Multi-pixel detector for increasing angular sensitivity and related methods

Abstract

An image sensor includes a multi-pixel detector. The multi-pixel detector includes a first pixel formed in a substrate and having a first photodiode region, a second pixel formed in the substrate adjacent to the first pixel and having a second photodiode region, and a microlens over both the first and second pixels. The microlens includes (a) a first height profile having N1 local maxima in a first cross-sectional plane perpendicular to an upper surface of the substrate and including both the first and second photodiode regions, and (b) a second height profile having N2>N1 local maxima in a second cross-sectional surface perpendicular to the first cross-sectional plane and the upper surface and including only one of the first and second photodiode regions.

Description

[0001] This application is a divisional application of Chinese invention patent with an application date of August 10, 2018, a priority date of September 25, 2017, application number "201810907835.3", and an invention title of "Multi-pixel detector for increasing angular sensitivity and related methods". Technical Field

[0002] This invention relates to the field of optical technology, and particularly to multi-pixel detectors and related methods for increasing angular sensitivity. Background Technology

[0003] Many digital cameras have autofocus capabilities. Autofocus can be fully automatic, allowing the camera to identify objects in a scene and focus on them. In some cases, the camera can determine which objects are more important than others and then focus on the more important ones. Alternatively, autofocus can use user input to specify which part or parts of the scene are of interest. Based on the above, the autofocus function identifies objects in one or more parts of the scene specified by the user and causes the camera to focus on such objects.

[0004] To achieve market adaptability, the autofocus function must be reliable and fast, so that the camera quickly focuses on the desired part or parts of the scene each time the user captures an image. Preferably, the autofocus function is fast enough that the user does not notice any delay between pressing the trigger button and image capture. Autofocus is particularly important for cameras that cannot be manually focused (such as compact digital cameras or camera phones).

[0005] To this end, camera manufacturers are developing image sensors with on-chip phase detection, which are image sensors that integrate phase detection capability by including phase detection autofocus (PDAF) pixels in the pixel array of the image sensor. Summary of the Invention

[0006] In one aspect, an image sensor includes a multi-pixel detector. The multi-pixel detector includes a first pixel formed in a substrate and having a first photodiode region, a second pixel formed adjacent to the first pixel in the substrate and having a second photodiode region, and a microlens above the first and second pixels. The microlens includes (a) a first height profile having N1 local maxima in a first cross-sectional plane perpendicular to the upper surface of the substrate and including both the first and second photodiode regions, and (b) a second height profile having N2>N1 local maxima in a second cross-sectional plane perpendicular to both the first cross-sectional plane and the upper surface and including only one of the first and second photodiode regions.

[0007] In a second aspect, a method for increasing the angular sensitivity of a multi-pixel detector to light incident thereon is disclosed. The multi-pixel detector has a substrate including a first photodiode region adjacent to a second photodiode region in a first direction. The method includes imparting a phase shift to the light so that the light has: an elliptical intensity distribution in a cross-sectional plane intersecting the substrate, parallel to the upper surface of the substrate, and including the first direction, the elliptical intensity distribution having (i) a first width in the first direction and (ii) a second width exceeding the first width in a second direction perpendicular to the first direction and in the cross-sectional plane.

[0008] In a third aspect, the method for forming a microlens includes forming a block of reflowable material on an upper substrate surface of a substrate, the block of reflowable material having an upper surface opposite to the upper substrate surface, and having a notch extending toward the upper substrate surface in the upper surface. The method also includes thermally reflowing the block of material to produce a microlens. Attached Figure Description

[0009] Figure 1 illustrates an exemplary image sensor with PDAF pixels in an exemplary use case.

[0010] Figure 2 This is a cross-sectional view of the multi-pixel detector of the pixel array in Figure 1.

[0011] Figure 3 yes Figure 2 A schematic diagram of the angular sensitivity of a multi-pixel detector.

[0012] Figure 4 This is a plan view of a multi-pixel detector including microlenses for increasing angular sensitivity in the embodiment.

[0013] Figure 5 , Figure 6 and Figure 7 yes Figure 4 A cross-sectional view of the PDAF pixel detector.

[0014] Figure 8 Compared to Figure 2 PDAF pixel detector Figure 4 A graphical depiction of the improved angular sensitivity of an embodiment of the PDAF pixel detector.

[0015] Figure 9A It has illumination incident on its microlenses. Figure 2 A top view of the PDAF pixel detector. Figure 9B and Figure 9C This illustrates the effect of the microlens on incident illumination.

[0016] Figure 10A It has illumination incident on its microlenses. Figure 4A top view of an embodiment of the PDAF pixel detector. Figure 10B and Figure 10C The effect of the microlens on incident illumination is shown in the embodiment.

[0017] Figure 11 In the embodiments Figure 4 A plan view of a notched block on the upper surface of the color filter of a multi-pixel detector.

[0018] Figure 12 yes Figure 11 A cross-sectional view of a block with a notch.

[0019] Figure 13 This is a flowchart of a method for increasing the angular sensitivity of a multi-pixel detector to incident light, as described in the embodiments.

[0020] Figure 14 This is a flowchart of the method for forming a microlens in the embodiment. Detailed Implementation

[0021] Figure 1 illustrates an exemplary image sensor 101 with PDAF pixels in an exemplary use case 190. The image sensor 101 includes a pixel array 100A and is implemented in a digital camera 180 for imaging scene 150. The digital camera 180 is, for example, a camera phone or a compact digital camera. The digital camera 180 uses the on-chip phase detection capability of the image sensor 101 to focus on scene 150. When focused, the digital camera 180 uses the image sensor 101 to capture a focused image 120 of scene 150, instead of a defocused image 130.

[0022] In one embodiment, the pixel array 100A includes at least one multi-pixel detector 200. Figure 2 This is a cross-sectional view of a multi-pixel detector 200. The multi-pixel detector 200 includes pixels 210 and 220, which have a shared color filter 240 and a microlens 250. Pixels 210 and 220 are formed in a shared substrate 202, which may be formed of one or more semiconductor materials.

[0023] Color filter 240 has a height 242 and a top surface 240T. Pixels 210 and 220 each have their own top surfaces 210T and 220T. Pixels 210 and 220 have a combined detector width 216X. Pixels 210 and 220 each have their own photodiode regions 212 and 222 separated by the inter-diode region 203 of the substrate 202. Vertical lines 215 represent planes equidistant from the photodiode regions 212 and 222 of each pixel 210 and 220. Photodiode regions 212 and 222 have a depth 222D.

[0024] The applicant has determined that the accuracy of PDAF depends on the angular sensitivity of the multi-pixel detector 200. Figure 3 Figure 300 shows the simulated angular sensitivity of the multi-pixel detector 200, where the detector width 216X equals 1.0 μm and the microlens 250 has a radius of curvature equal to 0.73 μm. Figure 300 includes the pixel response 310 of the left pixel 210 and the pixel response 320 of the right pixel 220 as a function of the incident light angle θ, where θ = 0 corresponds to light propagating perpendicularly to surfaces 210T and 220T. Pixel 220 has a peak response for positive incident light angles θ > 0. Pixel 210 has a peak response for negative incident light angles θ < 0. Both pixel responses 310 and 320 have a minimum value 312 and a maximum value 313.

[0025] The angular sensitivity of pixels 210 and 220 (or equivalently, multi-pixel detector 220) is measured as the ratio of a maximum value 313 to a minimum value 312. In the simulation that produces Figure 300, this maximum value 313 is equal to 1.65, and the minimum value 312 is equal to 0.41, thus the ratio of the two is equal to 4.02.

[0026] Figure 4 This is a plan view of the multi-pixel detector 400. The multi-pixel detector 400 is similar to the multi-pixel detector 200, except that the microlens 250 is replaced by the microlens 450.

[0027] Figure 5-7 It corresponds to each in Figure 4 The cross-sectional views of the multi-pixel detector 400 shown in the cross-sectional planes 5A / 5B, 6 and 7 are shown in the figure. Figure 4-7 A multi-pixel detector 400 is shown relative to a coordinate system 498 formed by axes x, y, and z, where the xy plane is parallel to the upper surface 240T. This is best viewed together in the following description. Figure 4-7 Although the cross-sectional plane 5A spans the photodiode region 212, it can equivalently span the photodiode region 222 or be located at the cross-sectional plane 5B between photodiode regions 212 and 222. Herein, or unless otherwise stated, any direction or plane represented by at least one of x, y, or z refers to coordinate system 498. The multi-pixel detector 400 has a detector width 216X in the x-direction and a detector width 216Y in the y-direction, which can be equal to the detector width 216X.

[0028] Figure 4The plan view includes pixels 210 and 220 below the surface 240T of the color filter 240. Microlens 450 is above the surface 240T. Photodiode regions 212 and 222 are shown with dotted lines to illustrate their position below the color filter 240 and relative to microlens 450. Microlens 450 has an outer edge 450E that can extend to cover the entirety of photodiode regions 212 and 222.

[0029] Microlens 450 in Figure 5 The cross-sectional view shows a height profile 550. Height profile 550 has regions 510 and 520, each with a peak height 511 and 521 above the upper surface 240T, and respective radii of curvature 512 and 522. Peak heights 511 and 521 may be equal or unequal. Similarly, radii of curvature 512 and 522 may be equal or unequal. Peak heights 511 and 521 are separated by a distance 517 in a direction parallel to the upper surface 210T. The ratio of distance 517 to detector widths 216X, 216Y may be 0.50 ± 0.05, so that microlens 450 can be sized to accommodate the dimensions of pixels 210 and 220. Microlens 450 has an intermediate height 515, which is greater than 0 and less than each of peak heights 511 and 521.

[0030] Microlens 450 in Figure 6 and Figure 7 The cross-sectional views show height profiles 650 and 750, respectively. Height profiles 650 and 750 have respective peak heights 611 and 711 above the upper surface 240T, and respective radii of curvature 612 and 712. Peak heights 611 and 711 may be equal or unequal. Similarly, radii of curvature 612 and 712 may be equal or unequal. Although Figure 6 and 7 Peak heights 611 and 711 are shown above the inter-diode region 203. One or both of peak heights 611 and 711 may be above photodiode regions 212 or 222, or more generally, eccentric relative to photodiode regions 212 and 222. Figure 6 and 7 Show Figure 5 The cross-sectional plane 5B. When cross-sectional plane 5A and cross-sectional plane 5B are placed side by side, the peak heights 611 and 711 can be equal to the peak heights 511 and 521, respectively.

[0031] Each radius of curvature in the xz plane (radii 612 and 712) can exceed all the radii of curvature in the yz plane (radii 512 and 522) so that the microlens 450 has greater focusing power in the yz plane than in the xz plane.

[0032] Figure 8This is a graphical depiction of the improved angular sensitivity compared to an embodiment of multi-pixel detector 400, which is different from multi-pixel detector 200. In this example, the microlens 250 of multi-pixel detector 200 has a radius of curvature equal to 0.73 μm. In this embodiment of multi-pixel detector 400, the microlenses 450 have radii of curvature 512 and 522 respectively equal to 0.40 ± 0.05 μm and radii of curvature 612 and 712 respectively equal to 0.58 ± 0.05 μm, and the detector widths 216X and 216Y are both equal to 1.0 μm.

[0033] Each as Figure 2 and 6 As shown, both multi-pixel detector 200 and multi-pixel detector 400 include right pixel 220. Figure 8 The pixel responses 820 and 840 of the right pixel 220 in multi-pixel detectors 200 and 400 are shown, respectively. Pixel responses 820 and 840 are functions of the incident light angle θ, where θ = 0 corresponds to light propagating perpendicularly to surfaces 210T and 220T, as shown. Figure 3 As shown. The incident light angle θ is within ±θ max Change, θ max Approximately 12.5°. Pixel response 820 and 840 across ±θ max The respective ranges are 825 and 845. Range 845 exceeds range 825, which is an indication that the PDAF detector 400 has higher angular sensitivity than the multi-pixel detector 200.

[0034] Figure 9A This is a schematic plan view of the multi-pixel detector 200, which includes microlens 250 and pixels 210 and 220. Illumination 900 is incident on microlens 250 at an incident angle θ and propagates toward pixels 210 and 220. Figure 3 The incident angle θ is shown.

[0035] Figure 2 A cross-sectional plane 9 is shown that is parallel to surfaces 210T and 220T and intersects with photodiode regions 212 and 222. Figure 9B and 9C This is a cross-sectional view of the multi-pixel detector 200 in the cross-sectional plane 9. When the incident angle θ is equal to 0, the microlens 250 imparts a phase shift to the illumination 900 so that it has an intensity distribution 902 in the cross-sectional plane 9, such as... Figure 9B As shown. The intensity distribution 902 has a diameter of 902W and is centered on pixels 210 and 220, so that the intensity distribution 902 is equally distributed between pixels 210 and 220.

[0036] When the incident angle θ equals 10°, the microlens 250 imparts a phase shift to the illumination 900 so that it has an intensity distribution 912 in the profile plane 9, such as Figure 9CAs shown. For the sake of discussion, intensity distribution 912 has a diameter 912W that differs from diameter 902W by an insignificant amount. Intensity distribution 912 is laterally offset from vertical line 215 by a distance Δx9 so that it is unequally distributed between pixels 210 and 220. Figure 9C Two regions 912a and 912b are shown in the intensity distribution 912, located within pixels 210 and 220, respectively. Regions 912a and 912b each have an area A. 912a and A 912b And has more than R 912 =A 912a / A 912b The diameters 902W and 912W can correspond to the full width at half maximum (FWHM) of the intensity distributions 902 and 912, respectively, or 1 / e 2 width.

[0037] Figure 10A This is a schematic plan view of the PDAF detector 400, which includes microlens 450 and pixels 210 and 220. Illumination 900 is incident on microlens 450 at an incident angle θ and propagates toward pixels 210 and 220. Figure 3 The incident angle θ is shown. Microlens 450 has a refractive index n and a complex amplitude transmittance t(x,y) = exp(-φ(x,y)), where the phase shift φ(x,y) = nk0d(x,y), where k0 = 2π / λ0, λ0 is the free-space wavelength included in illumination 900, and d(x,y) is a function of the thickness of microlens 450. The complex amplitude transmittance t(x,y) determines the phase shift attributed by microlens 450 to illumination 900 transmitted through it. Figure 5-7 The height profiles 550, 650, and 750 are graphical representations of the thickness function d(x,y) in the selected profile.

[0038] In this embodiment, the microlens 450 has a constant thickness d. g and spatially varying refractive index n g A biplane graded refractive index plate (x,y) where n g (x,y)·d g = n·d(x,y). In this example, the biplane graded refractive index plate imparts a phase shift φ. g (x,y)=φ(x,y).

[0039] Figure 10B and 10C yes Figure 6 and 7A cross-sectional view of the PDAF detector 400 in cross-sectional plane 9. When illumination 900 propagates through microlens 450 with normal incidence (θ equals 0), microlens 450 imparts a phase shift to illumination 900 so that it has an intensity distribution 1002 in cross-sectional plane 9, such as... Figure 10B As shown. Intensity distribution 1002 has a maximum width of 1002X and a height of 1002Y (in the x and y directions respectively), and is centered on pixels 210 and 220 so that intensity distribution 1002 is equally distributed between pixels 210 and 220. Intensity distribution 1002 and intensity distribution 902 may have equal areas.

[0040] When illumination 900 propagates through microlens 450 at an incident angle θ = 10°, microlens 450 imparts a phase shift to illumination 900 so that it has an intensity distribution 1012 in the cross-sectional plane 10, such as Figure 10C As shown. Intensity distribution 1012 has a maximum width 1012W and a height 1012Y (in directions x and y, respectively). For the sake of discussion, the maximum width 1012W and height 1012Y differ from the maximum width 1002X and height 1002Y by insignificant amounts, respectively. Widths 1002X and 1012W can correspond to the full width at half maximum (FWHM) or e- of intensity distributions 1002 and 1012, respectively. 2 width.

[0041] Intensity distribution 1012 with distance Δx 10 Laterally offset from vertical line 215 so that it is unequally distributed between pixels 210 and 220. Figure 10C Two regions 1012a and 1012b are shown in the intensity distribution 1012, located within pixels 210 and 220, respectively. Regions 1012a and 1012b each have an area A. 1012a and A 1012b And has more than R 1012 =A 1012a / A 1012b Compared to R 1012 More than R 912 The reason is that the maximum width 1002X of intensity distribution 1002 is smaller than the diameter 902W of intensity distribution 902. For example, when Δx9 = Δx 10 This inequality holds true when R is greater than R. 912 and R 1012 This difference demonstrates the higher angular sensitivity of the PDAF detector 400 compared to the multi-pixel detector 200, as shown in... Figure 8 The reason is that the change in the incident angle θ causes a large difference in pixel response between adjacent pixels 210 and 220.

[0042] Microlenses 450 can be formed by subjecting a notched block of material to a heat reflow process. Figure 11 and 12 An example of such a gapped block 1100 is shown. Figure 11 This is a plan view of the notched block 1100 on the upper surface 240T of the color filter 240. Figure 12 This is a cross-sectional view of the notched block 1100 in the yz plane of coordinate system 498. The microlens 450 and the notched block 1100 can be constructed from components having a T... min =140℃ and T max Glass transition temperature T between 180℃ g The material formation enables the formation of the microlens 450 via a thermal reflow process. For example, the notched block 1110 and the microlens 450 are formed using positive photoresist. In this embodiment, pixels 210 and 220, along with their respective photodiode regions 212 and 222, are located below the color filter 240. The photodiode regions 212 and 222 have respective widths 214X and 224X and a common length 214Y.

[0043] The notched block 1100 has spatial dimensions 1102X, 1102Y, and 1102Z. The notched block 1100 has a notch 1110 along the x-direction defined by sidewalls 1111 and 1112. The notch 1110 has a width 1116Y between sidewalls 1111 and 1112 and a depth 1116Z from the upper surface 1100T. The width 1116Y can be smaller than the spatial dimension 1102X. Although... Figure 11 The notch 1110 is shown as having a size of 1102X, and without deviating from its range, the notch 1110 may have a length less than 1102X.

[0044] The notched block 1100 has a thickness 1103Z within the notch 1110. The thickness 1103Z can be equal to 0 so that the notched block 1100 is only nominally a single block, since in this case it comprises two distinct and spatially separated blocks.

[0045] Figure 13 This is a flowchart illustrating a method 1300 for increasing the angular sensitivity of a multi-pixel detector to incident light. The multi-pixel detector has a substrate including a first photodiode region adjacent to a second photodiode region.

[0046] Method 1300 includes step 1310. Step 1310 is to impart a phase shift to the light so that the light has an elliptical intensity distribution in a cross-sectional plane that intersects the substrate, is parallel to the upper surface of the substrate, and includes a first direction. The elliptical intensity distribution has a first width in the first direction, and the first width is smaller than a second width in a second direction perpendicular to the first direction.

[0047] In the example of step 1310, the microlens 450 imparts a phase shift to the illumination 900 incident on the PDAF detector 400. Light transmitted through the microlens 450 is phase-shifted in the cross-sectional plane 10. Figure 6 and 7 The intensity distribution 1002 has a maximum width 1002X and a height 1002Y, the maximum widths 1002X and 1002Y corresponding to the first width and the second width in step 1310.

[0048] In step 1310, the phase shift can have (i) an absolute value with N1 local maxima in a first direction, and (ii) an absolute value with N2>N1 local maxima in a second direction. For example, the microlens 450 imparts a phase shift φ(x,y) to the incident illumination 900, wherein the phase shift φ(x,y) has one local maximum in the x-direction and two local maxima in the y-direction by means of the thickness function d(x,y).

[0049] Figure 14 This is a flowchart illustrating a method 1400 for forming a microlens. Method 1400 includes steps 1410 and 1420. Step 1410 involves forming a block of reflowable material on an upper substrate surface of a substrate. The block of reflowable material has an upper surface opposite to the upper substrate surface, wherein it has a notch extending toward the upper substrate surface. In an example of step 1410, the notched block 1100 is formed on the upper surface 240T of a color filter 240, as shown... Figure 11-12 As shown. The notched block 1100 includes a notch 1100 that extends from the upper surface 1100T toward the upper surface 240T.

[0050] For example, when the substrate includes a first photodiode region adjacent to the second photodiode region in the first direction, step 1410 may further include step 1412. Step 1412 is to form a block such that the notch has a width in a second direction perpendicular to the first direction and a length exceeding the width in the first direction. In the example of step 1412, a block 1100 with a notch and adjacent to the photodiode region in the x-direction is formed on the upper surface 240T of the color filter 240 such that the groove 1110 is longitudinally oriented and parallel to the x-direction.

[0051] Step 1420 is to thermally reflow the blocks of the material to form microlenses. In an example of step 1420, the notched block 1100 is thermally reflowed to form the microlens 450, Figure 4-7 .

[0052] Without departing from its scope, the features described above and those claimed below can be combined in various ways. The following examples only show some possible, non-limiting combinations:

[0053] (A1) The image sensor includes a multi-pixel detector. The multi-pixel detector includes a first pixel formed in a substrate and having a first photodiode region, a second pixel formed in the substrate adjacent to the first pixel and having a second photodiode region, and a microlens above the first pixel and the second pixel. The microlens includes (a) a first height profile having N1 local maxima in a first cross-sectional plane perpendicular to the upper surface of the substrate and including both the first and second photodiode regions, and (b) a second height profile having N2 > N1 local maxima in a second cross-sectional surface perpendicular to the first cross-sectional plane and the upper surface and including only one of the first and second photodiode regions.

[0054] (A2) In the image sensor as represented in (A1), the first height profile can have N1 = 1 local maximum and be characterized by a first radius of curvature R1. The second height profile can include a first local maximum and a second local maximum corresponding to respective radii of curvature R 21 < R1 and R 22 < R1.

[0055] (A3) In any of the image sensors as represented in (A2), the radius of curvature can satisfy R1 = 730 ± 70 nanometers.

[0056] (A4) In any of the image sensors as represented in one of (A2) and (A3), the radii of curvature R 21 and R 22 can be considered equal when they differ by within ten percent of (R 21 + R 22 ) / 2.

[0057] (A5) In any of the image sensors as represented in one of (A2) to (A4), the radii of curvature R 21 and R 22 can be equal to 580 ± 60 nm.

[0058] (A6) In any of the image sensors as represented in one of (A1) to (A5), the microlens can have a maximum thickness of 730 ± 70 nanometers.

[0059] (B1) represents a method for increasing the angular sensitivity of a multi-pixel detector to incident light, the multi-pixel detector having a substrate including a first photodiode region adjacent to a second photodiode region in a first direction. The method (B1) includes imparting a phase shift to the light such that the light has an elliptical intensity distribution in a cross-sectional plane intersecting the substrate, parallel to the upper surface of the substrate, and including the first direction, the elliptical intensity distribution having (i) a first width in the first direction and (ii) a second width exceeding the first width in a second direction perpendicular to the first direction and in the cross-sectional plane.

[0060] (B2) In method (B1), the phase shift may have (i) an absolute value with N1 local maxima in the first direction and (ii) an absolute value with N2>N1 local maxima in the second direction.

[0061] (C1) indicates a method for forming a microlens. Method (C1) includes forming a block of reflowable material on an upper substrate surface of a substrate, the block of reflowable material having an upper block surface opposite to the upper substrate surface, wherein a notch extends toward the upper substrate surface. The method further includes thermally reflowing the block of reflowable material to produce a microlens.

[0062] (C2) In method (C1), wherein the substrate includes a first photodiode region adjacent to the second photodiode region in a first direction, the forming step may include forming a block such that the notch has a width in a second direction perpendicular to the first direction and a length in the first direction exceeding the width.

[0063] (C3) In either of the representations of (C1) and (C2), the microlens may have the properties of a microlens of any image sensor represented by one of (A1) to (A6).

[0064] Modifications may be made to the image sensor, multi-pixel detector, and related methods described above without departing from their scope. Therefore, it should be noted that the manner in which these are included in the foregoing description or shown in the accompanying drawings should be understood as illustrative and not limiting. Herein, and unless otherwise indicated, the adjective “exemplary” means used as an example, illustration, or description. The appended claims are intended to cover all general and specific features described herein, and all linguistic claims regarding the scope of the method, image sensor, and detector should be considered to fall within their scope.

Claims

1. A method for increasing angular sensitivity of a multi-pixel detector to light incident thereon, the multi-pixel detector having a substrate including a first photodiode region adjacent to a second photodiode region in a first direction, the method comprising: imposing a phase shift on the light so that the light has an elliptical intensity distribution in a cross-sectional plane intersecting the substrate, parallel to an upper surface of the substrate, and including the first direction, the elliptical intensity distribution having (i) a first width in the first direction and (ii) a second width in a second direction perpendicular to the first direction and in the cross-sectional plane that exceeds the first width.

2. The method of claim 1, wherein, The phase shift has (i) an absolute value in the first direction having Nl local maxima and (ii) an absolute value in the second direction having N2>Nl local maxima.

3. A method for forming a microlens on an upper substrate surface of a substrate, the substrate including a first photodiode region adjacent to a second photodiode region in a first direction, the method comprising: forming a mass of reflowable material on the substrate surface, the mass of reflowable material having an upper mass surface opposite the upper substrate surface, the upper mass surface having a notch extending toward the upper substrate surface, the notch having a width in a second direction perpendicular to the first direction and a length in the first direction, the length exceeding the width; and thermally reflowing the mass of reflowable material to produce the microlens; wherein the microlens imposes a phase shift on illumination incident to a multi-pixel detector so that light transmitted through the microlens has an elliptical intensity distribution in a cross-sectional plane intersecting the substrate, parallel to an upper substrate surface of the substrate, and including the first direction, the elliptical intensity distribution having (i) a first width in the first direction and (ii) a second width in a second direction perpendicular to the first direction and in the cross-sectional plane that exceeds the first width. In the step of thermally reflowing, the microlens has:

4. The method of claim 3, wherein, a first height profile having Nl local maxima in a first cross-sectional plane perpendicular to the upper surface of the substrate and including both the first photodiode region and the second photodiode region; and a second height profile having N2>Nl local maxima in a second cross-sectional plane perpendicular to the first cross-sectional plane and the upper surface and including only one of the first photodiode region and the second photodiode region.

5. The method of claim 4, wherein, the first height profile has Nl = 1 local maxima and is characterized by a first radius of curvature Rl; and the second height profile has N2>Nl local maxima and is characterized by a second radius of curvature R2. ​ Said second height profile comprises a respective curvature radius R 21 < R1 and R 22 < first and second local maxima of R1.

6. The method of claim 5, wherein, The radius of curvature R1 = 730 70 nanometers.

7. The method of claim 5, wherein, The radius of curvature R 21 and R 22 In equal within ten percent.

8. The method of claim 5, wherein, the radius of curvature R 21 and R 22 equal to 580 60 nm.

9. The method of claim 4, wherein, The microlenses have 730 70 nanometers of maximum thickness.

Images