WO2022023170A1

WO2022023170A1 - Color splitter system

Info

Publication number: WO2022023170A1
Application number: PCT/EP2021/070560
Authority: WO
Inventors: Oksana SHRAMKOVA; Bobin VARGHESE; Laurent Blonde; Valter Drazic; Valerie Allie
Original assignee: Interdigital Ce Patent Holdings, Sas
Priority date: 2020-07-30
Filing date: 2021-07-22
Publication date: 2022-02-03

Abstract

In example embodiments, a color splitter system includes: an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2); an insert block in the outer block, the insert block having a second refractive index (n3); a first photodetector substantially centered under the outer block; and at least a second photodetector positioned at a lateral offset from the first photodetector. In some embodiments, ns is less than ns. Some embodiments include an isolation structure between the first photodetector and the second photodetector.

Description

COLOR SPLITTER SYSTEM

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] The present application claims priority from European Patent Application No. EP20305869, entitled “Color Splitter System,” filed 30 July 2020, which is hereby incorporated by reference in its entirety.

BACKGROUND

[0002] This section is intended to introduce the reader to various aspects of art, which may be related to various aspects of the present disclosure that are described and/or claimed below. This discussion is believed to be helpful in providing the reader with background information to facilitate a better understanding of the various aspects of the systems and methods described herein. Accordingly, it should be understood that these statements are to be read in this light, and not as admissions of prior art.

[0003] Image sensors usually include an array of pixels. It is desirable to achieve small pixel sizes, especially for mobile applications. Smaller pixel sizes can lead to smaller silicon chip area and potentially lower price for the same pixel count, or higher resolution image sensors with the same silicon area. On the other hand, smaller pixel size in the image sensor generally results in less received light by each pixel. Performance of the image sensor, characterized by parameters such as signal to noise ratio (SNR) and optical efficiency, directly depends on the amount of received light to each pixel.

[0004] When image sensors are in use, not all the received light contributes to the electric signal generated by each pixel. Each pixel may include an optical stack to collect light and a photosensitive part to convert the received photons into a readable electric signal. The optical stack in color image sensors usually includes a color filter, which is mostly light-absorbent for a part of the spectrum, to filter out the undesired band and to transmit only the desired band of the received light spectrum. The performance of the image sensor, characterized by parameters such as signal-to-noise ratio (SNR) and optical efficiency, is influenced by the amount of light received by each pixel. By using light-absorbent color filters in image sensors, about two-thirds of the light entering a pixel is wasted by the filtering process.

[0005] Various color filtering methods have been proposed to increase the light intake in image sensors. [0006] The use of white pixels, clear pixels or panchromatic pixels is one solution which uses the concept of increasing the amount of light intake for better efficiency. Although this solution takes in more light compared to, for example, the widely used Bayer filter array, it still filters out a big portion of the incident light using absorptive band pass filters for R, G and B channels. Color management exploiting the signals presents more difficulties to respect the color representation standards as it becomes spectrum dependent.

[0007] In some systems, the light-absorbent color-filter array is replaced with a transparent diffractive- filter array (DFA), and computational optics techniques are applied to enable color imaging with enhanced sensitivity. Absorption-free color imaging is realized in such systems by using a DFA to diffract incident light onto a conventional monochrome sensor array to create intensity distributions that are wavelength dependent. Potential disadvantages of the use of a DFA may include: lower pixel resolution as the resolution budget on the image sensor is spent for increasing the sensitivity of the color image sensor; the need for post-processing (computational techniques) to extract color information from the captured data; and losses due to diffraction efficiency.

[0008] Some systems use color deflectors to deviate sub-bands of the incoming light towards different photosensitive areas. Such systems, however, may have an aperture size of the splitter component that is small, compared to the pixel size, which creates the need for a focusing element. The aspect ratio between the dimensions of the color splitter element may be relatively large. For example, the length of such splitter elements may be more than four times the width (e.g. llw > 4). This increases the requirements on fabrication tolerances. Microfabrication techniques are usually adapted to processes with wide and thin structures and so smaller llw aspect ratios are appreciated. With the use of color deflectors, some feature dimensions may be much smaller than the pixel width. In addition, the distance between the splitting element and the surface of the detector increases the thickness of the pixel, which increases the fabrication complexity and risk for crosstalk.

[0009] Some sensor systems, such as the Foveon sensor, operate without absorptive color filters or diffractive elements. In such systems, the pixel structure is threefold, with three photodiodes being stacked vertically one on top of the other. The color selectivity uses the fact that each color band will get absorbed after a different depth in the silicon material. Blue radiation is absorbed in the first photodiode, and red in the last, green being absorbed in between. Potential disadvantages of such systems may include: the presence of color cross-talk as the absorption bands overlap somewhat; a low signal-to-noise ratio that limits the dynamic of that type of sensors which compares to Bayer color filter provided sensors up to only ISO 400. Moreover, the Foveon sensor pixel size is about 7 pm which makes them less suitable for smart phone or compact camera formats.

[0010] A method to split color-bands of the incident light by combining two or more dielectric materials with different refractive indexes was proposed in WO2019175062A1 , “Image sensor comprising a color splitter with two different refractive indexes.” In the topologies presented in that publication, nanojet beams originating from different edges (associated with different blocks/layers) of the microstructure, recombine and contribute to the formation of a spectrally-dependent nanojet beam deflection. This nanojet beam deflection can generate focused color images. The characteristics of the generated nanojet beams are controlled by the parameters of the corresponding blocks, such as the refractive index ratios between the blocks, dimensions of the blocks and angle of wave incidence.

SUMMARY

[0011] References in the specification to “one embodiment,” “an embodiment,” “an example embodiment,” and the like indicate that the embodiment described may include a particular feature, structure, or characteristic; but not every embodiment necessarily includes that particular feature, structure, or characteristic. Moreover, such phrases are not necessarily referring to the same embodiment. Further, when a particular feature, structure, or characteristic is described in connection with an embodiment, such feature, structure, or characteristic may be used in connection with other embodiments whether or not explicitly described.

[0012] A color splitter system according to some embodiments includes an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2); and an insert block in the outer block, the insert block having a second refractive index (n3). The first and second side edges are in contact with an ambient medium

[0013] In some embodiments, a first photodetector is substantially centered under the outer block; and at least a second photodetector is positioned at a lateral offset from the first photodetector.

[0014] In some embodiments, is less than n2. In some embodiments, the refractive index of the ambient medium (m) is less than n2.

[0015] Some embodiments further include at least one deep trench isolation structure between the first photodetector and the second photodetector.

[0016] Some embodiments further include a third photodetector positioned at a second lateral offset from the first photodetector, the second lateral offset being in an opposite direction from the first lateral offset.

[0017] Some embodiments further include a first deep trench isolation structure between the first and second photodetector and a second deep trench isolation structure between the second and third photodetectors, wherein the first and second deep trench isolation structures have a spacing of W^*, with

OB = W₁ — H^ahq^ , with W_i being half of the width of the outer block, W₂ being the width of the insert block, being the height of the outer block, H₂ being the height of the insert block, and

90° — sm^_1(n₁/n₂)

90° — sin^~1(n₃/n₂ )

Q_B2 - 2 - where m is the refractive index of the ambient medium.

[0018] In some embodiments, the height H₂ of the outer block satisfies

with W being half of the width of the outer block, W₂ being the width of the insert block,

being the height of the outer block, H₂ being the height of the insert block, and

90° — sm^_1(n₁/n₂)

where m is the refractive index of the ambient medium.

[0019] In some embodiments, the width (214^) of the outer block is between 700nm and 800nm and the height {H^ of the outer block is between 500nm and 700nm.

[0020] In some embodiments, the width [W₂) of the insert block is between 200nm and 300nm, and the height ( H₂ ) of the insert block is between 250nm and 450nm.

[0021] In some embodiments, the insert block is substantially laterally centered in the outer block.

[0022] In some embodiments, the insert block has an upper surface substantially even with an upper surface of the outer block.

[0023] In some embodiments, the first and second side edges are substantially vertical.

[0024] In some embodiments, the insert block has a substantially vertical third side edge and a substantially vertical fourth side edge.

[0025] In some embodiments, the outer block is on a substrate.

[0026] Some embodiments further include an antireflective coating between the outer block and the substrate.

[0027] An image sensor according to some embodiments includes a plurality of color splitter systems as described herein. In some embodiments, an image sensor comprises a two-dimensional array of color splitters arranged on a substrate, each of the color splitters comprising: an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2), the first and second side edges being in contact with an ambient medium; and an insert block in the outer block, the insert block having a second refractive index {m).

[0028] In some embodiments, a method comprises directing incident light on a color splitter, where the color splitter comprises: an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2), the first and second side edges being in contact with an ambient medium; and an insert block in the outer block, the insert block having a second refractive index (n3). The method may further include sensing light of a first color at a first photodetector substantially centered under the outer block and sensing light of a second color at a second photodetector positioned at a lateral offset from the first photodetector.

BRIEF DESCRIPTION OF THE DRAWINGS

[0029] FIG. 1 is a cross-sectional view illustrating an example topology of a nanojet-based dielectric color splitter for normal incidence according to some embodiments.

[0030] FIG. 2A is a cross-sectional view of an image sensor including a color splitter with a doublematerial microlens with an insert according to some embodiments.

[0031] FIG. 2B is a cross-sectional view of a color splitter with a double-material microlens with insert according to some embodiments, illustrating nanojet generation by double-material microlens for an inclined incidence of a plane wave.

[0032] FIG. 3A is a cross-sectional view illustrating geometry of a unit cell in an example system, the properties of which are investigated numerically herein.

[0033] FIG. 3B illustrates calculated power distribution in the unit cell for l =500nm.

[0034] FIG. 3C illustrates calculated power distribution in the unit cell for l =700nm.

[0035] FIGs. 4A-C are graphs illustrating transmittivity for the full visible spectrum at normal incidence and different depths of dsi. In FIG. 4A, dsi = 100nm. In FIG. 4B, dsi = 800nm. In FIG. 4C, dsi = 1500nm.

[0036] FIGs. 5A-5C are graphs illustrating the dependence of total transmittivity measured for two ports at different depths for three different RGB colors at normal incidence.

[0037] FIGs. 6A-6C are graphs illustrating transmittivity for the full visible spectrum at normal incidence and dsi = 100nm for different antireflection layer thicknesses. In FIG. 6A, dARc = 150nm. In FIG. 6B, dARc = 300nm. In FIG. 6C, dARc Port2 = 150nm and dARc Poro = 200nm.

[0038] FIGs. 7A-7C are graphs illustrating transmittivity for RGB colors as a function of the distance between the deep trench structures at normal incidence and dARc=200nm, dsplOOnm. In FIG. 7A, dDT=100nm. In FIG. 7B, dDT=150nm. In FIG. 7C, dDT=200nm. [0039] FIGs. 8A-8C illustrate transmittivity for RGB colors as a function of the distance between the deep trench structures at normal incidence and dARc=200nm, dsi=800nm. In FIG. 8A, dDT=100nm. In FIG. 8B, dDT=150nm. In FIG. 8C, dDT=200nm.

[0040] FIGs. 9A-9C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, and W*=440nm. In FIG. 9A, a=5°. In FIG. 9B, a=10°. In FIG. 9C, a=15°.

[0041] FIGs. 10A-10C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, and W*=360nm. For FIG. 10A, a=5°. For FIG. 10B, a=10°. For FIG. 10C, a=15°.

[0042] FIGs 11 A-11 C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, and W*=520nm. In FIG. 11 A, a=5°. In FIG. 11 B, a=10°. In FIG. 11C, a=15°.

[0043] FIGs. 12A-12C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, W*=440nm, and a=0°. In FIG. 12A, «3=1.6. In FIG. 12B, «3=1.8. In FIG. 12C, «3=2.0.

[0044] FIGs. 13A-13C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, W*=440nm, and «3=1.6. In FIG. 13A, a=5°. In FIG. 13B, a=10°. In FIG. 13C, a=15°.

[0045] FIGs. 14A-14C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, W*=440nm, and «3=1.8. In FIG. 14A, a=5°. In FIG. 14B, a=10°. In FIG. 14C, a=15°.

[0046] FIG. 15 is a schematic cross-sectional view illustrating a color splitter system according to some embodiments.

DETAILED DESCRIPTION

Overview of Example Embodiments.

[0047] The present disclosure relates generally to the field of image sensing devices and, more particularly, to the part of the color image sensor which guides a selected bandwidth of incident light towards the proper photosensitive area of the image sensor. Described are methods and components for on-chip optics in the image sensor to improve the optical efficiency by splitting the incoming light between the image sensor pixels. Such splitting may be performed as an alternative to filtering out part of the received light on each color subpixel.

[0048] Example embodiments include a double material nanojet-based color splitter with a deep-trench- isolation (DTI) structure to suppress crosstalk. Such embodiments may, for example, improve the quality of filtering and increase the sensitivity of a back-illuminated CMOS image sensor (IS). An example nanojet microlens comprises an outer block of dielectric material with refractive index n2 and at least one dielectric insert block having a refractive index different from n2. Microlenses having such a topology are referred to herein as double-material microlenses with an insert.

[0049] The use of a dielectric insert block in some embodiments may allow for reduction in the size of the color splitting element and of the optical crosstalk through the active silicon layer. By changing the position of the DTI, the functionality of separating blue, green and red light can be improved.

[0050] To maintain high sensitivity and to prevent cross-talk among adjacent pixels, example embodiments use DTI between pixels. The DTI structure, which may be filled with lower refractive index materials compared to Silicon photodiodes, is effective in reducing or blocking the electrical crosstalk in the deep quasi-neutral region as well as the optical crosstalk through the active silicon layer. In example embodiments, the DTI confines nanojet beams created inside the diffractive element within the pixel, preventing light intended for one photodetector from impinging on a different photodetector. To form a DTI structure between pixels, only one extra mask is needed. Photolithography and etch processes make it possible to achieve DTI with an aspect ratio higher than 1 :25.

[0051] In some embodiments, an aperture size of the color splitting element is close to the width of the image sensor pixel, removing the need for a focusing element on top of the color splitting element, and consequently, simplifying the pixel. Some embodiments have a relatively low pixel size. Some embodiments have a low thickness of the color splitting element to reduce the thickness of the optical stack to improve efficiency and potentially to reduce crosstalk. Some embodiments use a DTI structure between the pixels to provide better efficiency of the system.

[0052] The present disclosure relates generally to the field of image sensing devices and, more particularly, to the part of the color image sensor which guides a selected bandwidth of incident light towards the proper photosensitive area of the image sensor. Disclosed herein are methods and components for on-chip optics in the image sensor to improve the optical efficiency by splitting the incoming light between the image sensor pixels instead of filtering out part of the received light on each color subpixel. A new topology of double material nanojet-based color splitter with deep-trench-isolation (DTI) structures is proposed to suppress the crosstalk. Some embodiments can be used to improve the quality of filtering and to increase the sensitivity of a back-illuminated CMOS image sensor (IS).

[0053] Example embodiments employ a nanojet microlens having an outer block of dielectric material with refractive index n2 and at least one dielectric insert block having a refractive index different from n2. Hereafter, microlenses having such a topology are referred as double-material microlenses with insert. This transformation results in reduction of the size of the color splitting element and of the optical crosstalk through the active silicon layer. By changing the position of the deep-trench-isolation structures, the blue, green and red separation functionality can be improved.

[0054] Example embodiments employ a topology of diffractive elements to split color-bands of incident light by combining two or more dielectric materials with different refractive indexes. (The refractive indexes of constitutive parts are higher than the surrounding material.) The topology is arranged in such a way that the nanojet beams, originating from external and internal edges and having different intensities and angles of deviation, recombine and contribute to the formation of a spectrally-dependent nanojet beam deflection, as illustrated in FIG. 1. Numerical simulations have been performed to demonstrate that the proposed embodiments can generate focused beams for each color outside the diffractive element. The characteristics of the generated nanojet beams are controlled by the parameters of the constitutive parts, such as the refractive index ratios for the contacting materials for corresponding edges, sizes of the constitutive parts (height and width of the outer blocks, size/shape of the insert block), base angle and angle of wave incidence.

[0055] FIG. 1 is a cross-sectional view illustrating an example topology of a nanojet-based dielectric color splitter for normal incidence according to some embodiments.

[0056] FIG. 2A is a cross-sectional view of an image sensor including a color splitter with a doublematerial microlens with insert according to some embodiments.

[0057] FIG. 2B is a cross-sectional view of a color splitter with a double-material microlens with insert according to some embodiments, illustrating nanojet generation by double-material microlens for an inclined incidence of a plane wave.

[0058] Properties of example embodiments have been explored numerically via full-wave electromagnetic analysis of 2D microlenses, whose cross-section and focusing function are schematically represented in FIG. 2A. To estimate the absorption in the silicon layer, the calculations take into account that silicon is a lossy and dispersive material. The calculations have revealed that diffraction of a plane wave on a microlens based on the combination of different dielectric materials can result in a spectrally- dependent nanojet beam deviation. The calculations were based on the use of an inhomogeneous microlens, with the insert blocks having a lower refractive index than that of the outer block in which they are inserted (n2>n3>m)). The calculations demonstrate that for some particular parameters, the nanojet- based double-material microlens with an insert block can split colors. The position of focal spot, angle of deviation, intensity and shape of the nanojet beam can be controlled by a variation of the refractive indexes and sizes of the constitutive parts (outer block and insert block). Such elements may be used for the creation of color splitters without a classical focusing microlens on the top. Calculations have been performed of light transmission through the silicon substrate with deep trench isolation of pixels. The effect of DTI structures on the crosstalk blocking is considered. Described herein are example positions of deep trench structures that can be used to enhance the optical efficiency.

Example Topology.

[0059] The general topology of an example splitter is illustrated in FIG. 2A. The cross-sectional view of the nanojet microlens on the top may correspond to a cuboid, cylinder or prism embedded in a homogeneous dielectric host media with a refractive index m < n2. In an example, the nanojet microlens has a dielectric insert block with a refractive index n3<n2. The material and size of the insert block with refractive index can be selected based in part on the parameters of the outer block in order to manage the spectrally-dependent nanojet beam deflection outside the microlens. The nanojet beams associated with the edges of the nanojet microlens’ constitutive parts penetrate into the silicon substrate and recombine. The effect of size, position and refractive index of the insert block for such type of microlens on the nanojet beam’s spectrally-dependent deflection is analyzed. To reduce the effect of the silicon substrate (e.g. refraction of the nanojet beam inside the silicon substrate) and to increase the optical efficiency, example embodiments use DTI technology. The dependence of optical efficiency of example embodiments on the parameters (position and thickness) of DTs and thickness of SiNx anti-reflection layer has been analyzed.

[0060] The following description primarily uses examples in which the nanojet microlens structure has vertical edges parallel to the z-axis and top/bottom surfaces parallel to the xy-plane, which corresponds to a base angle equal to 90 degrees. However, in different embodiments, structures with different base angles can also be used. Variation of the base angle value provides additional degree of freedom in the control of the nanojet beam radiation direction. The total response of such system depends also on the angle of the local plane wave illumination. The effect of the plane wave angle of incidence is investigated.

[0061] Some equations are provided below that may be used to select combinations of materials and dimensions of the blocks for spectrally-dependent nanojet beam deflection. The splitting functionality of example embodiments has been investigated numerically. The effect of the parameters of the single color splitting element on the functionality of the system is considered. The set of equations to estimate the optimal position of the DT is also presented.

Characteristics of Nanojet Beams.

[0062] In a first approximation, the focal length of the nanojet microlens with insert can be characterized as the function of the size (width or radius) and index ratio of the media inside and outside the microstructure. Let us present a set of equations to estimate the position for nanojets generated by the system with n2>nz> .

[0063] The near-field pattern and position of nanojet hot spots are determined by the form, size, position regarding the outer block and values of refractive index of the insert block. This effect may be explained by the interference of, on the one hand, the nanojet beams associated with the top edge of the outer block of the nanojet microlens and, on the other hand, the nanojet beams associated with the top edge of the insert block. (For a case of electromagnetic wave incidence from the top of the microlens, see FIG. 2A). In this case, the intersection of nanojets associated with the edges of different constitutive parts leads to the forming of hot spots located out of the axis of symmetry of the nanojet microlens’ outer block. The total response of the inhomogeneous systems with dimensions larger than a few wavelengths of an incident wave represents the interplay between the nanojet and Fresnel diffraction phenomena.

[0064] It has been demonstrated that the beam-forming phenomenon is associated primarily with the edge of the system, and the nanojet beam radiation angle is defined by Snell’s law. See A. Boriskin, V. Drazic, R. Keating, M. Damghanian, O. Shramkova, L. Blonde, “Near field focusing by edge diffraction,” Opt. Lett., 2018. So, for the normal incidence of incident wave, the nanojet beam radiation angle for constitutive parts of the nanojet microlens can be determined as a function of the ratio between the refractive indexes of the surrounding media and material of the outer block of the lens (for the insert block, the “host medium” is the material of the outer block), and the base angle of the element. (Some embodiments use elements with the vertical edges, giving a base angle equal to 90°). For the outer block part of the nanojet microlens with refractive index m, the NJ1 beam radiation angle can be determined using the approximate formula:

where

is the critical angle of refraction. Therefore Q_B1 ~

point (hot spot) of the two identical and symmetrical nanojets N J1 generated by the external edges of the nanojet microlens (outer block) determines the focal length of the single material nanojet microlens. This focal length can be estimated as:

tail &_m ^2 j where 2l/l/i is the full width of the outer block of the nanojet microlens. For the case of normal incidence of electromagnetic wave, the focal point will be located on the axis of symmetry of the nanojet microlens. The maximal intensity of generated nanojet beam corresponds to the edge with the critical height, as described in B. Varghese, 0. Shramkova, V. Drazic, V. Allie, L. Blonde , “Influence of an edge height on the diffracted EM field distribution,” ICTON 2019, Angers, France. In example embodiments, the height Hi of the outer h_c block is close to the critical height ”² ”^l

[0065] To determine the total width of the outer block we take that F_L > H_± . So, we obtain: l tail Q _m

W > h2 ^{~ p} 1

[0066] The color splitting functionality may be understood as relating to the presence of nanojets of different types (with different angles of deviation and different intensity) inside the multi-material microlens. To generate the NJ2 , example embodiments use an insert with refractive index m.

[0067] In the case of a symmetrical inhomogeneous microlens with an insert for which P3<P2 and m<n2, the two additional similar nanojets (NJ2) will be generated by the internal edges of the microlens with the insert. The NJ2 beam radiation angle can be determined using the approximate formula:

90 °-sin ¹(n₃/n₂) -p. .

is the critical angle of refraction. Therefore Q_B2 _« - ₂ The proposed example ratio between the refractive indexes leads to a result in which ©o ^<¾ and NJ₂ is less intensive than the NJ1.

[0068] The size (width and height) of the insert may be selected based on parameters of the outer block and on the refractive index m. If n₂ < 2, it is desirable for the generated NJ2 not to cross the vertical edges of the microlens to avoid the additional nanojet refraction at the boundary between the material of outer block and host medium. So, parameters may be selected such that AA’<2l/l/i and

^W2 <

_n2 > 2 NJ2 will be reflected by the vertical wall due to the total internal reflection phenomenon. So, to get a maximal distance between NJi and NJ2 in the Silicon substrate, the width of the insert may be selected to provide favorable conditions to get AA’ as close as possible to the full width of the outer block. The maximal input of NJi may be observed when NJi does not cross the insert.

So, we obtain:

[0069] For the chosen size of the outer block we will observe at least two nanojet hot spots (crossings of NJi and NJ2, see FIG. 2A) symmetrically situated relative to the axis of symmetry inside the outer block.

For example embodiments, outside the microlens there may be four nanojets penetrating into the silicon substrate. First two nanojets (NJi) will cross the boundary between the element and substrate at points B and B’:

[0070] The nanojets of second type (NJ2) will cross the boundary between the element and substrate at points A and A’ :

OA = OA' = + H₁ tan Q_B2

(4b)

[0071] Inside silicon, the radiation angles of all these nanojets will be reduced due to the refraction phenomenon, and the nanojets of different types will be closer to each other. This may complicate the nanojet detecting inside the silicon substrate. For better separation of the nanojets, some embodiments use DTI. In an example, two symmetrically positioned (relative to the axis of symmetry of the single element) DTs may be located inside the substrate. To get the desirable color splitting function, each DT should be placed between the nanojets penetrating into the silicon substrate:

Where W* is the minimal distance between DTs as shown on FIG. 2A.

[0072] Let us now consider the effect of the angle of plane wave incidence on the properties of generated nanojet beam. The angle of electromagnetic wave incidence is indicated by a, as illustrated in FIG. 2B, which shows an example inhomogeneous microlens with an oblique plane wave incidence.

[0073] To get the approximate formula for nanojet beam radiation angles in the case of plane wave oblique incidence on the outer block with refractive index n2, it is noted that the radiation angles 0'_Bi and 0"BI for opposite edges of the system are not equal (see FIG. 2B, NJ’i and NJ”i). As a result, for the outer block we can have:

[0074] In a similar way, the nanojet beam radiation angles for the insert (NJ’2 and NJ’2) can be determined as:

[0075] A system optimized for normal incidence may have poor splitting functionality in the case of inclined incidence. To improve the efficiency for a wider range of angles of incidence, the parameters of some embodiments may be optimized taking into account that a>0. As in the case of normal incidence, the parameters of an insert may be calculated to avoid the crossing of NJ’2 and NJ”2 with the vertical edges of the microlens. For example, for a>0 , parameters may be selected such that W₂ < 2(W_j_ - H₁ΐahq'_B2). To determine the position of deep trench structures, considering that for inclined incidence OB¹OB’ and OA¹OA’:

OB = W₁ — H^ahq' _B2 , OB' = W₁ — H^ahq" _B2 ,

[0076] Finally, for a>0 Eq. (5) will take the form OB < ^- < OA'.

[0077] Some embodiments may employ a periodic array of such nanojet microlenses. In this case, inside the substrate and close to its surface, there is periodic alternation of the hot spots for the nanojets of the same type: nanojets of first type (NJi) will have their crossing points at the axis of symmetry of the microlenses; nanojets of second type (NJ2) will provide hot spots at the boundaries of the pitches. Upon changing the pitch of this system, the intensity of the hot spot for NJ2 can be adjusted.

Parametric Study.

[0078] To illustrate the features of example embodiments, a periodic array of 2D double-material microlenses with the insert has been investigated numerically computed using COMSOL software. The calculations assume that the system is illuminated by a linearly TM-polarized wave. To model wave propagation in a single unit cell of the array, on either side of the unit cell we use periodic boundary conditions with Floquet periodicity. To avoid non-physical reflection we model the open boundaries using the perfectly matched layer (PML) domains (FIG. 3A). To measure the changing of the incident light transmittivity we scan the power density between the deep trench structure (port 2 and port 3 in FIG. 3A) at some depth inside the silicon layer (dsi).

[0079] FIG. 3A is a cross-sectional view illustrating geometry of a unit cell in an example system, the properties of which are investigated numerically herein.

[0080] FIG. 3B illustrates calculated power distribution in the unit cell for l =500nm.

[0081] FIG. 3C illustrates calculated power distribution in the unit cell for / =700nm.

[0082] The color splitting functionality of an example embodiment is illustrated using the power distribution for two wavelengths in FIGs. 3B and 3C. For the red color band central wavelength ( =700 nm) main power is transmitted through Port 2. In a case of wavelength corresponding to the green color band (centered on l =500 nm) the main part of the light will be transmitted through Port 3.

[0083] The simulations presented below correspond to such parameters of the system selected for l =620 nm to provide red color splitting functionality. It is assumed that the host medium has a refractive index m=1.0. Using the formula provided above and additional numerical optimization an embodiment is selected using a microlens with «2=2.0, n3=1.4, Hi =600 nm, Hi =400 nm, \Ni =380 nm, I/I/2 =240nm. A layer of SiNx with a thickness of dARc=200nm and refractive index 2.04 as an antireflection layer is adopted. The DTI layers are constructed with S1O2 material with a refractive index of 1.5; W*=440nm, dDT=100nm. [0084] The proposed parameters of the example nanojet microlens element result in distance OB=33.6nm and OA=372nm.

[0085] FIGs. 4A-C are graphs illustrating transmittivity for the full visible spectrum at normal incidence and different depths of dsi. In FIG. 4A, dsi = 10Onm. In FIG. 4B, dsi = 800nm. In FIG. 4C, dsi = 1500nm.

[0086] FIGs. 4A-4C illustrate the dependence of integral power density on the wavelength for three different positions of the ports inside the silicon layer. It is possible to observe that at Port 2 we can register the maximal power at wavelengths corresponding to the red color, while other wavelengths are registered at Port 3.

[0087] FIGs. 5A-5C are graphs illustrating the dependence of total transmittivity measured for two ports at different depths for three different RGB colors at normal incidence.

[0088] It can be seen that the second port, Port 2, effectively registers red colored light. Green and blue colored light can be registered at Port 3. In some embodiments, photodiodes or other photodetectors for green and blue colors are placed at different depths, which may improve the ability to differentiate between them.

[0089] FIGs. 6A-6C are graphs illustrating transmittivity for the full visible spectrum at normal incidence and dsi = 100nm for different antireflection layer thicknesses. In FIG. 6A, dARc = 150nm. In FIG. 6B, dARc = 300nm. In FIG. 6C, dARc Port2 = 150nm and dARc Poro = 200nm.

[0090] As shown in FIGs. 6A-6C, the thickness of the antireflection layer affects the transmittivity of the incident light. It is possible to see that by taking different thicknesses of the antireflection layer for Port 2 and Port 3 it is possible to increase the portion of light transmitted through the Port 2 (and decrease it for the Port 3) at the red color wavelengths (FIG. 6C). Moreover, a uniformity of the distribution can be also improved.

[0091] FIGs. 7A-7C are graphs illustrating transmittivity for RGB colors as a function of the distance between the deep trench structures at normal incidence and dARc=200nm, dsplOOnm. In FIG. 7A, dDT=100nm. In FIG. 7B, dDT=150nm. In FIG. 7C, dDT=200nm.

[0092] FIGs. 8A-8C illustrate transmittivity for RGB colors as a function of the distance between the deep trench structures at normal incidence and dARc=200nm, dsi=800nm. In FIG. 8A, dDT=100nm. In FIG. 8B, dDT=150nm. In FIG. 8C, dDT=200nm.

[0093] The thickness and position of deep trench structures also affect the color splitting functionality of the device, as shown in FIGs. 7A-7C and 8A-8C. The high red color transmittivity through the Port 2 can primarily be observed starting from some critical distance between the deep trench structures. Increasing W* can significantly reduce the transmittivity through Port 3 corresponding to green and blue colors. The presented simulations were obtained for two different depths inside the silicon substrate. For FIGs. 7A-7C, dsi = 100nm. For FIGs. 8A-8C, dsi = 800nm. [0094] FIGs. 9A-9C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, and W*=440nm. In FIG. 9A, a=5°. In FIG. 9B, a=10°. In FIG. 9C, a=15°.

[0095] The numerical simulations presented in FIGs. 9A-9C demonstrate the transmittivity of RGB colors for three different angles of incidence, a. It is possible to observe that the portion of power transmitted through Port 3 and corresponding to the blue band dramatically drops with the angle of incidence. At a=15° the main part of blue color will be transmitted through Port 2. In the case of green color, for all three angles of incidence, the main part of the power will be transmitted through Port 3.

[0096] FIGs. 10A-1 OC illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, and W*=360nm. For FIG. 10A, a=5°. For FIG. 10B, a=10°. For FIG. 10C, a=15°.

[0097] FIGs. 11 A-11 C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, and W*=520nm. In FIG. 11 A, a=5°. In FIG. 11 B, a=10°. In FIG. 11C, a=15°.

[0098] Keeping the same parameters for the nanojet microlens while changing the position of the deep trench structures can redistribute the portion of power transmitted through the ports due to the limitation of the nanojet input. It also affects the power redistribution in the case of inclined incidence. The RGB color transmittivity through Ports 2 and 3 for W*=360nm and 520 nm and inclined incidence is presented in FIGs. 10A-10C and FIGs. 11 A-11C. It can be seen that by decreasing W* (FIGs. 10A-C) the green and blue color transmittivity through Port 3 is improved, but red color transmittivity through Port 2 will be decreased. Upon increasing W* (FIGs. 11 A-11C) the efficiency of the proposed color splitter may be improved for the red color, but blue and green color performance may become worse.

[0099] FIGs. 12A-12C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, W^*=440nm, and a=0°. In FIG. 12A, «3=1.6. In FIG. 12B, «3=1.8. In FIG. 12C, «3=2.0.

[0100] The effect of refractive index of an insert on the color splitting functionalities of the system has been investigated. The numerical simulations presented in FIGs. 12A-12C demonstrate the RGB colors’ transmittivity for three different values of refractive index m. By increasing the refractive index of the insert, the angles for NJ2 are decreased. As a result, the position of point A (see FIG. 2A for normal incidence) will be shifted, and for «3=1.6 we obtain OA=320nm; for «3=1.8 we get OA=257.7nm. Such parameters of the system satisfy the conditions of Eq. (5). To observe the effect of an insert on the characteristics of the proposed nanojet microlens, FIG.12C corresponds to the case of single material block with m= . The simulations show that the power transmitted through Port 2 rises with the refractive index of the insert. Correspondingly, the portion of power for green and blue colors registered at Port 3 will be decreased. Finally, in the case of a single material block, the main part of the power for the three colors will be transmitted through Port 2.

[0101] FIGs. 13A-13C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, W*=440nm, and «3=1.6. In FIG. 13A, a=5°. In FIG. 13B, a=10°. In FIG. 13C, a=15°.

[0102] FIGs. 14A-14C illustrate transmittivity for RGB colors as a function of the silicon layer depth for inclined incidence at dARc=200nm, dDT=100nm, W*=440nm, and «3=1.8. In FIG. 14A, a=5°. In FIG. 14B, a=10°. In FIG. 14C, a=15°.

[0103] In some embodiments, a splitter can be configured for inclined incidence by keeping the size of the insert unchanged and increasing the refractive index m. For example, such parameters of a full system (size of the insert and position of the deep trench structures) can correspond to a system configured for «3=1.6 and a=10°. The results presented in FIGs.13A-13C and FIGs.14A-14C show the RGB colors’ transmittivity of the microlenses with «3=1.6 and «3=1.8 for inclined incidence. It is possible to observe that for a>0, systems with higher refractive index m also demonstrate better effectiveness for the red color. The portion of the power transmitted through Port 3 and corresponding to the blue and green bands will almost be the same.

[0104] Example embodiments of a nanojet-based microlens with an insert operate to split incident light into different colors, particularly with particular parameters (e.g. Hi, F , Wi, W2 for some combinations of indexes of dielectric materials) as described herein. To split the colors, example embodiments use a dielectric microlens with an insert, where the outer block of this microlens is a dielectric material with refractive index n2, and is the refractive index of an insert (with n2>n3>ni, where m is the refractive of the host medium).

[0105] Some embodiments use deep trench isolation structures to suppress crosstalk. Example parameters governing the position of such structures, such as the value W*, are described above to provide enhancement of the optical efficiency.

[0106] Example embodiments may provide increased light intake due to the proposed light splitting structure (in contrast to the classical absorptive solutions). Example embodiments of a nanojet-based color- splitter may further provide one or more of the following benefits: simpler pixel architecture; relaxed need of a focusing lens on top of the photodiode in the pixel architecture, as the color splitter element provides focusing effect as well; less constrained fabrication process due to the lack of high aspect ratios and small feature sizes; reduced risk of crosstalk due to deep trench isolation technology; better optical efficiency due to thinner optical stack (less losses) and deep trench isolation structure; better angular performance with regards to the angle of the incident light; and the ability to split three colors. [0107] FIG. 15 is a schematic cross-sectional view illustrating a color splitter system according to some embodiments. FIG. 15 illustrates an outer block 1502 with a first side edge 1504, a second side edge 1506, and an upper surface 1508, the outer block having a first refractive index (n2). The first side edge and second side edge are in contact with an ambient medium (e.g. air) or other region having a refractive index m, with m < n2.

[0108] An insert block 1510 is provided in the outer block, the insert block having a second refractive index (n3). A first photodetector 1512 is substantially centered under the outer block. At least a second photodetector 1514 is positioned at a lateral offset from the first photodetector. In some embodiments, is less than n2. Some embodiments further include one or more isolation structures 1516, such as deep trench isolation structures, between the photodetectors.

[0109] FIG. 15 further illustrates features that may be present, individually or in combination, in some embodiments. For example in some embodiments, the insert block is substantially laterally centered in the outer block. In some embodiments, the insert block has an upper surface 1518 that is substantially even with the upper surface 1508 of the outer block. In some embodiments, the first and second side edges 1504, 1506 are substantially vertical. In other embodiments, first and second side edges 1504, 1506 have different non-vertical base angles. In some embodiments, the insert block has substantially vertical third (1520) and fourth (1522) side edges. (It should be understood that “vertical” is with reference to an associated base or other mounting, for example with respect to a surface 1524 on which the outer block is positioned.) In some embodiments, the outer block 1502 is on a substrate 1526. Some embodiments further include an antireflective coating 1528.

[0110] In some embodiments, a plurality of color splitter systems as illustrated in FIG. 15 are arranged together to form an image sensor. The color splitter systems may be arranged in a two-dimensional array. The image sensor may be a full-color image sensor. In some embodiments of an image sensor, no microlenses are positioned over the individual color splitter systems (although the image sensor may be a component of a system, such as a camera, with one or more objective lenses).

[0111] While the above examples refer primarily to the use of devices configured for visible light, other embodiments are configured for use with longer or shorter wavelengths, such as infrared or ultraviolet light, or for use with waves in other parts of the electromagnetic spectrum. Such embodiments may employ materials that are transparent to the wavelengths for which they are designed.

[0112] Although features and elements are described above in particular combinations, one of ordinary skill in the art will appreciate that each feature or element can be used alone or in any combination with the other features and elements.

Claims

CLAIMS What is Claimed:

1. A system comprising a color splitter, the color splitter comprising: an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2), the first and second side edges being in contact with an ambient medium; and an insert block in the outer block, the insert block having a second refractive index (n3).

2. The system according to claim 1, further comprising: a first photodetector substantially centered under the outer block; and at least a second photodetector positioned at a first lateral offset from the first photodetector.

3. The system of claim 1 or 2, wherein the second refractive index {m) is less than the first refractive index

(n ).

4. The system of any one of claims 1 -3, wherein the ambient medium has a third refractive index (m) that is less than the first refractive index (n₂).

5. The system of any one of claims 1-4, further comprising a third photodetector positioned at a second lateral offset from the first photodetector, the second lateral offset being in an opposite direction from the first lateral offset.

6. The system of claim 5, further comprising a first deep trench isolation structure between the first and second photodetector and a second deep trench isolation structure between the second and third photodetectors, wherein the first and second deep trench isolation structures have a spacing of W^* ,

w₂

OA = — + H₁tan0_B2 ,

OB = W₁ — H^ahq^ , with W being half of the width of the outer block, W₂ being the width of the insert block,

90° — sm^_1(n₁/n₂)

90° — sin^~1(n₃/n₂ ) B2 = - o - / where m is the refractive index of the ambient medium.

7. The system of any one of claims 1-6, wherein the height H₂ of the outer block satisfies

with W being half of the width of the outer block, W₂ being the width of the insert block, //_c being the height of the outer block, H₂ being the height of the insert block, and

90° — sm^_1(n₁/n₂)

where m is the refractive index of the ambient medium.

8. The system of any one of claims 1-7, wherein the insert block is substantially laterally centered in the outer block.

9. The system of any one of claims 1-8, wherein the insert block has an upper surface substantially even with an upper surface of the outer block.

10. The system of any one of claims 1-9, wherein the width (2 W_±) of the outer block is between 700nm and 800nm, the height {H^ of the outer block is between 500nm and 700nm, the width [W₂) of the insert block is between 200nm and 300nm, and the height [H₂) of the insert block is between 250nm and 450nm.

11. The system of any of claims 1-10, wherein the outer block is on a substrate, further comprising an antireflective coating between the outer block and the substrate.

12. An image sensor comprising a two-dimensional array of color splitters arranged on a substrate, each of the color splitters comprising: an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2), the first and second side edges being in contact with an ambient medium; and an insert block in the outer block, the insert block having a second refractive index (n3).

13. The image sensor of claim 12, wherein the ambient medium has a third refractive index (m) that is less than the first refractive index (n2).

14. A method comprising: directing incident light on a color splitter, the color splitter comprising: an outer block having, in cross section, a first side edge, a second side edge, and an upper surface, the outer block having a first refractive index (n2), the first and second side edges being in contact with an ambient medium; and an insert block in the outer block, the insert block having a second refractive index {m); and sensing light of a first color at a first photodetector substantially centered under the outer block; and sensing light of a second color at a second photodetector positioned at a lateral offset from the first photodetector.

15. The method of claim 14, wherein the ambient medium has a third refractive index (m) that is less than the first refractive index (n2).