NL1043763B1

NL1043763B1 - Method and sensor for measuring electrons

Info

Publication number: NL1043763B1
Application number: NL1043763A
Authority: NL
Inventors: Marc Van Eert Ir; Rob Siegerist Ir; Woutersen Sijmen; Coquereau Yannick
Original assignee: Technolution Bv
Priority date: 2020-08-14
Filing date: 2020-08-14
Publication date: 2022-04-13

Abstract

The invention relates to a method, system, chip or set of chips and devices such as TEM microscope, involving improvement by way of a method of detecting fast high contrast transitions at image capturing, the method comprising the steps of providing a chip (26) of at least one Bilicon photomultiplier array, where the array can be a matrix or a line of pixels, directly coupling each pixel comprising at least a photomultiplier thereof directly to a distinct amplifier circuit (28), reading out an integrated analog charge of each of the pixels, providing the amplified signal to an analogue to digital convertor (30), and subsequently processing the amplified signal in a digital domain. Light intensity may be measured by using the piling-up effect of the parallelization of each photomultiplier or avalanche photodiodes in each pixel, in particular actually creating a photomultipliers array or a diodes array as integration of charges.

Description

METHOD AND SENSOR FOR MEASURING ELECTRONS

[0001] The present invention relates to a sensor and method of sensing photons. Such sensors are known per sc and applied in a wide area of use, including e.g. automotive image retrieval or more in general applications of so-called LIDAR, acronym for light imaging, detection, and ranging, transmission electron microscopes and gamma spectroscopy.

[0002] While the known methods and devices, amongst which so-called LDR and silicon photomultiplier read out ASICs, are satisfactory for most of the fields they are used in, indications start arising that with contemporary developments in image retrieval, the known methods and devices are more and more falling short in desired performance.

[0003] One example in the latter respect is in the ares of gamma spectroscopy as is e.g. provided by the 2017 poster on SIPHRA or Silicon Photomultiplier Readout ASIC i5 titled "Using the SIPHRA ASIC with an SiPM array and seintiflators for gamma spectroscopy”. This disclosure indicates that "read out rates might be limited". Another examples is of application is in the field of automotive, in particular for enabling the practical use of antonomous vehicles. This is e.g. provided by the 2018 SensL technologies poster "SIPM for automotive 3D Imaging LIDAR systems”. The publication discloses the use amongst others of a 1 by 16 array of silicon photomultiplier. Such a set-up is used in situations with low reflectivity and in complex enviroments with amongst others solar ambient light noise. Yet another example of application is in the field of electron microscopes, in particular transmission electron microscopes, as may be known by e.g. the 27 oct. 2016 article "Transmission electron imaging in the Delft multibeam scanning electron microscope” of by Yan Ren and Pieter Kruit.

[0004] While many morte applications of electron sensing may be identified, all deal with the problem of an ever increasing desire and hence need for faster read-out rates or at least more sophisticated sensors in general, more in particular of how to make such available in an economic, Le. cost-wise widely accessible and preferably compact manner.

[0005] In the latter respect an impression and overview of evolution of high speed imaging devices over the past three decades ts provided by figure 1 of the 2017 article by Takeharu Goji Etoh et al, "the theoretical highest frame rate of silicon image

Se sensors”. The article identifies that highest sensing speeds may be attained by relatively costly and/or complex devices like Laser, Streak tube and holography, while the relatively simple and affordable Silicon Photomultiplicr sensors as available attain between 1077 and 1078 frames per second for devices with between 100 and 999 frames. Anticipated, ie, "under development" sensors would attain 10°8 frames per second for devices with up to 999 frames, and 109 frames per second for a device with less than 10 frames,

[0006] While the present invention is by no means limited to any particular one of various, partly above mentioned possible applications, one particular and exemplary area of application with contemporary, challenging requirements is that of the earlier mentioned transmission electron microscopes. The cited article discloses the conversion of electron into photons in a manner known per se, using a fluorescent material, the conversion is in this case desired for minutely analyzing biological tissue, in particular in a manner using the particular contrast that transmission electron beam microscope IS may provide at identifying small tissue particles, Departing from a single beam scanning electron microscope, hereinafter "SEM", the article indicates that acquiring a 30 image of a 400X400X 1000 um mouse cortical volume, with a typical 4 nm resolution, may take aboot 400 days, depending on the detection method applied. The disclosure suggests that it would help if this imaging acquisition time would be reduced io one or two days while still keeping the imaging quality, and further suggests that "Multibeam technology would be "the only option” to achieve this",

[0007] Having this problem of acquiring smal! biological tissue elements at a highest possible rate in mind, while proposing an affordable and compact solution, practicable within contemporary industrial setting, the present invention hence seeks to improve on a sensing means, preferably even where the contemporary single e-beam transmission microscope beam would be replaced by a system as proposed and under development with multiple e-beams, more in particular with 14 x 14 or 196 g-beams as in this example,

TECHNICAL BACKGROUND [voos] While the article on theoretical highest frame rate of silicon image sensors evaluates performance, it does not provide any design for the same, other than referring to academic efforts towards highest achiovable frame rates like the therein cited articles

Ze by Etoh et al. A recent article on academie achievements also provided by Takebare Goji Btoh et al, Light in flight imaging by a silicon image sensor: toward the highest theoretical Frame Rate was published May 15, 2019. The present invention however seeks to depart from readily available sensors, hence to provide a readily industrially applicable sensing device.

{00091 One such earlier application seems to be that of SensL, which by its above cited 2018 poster on SIPM for Automotive 3D imaging LIDAR Systems indicates that the latter should amongst others be low cost and fully solid state, i.e. without moving parts, hence relatively simple, 1 departs from an eve safe laser diode array and a 16 19 array silicon photomultiplier as receiver, associated with readout and control electronics. The poster concludes that Silicon photonwltipliers have emerged as a preferred sensor for aatomotiwe LIDAR applications. While in this application good range information results may be achieved by utilizing multiple, here 20 shots per measurement, the required frame rate is quite low with a mere 30 frames per second.

fepio] Further, the above cited gamma spectroscopy application of silicon photomultipliers, indicates to be capable of handling signals from bright scintillators with energies of tens of MeV, in that respect relating to the above cited microscope application with electron beams and scintillator. The publication teaches to connect silicon photo multipliers in DC mode to a readout ASIC, using a 16 pixel SiPM array in three different readout configurations. The SIPM outputs in this embodiment are DC- connected to SIPHRA inputs using 30 cm long coaxial cables, Three readout configurations were examined: all 16 SiPM pixels connected to one SIPHRA input, 16 pixels connected to 4 SIPHRA inputs (4 pixels per input), and one single pixel connected to one SIPHRA input. Concluding, the experiments of this set-up the disclosure, without indicating how, remarks that "higher readout rates might be achievable with analogue readout”. With readout usually meaning a conversion from analogue signals to another signal, like a spectrum, a counter or a count rate-meter, it is remarked that it seems difficult within contemporary technology to achieve state of art throughput with analogue systems without suffering stability and offset variation.

BRIEF SUMMARY OF THE INVENTION

[0011] The present invention hence additionally seeks to define a photons capturing device of the known type, allowing operation in at least a transmission electron ode microscope at a practically useable acquisition rate, preferably at reasonable costs and in a relatively modest volumetric envelope, In accordance to the present invention such may be essentially achieved by a method of detecting fast high contrast transitions at image capturing, the method comprising the steps of providing a chip of at least one silicon photomultiplier array, where the array can be a matrix or a line of pixels, directly coupling each pixel comprising at least a photomultiplier thereof directly to a distinct amplifier circuit, reading out an integrated analog charge of each of the pixels. The amplified signals are in a further elaboration of the invention be provided to an analogue to digital convertor or ADC, and subsequently processed in a digital domain.

1 [9012] The invention recognizes that sampling rate and dynamic range are key io certain applications like the application of TEM. B is recognized in this respect that most systems with a DC coupling are high frequency but low resolution, e.g. in comparison with AC coupling systems which are somewhat slower but with higher resolution. It is seen in practice to use AC coupling to increase signal resolution by especially removing leakage current from the detector, For obtaining an AC coupling, a DC coupling may be departed from, while inserting a capacitor in series between the sensor and the amphfier of the circuit, The present invention however, amongst others combines the advantage of a DC coupling, thereby measuring absolute light intensity, with a digital readout in which an AC coupling is performed by measuring the signal amplitude by subtracting the dark current from the average local light signal in the digital domain, thereby providing optimum dynamic range. Using 2 high resolution, typically above 12-bit, fast analogoe-to-digital convertor or ADC, say above 10 million samples per seconds or MSPS, enables to get rid of the offset and obtain sufficient contrast in the digital domain even if only a smaller range of the full scale range or FSR ofthe ADC isutilized, e.g. as depicted in Figure 8. The present invention hence combines precision and speed/hiph frequency. Where reading out is performed In analogue charges, the information is, amongst others for meeting contemporary industrial standards, transformed digitally for recording or archiving and e.g. later processing, including correction of signal, hence utilizing and allowing the convenience in contemporary digital domain processing. In doing so, the photomultiplier of the system, preferably embodied with a number of diodes, or a diode array, is in a preferred embodiment read out as a single value.

-5-

[0013] Where it is known that the gain-bandwidth product of the amplifier often is a lirmting factor to signal speed, and efforts to overcome the same are directed to dedicated and often scientific level only or otherwise highly expensive solutions, the present invention unexpectedly proposes the application in the present context of an elegant known per se, industrially applicable solution by performing amplification in two or more consecutive partial stages. In this respect it is in advance of other to be described factors remarked that where it may normally be considered that the gain- bandwidth product of the amplifier would become a limitation to the signal speed, which deals with an optimal, typically high frequency level, e.g. a S0MHz in a presently targeted application, as a detected photon would generate an output signal of less than 20ns rise time. Splitting the amplifier in two stages m the set up and goal according to the present invention allows for a higher gain bandwidth product that a single amplifier could allow in order to cope with the application needs. In yet further elaboration the new method may hold the feature that light intensity is measured by using the piling-up effect of the parallelization of each photomultiplier or avalanche photodiodes in each pixel, in particular actually creating a photomultipliers array or a diodes array as integration of charges, in particular in a biased photoconductive mode, and converting the amount of light received per pixel in an electrical signal with a short response time, more in particular with a quick decay, for time domain high dynamics, It will hence be evident that the invention uses an otherwise known, or known per se sensor in a quite different and unexpected manner, however leading to significantly improved measuring characteristics, Some counterintuitive measures may be involved in reaching at least part of such result, e.g. having generated in fact and using a pulse by an array of the sensor, and integrating the same in fact, as will effectively be explained in the description, Certainly, in fact the present invention may also, i.e. alternatively be defined as a method of measuring light tensity, in which light intensity is measured by using the piling-up effect of the parallelization of cach photomultiplier or avalanche photodiodes in each pixel, in particular actually creating a photomultipliers array or a diodes array as integration of charges, in particular in a biased photoconductive mode, and converting the amount of light received per pixel in an electrical signal with short persistence, more in particular with a quick decay for time domain high dynamics,

[00141 In further elaboration zach photomultiplier of the pixel array may be provided with a number of diodes, in particular in the form of a chip forming an array of arrays, alternatively denoted a pixel array of diodes arrays, more in particular in the order of at least hundreds per photomultiplier, At least part of the photomultiplier array, in particular the fll array may be readout as a single value, In this context either one or both values of mean light and ultimately instantaneous absolute amount of light may be determined. Also, the absolute amount of light may be retrieved by correcting the digital signal for a resulting offset by means of a digital calibration procedure at the system level. Factors that may be taken into account in this respect may include e.g. the sensor temperature and bias.

[0015] Favorably, amplification may be performed with a high amplification gain, in particular gain bandwidth product of at least 20GHz, more in particular determined by considering a useful range of the ADC input, in particular while performing the amplification in two or more consecutive partial stages. The amplification of the amplifier circuit, in particular at each amplification stage, may be set with an af least relatively high amplification gain, 1e, is performed in the order of hundreds of times. In line with the preceding, the present invention and method may rely on the provision of a fast Analogue to Digital Convertor or ADC in the sampling rate of 10 to 100MSES. For determining an actual light intensity difference, apreceding calibrated digital offset is subiracted from an acquired value, The ADC may herein again favorably be provided with a multiplexer, and may be embodied as one of a multi-channel with embedded analogue multiplexers, and as a multi ADC embedded in a single chip. In one favorably designed embodiment, the array may be comprised of at least 4 pixels or photomultipliers, arranged in line or in matrix.

[0016] An industrially interesting embodiment of the invention may hold that the detection concept is incorporated into an ASIC. The invention may also hold an embodiment with the incorporation comprising a set of chips, comprising at least a silicon photomultiplier array chip and an ADC chip which is functionally coupled thereto. In yet a further embodiment of the method of the invention, the invention may hold a signal acquiring system, comprising a photomultiplier array electrically connected to a front-end subsystem, e.g. by an interconnect, comprising a two-stage amplifier circuit for each pixel of the photomultipliers to be readout, supplying N, e.g. 64 amplified readout signals, ¢.g. via an interconnect, to a multi-channel ADC chip,

wa each amplified signal being digitized thereby, before being streamed out by a digital streamer or subsystem to a high level signal processing unit; an ASIC configured for either embodying the system or for executing the method of the invention; a set of chips for interconnected application, including an ASIC, configured for executing the method according to the invention in a manner distributed over cach of the chips of the set; and a device, in particular TEM microscope, improved by the inclusion of any one of a chip, chip set, system and method in accordance with the invention.

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Various aspects of the invention and an example of a possible embodiment of the invention are illustrated in the drawings in which: [00181 FIG. | by way of one possible application that is improved by the present invention, is a schematic representation of a transmission miecroscope system provided with a camera; 153 [00191 FIG. 215 a typical pulse shape of a silicon characteristic of a photomultiplier; [00201 FIG. JA is a known per se schematic lay out of a basis configuration of an -exeraplary type of silicon photomultiplier array, with FIG 3B providing a simplified, electrical schematic thereof [00211 FIG. 4 is a schematic representation of a conventional measuring set up or so called acquisition chain using a silicon photomultipliers;

[0022] FIG, 5A illustrates a problematic situation of piling up of the pulse of two photons such as is recognized to occur in situations with c.g. high contrast and fast changes, i.e. high dynamic, in which a single photon would create a signal 23 and the resulting pile-up signal! 24; [o023] FIG. 3B is an illustration of possible outcome when using conventional measuring techniques such as photo counting for acquiring pixel intensity at high rate; [00241 FIG, 6A is a schematic of a fairly standard, ie. known per se and simple transimpedance amplifier circuit foe2s] FIG. 6B anda represent modifications to the standard transimpedance amplifier circuit converting incident light to charges by way of biasing and for measuring the charges by way of direct and indirect coupling respectively;

~§~ {0026} FIG. 6D represents an otherwise basic amplification circuitry, adapted by splitting the amplifier m two stages, and unexpectedly applied as a DC type of measuring for imaging bigh contrast and high dynamic signal;

[0027] FIG. TA represents the output signal of a photon multiplier under high level of incident light on top of its own dark current signal (DS).

[0028] FIG. 7B represents the output of a biased photon multiplier under pulses of high level of incident lights, providing a local average signal of light (LS), highlighting the instants of analogue, locally average dark current signal (DS) as wel, and analogue light signal difference to be readout as the pixel amplitude (PA), with distinction of the integral AC level {ACS) of this signal.

[00281 FIG, 8 illustrates the effect in accordance with the invention of using a fast analogue to digital convertor or ADC, in order to get rid of the signal offset SO while obtaining sufficient contrast of the signal dynamic (SD) in a digital domain and while applying merely a small range of the full scale range or FSR of the ADC;

[0030] FIG. 9A and 9B provide signal characteristics of the DC coupled and AC coupled amplifier circuits of FIG 7A and 7B respectively, while a slow modulation of amplitude occurs in the input signal, and such modulation visible in FIG. 9A being filtered out in the AC coupled signal shown in FIG, 9B; 0031] FIG. 10 is an illustration of a possible signal acquiring architecture in accordance with the invention; (00321 FIG. 11 is a schematic representation of an alternative architecture according to the invention, in which an ASIC cooperates with the photomultiplier array; [00337 FIG. 12 is yet a further alternative architecture in which the same is accommodated in a single ASIC or flip-chip assembly.

DETAILED DESCRIPTION

[0034] FIG. 1 represents a state of the art multi beam electron-optical system 1 for a transmission electron microscope, forming one possible application of the present invention, Such a system comes with an otherwise not represented source for electron beams 3, such as a known Schottky source, which in a manner known per se, using e.g. extraction electrodes and a current limiting array creates a multiplicity of electron beams, ¢.g. in an array of 14 by 14 beams, In the represented system the beams 3 are condensed by an accelerator lens 2 and passed through a variable aperture $ via a

Ge second condenser lens 4, and scanned by scan coils 7 after having passed an intermediary lens 6, The scanned beams pass an uliva-high resolution lens § before being projected onto sample , which in this case sits right on top of a photon generator 9, e.g, embodied by a so called YAG or fluorescent material and forming part of a detector for capturing, at least detecting electrons transmitted through the sample.

[0035] The light beams 10 emitted by photon gensrator 9 are captured by a camera system 13 by projection of the light beams 10 through a window 12, here combined with a focusing lens. The projection may include redirecting the light beams 10 via a mirror 11.

16 [0036] FIG. 2 illustrates the typical signal as produced by contemporary silicon photomultipliers, henceforth SIPM, used for counting the number of pulses generated thereby as an indicator of incident light. The pulse signal generated typically comprises a very short rise phase 14 of several nanoseconds and a much longer decay or reset phase 15 of up to 100 nanoseconds, General teaching as per Wikipedia page in Silicon photomultiplier amongst others holds that signal decay time is inversely proportional to square root of photoelectrons number within an excitation event,

[0001] FIG. 3A is a schematic representation of a known array 16 of silicon photonultipliers SiPM. Typically such S8iPM come with a basic N-doped body layer 17, an Avalanche layer 18 and P-doped entrance windows 19 for receiving photons. The photon receiving windows may be provided with an anti-reflective layer 20 and are physically separated, here by a strip of the avalanche zone creating rectangular cells. The strip areas are used in this example for allocating a so-called quenching resistor 22 with which the window, or diode or PN-junction forms a cell in the one orientation, and a bias lead 21 on top of the strip oriented in transverse direction, each cell being a pixel ora sensor cell for the later image acquisition system. In order to increase the sensitive area, it is common to use photodiode array, Le, several photodiodes in parallel, in the vicinity of each cell. Such Array is electrically represented in FIG, 3B where each diode symbol represents an avalanche photo diode connected to a serial resistor also called quenching resistor. This is the parallelization of these APD which make possible to use a summed signal of all current circulating through the diode array between its cathode K and anode A. FIG 3B provides a simplified, electrical schematic of FIG. 34, where each dieds symbol represents an avalanche photo diode connected to a serial

~10~ resistor also called quenching resistor and where cathode K and anode A are the available electrical connections. {00021 FIG, 4 is a schematic representation of a common use of a light acquisition chain with silicon photomultipliers §, in general applicable to any camera or spectrometer as e.g. may be applied in particle physics. Light intensity LI of a light signal is captured by a sensor 5, coupled to an amplifier A, coupled to an analogue to digital convertor ADC, which provides a digital value DV for the measured light intensity. Main architectures in photo detection consist in either photo-counting or light tegration, and either AC or DC coupling, The choice of architecture is dependent on the information to be retrieved and on the nature of the incoming light.

[0003] While the transition from previously known photomultiplier tubes towards above described solid state single photon sensitive devices has opened tremendous possibilities for measuring high frequency signals such as in LiDAR applications, measuring at high speed with these contemporary sensors may still pose a problem in situations where high acquisition rate, Le. high frame rate is desired such as high speed camera, e.g. as used for research, typically for analyzing quick changing events, with more than 10.000 frames per second, yet not excluding high resolution imaging, and such as in high contrast situations. foooa] Hence, for a camera to provide ultimate time and intensity resolution of a highly dynamic amount of light, e.g, as in contrast between dark and bright spots of a sample, the present invention infegrates a photo detector array in a complete readout system with direct coupling. [oons] Where such may render the same global measuring architecture as provided by FIG. 4, in the present invention, the energy information or wavelength of the incoming light is now regarded not of interest and here assumed constant. Rather, the invention seeks to measure the amount of light received by cach pixel of the photon-detector at a very high frequency, say a frame rate of over 1 Mega Hertz, (>1 MHz), i.e. an interval time below tus, Tt may be clear that such a requirement brings about a barrier to using a photo-counting architecture, in which the state-of-the-art maximum throughput is 20 currently about 10 Meps, ie. 10 million counts per second. fooos] Tt is an insight underlying the present invention, that this state of the art limitation 1s mainly due to a what is called a relatively slow signal, i.e, what is here, in accordance with underlying insight, concluded as to mean low system throughput

~11- capability, that is to say at least due to above mentioned pile up of signal due to signal decay time, obtained after electrical conversion of the photon energy.

The present invention recognizes a circumstance that in order to measure light intensity by counting the numbers of incident photons, each photon raust be resolved, and that such is challenging and in fact also not possible in case of so-called pile-ups, Such pile up phenomenon is illustrated by FIG. 5A, which depicts a situation where at least two photons with a signal 23 arrive in a short interval.

As illustrated this situation caused a pulse pile-up 24 in the signal, at least compared to a non-piling up pulse signal 23, or 14 and 15 in FIG, 2. Since counting is performed by identifying a flank of a pulsed signal as presented by FIG. 2, the pile-up of two or more photons will eventually be seen as a single pulse signal, implying that at least one count is lost in case of each pile-up.

The present invention recognizes the latter as a cause of attaining the high pulse count rates as desired in contemporary applications.

In particular it further recognized and desires to solve measuring situations where the incoming photon rate is higher than the system throughput capability, and where hence the chance of pile-up occurrence is large. [00077 FIG.

SB is an illustration of possible outcome when using conventional measuring techniques such as photo counting for acquiring pixel intensity at high rate.

The top chart illustrates a periodic distribution of photons whereas the bottom chart represents a random distribution of photons, If photons can be counted per interval, the resulting countrate can be calculated, in the example of 100 ns interval, the resulting count rate is 10 million count per seconds as 1 count per interval is recorded (CPI). voos] With a conventional photo-counting system with for instance 10MHz= sampling rate and doe to the random nature of photon arrivals, this means that about 1 photon may be counted in every 100ns sampling period.

Contrary to a periodic signal,

where interval between events is constant, the number of incident photons per time interval can in the desired situation to be measured, very well be 0 (zero) or 3 (three) due to the randomness of the light generation process, No signal, ie. zero (0) counts per interval CPL, usually expressed as counts per second, abbreviated ops, would lead to 0 (zero) photon count per time interval or CPI measurement at any time, A maximum measureable signal can on the other hand also lead to a zero (0) count in a same time interval of for instance 100 ns.

An example is depicted in FIG. 5C, where both measuring channels see and count an averaged flux of 10 Mops, Nevertheless, whereas 1 count per interval is always reported with the periodic signal, zero (0) count can be

-12- reported In the same time interval for the random signal as illustrated between 0.205 and {.3us. As amongst others underlying the present invention, this recognized situation is concluded to mean that a proper measurement of the light intensity requires a longer integration time, for instance one micro second, to average the number of photons per interval Such a duration is however not convenient for an application where pixels intensity must be acquired at an even faster rate than 100s. [00097 An obvious measure to avoid pile-up or to correct for it, seems to be to keep the incoming photon rate low. However, on the other hand, in a sensing application such as is desired, and for even further improving high resolution TEM, however including other applications in this field as amongst others mentioned in the introductory part, the invention recognizes that the application requires short measurement times and a high dynamic in terms of signal amplitude, where 1 to about 10000 incident photons needs to be identified in this short measurement interval. This in fact leads to, or requires a high incoming photon rate.

iS {0010} The present invention recognizes that for such a latter mentioned pixel rate, photon counting may not be adapted, at least not adequately, and proposes a different approach of light integration. This new proposal still utilizes the same kind, state of the at silicon photomultiplier, however in a manner different from what is normally proposed. Usually, such ordinary silicon photomultiplier is used to detect a single 20 photon per pixel and per time interval, The generated signal of the incident photon is thus generally quickly converted in a digital trigger instant. In this respect the mvention recognizes that silicon photomultipliers as sensor cell may be used to convert incident light into charges by means of proper biasing. Hence it proposes to turn to analogue measuring or reading out of the photomultiplier, rather than digitally, In this respect #t 25 should be remarked that even in pulse counting mode the readout from the array can be analogue. Actually it might be represented as that the number of single photon avalanche diodes is of importance in the discretization of the light amount and in the requested signal dynamic to be readout. The photon multiplier, being based on an array of single photon avalanche diodes in parallel, is used and recognized to allow summing 30 ofthe response of cach diode so that the output signal is proportional to the amount of incident photons absorbed. The proposed analogue reading out method may, despite the common practice of counting pulses by way of detecting rise phase 14 and decay phase of signal, at hindsight, in principle still be derived from the basic signal

~13- representation of FIG.2, where the signal is represented as an accurate measurement of Voltage against time in ns. However, a pulse duration would under contemporary conditions, in which incidentally the photomultiplier has developed, normally typically be measured digitally.

[001] Once having recognized the possibility of such a new use and set-up, the vention recognizes the possibility to measure the charges with either direct (DC) or indirect (AC) coupling, in particular when using a known per se auplilier. One example of such is the simple transinpedance amplifier circuit proposed by the Analog Devices technical article MS-2624, here represented as FIG, 6A. Such standard set up converts light into charges, thus current, Le. under the electrical field provided by the diode biasing. The photodiode, in this circuit connected to ground, is in the left hand side of the drawing referenced by the pair of Z-shaped light arrows, and the circuit provides an output terminal and signal Vour at the right hand side of the diagram, Incidentally, it is further recognized by the present invention that the known per se acquisition chain of FIG. 4 can also be used to measure photon energy when the digital value thereof compares to that of the representation in FIG. 2 and using a high sampling rate. oai} FIG. 6B and 6C illustrate the above mentioned divect or DC coupling and the indirect or AC coupling respectively, showing a connection of the photodiode to bias voltage Veras. In case of the DC coupling of FIG. 6B, the photodiode output is directly coupled to an amplifier circuit, Hence without any capacitor as is included between the photodiode and the amplifier civeuit in the AC coupled configuration of FIG. 6C.

[00131 While the AC coupling of FIG. 6B should from an electrical standpoint be preferred in order to get rid of the biasing voltage and to reduce total noise of the circuit output, the present invention recognized that counter-intuitively, a direct coupling should be applied despite the fact that noise will be maintained, i.e. be included in the signal, hence will also be measured and, what's more, will also be amplified. Yet, it was in view of the new manner of measuring to be set, recognized by the present invention that such an AC coupled solution would otherwise loose DC supplied information such as dark current or mean light and ultimately instantaneous absolute amount of light, As it comes to higher level of incident light, the response signal does not allow for individual pulse counting anymore as one can understand for the signal shape given in FIG, 7A and 7B. The local average value becomes the integration of charges equivalent to the incident light level. Hence, in accordance with the latier insight and again further

“14 underlying the present invention, measuring should for attaining high measuring rates, against a prejudice in the art, where signal noise is normally avoided by way of melination, or not primary nature as it were, be performed using DC connection with, in fact despite the presence, at least initially, of "non cleaned" or raw signal in the 3 measurement taking, Where in the ordinary case of one signal as in Fig. 7A or Fig. 78, removing dark current can be profitable, such could in another case or in specific application, imply that a mean signal will also be lost, thereby reducing the contrast in total acquisition.

[0014] It is additionally recognized that the contemporary manner of measuring using AC coupling has a time constant which affects the measured signal amplitude. This will depend on the local wean amount of light received. In this respect, FIG. 9A illustrates the signal amplitude in case of DC coupled amplification. In contrast to such a type of signal, an AC coupled signal as in FIG, 9B does not provide information about the amount of light but only the local variation. In fact the AC coupled signal meanders iS above and below the zero current line, implying a loss of absolute light intensity information. [00151 On the other hand, while DC coupling is known per se, it is often used to know the absolute amount of light or to enable time over threshold measurements. This might be performed, at least makes sense when single photons must be detected and characterized such as time of arrival or incident energy. If also makes sense to measure amount of light in slow changing application. In fast changing application, as is recognized to be the case e.g. a TEM application, DC coupling would not be trivial because of the chances of signal saturation. It is however, further recognized by the invention that by proper gain tuning, this effect may be limited, and that in case of an application with focus on imaging high contrast and high dynamic, a DC coupling should nevertheless be preferred, be it while adding a special measure of gain tuning. foots] A DC coupling, of which basic cireuitry is provided in FIG. 6B, with the diode used in photoconductive mode, i.e. with reverse bias applied, is hence proposed to improve speed and linearity of the diode as opposite to photovoltaic mode such as implemented in FIG, 6. The reverse bias, however increases the dark current and noise current. In such implementation, the amplifier stage also contributes to signal offset error on top of dark current. To improve the contrast and signal resolution, the invention sets a high amplification gain in the order of 100 to 1000 in order to fit the signal of the

~15- sensor, e.g. nulti-pixel silicon photon multiplier, to the ADC input, the maximum gain being dictated by the maximum ameunt of incident light to be digitized without over- ranging the ADC input. Because any offset may be added to the amplified signal, the gain should be set by considering the remaining useful range of the ADC mput, e.g. as indicated by FIG, 8. To retrieve the absolute amount of light, the digital signal is corrected in a digital domain for the resulting offset by means of a (digital) calibration procedure at the system level, The invention favorably applies a trick to reduce the effect of the amplifier offset, splitting the amplification in two stages so that parasitic offsets like mput bias current and voltage offset are not amplified with the full gain factor. This is also valid for the noise level. Above that, the gain-bandwidth product of the amplifier would become a limitation to the signal speed which is dealing with an optimal, typically high frequency level, e.g. a SO0MH2z in a presently targeted application, as a detected photon would generate an output signal of less than 20ns rise true. Splitting the amplifier in two stages in the set up and goal according to the present vention allows for a higher gain bandwidth product that a single amplifier could allow in order to cope with the application needs. It is being recognized that amplifiers have a gain-bandwidth prodact (GBW) limit, nowadays in the range of 1 to 10 GHz, the vention according to yet further insight applies a measure of splitting the gain in two stages as schematically represented by FIG. 6D, recognizing that the GBW product of cach of the split amplifiers allows for a larger bandwidth, Esch gain bandwidth product being for instance 1000MHz (gain=20, Bandwidth = S0MHz) with a total chain gain bandwidth product of 20GHz (gain=400, bandwidth = SOMHz)}. Offset and gain correction however, may according to the present invention preferably be applied to the digital domain, removing complexity from the analogue path. Figure 8B illustrates the effect of a further measure in accordance with the invention, with a signal with relatively strong, ie. useful signal dynamic SD, however with high offset SO, The invention, in a further elaboration proposes to get rid of the offset SQ, so as to renin with sufficient signal contrast, at least in the digital domain, utilizing only a relatively small range SD of the full scale range or FSR of the analogue to digital convertor or ADC, it being provided that a fast ADC is utilized, Le. working at least at about 20MHz to be able to acquire the desired pixel rate of 100ns.

[0017] FIG, 10 provides an example of a signal acquiring architecture in accordance with the invention, in which the photomuitpler array 26 is wired via an interconnect 27

16. to a front-end subsystem 28 conyrising of a 2-stage amplifier circuit for each pixel to be readout, supplying N, e.g. 64 amplified channels via interconnect 29 to a multi- channel ADC chip 30, each amplified signal being further digitized thereby, before being streamed out by a fully digital streamer or subsystem 31, for communication to a high level unit not represented in the illustration.

[00181 FIG. 11 represents the architecture of FIG. 10, at least the principles thereof, or at least the method of detecting fast contrast transitions at image capturing according to the invention, carried over into a so-called ASIC cmbodiment with DC signal coupling. Such figure highlight the fact that the required fonctions of amplifiers, multiplexer, ADC and transceiver can be integrated in a chip for higher integration level.

[0019] Figure 12 represents yet a further possible embodiment of the principles of the present invention, wherein the principle or method according to the vention is in application executed by means of a completely integrated ASIC comprising the photomultipliers array and the ASIC embodiment described above. One possible embodiment of such an ASIC instead of an above exemplified discrete implementation would require nulti-channel amplifier/front-end and a multi-channel ADC in one chip, each of e.g. 64 channels. Such ASIC embodiment features a matching input impedance for reducing noise, transimpedance amplifiers and a high gain with a factor preferably being of at least about 100, thereby allowing a synchronous sampling of said multiple signals.

[0020] The present invention, apart from what has been described above, also relates to all details in the figures, at least Tor as far as these are directly and unambiguously retrievable by a skilled person, and to everything that is described in the following set of claims.

CLAUSES i. Method of detecting fast high contrast transitions at image capturing, the method comprising the steps of providing a chip (26) of at least one silicon photomultiplier array, where the array can be a matrix or a line of pixels, directly coupling cach pixel comprising at least a photorultiplier thereof directly to a distinct amplifier circuit (28), reading ont an integrated analog charge of cach of the pixels, providing the amplified signal to an analogue to digital convertor or ADC (30), and subsequently processing the amplified signal in the digital domain.

2 Method according to clause 1, m which Light intensity is measured by using the piling-up effect of the parallelization of each photomultiplier or avalanche photodiodes in each pixel, in particular actually creating a photomultipliers array or a diodes array as integration of charges, in particular m a biased photoconductive mode, and converting the amount of light received per pixel in an electrical signal with a short response time, 1S more in particular with a quick decay for time domain high dynamics.

3. Method of measuring light intensity, in which light intensity is measured by using the piling-up effect of the parallelization of cach photomultiplier or avalanche photodiodes in each pixel, in particular actually creating a photomultipliers array or a diodes array as integration of charges, in particular in a biased photoconduetive mode, and converting the amount of light received per pixel in an electrical signal with short persistence, more in particular with a quick decay for time domain high dynanies.

4. Method in accordance with any of the preceding clauses, in which each photomultiplier of the pixel array is provided with a number of diodes, in particular in the form of a chip forming an array of arrays, alternatively denoted a pixel array of diodes arrays, more in particular in the order of at least hundreds per photomultiplier,

5. Method in accordance with any of the preceding clauses, in which at least part of the photomultiplier array, in particular the full array is readout as a single value,

6. Method according to any of the preceding clauses, in which either one or both values of mean hight and ultimately instantaneous absolute amount of light is determined.

7. Method according to any of the preceding clauses, in which absolute amount of light is retrieved by correcting the digital signal for a resulting offset by means of a digital calibration procedure at the system level.

„18.

8. Method in accordance with any of the preceding clauses, in which the amplification is performed with a high amplification gain, in particular gain bandwidth product of at least 20GHz, more in particular determined by considering a useful range of the ADC input. 9 Method in accordance with any of the preceding clauses, in which the amplification is performed in two or more consecutive partial stages.

10. Method according to any of the preceding clauses, in which the amplification of the amplifier circuit, in particular at cach amplification stage, is set with an at least relatively high amplification gain, Le. is performed in the order of hundreds of times.

11. Method in accordance with any of the preceding clauses, in which the method relies on the provision of a fast Analogue to Digital Convertor (ADC) in the sampling rate of 10 to 100MSPS,

12. Method in accordance with any of the preceding clauses, in which the actual light intensity difference is measured by substraction of acquired value with a formerly calibrated digital offset. 13, Method according to the preceding clause, in which the ADC is provided with a multiplexer. 14, Method according to the preceding clause , in which the ADC is embodied as one of a multi-channel with embedded analogue multiplexers, and a multi ADC embedded mn a single chip.

18. Method in accordance with any of the preceding clauses, in which the array is comprised of at least 4 pixels or photomultipliers, arranged in Hne or in matrix.

16. Method according to any of the preceding clauses, in which the detection concept is incorporated into an ASIC.

17 Method in accordance with the preceding clauses, in which the incorporation comprise a set of chips, comprising at least a silicon photomultiplier array chip and an ADC chip which is functionally coupled thereto.

18. Signal acquiring system, comprising a photomultiplier array (26) electrically comected to a front-end subsystem (28), e.g. by an interconnect {27}, comprising a two-stage amplifier circuit for each pixel of the photomultipliers to be readout, supplying N, e.g. 64 amplified readout signals, e.g. via an interconnect (29), to a nuli channel ADC chip (30), each amplified signal being digitized thereby, before being

“19. streamed out by a digital streamer or subsystem (31) to a high level signal processing unit.

19. ASIC configured for either embodymg the system or for executing the method in accordance with any of the preceding clauses.

20. Set of chips for interconnected application, including an ASIC, configured for executing the method according to the preceding toethod clauses in a manner distributed over each of the chips of the set,

21. Imaging device such as a microscope, in particular a TEM microscope, improved by the inclusion of any one of a chip, chip set, system and method in accordance with any of the preceding clauses.

Claims

<20

CONCLUSIONS

A method for detecting high-contrast fast transitions in image capture, the method comprising the steps of providing a chip (26) with at least one silicon photomultiplier array, the array being a matrix or array or array of pixels and which each pixel having at least one photomultiplier is coupled directly to a separate amplifier output (28), with an integrated analog charge read from each of the pixels, and the amplified signal sent to an analog to digital converter or ADC ( 30) 1D is passed through, and then processed in the digital domain.

Method according to claim 1, wherein Uchtmienstteil is measured using the stacking effect of the parallelization of each photomultiplier or avalanche photodiode in each pixel, in particular actually creating a photomultiplier array or a diode array as integration of charges 13 in particular in a biased photoconductive mode, and converting the amount of light received per pixel into an electrical signal with a short response time, more specifically with a fast decay for hope dynamics in the time domain,

3. Method for measuring the hobt density, in which the light intensity is measured by using the stacking effect of the parallelization of each photomultiplier or avalanche photodiodes in each pixel, in particular actually creating a photomultiplier array or ven diodes- array as an integratic of charges, especially in a biased photoconductive mode, and converting the amount of received light per pixel into an electrical signal with short persistence, especially with a fast decay for time-domain high dynamics.

A method according to any one of the preceding claims, wherein each photomultiplier of the pixel array is provided with a plurality of diodes, in particular in the form of a zen chip forming an array of arrays, also referred to as a pixel array or array of diodes. arrays, more typically on the order of at least hundreds per photomultiplier.

A method according to any one of the preceding claims, wherein at least a portion of the photomultiplier sequence, in particular the entire sequence, is read as a single value,

21-6. A method according to any one of the preceding claims, wherein either one or both values of average light and final instantaneous absolute amount of light is determined.

The method of any preceding claim, wherein absolute S amount of light is obtained by correcting the digital signal for a resulting offset by a system level digital calibration procedure.

Method according to one of the preceding claims, wherein the amplification is performed with a high gain gain, in particular a gain bandwidth product of at least 20 GHz, more particularly i0 by taking into account a useful range of the ADC input .

Method according to one of the preceding claims, wherein the reinforcement is carried out in two or more successive partial steps. A method according to any one of the preceding claims, wherein the gain of the amplifier circuit, in particular at each amplification stage, is set with 13 centimetres at least relatively high amplification gain, ie is performed on the order of hundreds of times, il, Method according to one of the preceding claims. preceding claims, wherein the method is based on providing a fast Analog-to-Digital Converter (ADC) with a sample rate of 10 to 100 MSPS,

A method according to any one of the preceding claims, wherein the actual difference in light intensity is measured by decreasing the obtained value by a previously calibrated digital offset.

A method according to the preceding claim, wherein the ADC comprises a multiplexer.

The method of the preceding claim, wherein the ADC converter is configured as either a multi-channel converter with embedded analog multiplexers or a multi-ADC embedded in a single chip.

Method according to one of the preceding claims, wherein the array consists of at least four pixels or photomultipliers arranged in line or in matrix. Method according to one of the preceding claims, wherein the detection concept is incorporated in an ASIC.

Im

A method according to the preceding claims, wherein the embodiment comprises a set of chips comprising at least a silicon photomultiplier array chip and an ADC chip operably coupled thereto.

A signals acquisition system comprising a photomultiplier array (26) electrically connected to a front-end subsystem (28), e.g. by a zen antconnect (27), comprising a two-stage amplifier circuit for each pixel of the photomultipliers to be read, and wherein N, e.g., 64 amplified readout signals, e.g. via a link (29), are applied to a multi-channel ADC chip {30}, each amplified signal being digitized thereby before passing through a digital streamer or subsystem (31) to a high level signal processing unit is streamed.

An ASIC adapted to either incorporate the system or to perform the method according to one of the preceding claims,

A set of interconnected application chips, including an ASIC, 153 configured to perform the method of any of the preceding method claims, distributed in a manner among each of the chips of the set. 21, Imaging device such as a microscope, especially a TEM. microscope improved by incorporating a chip, chipset, system or method in accordance with any one of the preceding claims.