ZuOrdnungsvorschrift can be stored in a programmable memory unit of the detector so that it can be changed at any time by re-programming or reprogramming. By integrating active components into the detector cells, in a preferred embodiment, noisy cells can also be identified and switched off in a targeted manner. Preferably, this includes each detector cell in addition to the Detector element and an active electronics or a switch for controlled shutdown of the cell.

In an advantageous embodiment of the detector, a buffer memory is integrated in each detector cell, which enables a buffering of the detection events in the cell. As a result, the dead times can be reduced since a second detection event can be detected even before the preceding detection event has been read out.

The assignment of the detector cells to the readout channels is in the proposed detector

Preferably, each locally approximates a distribution function which, in an ideal case, discretizes a quantity which is not finite by detector cells

Receiving area at each point a unique

Determining the position of a single incident photon or particle.

Each read-out channel is assigned a position around the detector-receiving surface or on the detector-receiving surface, these N positions spanning an area in which the detector-receiving surface is located. The assignment of the detector cells to the readout channels is then locally approximated to the chosen distribution function. The

Distribution function allocates signal components of the detector cells to each readout channel preferably as a linear or nonlinear function of the relative position of the respective detector cell to the position associated with the respective readout channel. The approach is accomplished by viewing areas comprising multiple detector cells. In these areas, the assignment of the individual detector cells to the

Selected channels selected so that over the respective area considered approximately a division of the Signal components to the readout channels, as obtained by the distribution function for a detector cell arranged in the center of gravity of the area. In one embodiment of the detector, in which the individual detector cells have approximately rectangular areas and form a rectangular arrangement with vertical rows and columns, preferably a total of N = 4 readout channels are used, which correspond to the corners of the rectangular arrangement. A use of only N = 3 readout channels is also possible with a rectangular arrangement, a

However, using four readout channels results in less noise. However, use of a total of N = 3 readout channels is advantageous, for example, for a triangular arrangement of detector cells.

The individual detector cells may be avalanche photodiodes in the proposed photodetector, preferably SPADs. In the case of a detector for the detection of particle distributions, the

Detector cells, for example, be MAPS (monolithic active pixel sensors). In a further advantageous embodiment, a scintillator of a plurality of scintillation crystals is arranged for the detection of X-ray or gamma quanta over the detector-receiving surface, which converts the incident X-ray or gamma quanta into optical photons detectable with photodiodes as detector elements. In this case, the scintillator can, for example, be subdivided into individual columns, as is known from the prior art for achieving a high spatial resolution.

The proposed detector can be in the embodiment as a photodetector, for example, in a Use gamma detector in conjunction with scintillation crystals. Application examples for this are the already mentioned PET as well as applications in the material sciences. Also in the field of research, such a photodetector can be used for applications in which the high spatial resolution is required with as few electronic readout channels as possible.

Brief description of the drawings

The proposed detector will be briefly explained again with reference to an embodiment in conjunction with the drawings. Hereby show:

Fig. 1 shows two examples of an assignment of

individual detector cells of the proposed

Detector for a total of four readout channels;

Fig. 2 in four partial images another example of an assignment of the individual detector cells of the proposed detector to the four readout channels;

Fig. 3 shows an example of a simulation ("Flood Map") with oblique arrangement of a scintillator in connection with the proposed detector;

4 is a schematic representation of an arrangement of a plurality of adjacent detectors;

Fig. 5 is a schematic representation of a plan view of an embodiment of the proposed detector;

Fig. 6 is a schematic representation of three

different concepts for the determination of

Hit sum per readout channel in the proposed detector;

Fig. 7 is an example of a construction of the proposed detector; Fig. 8 is an example of an OR structure of Fig. 7;

9 shows possible implementations of the group OR structure; and

10 shows possible implementations of the multiplicity counter.

Ways to carry out the invention

In the following example, the detector is designed as a silicon photomultiplier (SiPM), in which the detector receiving surface is composed of many individual cells, referred to as detector cells in the present patent application. The detector cells are in turn formed in a known manner by avalanche photodiodes with series resistor. In which

proposed photodetector is any for the

Detection used detector cell at least one

Selection channel assigned. In the present example of a rectangular detector receiving surface, N = 4

Readout channels used. The assignment of the detector cells to the readout channels is selected such that the position of the center of gravity of a photon distribution impinging on the detector surface can be determined from the signals of the readout channels via a center of gravity formation. If a range of detector cells is hit by photons, then signals are produced at the N outputs, which correspond in each case to the number of cells of the area hit by the photons, which are assigned to the respective read-out channel. From these signals can then due to the cell allocation by focusing on the position of the

Be recalculated area. The assignment of the cells to the read-out or output channels takes place in such a way that this is achieved locally as well as possible-within the discretization accuracy by the finite size of the individual cells. Focusing is only one preferred example based on a special one

Distribution function. The selected positions or coordinates {xKanai, i,

Y ^" Kanai, i}, to which the readout channels i (i = l..N) have been assigned, are added in a weighted manner to the signals (Signali) measured there, the whole being normalized to the overall signal:

{Xrek, Vrek} =

Sum [{x _K anai, i, yKanai, i} * Signali] / Sum [Signali] where {xrek, yrek} corresponds to the coordinate of the position to be determined.

Other non-linear distribution functions over the receiving surface can also be selected if, for example, a higher spatial resolution in the center of the receiving surface than at the edges is desired. The assignment is dependent on the selected

Distribution function performed, these

Distribution function is then approximated by the local map as well as possible. If a distribution function is selected which has the form of a sinh (hyperbolic sine) in the directions parallel to the edges of the detector receiving surface, advantageously an equal spatial error over the entire receiving surface and a higher average spatial resolution (with respect to a distribution function for center-of-gravity formation) given noise).

Fig. 1 shows two examples of this

Assignment of the detector cells 2 of a detector receiving surface 1 to the four readout channels for two different discretizations. In the upper part of the figure, the receiving surface is for illustration only from 16 x 16 detector cells 2, in the lower part of the figure from 32 x 32 detector cells 2. In photodetectors used in practice, the number of cells can be even higher, for example between 40 x 40 and 160 x 160 cells or more. The different assignment of the individual detector cells 2 to the four readout channels is indicated by the different representation of the cells. With such an assignment of the detector cells 2 to the four readout channels, a distribution function is approximated, with which the position of the center of gravity of the incident photon distribution over a center of gravity of the signals of the four readout channels

can be determined. This leads to a nearly identical spatial resolution over the entire receiving area, whereby the determination error for each area of the detector receiving area 1 is approximately the same.

2 shows in the four partial illustrations a to d the respective assignment of the detector cells 2 to one of the readout channels for a size of the detector receiving surface 1 of 80 × 80 cells. The dots in the respective sub-images mark the cells that are assigned to the respective channel. Each cell is assigned exactly one channel, so that one

Overlaying the four-part image would result in a completely black area.

The following is an example of a procedure for the assignment of the detector cells to the

Readout channels explained with the assignments of Figures 1 and 2 were generated. In this case, the following steps are carried out: 1) For each detector cell (pixel), in a first step, the ideal percentage distribution Ii of a (unit) signal on the N readout channels calculated (i = 1 ... N). This is the chosen

Distribution function across the surface of the cell

integrated. The sum of the N components in this example gives the value 1 due to the standard signal.

2) The individual detector cells are initially not assigned to any channel.

3) As the initial block size M x M, a size of 2 x 2 detector cells is set.

4) Start with a cluster in a corner of the detector receiving surface. 5) The sum of the M x M distribution shares Ii is calculated for this cluster. Ii is usually not integer. Already assigning pixels in the cluster to a readout channel is added up in Fi.

Fi is an integer for all i = 1 ... N.

6) As long as any Ii is greater than the associated Fi by more than 1, another pixel must be assigned to channel i in the cluster:

an unassigned pixel in the cluster is randomly selected and assigned to channel i, - Fi is incremented by 1.

This step is repeated until all differences Ii - Fi are less than 1.

7) The cluster is shifted by M to the right / left or up / down and from step 5) the

Repeat the process until the entire detector reception area has been completed. Of course, this can in principle also elsewhere Receiving area to be started or the processing of the total area according to another scheme.

8) In the next step, the cluster size is increased, preferably doubled, the maximum size being limited by the size of the detector receiving area, and started again at step 4). This takes place until all detector cells or pixels are assigned to a readout channel.

The approach to the desired distribution function is performed the better, the more individual detector cells are available on the detector receiving surface. This also applies to the later

Error in the positioning, which also with

increasing size of each illuminated area decreases, since then the statistics will be better. With a detector of 8mm edge length and cells of 50μιτι x 50μιτι (square) can be about 10 blocks of 0.8mm

Differ edge length.

A particular advantage of the proposed

Photodetector is that the spatial resolution and determination accuracy is independent of the position of a scintillator on the receiving surface. Fig. 3 shows an example of this over the

Edges of the detector-receiving surface 1 twisted

Scintillator 3 with 7 7 individual scintillator crystals. The ones emitted by the crystals

Photons were simulated with a finite number of photons and the assignment of the

Detector cells adopted to the readout channels.

Here, a detector receiving surface was simulated with 100 x 100 cells. Out of the over the assignment

calculated impact locations 4 of this so-called flood map is apparent that despite the twisting the

Positions of individual scintillator crystals are good can be resolved. The photodetector is thus very tolerant of a misalignment of any scintillator used in conjunction with the detector.

Figure 4 shows an example of an arrangement of several adjacent detectors with triangular

Detector Receiving Face 1. Each of these detectors has three readout channels which correspond to the corners of the

Detector receiving surfaces 1 are assigned. In an advantageous embodiment, in each case one or more readout channels of respectively adjacent detectors are connected to one another, so they are used together. Thus, in the example of Figure 5 each of the

Corner point 6 associated readout channel shared by all six adjacent detectors. The same applies to the other corner points. This can additionally reduce the number of readout channels in such an arrangement.

Figure 5 shows a schematic representation of an exemplary construction of the proposed photodetector in plan view, which is implemented in a CMOS chip. In each pixel 2 is a single photon avalanche diode (SPAD) 7 as a detector element and electronics 8 for readout. FIG. 5 shows an exemplary chip on which other electronics 9 and connection pads 10 are also arranged on the edge.

The typical size of a detector cell or a pixel 2 is about 50 x 50 μιτι ² , the size of a chip up to 1 cm ² , so that about 200 x 200 = 40000 pixels 2 per chip can be arranged. A detection event, also referred to as a hit, in a SPAD cell results in signal activity in the associated electronics 8. This can then do more Trigger processes, in particular the reading of the entire detector. Here it can then be determined which detector cells are hit.

In principle, the 1/0 hit information of each detector cell could be read out. However, this is far too slow due to the large amount of data. In the proposed detector, therefore, the hits of the groups associated with the individual readout channels are summed separately. The CMOS implementation here quite simply offers the possibility of the

To make assignment programmable. In the following, different possibilities of such a readout in the proposed photodetector are exemplary

shown. The representations of FIG. 6 each show, by way of example only, three of the detector cells or pixels 2 of the detector.

In the example of FIG. 6a, readout is realized by means of a shift register 11. All

Hits of the detector cells or pixels 2 of a column (or line) are loaded into a shift register 11 and clocked out. A serially arriving hit increases a channel allocation counter 12, four counters 12 of which are shown in this example for four readout channels. The assignment of

Detector cells to the readout channels is in one

Memory 13 stored in the periphery of the chip. This solution is very compact in the single pixel, but requires a clock and is relatively slow, as many

Clocks are required until all the pixels of the detector are read out.

In the example of Figure 6b, there are multiple adders 14 (binary or otherwise encoded) in each pixel 2, which pass their results to the next pixel. Each read channel is an adder available. In the case of a hit, a "1" is added in one of the adders The adder's selection (= channel assignment), which adds up the pixel's hits, is stored in a memory in the pixel itself This solution requires a lot of logic per pixel, but is faster than the readout with a shift register and works without a clock.

In the embodiment of FIG. 6c, in each pixel 2 there is only one adder 14, which successively adds up the individual channels. The selection of the assignment takes place in the pixel, the selection of each

edited channel via a control signal with which the pixels 2 are driven on reading. The control signal transmits to the pixels 2 the respective channel which is currently being read out. In the pixel is

in turn, the channel saved to the pixel

assigned. This solution is more compact, but somewhat slower than the solution of Figure 6b. A

Simplification and acceleration of this adder approach can be achieved by hierarchical addition with increasing bit widths. This also reduces the number of steps to be traversed. Thus, for example, a binary tree can be realized in which the first level requires only one bit addition.

The proposed photodetector can be implemented very advantageously in a chip. This

allows low cost manufacturing as well as lower overall power consumption than an analog photodetector with an additional chip.

Due to the possibility of programmable channel assignment, an adjustment of the active area, a correction of the crystal geometry or even a

Correction of edge effects done. Such a photodetector has the advantage of better time resolution over an analog detector, since the capacitances and paths are lower and the signals are higher. The possibility of integrating additional active electronics into the individual cells makes it possible to switch off defective cells, so that the rate of false hits can be reduced. Also rushing (but working) pixels can be excluded from the trigger. The trigger, ie the start signal for reading the cells, can be made to an exactly programmable multiplicity. For example, it can be determined that at least one

given number of events within a certain time window must be detected by the pixels to trigger the trigger. This can very effectively exclude noise as a start signal.

Finally, FIGS. 7 to 10 show a

Example of the construction of a digital photodetector chip according to the present invention as well as possible implementations of individual circuit parts. FIG. 7 shows in the upper part the circuit in each pixel, in the lower part the circuit in the periphery of the chip.

In each unit cell or each pixel 2 of the photodetector chip is a single photon avalanche photodiode (SPAD), which delivers a signal pulse upon the arrival of a photon. The signal processing part 15 contains, in addition to the SPAD, for example, a fast discriminator to generate a digital signal level, a quench circuit to quickly make the SPAD operational again after a hit, and the ability to turn off defective cells. The hit signals 16 in the pixels must now trigger the readout of an event. To read wrong by Preventing noise hits, for example, requires several hit signals in different pixels within a short time interval. This can be implemented by the hit signals 16 generating in each pixel a digital pulse of adjustable width, eg in a monoflop 17. These short trigger signals 18 of all pixels are transported in an OR structure 19 to the periphery of the chip, where a multiplicity is determined. An implementation of this part is described separately below. In order to exclude functioning but noisy pixels from the trigger, the trigger part can be turned off pixel by pixel with a switch-off device 20. The pixel configuration bits, which for example deactivate the SPAD or exclude it from the trigger, can be stored in a configuration memory 21 in the pixel. Parallel to the trigger path, the hit information is stored in each pixel in memory 22. In order to reduce the dead time, ie to be able to accept hits, if the readout has not yet ended, a plurality of memories 22 can be used for storing the hit information. The time interval in which SPAD signals are accumulated for a specific event, the selection of the memory 22 and the reset are performed by global control signals 23

controlled. For testing purposes and for determining

In order to be able to determine which SPADs triggered a trigger, the trigger signal can also be stored in a further buffer 24. In order to read out the data pending in the pixels, control signals 23 select one of the stored bits with the aid of an ultiplexer 25. The value is loaded into a flip-flop 26 which is connected to flip-flops of other pixels to a shift register. With the aid of a clock signal, the bits are conveyed sequentially to the edge of the chip. In a compact embodiment is only a flip-flop per pixel available, so that eg the hit buffer 22 and the trigger buffer 24 must be read out one after the other. Several flip-flops per pixel would be possible. To speed up the single readout, the input of the flip-flop in one pixel may each be connected to the next but one pixel so that the shift register is effectively half as long but two bits wide. This method can work on more bits

be extended. The circuit in the periphery must be adapted accordingly.

Triggers of an interesting event are enough timely triggers. These are passed through the OR structure 19 to the periphery. There, the multiplicity (ie the number of simultaneously active triggers) is determined in a multiplicity unit 27. If a threshold 28 is exceeded, an event is triggered. This part is more specific below

described. A valid trigger 29 starts a

State machine (main controller 30), which controls the operations in the chip by means of control signals 23. To determine the time of the event, an on-chip time-to-digital converter (TDC) 31 may be used

Timestamps for the signal (trigger 29) determine or the signal is transmitted via terminal pads 32 to an external circuit for timing. The main controller 30 must read the bits stored in the pixels for the trigger after a valid trigger 29.

For this purpose, an address counter 33 is started, which selects data from a memory (RAM) 34. There is stored for each clock of the pixel shift register, which

Read out channel ('color') to that at the output of the

Shift register applied bit belonging pixels

assigned. In this way, the assignment is freely programmable in the RAM 34. There are several counters 35 available per column, each counting the pixels of a readout channel. Depending on the value of the memory 34 For this purpose, one of these counters 35 is increased. Pixels can also remain unused if no counter is activated by suitable control. After all the pixels are clocked out, digital adders 36 compute the sums of the values in the individual shift register groups. The results can

are first stored in a FIFO 37, so that a new pixel read cycle can begin immediately, so even before the serial removal of the data with a serializer 38 is completed. Usually, the chip has an interface 39, with which the operating parameters (time windows, multiplicities, color assignment in the RAM, etc.) are set and the configuration bits in the

Configuration memory 21 of the pixels set and possibly can be read.

An example of the OR structure 19 of FIG. 7 will now be described with reference to FIG. A simple multiplicity determination can be made by each pixel sends a unit current in a node in a hit. A measurement of the current there allows a measurement of the trigger multiplicity. This method is simple but costs a lot of power if the speed needs to be high (to maintain the time information of the hits). Another solution that works with binary signal levels is shown in FIG. 8

outlined. The pixels 2 are divided into groups, which may preferably be columns, double columns, etc. All pixels 2 of a group send their hits (trigger signals 18) to the OR block 40 of the group which generates a binary output 'ColOR' 41 or 'RowOR ¹ 47 as soon as at least one input is active. One possible implementation of the OR blocks 40 is described below. By combining many pixels in a group and the reduction of the hit information to one bit, no precise multiplicity determination is possible anymore. With enough small ones However, in real applications, groups are likely to have a double hit in a group. Optionally, this problem can be further mitigated by assigning each pixel to another (possibly disjunctuated) group, eg

Lines. Will also be a hit bit with a

determined further OR block 40, the probability that a hit occurs in an already used 'columns ¹ AND ¹ lines ¹ block is very low. The binary signals of the groups must now be 'counted' in circuit blocks (multiplicity counter 42). In this case, in most applications it is sufficient to determine only small multiplicities exactly, so that the multiplication signal 43 can have a small bit width (see below). The multiplicity 43 is compared in a comparator 44 with a minimum multiplicity 45 and thus triggers a valid trigger 29. If another group structure is used, the second multiplier signal is processed accordingly and the two decisions combined. Usually one gets a full multiplicity in one of the two

Demand group structures and use an OR gate 46, but other combinations are possible. FIG. 9 shows possible implementations of the

Group OR structure (OR block 40). The OR structure must generate an output signal for at least one hit in one of N inputs. This

corresponds e.g. a logical OR function with many inputs. An embodiment, as in

Part A of FIG. 9 uses a 'wired-OR': each input is a transistor 48

assigned (e.g., an NMOS), the stream of a

central node 49 takes. Such a current flow is detected by a receiver 50. In the simplest case, the receiver 50 consists only of elements which try to connect the node 49 to a positive level hold. The current flow counteracts this sufficiently so that a low level on the node 49 indicates an active input. To achieve a higher speed, the voltage swing at the node 49 should be limited and the receiver 50, for example, have a low input impedance, such as through a cascode.

A purely digital embodiment is shown in submap B of Figure 9: The OR function is generated by a sequence of cascaded gates. A first gate stage 51 with two or more inputs reduces the number of signals from N to N / 2 or

less, further steps 52 ... 53 reduce to a single signal. This implementation requires only logarithmically few throughput times to the output, the individual gates have a small fan-out and fanin and thus a short turnaround time, and all

Passage paths from one input to the output are the same length, so that the cycle time is independent of the input. Since OR gate z.T. can not be implemented directly (in CMOS, gates are usually inverting), alternating NAND and NOR gates can also be used. Finally, FIG. 10 shows possible implementations of the multiplicity counter 42. The multiplicity counter 42 should determine how many of the N binary inputs are active. Since only small multiplicities M (M <~ 5) are relevant in real applications, it is sufficient to encode the information with a few (K) bits and combine large multiplicities into one code, e.g. Use codes for M = 0, M = 1, M = 2, M> = 3. The codes do not have to be binary.

Groups of inputs are first encoded with a first 'adder' ADD1 54 in the K bit wide code 55

transformed. The other 'adders' ADDK 56 ..57 now each process two K bit wide inputs to one Again, a binary tree with constant cycle times should be implemented for all inputs. The adders ¹ ADD1 and ADD may be constructed from conventional digital circuits.

The work leading to this invention has been completed under the Seventh Framework Program Grant Agreement No 241711

European Union (FP7 / 2007-2013).

LIST OF REFERENCE NUMBERS

1 detector receiving surface

2 detector cells or pixels

3 scintillator

4 simulated places of incidence

5 readout channels

6 corner point

7 SPAD

8 readout electronics in pixel

9 readout electronics

10 connection pads

11 shift registers

12 counters

13 memory

14 adders

15 signal processing part

16 hit signal

17 monoflop

18 trigger signal

19 OR structure

20 Shutdown device for triggers

21 Configuration memory

22 memory

23 control signals

24 buffers

25 multiplexers

26 flip-flop

27 Multiplicity unit

28 trigger threshold

29 valid trigger

30 main control

31 time digital converter

32 connection pads

33 address counter

34 memory counter

adder

FIFO

serializer

interface

OR block

binary output signal ^• ColOR ultiplizitätszähler

multiplicity signal

comparator

Mindestmultiplizität

OR gate

binary output signal ¹ RowOR

transistor

central node

receiver

first gate level

second gate level

another gate level

first adders

code

second adders

more adders

output

Claims

claims

Location-sensitive detector for the detection of

Photon or particle distributions, with

- A detector-receiving surface (1) by a plurality of detector cells (2) with individual

Detector elements (7) is formed, and

a number N of read-out channels (5) for the detector cells (2), which is smaller than the number of detector cells (2),

- A read-out device for reading the

Detector elements by which each used for detection detector cell (2) according to a predetermined ZuOrdnungsvorschrift at least one of

Readout channels (5) is assigned, thereby

different groups of detector elements (7) are formed,

- wherein the read-out device has one or more counting devices, the detection events occurring at the detector elements (7) count separately for the different groups of detector elements (7), and a count result for each group in a memory buffering and / or according to the allocation issue or read out the regulation via the readout channels (5),

- And wherein the ZuOrdnungsvorschrift is selected such that from the signals of the readout channels (5) the position of a on the detector-receiving surface (1) incident photon or

Particle distribution can be determined.

Detector according to claim 1,

characterized,

that the detector elements (7) with the readout device are integrated on a CMOS chip.

Detector according to claim 2,

characterized,

the detector elements (7) are SPADs or MAPS.

Detector according to one of claims 1 to 3,

characterized,

that the ZuOrdungsvorschrift is stored in one or more programmable memory units of the detector.

Detector according to one of claims 1 to 4,

characterized,

in that each detector cell (2) has a detector element (7) as well as a part of the read-out device.

Detector according to one of claims 1 to 5,

characterized,

in that each detector cell (2) has active electronics for the controlled deactivation of the detector cell (2).

Detector according to one of claims 1 to 6,

characterized,

the read-out device has a shift register (11) for each column or row of the detector and a counter (12) for each read-out channel (5), the counter (12) being provided via the shift registers (11).

incoming information about detection events in the detector cells (2) according to the ZuOrdnungsvorschrift be distributed to the counter (12). Detector according to one of claims 1 to 6,

characterized,

each detector cell (2) has a number corresponding to the number of read-out channels (5)

Adders (14) and a memory (13) in which an assignment of the detector cell (2) to at least one of the adders (14) is stored, each of the adders (14) one of

Readout channels (5) is assigned and the adder (14) of each readout channel (5) are interconnected.

Detector according to one of claims 1 to 6,

characterized,

in that each detector cell (2) has an adder (14), a memory (13) in which an assignment of the detector cell (2) to at least one of

Read out channels (5) is stored, and has an input for a control signal via which the detector cells (2) a read-out time for each of the readout channels (5) is communicated, wherein the adder (14) of each column or row of the detector serially or via a hierarchical adder network are interconnected.

Detector according to one of claims 1 to 9,

characterized,

that the assignment of the detector cells (2) to the readout channels (5) each locally

In an ideal case of a receiving surface which is not discretized by detector cells of finite size, an unambiguous determination of the

Position of a single impinging photon or particle. Detector according to one of Claims 1 to 10, characterized

that each readout channel (5) is assigned a position around the detector receiving surface (1) or on the detector receiving surface (1), the positions spanning an area in which the detector receiving surface (1) lies, and

Approximation function is approximated that assigns each readout channel (5) signal components of the detector cells (2) as a linear or non-linear function of a relative position of the respective detector cell (2) to the position associated with the respective readout channel (5).

Detector according to claim 11,

characterized,

in that the positions assigned to the readout channels (5) are corner points of the detector receiving surface (1).

Detector according to one of Claims 1 to 12, characterized

that the assignment of the detector cells (2) to the

Readout channels (5) is selected such

that from signals of the readout channels (5) via a

Focusing the position of the center of gravity of the detector receiving surface (1)

incident photon or particle distribution can be calculated.

Detector according to one of Claims 1 to 13, characterized

that each detector cell (2) used for the detection clone only ever one of the readout channels (5) assigned .

Detector according to one of Claims 1 to 1, characterized

in that the detector cells (2) have rectangular receiving surfaces and form a rectangular arrangement with mutually orthogonal rows and columns, wherein the detector has a total of four

Has readout channels (5).

Detector according to claim 15,

characterized,

Distribution function approximated in

Directions parallel to edges of the detector receiving surface (1) the shape of a sine

Hyperbolic has.

Detector according to one of Claims 1 to 14, characterized

the detector cells (2) are triangular

Form arrangement, wherein the detector has a total of three readout channels (5).

Arrangement of a plurality of adjacently arranged detectors according to one or more of

preceding claims, wherein one or more readout channels (5) of adjacent detectors are connected to each other.

Use of the detector or arrangement according to one or more of the preceding.

Claims as a photodiode in a

Detector for positron emission tomography