HIGH SENSITIVITY SINGLE-PHOTON AVALANCHE DIODE ARRAY

Abstract

The present invention relates to a photodetector array for capturing image data, comprising: photodetector cells arranged on a substrate, each including a single-photon avalanche diode, wherein the active areas of the photodetector cells are neighbored along a hexagonal grid; microlenses, having a hexagonal or circular shape, each arranged on one photodetector cell to focus light onto the photodiode.

Claims

1-12. (canceled)

13. Photodetector array for capturing image data, comprising: photodetector cells arranged on a substrate, each including a single-photon avalanche diode, wherein the active areas of the photodetector cells are neighbored along a hexagonal grid; microlenses, having a hexagonal or circular shape, each arranged on one photodetector cell to focus light onto the photodiode.

14. Photodetector array according to claim 13, wherein the active area of the photodiode has a circular shape.

15. Photodetector array according to claim 14, wherein at least one of the microlenses has a circular shape, wherein the radius of the microlense is by a factor of 3 to 20 larger than the radius of the active area of the photodiode.

16. Photodetector array according to claim 13, wherein particularly the radius of the active area is lower than 9 μm, preferably lower than 5 μm, wherein particularly the radius of the active area is more than 3 μm.

17. Photodetector array according to claim 13, wherein the active area of the photodiode is configured with an area of 5-33% of the whole photodiode area.

18. Photodetector array according to claim 13, wherein the microlenses are printed on the substrate with a ratio between a sag height and a residual height of between 0.05 and 0.9.

19. Photodetector array according to claim 13, wherein each of the photodetector cells is coupled with an individual read-out line for reading out a pixel data.

20. Photodetector array according to claim 19, wherein the read-out lines are arranged straight along an arrangement direction of the photodetector cells between neighboring rows of photodetector cells.

21. Photodetector array according to claim 20, wherein the power lines extend perpendicular to the read-out lines and run in a serpentine manner around the active areas of the photodetector cells.

22. Microscopic system with an optical system and a photodetector array according to claim 13.

23. Microscopic system according to claim 22, which uses an image scanning microscopy approach.

24. Microscopic system according to claim 23, which places the photodetector array at a pinhole plane of the microscopic system.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Embodiments are described in detail in conjunction with the accompanying drawings, in which:

[0029] FIG. 1 schematically shows a microscope system using a photodetector array.

[0030] FIG. 2 a top view on the photodetector array according to an embodiment of the present invention.

[0031] FIGS. 3a and 3b cross-sectional views of two adjacent pixels with deep-trench isolation between the photodetector cells according to alternative embodiments.

[0032] FIGS. 4a-4c simulation results for effective fill factors with square and circular microlenses for different sizes of active areas.

DESCRIPTION OF EMBODIMENTS

[0033] FIG. 1 schematically shows a microscopic system 1 as an exemplary application for a photodetector array. The microscopic system 1 has an optical objective 2 with a high numerical aperture to collimate light from a sample 3 into the microscope body 4. The collimated light L is projected onto a photodetector array 5 by means of a focusing tube lense 6 having a low numerical aperture. Further lenses 6a are included to cross light L on a pinhole 8. Illumination is made by an illumination device 7 which couples illumination light in onto the sample 3 by means of a semitransparent mirror 9.

[0034] For microscopy applications, a high sensitivity is required with a low noise, a low dark count rate, a high dark count rate uniformity and a low cross-talk between the neighboring pixels. For image-scanning microscopy which enables a theoretical image resolution improvement by a factor of 2 and an improved light collection, a fast photo-counting photodetector array with less than 1 μs integration time is required.

[0035] As the illumination intensity has to be restricted to reduce phototoxicity and photobleaching, a very sensitive detector array 5 is required. The most sensitive photodetector arrays use as photodiodes single-photon avalanche diodes, which have a photon counting ability and a high signal-to-noise ratio.

[0036] In the following, the photodetector array 5 is described using photodetector cells each provided with a single-photon avalanche diode.

[0037] FIG. 2 shows a top view on an arrangement of photodetector cells 11 of an embodiment of a photodetector array 10. The photodetector cells 11 are arranged in a semiconductor substrate as known in the art. The photodetector cells 11 each form a pixel of the photodetector array 10 and comprise single-photon avalanche diodes (SPAD) 12. The photodetector cells 11 are arranged in a hexagonal grid and therefore have a staggered arrangement which allows to compact the photodetector cells 11 with its small active areas 13 arranged in the center of the photodetector cells 11.

[0038] Particularly, the portion of the active area 13 of the photodetector cells compared to the area of the photodetector cell 11 is between 5 to 33%. This allows a larger distance between the active areas 13 of the photodetector cells so that the cross-talk between active areas 13 of neighboring photodetector cells 11 can be kept quite low.

[0039] Further, the relatively small active area 13 is beneficial as the photodiode capacitance and the charge flow through the photodiodes is significantly reduced.

[0040] The reduction of the photodiode active area 13 requires a large recovery of the fill factor by means of microlenses 14 for each of the photodiodes to maintain or to increase the sensitivity of the photodetector cells 11 of the photodetector array 10. The microlenses 14 may be formed as a single microlens layer on top of the semiconductor substrate. The area of light focused by the microlenses 14 which lies within the active area 13 is indicated by 16.

[0041] FIG. 2 shows the arrangement of the photodiodes with a circular active area 13 in a total photodetector cell area which is hexagonal or circular wherein on top of each of the photodetector cells 11 a microlens 14 is arranged. The microlenses 14 therefore also have a staggered arrangement corresponding to the arrangement of the photodetector cells 11. In difference to the active areas 13 of the photodetector cells 11, the microlenses 14 do not need to have to suppress optical or electrical interference. Therefore, none or only a small gap 15 between neighboring microlenses 14 may be provided.

[0042] In FIGS. 3a and 3b, cross-sectional views of two embodiments for the arrangement of two neighboring photodetector cells 11 are shown. The arrangement shows two neighboring photodetector cells 11 with its semiconductor depletion region divided by a deep-trench isolation 22 in between. On the surface S of the semiconductor substrate 20, a microlens array 21 is formed having a residual height R with a dome-shaped microlens 14 on top having a sag height H and being distanced from neighboring dome-shaped microlens 14 by the gap 15. This allows to form a focusing microlens 14 which collects and directs incoming photons onto the active area of the respective photodetector cell 11 underneath.

[0043] FIG. 3b shows a similar design, wherein between the microlenses 14 a reflective material 17 is included. Furthermore, the deep-trench isolation 22 may be made by a reflective material so that incoming photons may be reflected once or several times on the sidewalls formed by the reflective material.

[0044] FIG. 4a shows the effective fill factors for circular and square microlenses used with pixels with 1 μm active area radius and a native fill factor of 0.7% and 0.6% for a hexagonal and square grid of pixels, respectively. One can observe a much higher effective fill factor for circular microlenses on a hexagonal grid when compared to the square microlenses on a square grid. FIG. 4a shows the curves for two f numbers indicating different collimation of incident light.

[0045] FIG. 4b shows the effective fill factors for circular and square microlenses used with pixels with 3 μm active area radius and a native fill factor of 6.7% and 5.8% for a hexagonal and square grid of pixels, respectively.

[0046] FIG. 4c shows the effective fill factors for circular and square microlenses used with pixels with 6 μm active area radius and a native fill factor of 27% and 23.4% for a hexagonal and square grid of pixels, respectively.

[0047] As it is clear from FIG. 4b and FIG. 4c, the circular microlens on a hexagonal grid is also more robust to variations in microlens residual height. An active area radius of less than 3 μm is not desirable due to lateral shifts of microlenses with respect to the photodetector substrate. Lateral shifts decrease the robustness to microlens residual height variations.

[0048] The staggered arrangement of circular microlenses yields for a fill factor of 90.6% compared to a fill factor of 100% of squared microlenses of a rectangular arrangement. However, the photon loss of the rectangular-shaped microlenses is high, so that the staggered arrangement of circular or hexagonal microlenses 14 effectively directs more photons onto the active area 13 of the photodetector cell 11 than rectangular-shaped microlenses. Furthermore, circular or hexagonal microlenses turned out to be more robust with respect to lateral shifts and variations in microlens residual height R occurring in the microlens production, as clear from FIG. 4.

[0049] The photodetector active area 13 is provided with a circular shape which may reduce the edge electrical field compared to squared active areas. Preferably, the radius of the active area 13 of the photodetector cell 11 is smaller than 9 μm, more preferred smaller than 5 μm. The smaller the active area 13 is, the lower is the chance to capture impurities in the semiconductor material which further reduces the count of noisy pixels.

[0050] In contrast thereto, the size of the active area 13 of the photodetector cell should be not smaller than 3 μm due to a lateral shift variations for the placement of microlenses on top of the photodetector array 10 as well as due to microlens residual height variations.

[0051] The substantial area difference between the active area 13 and the photodetector cell 11 allows to accommodate the pixel electronics around the active area 13 which is composed of a quenching transistor, a gate for capacitive isolation of the photodiode, an active recharge mechanism and a photodiode disabling memory. The gate for capacitive isolation of the photodiode allows to have a capacitive isolation to reduce the cross-talk and after-pulsing due to a smaller amount of charge flowing through the photodiode. The quenching transistor and the gate for capacitive isolation need to be positioned as close as possible to the photodiode output contact so that the connections do not increase the capacitance significantly. The electronics for active recharge mechanism and a photodiode disabling memory can be placed farther away from the photodiode output contact since an increased capacitance do not affect the performance greatly.

[0052] As further shown in FIG. 2, the metal connections extend in a rectangular fashion between the substrate surface and the microlens layer. The read out lines 18 run along straight lines along the arrangement direction of the photodetector cells 11 between two neighboring rows of photodetector cells 11, respectively. The power lines 19 substantially extend perpendicular to the read out lines but are guided in a serpentine way around the active areas 13 of the photodetector cells 11.

[0053] Since high-performance applications require to read all photodetector signals in parallel within a small integration time and without delays between the read-outs of the individual photodiodes, each photodiode needs to have a single connection to the outside of the array. These connections should be as short as possible to enable high-speed and low-jitter data transmission so that read out lines 18 are used for the photodiode outputs, while the serpentine power lines 19 are used for the power supply of the photodetector cells 11.