Optical sensor device and method for producing an optical sensor device

Abstract

The invention relates to an optical sensor device (100) comprising a layer stack (110) having at least one substrate (111) which has at least a first optoelectronic component (112a) and a second optoelectronic component (112b). The layer stack (110) further comprises at least one dielectric layer (114), which is provided above the substrate (111); at least one metallization structure (120) having an uppermost metal layer (121), the at least one metallization structure (120) being provided between the first optoelectronic component (112a) and the second optoelectronic component (112b); and at least one passivation layer (115), which is provided above the dielectric layer (114). The layer stack (110) has at least one window (130) above the at least one metallization structure (120) and / or above the first optoelectronic component (112a) and / or the second optoelectronic component (112b). The invention further relates to a method for producing such an optical sensor device (100).

Description

[0001] Optical sensor device and manufacturing process for an optical sensor device

[0002] The invention relates to an optical sensor device and a method for manufacturing an optical sensor device.

[0003] SPADs (single-photon avalanche diodes) are highly sensitive photon-receiving elements that, due to the avalanche effect, provide a large amount of charge (approximately 10⁵–10⁶ electrons) with high temporal resolution upon activation. The activation efficiency, which describes the number of free charge carriers released per incident photon, is also known as the photon detection efficiency (PDE). It can be determined from the photon detection probability (PDP) and the fill factor (active area per total area, FF) according to the relationship PDE = PDP * FF.

[0004] A high photon detection efficiency (PDE) is crucial for the sensitivity and range of the sensor.

[0005] As a side effect, a secondary photon can be emitted when a primary photon triggers a charge carrier avalanche. If a first SPAD in a SPAD array emits such a secondary photon, it can be absorbed by a second SPAD adjacent to that first SPAD in the array. The secondary photon can reach the adjacent second SPAD via various pathways: directly, through reflection from the substrate on which the SPAD array is mounted, or through reflection in the so-called backend-of-line (BEOL) on the top side. These secondary photons trigger an avalanche effect in the second SPAD, generating a false signal. This results in a reduced dynamic range and blurred images. The effect is called optical crosstalk. The probability of crosstalk should be kept as low as possible.It increases with increasing PDE, because secondary photons are also converted more efficiently due to the indistinguishability of photons.

[0006] In the standard literature, in addition to the direct path, reflection at the back side is mainly considered as a coupling path, although this plays a subordinate role due to the small average path length for photons in the visible range.

[0007] The development of LiDAR cameras requires SPAD arrays which have a high photon detection efficiency (PDE).

[0008] Due to this requirement, the probability of crosstalk in such systems also increases in the same way as described above. Furthermore, the PDE is strongly wavelength-dependent due to constructive and destructive interference at the interfaces in the BEOL stack. Because of manufacturing tolerances in the layer thickness, the PDE maxima and minima do not always form for the same wavelengths, and their spectral position can differ even within a single wafer due to small local variations in BEOL thickness.

[0009] Several solutions for reducing crosstalk in SPAD arrays are already known. As described by Rech et al. (“Optical crosstalk in single photon avalanche diode arrays: a new complete model”, 2008 OSA, 9 June 2008 / Vol. 16, No. 12 / OPTICS EXPRESS 8383), the optical crosstalk in a SPAD array can be reduced by increasing the distance between the individual SPADs.

[0010] However, for increased resolution and a high fill factor or high photon detection efficiency (PDE), a dense arrangement of SPADs in an array is favored.

[0011] In Otake et al. (“A Back Illuminated 10pm SPAD Pixel Array Comprising Full Trench Isolation and Cu-Cu Bonding with Over 14% PDE at 940nm”), buried full trench isolation (FTI) is used to prevent optical crosstalk. In this process, the individual pixels of a SPAD array are separated from each other by metal-filled trenches that extend from the surface to the underside of the entire silicon substrate.

[0012] US patent 11,616,152 B2 also discloses optical isolation structures for preventing crosstalk in a SPAD array. In US patent 11,616,152 B2, the optical isolation structures can be formed in a ring around the respective SPAD. The isolation structure can be a hybrid isolation structure containing both a metal filler that reflects light and a dielectric filler with a low refractive index that reflects light.

[0013] A metal filler is used to reflect light back, preventing photons from traveling from a first SPAD to a neighboring second SPAD of the array and thus causing optical crosstalk.

[0014] A boundary layer to a dielectric filler with a different, in this case lower, refractive index can reflect light back into the SPAD, in particular causing total internal reflection, and keep the light in the active region of the SPAD.

[0015] The isolation structure can be configured as a single trench or have a deep trench at the rear and a deep trench isolation section at the front. The isolation structure can also contain a color-filtering material. Furthermore, US patent 11,616,152 B2 discloses microlenses for focusing light onto the individual SPADs of the array, as well as the application of a special antireflection layer. This results in improved light collection and disrupts the conditions for the development of interference effects within the SPAD array.

[0016] The fabrication of insulating trenches, as disclosed in Otake et al. and US 11,616,152 B2, requires expensive, advanced technological processes that are not available in every conventional CMOS fabrication process.

[0017] Besides the challenge of preventing optical crosstalk in a SPAD array while simultaneously optimizing photon detection efficiency and fill factor, a minimum metal density is always required for wafer fabrication, especially for chemical-mechanical planarization / polishing (CMP). However, no metal or metal-fill pattern can be placed in the active sensor area itself to avoid shadowing.

[0018] The object of the present invention is therefore to provide a solution for an improved optical sensor that eliminates optical crosstalk, wherein the optical sensor should have a high photon detection efficiency and a high fill factor, and wherein the solution should be cost-effective and not entail major technical changes to a standard method for manufacturing an optical sensor.

[0019] To solve the problem, a device and a method according to the independent claim are proposed.

[0020] Further advantageous embodiments of the invention can be found in the dependent claims, the description, and the figures.

[0021] The proposed solution involves an optical sensor device comprising a stack of layers.

[0022] The layer stack comprises at least one substrate which includes a plurality of optoelectronic components. In particular, the substrate is a semiconductor material. In particular, the substrate is silicon. In particular, the layer stack includes at least one first optoelectronic component and one second optoelectronic component. In particular, the at least one first optoelectronic component and the second optoelectronic component are each SPADs.

[0023] In particular, the layer stack has a gate oxide layer and / or a salicide protection layer. Specifically, the layer stack has a field oxide layer.

[0024] In particular, the layer stack comprises at least one dielectric layer arranged above the substrate within the layer stack. In particular, the layer stack comprises at least a first dielectric layer and a second dielectric layer separated from each other by an intermediate layer. In particular, the at least one dielectric layer is made of silicon dioxide. In particular, the intermediate layer is made of silicon nitride.

[0025] In particular, the layer stack comprises at least one metallization structure. Specifically, the at least one metallization structure has an uppermost metal layer.

[0026] In particular, the uppermost metal layer is a contact point for electrically connecting the optical sensor device. Specifically, the uppermost metal layer is a pad structure for bonding the optical sensor device.

[0027] In particular, at least one metallization structure is arranged between the first optoelectronic component and the second optoelectronic component.

[0028] In particular, at least one of the at least one metallization structure comprises a plurality of superimposed metal layers. In particular, the at least one metallization structure comprises a plurality of plugs. In particular, the plugs are made of metal. In particular, the superimposed metal layers are connected to each other via the plugs.

[0029] In particular, the layer stack comprises a polysilicon layer. Specifically, the polysilicon layer serves as an etch stop during the metallization process. Specifically, the metallization structure is arranged on the polysilicon layer.

[0030] In particular, the layer stack comprises a passivation layer. Specifically, the passivation layer is arranged above the at least one dielectric layer. Specifically, the passivation layer is arranged on the top surface of the layer stack. Specifically, the passivation layer is made of silicon nitride.

[0031] In particular, the layer stack has at least one window over the at least one metallization structure. In particular, the layer stack additionally or alternatively has at least one window over the first optoelectronic component and / or at least one window over the second optoelectronic component. In particular, the at least one window extends over a surface that includes a projection plane of the first optoelectronic component and / or a projection plane of the second optoelectronic component and / or a projection plane of the at least one metallization structure.

[0032] In particular, at least one window above the metallization structure is designed to at least partially expose the uppermost metal layer of the metallization structure. Specifically, the metallization structure is arranged between the first optoelectronic component and the second optoelectronic component.

[0033] In particular, the at least one window above the first optoelectronic component and / or the at least one window above the second optoelectronic component each have a convex curvature. Specifically, the convex curvature is designed to focus light onto the respective optoelectronic component.

[0034] In particular, at least one window is arranged above the at least one metallization structure, while no window is arranged above the first optoelectronic component and / or the second optoelectronic component.

[0035] In particular, at least one window is arranged above the first optoelectronic component and / or the second optoelectronic component, while no window is arranged above the at least one metallization structure.

[0036] In particular, at least one window is arranged above the metallization structure, as well as above the first optoelectronic component and the second optoelectronic component.

[0037] By arranging at least one metallization structure between the first optoelectronic component and the second optoelectronic component, optical crosstalk between the first optoelectronic component and the second optoelectronic component is largely prevented if no photons can pass through the layer stack below, above or through gaps in the metallization structure from one optoelectronic component to the other.

[0038] For this purpose, the metallization structure begins on the substrate in which the optoelectronic components are arranged. Furthermore, the metallization structure has a continuous, opaque structure.

[0039] If the dielectric layer and / or the passivation layer is also applied above the metallization structure, a raised area of ​​this material forms above the metallization structure. This raised area allows photons to travel from one optoelectronic component to another, resulting in crosstalk. To prevent this, a window is placed above the metallization structure, exposing the top metal layer. This ensures that no material from the layer stack remains above the metallization structure through which photons can pass.

[0040] Alternatively or additionally, at least one window is arranged above the first optoelectronic component and / or above the second optoelectronic component. If there is a raised area of ​​material above the at least one metallization structure arranged between the first and second optoelectronic components, the at least one window above the first optoelectronic component and / or the at least one window above the second optoelectronic component is designed such that the at least one window above the first optoelectronic component and / or the at least one window above the second optoelectronic component exposes the edge of the uppermost metal layer of the at least one metallization structure facing the first and / or the second optoelectronic component.In this way, a path through the material of the layer stack between the optoelectronic components is blocked for photons.

[0041] An arrangement of at least one window above the first optoelectronic component and / or above the second optoelectronic component disrupts the conditions for destructive / constructive interference within the layer stack and thus leads to improved, more homogeneous quantum efficiency.

[0042] The convex curvature of the window above the first optoelectronic component and / or above the second optoelectronic component also acts as a converging lens, focusing incoming light onto the respective optoelectronic component. This increases detection efficiency because the more effective sub-areas of the SPAD in the center are targeted.

[0043] In one embodiment, the at least one window in a projection plane of the first optoelectronic component and / or the at least one window in a projection plane of the second optoelectronic component each has a predetermined width. In particular, the width of the at least one window corresponds to a value of at least 3 pm.

[0044] In one embodiment, the width of the at least one window in the projection plane of the first optoelectronic component and / or in the projection plane of the second optoelectronic component corresponds to a value of lOpm. In one embodiment, the at least one window arranged above the first optoelectronic component is bounded on at least one side by the at least one metallization structure.

[0045] In one embodiment, the at least one window arranged above the second optoelectronic component is limited at least on one side by the at least one metallization structure.

[0046] If the at least one window located above the first optoelectronic component and / or the at least one window located above the second optoelectronic component borders the at least one metallization structure and exposes the edge of the uppermost metal layer of the at least one metallization structure facing the respective optoelectronic component, then the fill factor FF of the respective optoelectronic component is effectively increased, and thus also the photon detection efficiency PDE.

[0047] In one embodiment, the metallization structure comprises a plurality of metal layers and plugs.

[0048] In particular, the metal layers contain aluminum. In particular, the metal layers are made of copper.

[0049] In particular, the plugs are made of an electrically conductive material. In particular, the plugs are made of metal. In particular, the plugs are made of tungsten. In particular, the plugs are made of copper. In particular, the plugs contain aluminum.

[0050] In one embodiment, the layer stack of the optical sensor device has a poly-silicon layer.

[0051] In particular, the polysilicon layer is arranged on the field oxide layer. Specifically, the field oxide layer is arranged on the substrate of the layer stack.

[0052] In one embodiment, the metallization structure is arranged on the poly-silicon layer.

[0053] In particular, the metallization structure is composed of a plurality of metal layers and plugs. Specifically, a plug of the metallization structure is based on the polysilicon layer. In one embodiment, the metallization structure is arranged in the optical sensor device such that a light path through the layer stack below the uppermost metal layer of the metallization structure is blocked by the metallization structure.

[0054] In particular, the metallization structure is composed of multiple metal layers and plugs. Specifically, the metallization structure is arranged on the polysilicon layer.

[0055] The invention further relates to a method for manufacturing the optical sensor device, comprising the steps:

[0056] • Providing a layer stack with at least one substrate, comprising at least one first optoelectronic component and one second optoelectronic component, at least one dielectric layer, and at least one metallization structure with a top metal layer,

[0057] • Ablation or removal of material from the layer stack (110) by ablation or removal in at least one local area above the at least one metallization structure and / or above the first optoelectronic component and / or the second optoelectronic component, wherein a mask is applied to the layer stack and the local ablation or removal is carried out using the mask, such that

[0058] - a top metal layer (metal top or metal top 1) of the at least one metallization structure, which is arranged between the first optoelectronic component and the second optoelectronic component, is at least partially exposed, and / or

[0059] - at least one window with a convex curvature is formed over the first optoelectronic component and / or the second optoelectronic component, such that the convex curvature has a lens structure designed to focus light in the direction of the respective optoelectronic component.

[0060] In particular, the local ablation or removal is carried out using the mask to form the at least one window over the first optoelectronic component and / or the at least one window over the second optoelectronic component, such that the convex curvature is automatically formed on the underside of the respective window. Specifically, a mask is used for the local ablation or removal of material, wherein a value for the width of the at least one window over the first optoelectronic component and / or a value for the width of the at least one window over the second optoelectronic component is predetermined such that, during the local ablation or removal of material, the convex curvature is automatically formed in the window in the projection plane of the first optoelectronic component and / or in the window in the projection plane of the second optoelectronic component.

[0061] In particular, the removal or ablation of the material is carried out using a suitable mask via an etching process. This can specifically involve passivation to expose contact points (pads).

[0062] In one embodiment, the mask is chosen such that the window over the first optoelectronic component and / or the window over the second optoelectronic component each have a width of at least 3 pm after ablation or removal of the material from the layer stack.

[0063] By having a predetermined width of at least 3pm for the windows above the optoelectronic components, a convex curvature automatically results at the bottom of the respective window during the removal of the layer stack material, which serves as a converging lens for the respective optoelectronic component below.

[0064] In one embodiment, the mask is chosen such that the window over the first optoelectronic component and / or the window over the second optoelectronic component has a width of 10pm.

[0065] In particular, the mask is chosen such that the window over the first optoelectronic component and / or the window over the second optoelectronic component each have a width of 10 pm.

[0066] If the windows above the optoelectronic components have a width of approximately 10 pm, a convex curvature automatically forms at the bottom of each window during the removal of the layer stack material using a suitable mask, particularly by means of an etching process. This curvature is shaped such that incident light is focused onto the corresponding optoelectronic component below. In one embodiment, the mask is selected such that the window above the first optoelectronic component borders at least one metallization structure on at least one side.

[0067] In one embodiment, the mask is chosen such that the window above the second optoelectronic component borders on at least one metallization structure on at least one side.

[0068] In particular, the mask is chosen such that the window extends over a surface which includes a projection plane over the first optoelectronic component and / or a projection plane over the second optoelectronic component and / or a projection plane over the at least one metallization structure which is arranged between the first optoelectronic component and the second optoelectronic component.

[0069] In the optical sensor device according to the invention, an optical path through the layer stack (the backend-of-line (BEOL)) is disturbed, particularly in the passivation area, by using a conventional passivation aperture in the sensor area.

[0070] The sensor device according to the invention reduces optical crosstalk and / or, depending on the embodiment, can prevent unwanted constructive and destructive interference.

[0071] The optimization of the various parameters (photon detection efficiency / interference or crosstalk) can be achieved using a simple design / layout methodology and is feasible in all technologies.

[0072] By arranging a metallization structure to prevent crosstalk between the first optoelectronic component and the second optoelectronic component, the problem of shading, which occurs when a metallization structure is placed in the active sensor area, is avoided.

[0073] The production of the optical sensor device according to the invention does not require any additional process steps or non-standard masks compared to the prior art.

[0074] Furthermore, the optical sensor device according to the invention is area-neutral compared to the prior art, since the existing gap between the optoelectronic components, which is already present for electrical insulation, can be used for the metallization structure or the etching trench (material removal) above the metallization structure. Further advantageous embodiments, features, and functions of the invention are explained in connection with the examples shown in the figures.

[0075] This shows:

[0076] Fig. 1 schematic representation of an optical sensor device according to the prior art;

[0077] Fig. 2 schematic representation of an optical sensor device according to the invention;

[0078] Fig. 3 optical sensor device according to the invention in a second embodiment;

[0079] Fig. 4 optical sensor device according to the invention in a third embodiment;

[0080] Fig. 5 optical sensor device according to the invention in a fourth embodiment;

[0081] Fig. 6 optical sensor device according to the invention in a fifth embodiment;

[0082] Fig. 7 optical sensor device according to the invention in a sixth embodiment;

[0083] Fig. 8 optical sensor device according to the invention in a seventh embodiment; and

[0084] Fig. 9 shows the measured quantum efficiency for the optical sensor devices according to Fig. 1-8.

[0085] Fig. 1 shows a schematic representation of an optical sensor device 100 according to the prior art.

[0086] The optical sensor device 100 is constructed in the form of a layer stack 110.

[0087] The layer stack comprises a substrate 111 in which a plurality of optoelectronic components 112 are embedded. A first optoelectronic component 112a and a second optoelectronic component 112b are configured as SPAD diodes.

[0088] In this embodiment, secondary photons that can be emitted by a first SPAD diode 112a can reach a second SPAD diode 112b within the layer stack 110 and vice versa.

[0089] The photons can travel through the layer stack 110 from one SPAD diode 112 to the other SPAD diode via different paths, as shown by the dashed arrows in Fig. 1:

[0090] (1) Reflection at an intermediate layer 113 (e.g. SiN) of the layer stack 110

[0091] (2) Reflection at an interface between a dielectric layer 114 (e.g. SiO2) and a passivation layer 115 (e.g. SiN)

[0092] (3) Multiple reflections within the passivation layer 115

[0093] In this way, crosstalk can occur between the optoelectronic components 112a and 112b. Besides the crosstalk, which leads to photon misdetections, constructive and destructive interference occurs in the BEOL stack of the optical sensor device according to the prior art, and this interference is strongly wavelength-dependent. This is illustrated in Fig. 1 by a plot of the measured normalized quantum efficiency (QE), which is proportional to the photon detection efficiency (PDE), as a function of wavelength.

[0094] To manufacture the optical sensor device 100 according to Fig. 1, a plurality of optoelectronic components 112 are arranged on a substrate 111, which is usually made of silicon, and a gate oxide and / or salicide protection layer 116 is applied over it.

[0095] A first dielectric layer 114, an intermediate layer 113 and a second dielectric layer 114 are stacked on top of each other on the gate oxide and / or salicide protection layer 116.

[0096] The dielectric layer 114 is typically a silicon dioxide layer (SiO₂). The intermediate layer 113, which separates the first dielectric layer 114 from the second dielectric layer 114, is typically made of silicon nitride (SiN).

[0097] Subsequently, a passivation layer 115, which is usually also a silicon nitride (SiN) layer, is applied to these stacked layers.

[0098] Fig. 2 shows a first embodiment of an optical sensor device 100 according to the invention. This embodiment differs from the prior art shown in Fig. 1 in that the optical sensor device 100 additionally has a metallization structure 120 with an uppermost metal layer 121. The metallization structure 120 is arranged between the first optoelectronic component 112a and the second optoelectronic component 112b.

[0099] In the illustrated embodiment, the layer stack 110 comprises a gate oxide and / or salicide protection layer 116 and field oxide layers 117. A poly-silicon layer 118 is arranged on the field oxide layer 117 between the first optoelectronic component 112a and the second optoelectronic component 112b, on which the metallization structure 120 is arranged.

[0100] In the illustrated embodiment, the metallization structure 120 comprises a plurality of metal layers and plugs made of an electrically conductive material. Here, the lowest plug of the metallization structure 120 rests on the polysilicon layer 118. The metallization structure 120 is arranged such that the plugs and the metal layers are stacked alternately on top of each other. The uppermost metal layer 121 of the metallization structure 120, which in this example serves as a contact point, is located in the layer stack 110 within the dielectric layer 114 below the passivation layer 115.

[0101] By removing the passivation layer 115 and the portion of the dielectric layer 114 located above the metallization structure 120, a window 130 is created in the optical sensor device 100, which at least partially exposes the contact point 121.

[0102] If the first optoelectronic component 112a (first SPAD 112a) emits secondary photons, these cannot pass through the metallization structure 120 to the second optoelectronic component 112b (second SPAD 112b) and vice versa.

[0103] Because the top surface of the uppermost metal layer 121 of the metallization structure 120 is at least partially exposed by creating the window 130, so that at least one area is formed on the surface of the uppermost metal layer 121 in which the dielectric layer 114 above the metallization structure 120 has been completely removed, photons cannot pass above the metallization structure 120 from one optoelectronic component 112a, 112b to the other optoelectronic component 112a, 112b, as shown by the dashed arrows in Fig. 2.

[0104] In this way, crosstalk or optical crosstalk within the optical sensor device 100 is prevented.

[0105] However, as a measurement of the normalized quantum efficiency (QE) as a function of the wavelength of the light entering the optical sensor device 100 shows (see diagram in Fig. 2), strongly wavelength-dependent constructive and destructive interferences still occur in the BEOL stack in the same way as in the optical sensor device according to the prior art shown in Fig. 1.

[0106] By means of the embodiment of the optical sensor device 100 according to the invention shown in Fig. 2, crosstalk can thus be prevented, while the presence of strongly wavelength-dependent constructive and destructive interferences in the BEOL stack remains unchanged compared to the embodiment shown in Fig. 1 according to the prior art.

[0107] The manufacturing process for producing an optical sensor device 100 according to Fig. 2 differs from that for a device according to the prior art according to Fig. 1 in that a poly-silicon layer 118 and the metallization structure 120 are arranged between the first optoelectronic component 112a and the second optoelectronic component 112b on the field oxide layer 117, which is embedded in the first dielectric layer 114, the intermediate layer 113 and the second dielectric layer 114, and subsequently material is locally removed using a mask.

[0108] In the embodiment shown in Fig. 2, the metallization structure 120 comprises three metal layers, wherein the uppermost metal layer 121 corresponds to a contact point.

[0109] The second dielectric layer 114 completely covers the uppermost metal layer 121. Likewise, the passivation layer 115 is applied to the entire surface of the optical sensor device 100.

[0110] Subsequently, a mask is used and the passivation layer 115 and the dielectric layer 114 are locally removed by an etching process until the surface of the uppermost metal layer 121 is at least partially exposed.

[0111] Fig. 3 shows a second embodiment of the optical sensor device 100 according to the invention.

[0112] The embodiment shown in Fig. 3 is manufactured in an equivalent manner to the embodiment shown in Fig. 2.

[0113] In contrast to the embodiment shown in Fig. 2, the metallization structure 120 shown in Fig. 3 comprises a fourth metal layer.

[0114] Due to the increased metallization structure 120, a raised area 119 forms above the metallization structure 120 during the manufacturing process when the dielectric layer 114 and the passivation layer 115 are applied.

[0115] In the embodiment shown in Fig. 3, instead of removing the passivation layer 115 and part of the dielectric layer 114 above the metallization stack 120, a window 130 is formed above each of the first optoelectronic component 112a and the second optoelectronic component 112b by using a corresponding adapted mask in the projection plane of these optoelectronic components 112a and 112b.

[0116] Here, the width 131 of the windows 130 corresponds to a specific predefined value. Specifically, the width 131 of the windows 130 is always greater than 3 pm. Specifically, the width 131 of the windows 130 is always approximately 10 pm.

[0117] Due to this predetermined value for the width 131 of the windows 130, a convex curvature 132 automatically forms in the dielectric layer 114 in the projection plane above the optoelectronic components 112 when the passivation layer 115 and the portion of the dielectric layer 114 are locally removed. Because of these convex curvatures 132, the dielectric layer 114 has a lens structure, so that light entering the optical sensor device 100 through the windows 130 is focused by means of the convex curvatures 132.

[0118] As a measurement of the normalized quantum efficiency (QE) as a function of the wavelength of the light incident on the optical sensor device 100 shows (see diagram in Fig. 3), interference of the light within the optical sensor device 100 exhibits a smoothed dependence on the wavelength without local minima or maxima. This effect arises because the volume in which photons can interfere with each other has been reduced by the windows 130 above the optoelectronic components 112.

[0119] In contrast to the embodiment shown in Fig. 2, the metallization structure 120 in the variant shown in Fig. 3 comprises a further metal layer, so that the surface of the uppermost metal layer 121 of the metallization structure 120 arranged in the layer stack 110 on the field oxide layer 117 is located at least approximately on the same plane as the boundary layer between dielectric layer 114 and passivation layer 115, wherein, due to the manufacturing process of the layer stack 110, a local elevation 119 of the layer stack 110 is formed above the metallization structure 120 by the application of the dielectric layer 114 and the passivation layer 115.

[0120] Photons emitted from the first optoelectronic component 112a cannot pass through the metallization structure 120 to the second optoelectronic component 112b and vice versa.

[0121] As indicated by the dashed arrows in Fig. 3, photons can, however, pass through the raised section 119 of the layer stack 110 above the metallization structure 120. In this way, it is possible for photons to travel from the first optoelectronic component 112a to the second optoelectronic component 112b, or vice versa, by reflection at an interface.

[0122] In this embodiment, there is therefore a probability of crosstalk between the optoelectronic components 112.

[0123] The embodiment of the optical sensor device 100 according to the invention shown in Fig. 3 therefore prevents constructive and destructive interference with strong oscillation within the optical sensor device 100, which is strongly dependent on the wavelength of the incident light. However, crosstalk between the individual optoelectronic components 112 is not effectively prevented, since photons can pass from the first optoelectronic component 112a to the second optoelectronic component 112b, or vice versa, via the raised area 119 above the metallization structure 120.

[0124] This disadvantage is eliminated in the embodiment shown in Fig. 4.

[0125] The embodiment shown in Fig. 4 differs from the variant shown in Fig. 3 only in that the distance between the optoelectronic components 112 and the metallization structure 120, as well as the width 131 of the windows 130 above the optoelectronic components 112, are selected such that the windows 130 towards the metallization structure 120 are limited by the metallization structure 120. Using a suitably adapted mask, the passivation layer 115 and partially the dielectric layer 114 are locally removed such that material is removed above the respective optoelectronic components 112, but the elevation 119 of the layer stack 110 formed by the dielectric layer 114 and the passivation layer 115 remains at least partially above the metallization structure 120.By forming the windows 130 above the optoelectronic components 112, the edges of the uppermost metal layer 121 of the metallization structure 120 facing the optoelectronic components 112 are exposed. However, the surface of the uppermost metal layer 121 is not, or not completely, exposed by the removal of the material.

[0126] The width 131 of the windows 130 is again chosen such that when removing the material to create the windows 130, a convex curvature l32 is formed in the projection plane above the optoelectronic components 112, which focuses light entering through the windows 130 onto the respective optoelectronic component 112a, 112b.

[0127] As in the embodiment shown in Fig. 3, the shaping of the windows 130 with concave curvature 132 above the optoelectronic components 112 results in a smoothed dependence on the wavelength of the light incident on the optical sensor device 100 for a measured normalized quantum efficiency (QE) as a function of the wavelength of the light incident on the optical sensor device 100, without local minima or maxima for interference of the light within the optical sensor device 100.

[0128] Because the windows 130 extend to the edges of the metallization structure 120, photons (solid arrows) incident on the optical sensor device 100 near the metallization structure 120 do not have to pass through the passivation layer 115 on their way to the respective optoelectronic component 112, as in the embodiment shown in Fig. 3, but only through a significantly smaller portion of the dielectric layer 114. This effectively increases the fill factor (FF) of the optical sensor device 100 by enlarging the area through which photons can penetrate the SPADs.

[0129] By exposing the edges of the uppermost metal layer 121 of the metallization structure 120 that point towards the optoelectronic components 112, the connection between the first optoelectronic component 112a and the second optoelectronic component 112b via the dielectric layer 114 and the passivation layer 115 is interrupted. As a result, photons emitted by the first optoelectronic component 112a can no longer pass above the metallization structure 120 into the optoelectronic component 112b, and vice versa (see dashed arrows in Fig. 4). Likewise, these photons cannot pass through the metallization structure 120 itself.

[0130] In this way, crosstalk between the optoelectronic components 112 is prevented.

[0131] The embodiment shown in Fig. 5 differs from the variant shown in Fig. 4 only in that a different mask was used for removing the material to form windows 130 above the optoelectronic components 112, so that material was removed simultaneously above the optoelectronic components 112 and above the metallization structure 120. Effectively, the embodiment shown in Fig. 5 thus has an opening that extends over the first optoelectronic component 112a, the metallization structure 120, and the second optoelectronic component 112b.

[0132] While the ablation process is halted by the exposure of the top metal layer 121 of the metallization structure 120, the material of the dielectric layer 114 above the optoelectronic components 112 continues to be ablated. This creates windows 130 above the optoelectronic components 112, which extend deeper into the material of the optical sensor device 100 than to the top surface of the top metal layer 121. Thus, the ablation of the material forms a structure with two different heights hl and h2 above the metallization layer 120 and the optoelectronic components 112.

[0133] In principle, it is possible for the optical sensor device 100 to have further windows 130 by removing material using a suitable mask. These windows may have a third height h3 or a further height that differs from heights hl and h2. By simultaneously removing material above the metallization structure 120 and the optoelectronic components 112, the windows 130 created above the optoelectronic components 112 abut the edges of the uppermost metal layer 121 adjacent to the optoelectronic components 112.

[0134] Thus, the removal process leaves no material on the sides of the uppermost metal layer 121 and above the metallization structure 120 through which photons could pass from the first optoelectronic component 112a to the second optoelectronic component 112b or vice versa.

[0135] This completely prevents crosstalk.

[0136] As in the embodiment shown in Fig. 4, the fill factor (FF) of the optical sensor device is significantly increased compared to the device shown in Fig. 1 according to the prior art by extending the windows 130 to the edges of the uppermost metal layer 121.

[0137] When selecting the mask for removing material to form the windows 130, care is taken to ensure that the width 131 of the windows 130 has a predetermined value on both sides of the uppermost metal layer 121, where the width 131 of the windows 130 corresponds to an amount >3 pm. In particular, the width 131 of the windows 130 is approximately 10 pm.

[0138] Due to the predetermined value for the width 131 of the windows 130, a convex curvature 132 is formed in the projection plane of the optoelectronic components 112 in the dielectric layer 114 above the optoelectronic components 112 when the material is removed. This convex curvature 132 has the property of a converging lens. When light enters through the windows 130, it is focused by the convex curvature 132 onto the respective underlying optoelectronic component 112a, 112b.

[0139] The embodiment shown in Fig. 5 thus prevents crosstalk between the first optoelectronic component 112a and the second optoelectronic component 112b.

[0140] Increasing the effective fill factor also increases the photon detection efficiency (PDE), which is proportional to the fill factor (FF).

[0141] As a measurement of the normalized quantum efficiency (QE) as a function of the wavelength of the light incident on the optical sensor device 100 shows, a homogenized dependence on the wavelength without local minima or maxima for interference of the light within the optical sensor device 100 is obtained here in an equivalent manner as in the embodiments of Figs. 3 and 4.

[0142] This means that the PDE is no longer strongly wavelength-dependent due to constructive and destructive interference in the BEOL stack.

[0143] The embodiment shown in Fig. 6 differs from the variant shown in Fig. 5 in that a lower metal layer 122 of the metal structure 120, located below the uppermost metal layer 121, projects into the window 130 arranged in the projection plane of the first optoelectronic component 112a.

[0144] Due to the arrangement and dimensions of the underlying metal layer 122, during the manufacturing process of the window 130, which in the illustrated embodiment extends over the projection plane of the first optoelectronic component 112a, the metallization structure 120, and over the projection plane of the second optoelectronic component 112b, at least the edge of the underlying metal layer 122 facing the first optoelectronic component 112a is exposed. In this way, when material is removed to manufacture the window 130, a recess is created which has three different heights hl, h2, h3.

[0145] It is also possible that the deeper metal layer 122 also protrudes into the window area of ​​the window 130 in the projection plane of the second optoelectronic component 112b.

[0146] Similarly, in other design variants, deeper, recessed metal layers of the metallization structure 120 can project into the window 130, either alternatively or additionally. Thus, it is possible for the recess of the window 130 to have more than three different heights.

[0147] The embodiment shown in Fig. 7 differs from the variant shown in Fig. 5 in that the optical sensor device 100 has a first metallization structure 120a and a second metallization structure 120b.

[0148] The second metallization structure 120b is equivalent to that shown in Fig. 5, consisting of a plurality of metal layers and plugs, and is arranged between the first optoelectronic component 112a and the second optoelectronic component 112b.

[0149] In the illustrated embodiment, the first metallization structure 120a comprises only the uppermost metal layer 121. In the embodiment shown in Fig. 7, the optical sensor device 100 has both a window 130 above the first metallization structure 120a, which at least partially exposes the uppermost metal layer 121 of the first metallization structure 120a, and a further window 130, which extends, equivalently to Fig. 5, over the projection plane of the first optoelectronic component 112a, the second metallization structure 120b and over the projection plane of the second optoelectronic component 112b, whereby the uppermost metal layer 121 of the second metallization structure 120b is exposed.

[0150] Fig. 8 shows a further embodiment of the optical sensor device 100. The optical sensor device 100 has a window 130 with a convex curvature 132 in the respective projection plane of the first optoelectronic component 112a and the second optoelectronic component 112b, which serves as a converging lens for the respective underlying optoelectronic component 112.

[0151] Furthermore, the optical sensor device 100 has a third window 130 above a metallization structure 120 arranged between the first optoelectronic component 112a and the second optoelectronic component 112b, which at least partially exposes the uppermost metal layer 121 of the metallization structure 120.

[0152] This arrangement of the windows 130 and the metallization structure 120 prevents crosstalk between the first optoelectronic component 112a and the second optoelectronic component 112b. Furthermore, destructive and constructive interference in the material above the optoelectronic components 112 is minimized, so that a measurement of the normalized quantum efficiency (QE) as a function of the wavelength of the incident light shows smoothed behavior.

[0153] In the embodiment shown in Fig. 8, the width 131 of the window 130 above the metallization structure 120 is wider than the top surface of the uppermost metal layer 121 of the metallization structure 120, so that the top surface of the uppermost metal layer 121 is completely exposed and funnels 133 form at the edges of the uppermost metal layer 121 during the manufacture of the window 130 above the metallization structure 120.

[0154] In the variant shown here, the window 130 has a funnel 133 on each side facing the uppermost metal layer 121. However, it is also possible that, by positioning the window 130 differently or by choosing a different width 131 of the window 130, a funnel 133 is formed only on one side facing the uppermost metal layer 121. In the embodiment shown here, the window 130 has a first funnel 133a and a second funnel 133b above the metallization structure 120.

[0155] In this embodiment, the underside of the uppermost metal layer 121 is chosen to be wider than the top side of the uppermost metal layer 121.

[0156] In the illustrated example, the window 130 is positioned above the metallization structure such that, in the projection plane, it does not extend beyond the edge of the underside of the uppermost metal layer 121, so that the first funnel 133a terminates on a chamfered edge of the uppermost metal layer 121. On the other side of the uppermost metal layer 121, in this example, the window 130 extends beyond the edge of the underside of the uppermost metal layer 121 in the projection plane, so that during the fabrication of the window 130, deeper material of the dielectric layer 114 is also removed.

[0157] In this example, a deeper metal layer 122 of the metallization structure 120, which is composed of a plurality of metal layers and plugs, is wider on one side.

[0158] Due to the design of the metallization structure 120 and the positioning and width 131 of the window 130 above it, a second funnel 133b is created when material is removed to manufacture the window 130, which ends on the top of the lower metal layer 122 and thus partially exposes the top of the lower metal layer 122.

[0159] The embodiment shown in Fig. 8 also has windows 130 above the optoelectronic components 112, which, due to the specifically selected width 131 of the windows 130, each have a convex curvature 132. These curvatures 132 serve to focus incident light onto the respective optoelectronic component 112 below.

[0160] Because the top surface of the uppermost metal layer 121 of the metallization structure 120 arranged between the two optoelectronic components 112 shown here is at least partially exposed, and the structure of the metallization prevents a light path within the material of the optical sensor device between the two optoelectronic components 112, crosstalk between the first optoelectronic component 112a and the second optoelectronic component 112b is prevented.

[0161] Furthermore, the removal of material above the optoelectronic components 112 results in a smoothed behavior of the measured normalized quantum efficiency (QE) as a function of the wavelength of the incident light. The embodiments of the optical sensor device according to the invention shown in Figures 2 to 8 can be combined with one another.

[0162] It is thus possible that several optical sensor devices are arranged on a common substrate, wherein several optical sensor devices of a single embodiment according to the invention are arranged on this substrate according to one of the Figs. 2 to 8.

[0163] However, it is also possible that several optical sensor devices, at least partially different from each other, are arranged on a common substrate according to the embodiments shown in Figs. 2 to 8.

[0164] Fig. 9 shows the normalized quantum efficiency (QE) measured for the devices depicted in Figs. 1 to 8 as a function of the wavelength of the incident light. The solid line represents an exemplary measurement curve for the normalized quantum efficiency for the devices shown in Figs. 1 and 2. The dotted line represents an exemplary measurement curve for the normalized quantum efficiency for the devices shown in Figs. 3 to 8.

[0165] For the devices according to Fig. 1 and 2, the measurement curve for the normalized quantum efficiency shows a strongly oscillating behavior due to constructive and destructive interference in the dielectric layer stack (BEOL stack) above the optoelectronic components of the optical sensor device.

[0166] For the devices shown in Figures 3 to 8, the measurement curve for the normalized quantum efficiency, on the other hand, shows a smoothed dependence on the wavelength without strong local minima or maxima. This is achieved by removing material to create windows above the optoelectronic components.

[0167] Reference symbol list

[0168] 100 Optical sensor device

[0169] 110 layer stacks

[0170] 111 Substrat

[0171] 112 optoelectronic component

[0172] 112a first optoelectronic component

[0173] 112b second optoelectronic component

[0174] 113 Intermediate shift

[0175] 114 dielectric layer

[0176] 115 Passivation layer

[0177] 116 Gate oxide and / or salicide protection layer

[0178] 117 Field oxide layer

[0179] 118 Poly-silicon layer

[0180] 119 local increase in the layer stack

[0181] 120 Metallization structure

[0182] 120a first metallization structure

[0183] 120b second metallization structure

[0184] 121 uppermost metal layer of the metallization structure

[0185] 122 Deeper metal layer of the metallization structure

[0186] 130 windows

[0187] 131 Width of the window

[0188] 132 convex curvature

[0189] 133 funnels

[0190] 133a first funnel

[0191] 133b second funnel hl first height h2 second height h3 third height

Claims

Claims 1. Optical sensor device (100) comprising a layer stack (110) with • at least one substrate (111) comprising at least one first optoelectronic component (112a) and one second optoelectronic component (112b), • at least one dielectric layer (114) which is arranged above the substrate (111), • at least one metallization structure (120) with a top metal layer (121), and • at least one passivation layer (115) arranged above the dielectric layer (114), wherein the layer stack (110) has at least one window (130) above the at least one metallization structure (120) and / or above the first optoelectronic component (112a) and / or the second optoelectronic component (112b), wherein • the at least one window (130) above the at least one metallization structure (120) is designed to at least partially expose the uppermost metal layer (121) of the at least one metallization structure (120), which is arranged between the first optoelectronic component (112a) and the second optoelectronic component (112b), and / or • the at least one window (130) above the first optoelectronic component (112a) and / or the at least one window (130) above the second optoelectronic component (112b) have a convex curvature (132), wherein the convex curvature (132) is designed to focus light onto the respective optoelectronic component (112).

2. Optical sensor device (100) according to claim 1, characterized in that the at least one window (130) in a projection plane of the first optoelectronic component (112a) and / or the at least one window (130) in a projection plane of the second optoelectronic component (112b) each has a predetermined width (131), wherein the width (131) of the window (130) corresponds to an amount of at least 3pm.

3. Optical sensor device (100) according to claim 2, characterized in that the width (131) of the at least one window (130) corresponds to an amount of 10pm.

4. Optical sensor device (100) according to one of the preceding claims, characterized in that that the at least one window (130) arranged above the first optoelectronic component (112a) is bounded at least on one side by the at least one metallization structure (120).

5. Optical sensor device (100) according to one of the preceding claims, characterized in that the at least one window (130) arranged above the second optoelectronic component (112b) is limited at least on one side by the at least one metallization structure (120).

6. Optical sensor device (100) according to one of the preceding claims, characterized in that the metallization structure (120) comprises a plurality of metal layers and plugs.

7. Optical sensor device (100) according to one of the preceding claims, characterized in that the layer stack (110) comprises a poly-silicon layer (118).

8. Optical sensor device (100) according to claim 7, characterized in that the metallization structure (120) is arranged on the poly-silicon layer (118).

9. Optical sensor device (100) according to one of claims 6 to 8, characterized in that the metallization structure (120) is arranged in the optical sensor device (100) such that a light path through the layer stack (110) below the uppermost metal layer (121) of the metallization structure (120) is blocked by the metallization structure (120).

10. Method for manufacturing an optical sensor device (100) according to one of the preceding claims, comprising the steps: • Providing a layer stack (110) with at least one substrate (111) comprising at least one first optoelectronic component (112a) and a second optoelectronic component (112b), at least one dielectric layer (114), and at least one metallization structure (120) with a top metal layer (121), • Ablation or removal of material from the layer stack (110) in at least one local area above the at least one metallization structure (120) and / or above the first optoelectronic component (112a) and / or the second optoelectronic component (112b), wherein a mask is applied to the layer stack (110) and the local ablation or removal is carried out using the mask, such that - the uppermost metal layer (121) of the at least one metallization structure (120) which is arranged between the first optoelectronic component (112a) and the second optoelectronic component (112b) is at least partially exposed, and / or - at least one window (130) with a convex curvature (132) is formed over the first optoelectronic component (112a) and / or the second optoelectronic component (112b) such that the convex curvature (132) has a lens structure which is designed to focus light in the direction of the respective optoelectronic component (112).

11. Method according to claim 10, characterized in that the mask is selected such that the window (130) over the first optoelectronic component (112a) and / or the window (130) over the second optoelectronic component (112b) each have a width (131) of at least 3pm.

12. Method according to one of claims 10 to 11, characterized in that the mask is selected such that the window (130) over the first optoelectronic component (112a) and / or the window (130) over the second optoelectronic component (112b) has a width (131) of 10pm.

13. Method according to one of claims 10 to 12, characterized in that the mask is selected such that the window (130) above the first optoelectronic component (112a) borders at least one side on the at least one metallization structure (120).

14. Method according to one of claims 10 to 13, characterized in that the mask is selected such that the window (130) above the second optoelectronic component (112b) borders at least one side on the at least one metallization structure (120).

Images