Optical spectroscopy has been widely used to detect and quantify characteristics and concentrations of physical, chemical, or biological targets. A limitation to this technology is spectrometer size and cost. More specifically, spectrometers traditionally use a light dispersion method. A light dispersion system, which may include a prism or diffraction grating, disperses incoming light from a target sample into an optical spectrum, i.e., into components of different wavelengths. This optical spectrum is then scanned by an optical detector to investigate the spectral characteristics. Spectrometer size needs to exceed a volume required to accommodate both this light dispersion and the light dispersion system itself, which also contributes to spectrometer cost.
Spectrometers that use non-dispersion methods have been developed. For example, one type of spectrometer uses a metallic-dielectric layered structure (with nanoscale metallic embossing structures on a metal film) to filter incoming light. Based on the working principles of surface plasmon polariton (SPP), this metallic-dielectric filter selects a narrow wavelength band of light to pass through it, while blocking the rest of the light spectrum with surface plasmon. Since such a system does not require a bulky light-dispersion system, the spectrometer size is significantly reduced. However, fabricating SPP-based metallic-dielectric filters requires a complex wafer manufacturing process so spectrometer cost remains high.
In one embodiment, a stacked-filter pixel is disclosed. A stacked-filter pixel has a substrate and a color filter stack. The substrate includes a photodetector element electronically coupled to pixel circuitry. The color filter stack has a first color filter, a second color filter, and a first inter-filter layer therebetween. The second color filter is located between the photodetector element and the first color filter and has a second passband that partially overlaps a first passband of the first color filter in a first overlapping wavelength range. The first inter-filter layer is at least partially transparent to the second passband.
In one embodiment, a stacked-filter image-sensor spectrometer is disclosed. The stacked-filter image-sensor spectrometer includes an image sensor, a first color filter array, and a second color filter array. The image sensor has a pixel array including a plurality of pixels. The first color filter array has a plurality of first color filters, wherein each first color filter is located above at least one pixel. The second color filter array is located between the first color filter array and the image sensor and has a plurality of second color filters. Each second color filter is located above at least one pixel. Each of the plurality of pixels has thereabove a compound color filter formed of one of the second color filters and one of the first color filters; the second filter has a second passband that partially overlaps a first passband of the first color filter in a first overlapping wavelength range.
In the present disclosure, a stacked filter image-sensor spectrometer uses a multi-layer stacked color filter to select a relatively narrow wavelength band for detection. The color filter works by light absorption, and is fabricated with standard photolithography processes. Cost of the final spectrometer product is accordingly low as compared to the prior art.
Substrate 110 includes a frontside 110F and a backside 110B. To implement electromagnetic radiation (e.g., one or more of visible and near-infrared) detection, pixel 100 includes a color filter 130 disposed under a microlens 140, which focuses incident light onto photodiode regions 112. Color filter 130 may transmit any color with a wavelength within the visible wavelength range.
Compound color filter 230 may include an inter-filter layer 233 situated directly between first color filter 231 and second color filter 232. Inter-filter layer 233 for example strengthens the structural integrity of the first and second color filters 231 and 232. It may also serve as a barrier to prevent or reduce diffusion of dyes and color pigments between color filters 231, 232. Inter-filter layer 233 may be at least partially transparent to the passband of second color filter 232. If inter-filter layer 233 lacks such transparency, second color filter 232 may not transmit any light incident on microlens 140. Herein, a color filter's passband refers to a range of wavelengths that the filter transmits above a specified value, such as a full-width half-maximum transmission value.
Inter-filter layer 233 may be made of various materials with optically transparent properties, including photoresist, resins, polymers, dielectrics, thin sheet of metals, etc. Inter-filter layer 233 may be formed of a dielectric material that is optically transparent, physically and chemically stable, and is able to stop diffusion of dyes and color pigments. Inter-filter layer 233 is, for example, made of a material that has a refractive index n233 that is similar to the refractive index of the two color filters above and below it, in order to take advantage of index matching to reduce interface optical loss. Generally speaking, color filter materials are photoresist/resin with a refractive index of around 1.7, whereas the refractive index of commonly used dielectric in semiconductor devices is either too high or too low. For example, the refractive index of silicon oxide is too low (n≈1.46), while the refractive index of silicon nitride is too high (n≈2.0). Care should be taken to select the proper material for the inter-filter layer, so that its refractive index is close to the color filters, e.g., an arithmetic mean or geometric mean of the refractive indices of the top and the bottom color filters.
Inter-filter layer 233 has a thickness 233T. In an embodiment, inter-filter layer 233 a single-layer or multi-layer designed to be an anti-reflective coating between color filters 231 and 232 at a wavelength λc, transmitted by both filters 231 and 232. For example, inter-filter layer 233 is a single-layer thin film with refractive index n233 equal to a geometric average of the refractive indices of filters 231 and 232, where thickness 233T equals 0.25 λc/n233.
A candidate of this inter-filter material may include a glass substance, e.g., silicon oxide, which is doped with metal oxide dopants, e.g., titanium oxide, or zirconium oxide, such that the resulting refractive index is around 1.7. Another candidate may include silicon oxynitride, with its refractive index tuned to be around 1.7. Yet another candidate may be a transparent polymer, such as polycarbonate, with a refractive index of around 1.7.
First color filter 231 may be characterized by a first passband. For example, the first color filter 231 includes chemical dye and/or pigment that absorbs certain wavelengths of light, thereby permitting transmission of light within a certain range of wavelengths complementary to the absorbed wavelengths. This type of color filter is based on absorption and is different from other filtering such as destructive interference (dichroic filter) and surface plasmon polariton. Second color filter 232 works similarly as the first color filter 231, and may be characterized by a second passband. The first and second passbands may be different, but they share a common overlapping wavelength range. For example, the first passband may be 500-550 nm, whereas the second passband may be 525-575 nm; thus the common overlapping wavelength range is 525-550 nm (i.e., the lower bound of the second passband and the upper bound of the first passband).
Compound color filter 230 may be characterized by a net passband. At least one of the first passband, the second passband, and the net passband may correspond to near-IR or IR wavelengths, e.g., wavelengths exceeding 0.75 micrometers. At least one of the first passband, the second passband, and the net passband may span visible and near-IR wavelengths. The net passband may have a center wavelength equal to wavelength λc, introduced above as a design wavelength for when inter-filter layer 233 is a single-layer antireflective coating.
Transmission spectra 310 and 320 are Lorentzian functions with respective center wavelengths λ1 and λ2 and full-width half-maxima (FWHM) FWHM1 and FWHM2. In this example, FWHM1=FWHM2=50 nm, λ1=500 nm, and λ2=λ1+α·FWHM1=525 nm, where α=1/2. Transmission spectra 310 and 320 have overlapping passbands, which may be defined by the center wavelength and a linewidth such as a FWHM width, a e−1 linewidth, or other conventions known in the art. Transmission spectra 310 and 320 have FWHM passbands of 500±25 nm and 525±25 nm respectively, which overlap as illustrated in
As spectra describable by a standard continuous probability distribution, transmission spectra 310 and 320 may represent idealizations of actual transmission spectra of color filters 231 and 232, which may be asymmetric.
Net transmission does not always have a sharper appearance than the two transmission spectra. However, in the present example, net transmission spectrum 315 does have a sharper appearance than both transmission spectra 310 and 320 under the condition that λ2=λ1+α·FWHM1 where α is less than approximately 0.67 and FWHM1=FWHM2. By overlapping the first and second color filters 231 and 232, and also purposefully controlling spectral parameters such as λ1, λ2, α, FWHM1, FWHM2, etc, a narrower wavelength range optical transmission may be achieved, albeit the maximum transmission level is generally reduced compared to transmission spectra 310 and 320. Narrower wavelength range optical transmission is a desired feature in spectroscopic analysis because it allows for more accurate wavelength related analysis.
The practice of overlapping two color filters that share a common overlapping wavelength range to achieve a relatively narrower wavelength range of optical transmission may be further extended to overlapping three or more color filters. See, e.g,.
For example, and as shown in
In an embodiment, color filters 231 and 232 have Gaussian (or approximately Gaussian) transmission spectra having respective center wavelengths λ1 and λ2 and respective passband spectral widths σ1 and σ2, which denote a standard deviation of their respective Gaussian transmission spectrum. Compound color filter 230 has a center wavelength
and a spectral width σ3=σ1σ2/√{square root over (σ12+σ22)} that is less than both σ1 and σ2. For example, when σ1=σ2, σ3=σ1/√{square root over (2)}. More generally, when
which illustrates that σ3 is less than both σ1 and σ2.
For any given wavelength, its passband spectral widths centering on that wavelength may vary, depending on the color filter's material composition and thickness. Purposeful selection of relatively narrow passband spectral widths σ1 and σ2 will help to narrow the resulting compound color filter's passband spectral width σ3. For example, an appropriate blue (centered around 470 nm) filter material composition and thickness may be selected so that its spectral width σ≈70 nm; similarly a green filter (centered around 540 nm) may be selected have its spectral width σ≈80 nm, and a red filter (centered around 660 nm) may be selected to have spectral width σ≈67 nm. Generally speaking, the filter of a specific color may be purposefully selected so that its spectral width a is around a relatively narrow range of 70-80 nm. Then, the resulting compound filter (e.g., a two-layer filter) may have its spectral width σ around 50-55 nm, which is around 60%-80% of any single color filter that is a component of the compound filter.
Stacked-filter image-sensor spectrometer 400 may also include at least one of a light collector 404 and a diffuser 406 in front of image sensor 410. Light collector 404 is for example a lens or an axicon.
Memory 432 may also store image sensor calibration data 434, which is based on properties of pixel array 405A such as each pixel's gain and transmission function of its color filter, e.g., compound color filter 230 or 280. Sensor calibration data 434 may be based on other properties of image sensor 410 without departing from the scope hereof. In an embodiment, data processor 430 receives calibration data 434 directly from image sensor 410. Alternatively, data processor 430 may include a calibration data generator 444 that generates calibration data 434 by processing data received from image sensor 410.
Readout circuitry 414 may include one or more of amplification circuitry, analog-to-digital (“ADC”) conversion circuitry, and other circuits. Control circuitry 412 is coupled to pixel array 405A to control operational characteristics of pixel array 405A. For example, control circuitry 412 may generate a shutter signal for controlling data acquisition.
Pixel array 405A is a two-dimensional array of individual pixels Pi (e.g., pixels P1, P2 . . . , Pn)having X pixel columns and Y pixel rows. Each pixel Pi, is for example either stacked-filter FSI pixel 200 or stacked-filter BSI pixel 250. As pixel array 405A includes stacked-filter pixels and is part of image sensor 410, image sensor 410 is an example of a stacked-filter image sensor. As illustrated in
In operation of stacked-filter image-sensor spectrometer 400, light 491 emitted or reflected from target sample 490 is incident on pixel array 405A. A plurality of pixels of pixel array 405A generates a photocurrent to readout circuitry 414. Readout circuitry 414 generates and outputs spectral data 419 to memory 432 of data processor 430.
In an embodiment, light collector 404 is positioned to image sample 490 onto pixel array 405A. In a different embodiment, light collector 404 is positioned to maximize the amount of light emitted or reflected from sample 490, regardless of whether sample 490 is imaged onto pixel array 405A. In such a configuration, optimized light collection may result in sample 490, light collector 404, and image sensor 410 being longitudinally positioned in a non-imaging configuration. Such a non-imaging configuration may be beneficial for reducing the volume of spectrometer 400. For example, a longitudinal distance between light collector 404 and image sensor 410 is less than a focal length of light collector 404, such that light collector 404 cannot form an image on image sensor 410.
Spectrum generator 441 is capable of constructing an optical spectrum 450 from spectral data 419 and, optionally, calibration data 434. Spectrum processor 443, if included, is capable of further processing and analyzing spectral data 419, for example, to ascertain physical, chemical, or other attributes of sample 490.
Each filter pair 231(i) and 232(i) combines to yield a net transmission spectrum, such as net transmission spectrum 315, characteristic of a pixel 505(i), and hence what is referred to herein as a detector group of stacked-filter image-sensor spectrometer 400 characterized by a net transmission spectrum. The number of candidate color filter types available to each filter pair 231(i) and 232(i) determines the number of possible detector groups of spectrometer 400. Specifically, the maximum number of detector groups Ng equals the combination of k color filter layers and n color filter types:
For k=2 CFA layers as shown in
For any same color filter pair situation, the resulting compound filter's passband spectral width will be relatively wider than a pair of different color filters.
In
Image-sensor spectrometer 400 is shown with ten detector groups (represented by respective patterns of P1-P10) for illustrative purposes, and may have more or fewer detector groups without departing from the scope hereof. Hereinafter, P1-N
Detector groups P1-N
Detector groups P1-N
CFA 732A includes a plurality of color filters 732 that span more than one pixel 705. Optical properties of color filters 732, e.g., candidate passbands and transmission spectra, are similar to those of color filters 231. CFA 732A may also include a color filter that is beneath one and only one pixel 705. CFA 531 may include a color filter spanning more than one pixel 705.
One color filter 732(i) may span any number of contiguous pixels 705 forming different shapes. For example, a plurality of contiguous pixels 705 form an m x n array of pixels are m and n are positive integers. Alternatively, a plurality of contiguous pixels 705 may form a non-rectangular shape, for example one formed by intersecting rectangles such as an L-shape or a cross. In an embodiment, the plurality of contiguous pixels 705 is less than twenty-five percent of the total number of pixels in pixel array 705A.
In an embodiment, no color filter 732(i) of CFA 732A spans every pixel 705 of pixel array 705A, as CFA 732A includes other color filters 732(j≠i), that each cover at least one pixel 705 of pixel array 705A. When pixel array 705A is planar, CFA 732A may also be planar, such that two color filters 732 therein are coplanar. Hence the two color filters, such as 732(1) and 732(2) of
The above description of illustrated embodiments of the invention, including what is described in the abstract, is not intended to be exhaustive or to limit the invention to the disclosed forms. While specific embodiments of, and examples for, the invention are described herein for illustrative purposes, various equivalent modifications are possible within the scope of the invention, as those skilled in the relevant art will recognize. These modifications can be made to the invention in light of the above detailed description.
Features described above as well as those claimed below may be combined in various ways without departing from the scope hereof. The following examples illustrate some possible, non-limiting combinations:
(A1) A stacked-filter pixel has substrate and a color filter stack. The substrate includes a photodetector element electronically coupled to pixel circuitry. The color filter stack has a first color filter, a second color filter, and a first inter-filter layer therebetween. The second color filter is located between the photodetector element and the first color filter and has a second passband that partially overlaps a first passband of the first color filter in a first overlapping wavelength range. The first inter-filter layer is at least partially transparent to the second passband.
(A2) In the stacked-filter pixel denoted by (A1), the color filter stack may have a net passband with a spectral width narrower than a spectral width of at least one of the first passband and the second passband.
(A3) In any stacked-filter pixel denoted by one of (A1) and (A2) in which the first passband has a first center wavelength and the second passband has a second center wavelength, a difference between the first center wavelength and the second center wavelength may be less than three-quarters of the width of the first passband.
(A4) Any stacked-filter pixel denoted by one (A1) through (A3) may further include, between the first color filter and the second color filter, a first inter-filter layer that is at least partially transparent to the second passband.
(A5) In any stacked-filter pixel denoted by one of (A1) through (A4), the color filter stack may further include a third color filter having a third passband at least partially overlapping the first overlapping wavelength range, the second color filter being between the first color filter and the third color filter.
(A6) The stacked-filter pixel denoted by (A5) may further include a second inter-filter layer between the second color filter and the third color filter, wherein the second inter-filter layer is at least partially transparent to the third passband.
(A7) In any stacked-filter pixel denoted by one of (A5) and (A6), in which the first, second, and third passbands have a first, second, and third center wavelength respectively, a difference between the first center wavelength and the second center wavelength may be less than three-quarters of the width of the first passband, and a difference between the second center wavelength and the third center wavelength may be less than one quarter of the width of the second passband.
(B1) A stacked-filter image-sensor spectrometer includes an image sensor, a first color filter array, and a second color filter array. The image sensor has a pixel array including a plurality of pixels. The first color filter array has a plurality of first color filters, wherein each of the first color filters is located above at least one pixel. The second color filter array is located between the first color filter array and the image sensor and has a plurality of second color filters. Each of the second color filters is located above at least one pixel. Each of the plurality of pixels has thereabove a compound color filter formed of one of the second color filters and one of the first color filters, the second filter having a second passband that partially overlaps a first passband of the first color filter in a first overlapping wavelength range.
(B2) In the stacked-filter image-sensor spectrometer denoted by (B1), each compound color filter may have a net passband that is one of a plurality of net passbands, a first sub-plurality of net passbands spanning a first spectral range, a second sub-plurality of net passbands spanning a second spectral range that does not overlap the first spectral range, the first and second sub-pluralities of net passbands constituting the plurality of net passbands.
(B3) In any stacked-filter image-sensor spectrometer denoted by one of (B1) and (B2), each compound color filter may have a net passband that is one of a plurality of net passbands that includes a first net passband having a first spectral width and a second net passband having a second spectral width that exceeds the first spectral width by at least a factor of five
(B4) In any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B3), (a) one of the plurality of first color filters may be located above more than one pixel and (b) one of the plurality of second color filters may be located above more than one pixel.
(B5) In any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B4), each compound color filter may have a net passband with a spectral width narrower than a spectral width of at least one of the first passband and the second passband
(B6) Any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B4) may further include a light collector above the image sensor, the first and second color filter arrays located between the light collector and the image sensor, a distance between the light collector and the image sensor being less than a focal length of the light collector.
(B7) In any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B6), each compound color filter may include a first inter-filter layer located between the first color filter and the second color filter and being at least partially transparent to the second passband
(B8) In any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B7), the first passband having a first center wavelength, the second passband having a second center wavelength, a difference between the first center wavelength and the second center wavelength may be less than three-quarters of the width of the first passband
(B9) In any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B8), the color filter may further include a third color filter, the second color filter being between the first color filter and the third color filter, and the third color filter having a third passband at least partially overlapping the first overlapping wavelength range.
(B10) The stacked-filter image-sensor spectrometer denoted by (B9) may further include a second inter-filter layer between the second color filter and the third color filter, wherein the second inter-filter layer is at least partially transparent to the third passband.
(B11) In any stacked-filter image-sensor spectrometer denoted by one of (B10) and (B11), in which the first, second, and third passbands have a first, second, and third center wavelength respectively, a difference between the first center wavelength and the second center wavelength may be less than three-quarters of the width of the first passband, and a difference between the second center wavelength and the third center wavelength may be less than one quarter of the width of the second passband.
(B12) In any stacked-filter image-sensor spectrometer denoted by one of (B1) through (B11), each first color filter may be located above at most a first number of contiguous pixels less than one-quarter of pixels in the pixel array, and each second color filter may be located above at most a second number of contiguous pixels less than one-quarter of pixels in the pixel array.
Changes may be made in the above methods and systems without departing from the scope hereof. It should thus be noted that the matter contained in the above description or shown in the accompanying drawings should be interpreted as illustrative and not in a limiting sense. The following claims are intended to cover all generic and specific features described herein, as well as all statements of the scope of the present method and system, which, as a matter of language, might be said to fall therebetween.