U.S. patent application number 13/464955 was filed with the patent office on 2012-08-30 for optimized light guide array for an image sensor.
Invention is credited to Hiok Nam TAY.
Application Number | 20120217377 13/464955 |
Document ID | / |
Family ID | 43513941 |
Filed Date | 2012-08-30 |
United States Patent
Application |
20120217377 |
Kind Code |
A1 |
TAY; Hiok Nam |
August 30, 2012 |
OPTIMIZED LIGHT GUIDE ARRAY FOR AN IMAGE SENSOR
Abstract
An image sensor has a plurality of pixels in a pixel array. Each
pixel includes a photoelectric conversion unit below an insulating
layer and a light guide to transmit light to the photoelectric
conversion unit. Across five or more pixels arrayed in a direction,
the light guides have a spacing between them that varies
non-monotonically across the five or more pixels. A width of the
light guide and/or a horizontal pitch between consecutive light
guides may vary non-monotonically across same. A light guide of a
pixel that detects light of shorter wavelengths only may be
narrower than a light guide of another pixel that detects light of
longer wavelengths. A color filter may be coupled to the light
guide. A width of a gap between consecutive color filters may vary
non-monotonically across same. A pitch between the gaps may vary
non-monotonically across same.
Inventors: |
TAY; Hiok Nam; (Singapore,
SG) |
Family ID: |
43513941 |
Appl. No.: |
13/464955 |
Filed: |
May 5, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
12941004 |
Nov 5, 2010 |
|
|
|
13464955 |
|
|
|
|
61259180 |
Nov 8, 2009 |
|
|
|
61258581 |
Nov 5, 2009 |
|
|
|
Current U.S.
Class: |
250/208.1 |
Current CPC
Class: |
H01L 27/14627 20130101;
H01L 27/14685 20130101; H01L 27/14621 20130101; H01L 27/14632
20130101; H01L 27/14625 20130101; H01L 27/14687 20130101 |
Class at
Publication: |
250/208.1 |
International
Class: |
H01L 27/146 20060101
H01L027/146 |
Claims
1. An image sensor comprising a pixel array supported by a
substrate, the pixel array comprising a plurality of pixels, each
pixel comprising: a photoelectric conversion unit; a light guide;
and, a color filter coupled to transmit light to said photoelectric
conversion unit via said light guide, wherein a gap exists between
the color filters of each pair of consecutive pixels among five or
more pixels arrayed side-by-side in a direction, the five or more
pixels being among the plurality of pixels, wherein the gap has a
width, the width varying non-monotonically across the five or more
pixels.
2. The image sensor of claim 1, wherein the width alternates
between increasing and decreasing across the five or more
pixels.
3. The image sensor of claim 1, wherein the width varies by 0.1 um
or more across sixteen pixels arrayed side-by-side in the
direction.
4. The image sensor of claim 3, wherein the width alternates
between increasing and decreasing across the five or more
pixels.
5. The image sensor of claim 3, wherein the color filter comprises
a colorant.
6. The image sensor of claim 1, wherein the width varies by 0.2 um
or more across sixteen pixels arrayed side-by-side in the
direction.
7. The image sensor of claim 6, wherein the width alternates
between increasing and decreasing across the five or more
pixels.
8. The image sensor of claim 6, wherein the color filter comprises
a colorant.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. patent
application Ser. No. 12/941,004 filed on Nov. 5, 2010, which claims
priority to U.S. Provisional Patent Application No. 61/258,581
filed on Nov. 5, 2009 and U.S. Provisional Patent Application No.
61/259,180 filed on Nov. 8, 2009.
BACKGROUND OF THE INVENTION
[0002] 1. Field of the Invention
[0003] The subject matter disclosed generally relates to structures
and methods for fabricating solid state image sensors.
[0004] 2. Background Information
[0005] Photographic equipment such as digital cameras and digital
camcorders may contain electronic image sensors that capture light
for processing into still or video images. Electronic image sensors
typically contain millions of photoelectric conversion units such
as photodiodes.
[0006] Solid state image sensors can be either of the charge
coupled device (CCD) type or the complimentary metal oxide
semiconductor (CMOS) type. In either type of image sensor,
photoelectric conversion units are formed in a substrate and
arranged in a two-dimensional array. Image sensors typically
contain millions of pixels each comprising a photoelectric
conversion unit to provide a high-resolution image. To improve an
efficiency of light capturing, certain image sensors have
light-guides (or waveguides) to direct light towards the
photoelectric conversion units. The light-guides may comprise a
light transmissive material, for example silicon nitride such as
Si.sub.3N.sub.4, having a refractive index higher than that of a
surrounding insulating material, for example silicon oxide, so that
there is a total internal reflection at sidewalls of the
light-guides to keep light from exiting. Alternatively, the
light-guides may have a metal coating on sidewalls to provide the
reflection and are filled with a transparent material, for example
silicon oxide or an organic resin or spin-on-glass (SOG). A pixel
may comprise more than one light guides, one stacked above another,
to form a cascaded light guide. The light guides at any one height
from the substrate are typically at a given pitch from one another,
and share a common horizontal cross-sectional profile at any one
given height from the substrate. Keeping a constant pitch provides
a uniform sampling of the image projected on the face of the image
sensor along left-to-right and top-to-bottom directions (parallel
to the plane of the pixel array), thus better matched to how pixels
are arrayed on displays such as computer displays and prints.
BRIEF SUMMARY OF THE INVENTION
[0007] According to a first aspect of the disclosure, an image
sensor comprises a pixel array that comprises a plurality of
pixels, where each pixel comprises (a) a photoelectric conversion
unit below an insulating layer and below a plurality of wires that
are also embedded in the insulating layer, and (b) a light guide
embedded in the insulating layer and between the plurality of wires
to transmit light to the photoelectric conversion unit, where a
horizontal spacing between the light guides of each pair of
consecutive pixels among five or more pixels that are arrayed
side-by-side in a direction and that are among the plurality of
pixels varies non-monotonically across the five or more pixels. The
light guide may contain a dye or a color pigment. The color pigment
may be an organic pigment or an inorganic pigment or an
organometallic pigment.
[0008] In the first aspect, it is desirable that a horizontal pitch
between the light guides of each pair of consecutive pixels among
the five or more pixels varies non-monotonically across the five or
more pixels. It is further desirable that the horizontal pitch
varies by 0.1 um or more across the five or more pixels. It is
still further desirable that the horizontal pitch varies by 0.2 um
or more across the five or more pixels.
[0009] In the first aspect, the pixel may further comprise a color
filter coupled to transmit light to the photoelectric conversion
unit via the light guide, there being a gap between the color
filters of each pair of consecutive pixels among the five or more
pixels, the gap having a width that varies non-monotonically across
the five or more pixels. It is desirable that the gap width varies
by 0.1 um or more across sixteen pixels arrayed side-by-side in a
direction. It is more desirable that the gap width varies by 0.2 um
or more across the sixteen pixels.
[0010] In the first aspect, the pixel may further comprise a color
filter coupled to transmit light to the photoelectric conversion
unit via the light guide, there being a gap between the color
filters of each pair of consecutive pixels among the five or more
pixels, and a gap pitch varies non-monotonically across the five or
more pixels, the gap pitch being a horizontal distance (in a plane
parallel to the plane of the photoelectric conversion units)
between each pair of consecutive centerlines of the gaps. It is
desirable that the gap pitch varies by 0.1 um or more across
sixteen pixels arrayed side-by-side in a direction. It is more
desirable that the gap pitch varies by 0.2 um or more across the
sixteen pixels.
[0011] The gap may contain air or a gas. Alternatively, the gap may
contain a liquid or solid material that has a refractive index that
is at least 20% less than that of the color filter. It is also
desirable that the gap has a width not more than 0.45 um between
the adjacent color filters. It is further desirable that the gap is
topped with a convex ceiling. It is still further desirable that
there is at least 0.6 um from a bottom of the color filter to a top
of the convex ceiling.
[0012] According to a second aspect of the disclosure, an image
sensor comprises a pixel array that comprises a plurality of
pixels, where each pixel comprises (a) a photoelectric conversion
unit below an insulating layer and below a plurality of wires that
are also embedded in the insulating layer, and (b) a light guide
embedded in the insulating layer and between the plurality of wires
to transmit light to the photoelectric conversion unit, where the
light guide has a width that varies non-monotonically across the
five or more pixels. It is further desirable that the width of the
light guide is smaller for a pixel among the five or more pixels
that is configured to detect lights of shorter wavelengths only
than for another pixel among the five or more pixels that is
configured to detect lights of longer wavelengths. It is also
further desirable that the width of the light guide is smaller for
a blue pixel than for a red pixel. It is also further desirable
that the width of the light guide is smaller for a blue pixel than
for a green pixel. It is also further desirable that the width of
the light guide is smaller for a green pixel than for a red
pixel.
[0013] In the above, it is desirable that the color filter
comprises a colorant. The colorant may be a dye or a color pigment.
The pigment may be an organic pigment, an inorganic pigment, or a
organometallic pigment.
[0014] According to a third aspect of the disclosure, it is
provided a method for detecting an image using an image sensor, the
method comprising (a) providing a plurality of photoelectric
conversion units below an insulating layer and a plurality of wires
that are embedded in the insulating layer, (b) providing a
plurality of light guides embedded in the insulating layer and
between the plurality of wires to transmit light to the
photoelectric conversion unit, where a horizontal spacing between
the light guides of each pair of consecutive pixels among five or
more pixels that are arrayed side-by-side in a direction and that
are among the plurality of pixels varies non-monotonically across
the five or more pixels.
[0015] In the third aspect, it is desirable that the image sensor
has any of the desirable features from the first aspect.
[0016] According to a fourth aspect of the disclosure, an image
sensor comprises a pixel array that comprises a plurality of
pixels, where each pixel comprises (a) a photoelectric conversion
unit below an insulating layer and below a plurality of wires that
are also embedded in the insulating layer, and (b) a light guide
embedded in the insulating layer and between the plurality of wires
to transmit light to the photoelectric conversion unit, where a
horizontal spacing between the light guides of each pair of
consecutive pixels among five or more pixels that are arrayed
side-by-side in a direction and that are among the plurality of
pixels alternates between increasing and decreasing across the five
or more pixels. The light guide may contain a dye or a color
pigment. The color pigment may be an organic pigment or an
inorganic pigment or an organometallic pigment.
[0017] In the fourth aspect, the pixel may further comprise a color
filter coupled to transmit light to the photoelectric conversion
unit via the light guide, there being a gap between the color
filters of each pair of consecutive pixels among the five or more
pixels, the gap having a width that alternates between increasing
and decreasing across the five or more pixels. It is desirable that
the gap width varies by 0.1 um or more across sixteen pixels
arrayed side-by-side in a direction. It is more desirable that the
gap width varies by 0.2 um or more across the sixteen pixels.
[0018] In the fourth aspect, the pixel may further comprise a color
filter coupled to transmit light to the photoelectric conversion
unit via the light guide, there being a gap between the color
filters of each pair of consecutive pixels among the five or more
pixels, and a gap pitch alternates between increasing and
decreasing across the five or more pixels, the gap pitch being a
horizontal distance (in a plane parallel to the plane of the
photoelectric conversion units) between each pair of consecutive
centerlines of the gaps. It is desirable that the gap pitch varies
by 0.1 um or more across sixteen pixels arrayed side-by-side in a
direction. It is more desirable that the gap pitch varies by 0.2 um
or more across the sixteen pixels.
[0019] According to a seventh aspect of the disclosure, an image
sensor comprises a pixel array that comprises a plurality of
pixels, where each pixel comprises (a) a photoelectric conversion
unit below an insulating layer and below a plurality of wires that
are also embedded in the insulating layer, and (b) a light guide
embedded in the insulating layer and between the plurality of wires
to transmit light to the photoelectric conversion unit, where the
light guide has a width that alternates between increasing and
decreasing across the five or more pixels. It is further desirable
that the width of the light guide is smaller for a pixel among the
five or more pixels that is configured to detect lights of shorter
wavelengths only than for another pixel among the five or more
pixels that is configured to detect lights of longer wavelengths.
It is also further desirable that the width of the light guide is
smaller for a blue pixel than for a red pixel. It is also further
desirable that the width of the light guide is smaller for a blue
pixel than for a green pixel. It is also further desirable that the
width of the light guide is smaller for a green pixel than for a
red pixel.
[0020] According to a eighth aspect of the disclosure, it is
provided a method for detecting an image using an image sensor, the
method comprising (a) providing a plurality of photoelectric
conversion units below an insulating layer and a plurality of wires
that are embedded in the insulating layer, (b) providing a
plurality of light guides embedded in the insulating layer and
between the plurality of wires to transmit light to the
photoelectric conversion unit, where a horizontal spacing between
the light guides of each pair of consecutive pixels among five or
more pixels that are arrayed side-by-side in a direction and that
are among the plurality of pixels alternates between increasing and
decreasing across the five or more pixels.
[0021] In the eighth aspect, it is desirable that the image sensor
has any of the desirable features from the fourth aspect.
[0022] According to a ninth aspect of the disclosure, an image
sensor comprises a pixel array that comprises a plurality of
pixels, where each pixel comprises (a) a photoelectric conversion
unit below an insulating layer and below a plurality of wires that
are also embedded in the insulating layer, and (b) a light guide
embedded in the insulating layer and between the plurality of wires
to transmit light to the photoelectric conversion unit, where a
horizontal spacing between the light guides of each pair of
consecutive pixels among five or more pixels that are arrayed
side-by-side in a direction and that are among the plurality of
pixels has a wider horizontal spacing that follows immediately
after a narrower spacing and that is itself followed immediately by
another narrower spacing across the five or more pixels. The light
guide may contain a dye or a color pigment. The color pigment may
be an organic pigment or an inorganic pigment or an organometallic
pigment.
BRIEF DESCRIPTION OF THE DRAWINGS
[0023] FIG. 1A is an illustration showing a cross-section of four
image sensor pixels of an embodiment of the present invention;
[0024] FIG. 1B is a ray tracing diagram for the same cross-section
of four image sensor pixels of FIG. 1A;
[0025] FIG. 2A is an illustration showing a top view of fifteen
pixels within an array;
[0026] FIG. 2B shows the same top view of fifteen pixels;
[0027] FIG. 3 is an illustration showing a cross-section of four
image sensor pixels of an alternative embodiment;
[0028] FIG. 4 is an illustration showing a cross-section of four
image sensor pixels of an alternative embodiment that is the best
mode;
[0029] FIG. 5 is an illustration showing a cross-section of four
image sensor pixels of an alternative embodiment;
[0030] FIG. 6 is a schematic of an image sensor;
[0031] FIG. 7 is a ray tracing diagram for the third embodiment
shown in FIG. 4.
[0032] FIG. 8A is a illustration of a primary-color Bayer
pattern;
[0033] FIG. 8B is an illustration of a primary-color Bayer pattern
in 45 degree rotation;
[0034] FIG. 9 is a schematic of a 4T pixel;
[0035] FIG. 10 is a schematic of a 3T pixel.
DETAILED DESCRIPTION
[0036] Disclosed is an image sensor that has a pixel array that
comprises a plurality of pixels that each includes a photoelectric
conversion unit. Each of the pixels includes a light guide that is
embedded in an insulating layer and between wires, also embedded in
the insulator layer, to transmit a light to the photoelectric
conversion unit. The light guide of a pixel that detects lights of
shorter wavelengths may have a smaller width at its bottom ("bottom
width") than a light guide of another pixel that detects lights of
longer wavelengths only. A vertical centerline at a bottom of the
light guide may have a larger distance to that of the light guide
of an adjacent pixel on a lateral side than that of the light guide
of an adjacent pixel on an opposite lateral side. A spacing between
a bottom of the light guide and a bottom of the light guide of an
adjacent pixel on a lateral side ("bottom spacing") may be larger
than that between the bottom of the light guide and a bottom of the
light guide of an adjacent pixel on an opposite lateral side. The
pixel may comprise a color filter that comprises a color material.
A gap may exist between the color filters of each pair of
side-by-side pixels among the plurality of pixels. A width of the
gap ("gap width") may differ from a pair of side-by-side pixels to
another pair of side-by-side pixels. A pitch of the gap ("gap
pitch") across three or more pixels arrayed side-by-side in a
direction may vary by 15% or less. Having one or more of the above
technical features in an image sensor, an in particular a color
image sensor, permits higher layout densities of integrated circuit
features (e.g. gate electrodes, poly contacts, wires and diffusion
contacts) under the insulating layer as well as within the
insulating layer. A substrate supports the photoelectric conversion
unit and may be a semiconductor substrate lightly doped to a first
conductivity type, preferably p-type, and further preferably having
a doping concentration between 5 e14/cm.sup.3 and 5 e15/cm.sup.3.
The substrate 106 may be a p-epi layer on a heavily doped
p-substrate having doping concentration in excess of 1
e19/cm.sup.3. For example, substrate 106 may be of silicon doped
with boron to the concentration between 5 e14/cm.sup.3 and 5
e15/cm.sup.3, such as a conventional p-epi layer on a heavily doped
p+ substrate (not shown).
[0037] Referring to the drawings more particularly by reference
numbers, FIG. 6 illustrates an image sensor 10 comprising an array
12 of pixels 14 connected to a row decoder 20 by a group of control
signals 22 and to a light reader circuit 16 by a output signals 18
generated from the pixels 14. A light reader circuit 16 samples
output signals 18 generated from pixels 14 and may perform
subtraction and amplification on samples of the output signals 18
to generate analog signal(s) to be provided to an analog-to-digital
converter (ADC) 24. The ADC 24 converts the analog signal(s) to
digital image data on ADC output bus 66. If the image sensor 10 is
a color image sensor, the pixel array 12 includes a color filter
array that comprises color filters arrayed in two dimensions in
such a way that there is one color filter for each pixel 14.
[0038] FIG. 8A illustrates an example of a color filter array that
may be disposed over and as part of the pixel array 12. FIG. 8A
shows a Bayer primary color pattern that comprises a repeated
two-dimensional array of a two-by-two block (within dashed line) of
color filters each having one of a green color (G), a red color (R)
and a blue color (B). A pair of green color filters is disposed
along one diagonal of the two-by-two block. A pair of a red color
filter and a blue color filter is disposed along the other
diagonal. In this embodiment of the color filter array, the color
filters are arrayed side-by-side from left to right of the page and
from top to bottom of the page.
[0039] FIG. 8B illustrates an alternative embodiment of the color
filter array shown in FIG. 8A. In this variation, the directions in
which color filters are arrayed are rotated 45 degrees with respect
to the left-to-right and top-to-bottom directions as well as with
respect to the direction of bottom-to-top-of-the-image scan (shown
in upward-pointing arrow). The pixels 14 are arrayed likewise in an
embodiment of image sensor 10 that uses this color filter
array.
[0040] FIG. 9 shows a schematic for an embodiment of a pixel 14 of
the pixel array 12. The pixel 14 includes a photoelectric
conversion unit 102. By way of example, the photoelectric
conversion unit 102 may be a photodiode. The photoelectric
conversion unit 102 may be connected to a reset switch 112 via a
transfer gate 117. The photoelectric conversion unit 102 may also
be coupled to a select switch 114 through an output (i.e.
source-follower) transistor 116. The transistors 112, 114, 116, 117
may be field effect transistors (FETs). A gate of the transfer gate
112 is connected to a TF(n) line 121. A gate of the reset switch
112 is connected to a RST(n) line 118. A drain node of the reset
switch 112 may is connected to an IN line 120. A gate of the select
switch 114 is connected to a SEL line 122. A source node of the
select switch 114 is connected to an OUT line 124. The RST(n) line
118, SEL(n) line 122, and TF(n) line 126 may be shared for an
entire row of pixels in the pixel array 12. Likewise, the IN 120
and OUT 124 lines may be shared for an entire column of pixels in
the pixel array 12. The RST(n) line 118, SEL(n) line 122 and TF(n)
line 121 are connected to the row decoder 20 and are part of the
control lines 22. The OUT(m) line 124 is connected to the light
reader 16 and is part of the vertical signal lines 18.
[0041] FIG. 1A shows an embodiment of four adjacent pixels 14
arrayed side-by-side in a direction in a pixel array 12 of a color
image sensor 10 on a substrate 106. Each pixel 14 shown includes a
photoelectric conversion unit 102a or 102b that converts photonic
energy into electrical charges. For a pixel array 12 that uses a 4T
pixel architecture (such as shown in FIG. 9) or a variant thereof
(such as sharing reset switch 112, select switch 114 and output
transistor 116 among multiple pairs of photodiode 102 and transfer
switch 117), gate electrodes 104a, 104b may each be a gate
electrode of a different transfer switch 117 to transfer the
charges. Alternatively, for a pixel array 12 that uses a 3T pixel
architecture (such as shown in FIG. 10), gate electrodes 104a, 104b
may be gate electrodes of different reset switches 112 to reset the
photoelectric conversion units 102a, 102b, respectively. The gate
electrode 104c may be a gate electrode of a transistor serving a
different function within the pixel array 12, for example a reset
switch 112, or a select switch 114, or an output transistor 116.
The gate electrodes 104a, 104b, 104c and conversion units 102a,
102b are formed on or in the substrate 106. The gate electrodes
104a, 104b, 104c, the photoelectric conversion units 102a, 102b and
the substrate 106 may be covered under a protection layer 230 that
comprises a silicon nitride and that has a thickness between 200 to
1000 Angstrom. The protection layer 230 insulates the substrate 106
from metallic ions and moisture. A layer of insulation layer 110
covers the substrate 106. Wires 108 are embedded in the insulating
layer 110 and above gate electrodes 104a, 104b, 104c. The wires 108
may be conductive interconnect wires that comprise aluminum or
copper. Other interconnect wires (not shown) may be formed in other
routing planes, each plane comprising multiple interconnecting
wires (that may be metallic) and being at a different heights above
the gate electrodes 104a, 104b, 104c than wires 108. A wire 108 may
be connected by means of a conductive (such as metallic) via that
crosses to an interconnect wire on an adjacent routing plane at a
different height.
[0042] Photoelectric conversion units 102a, 102b may be paired with
lower light guides 316a, 316b, respectively, that are embedded in
the insulating layer 110 and between the wires 108. The lower light
guides 316a, 316b may comprise a transmissive material, e.g. a
silicon nitride such as Si.sub.3N.sub.4, that has a higher index of
refraction than the insulating layer 110 (being the external
material), e.g. a silicon oxide, and use total internal reflection
between the light guide and the external material to help keep
light from exiting the lower light guides. Alternatively, the lower
light guide may be filled with a transparent material such as
spin-on-glass (SOG) or a transparent resin of an organic material
and may even comprise a color material (such as an organic or
inorganic pigment or an organometallic pigment) and have reflective
metal coatings on its lateral walls (reflective metal coating type)
to reflect light inwards to help keep light from exiting the lower
light guide.
[0043] Upper light guides 130 may be located above lower light
guides 316a, 316b and may comprise either the same material(s) as
the lower light guides 316a, 316b or different material(s). The
upper light guide 130 and the lower light guide 316a, 316b may be
both of the total internal reflection type, or both of the
reflective metal coating type, or one of them may be of one type
while the other one being of the other type. A top end of the upper
light guide 130 is wider than a bottom end, where the upper light
guides 130 meet the lower light guides 316a, 316b.
[0044] Color filters 114a, 114b are located above the upper light
guides 130. The color filter 114a, 114b may each comprise a
different color material, or colorant, such as a dye or a organic
or inorganic or organometallic pigment. The color filter may
comprise a resin in which the dye is dissolved or the organic or
inorganic or organometallic color pigment is suspended, where the
resin may be organic or comprise a polymer that has at least an
organic group such as methyl, ethyl or phenyl (an example being
silicone). Alternatively, the color filter may comprise a
transmissive inorganic material (e.g. silicon nitride) having
particles of a color pigment (e.g. an inorganic color pigment such
as iron oxide, a cobalt or manganese or zinc or copper pigment, or
an organometallic pigment, or a complex inorganic color pigment)
dispersed therein. The color filters 114a, 114b exhibit different
colors in white light. Preferably, each has a highest transmittance
and a least transmittance of at least 50% and at most 10%,
respectively, between wavelengths (in air) of 400 nm to 700 nm.
Alternatively, a ratio between the highest and least transmittances
shall be more than 4-to-1.
[0045] As shown in FIG. 1A, adjacent color filters 114a, 114b have
gaps between them. Gaps 422a, 422b between the color filters 114a,
114b have widths of 0.45 um or less. The gaps 422a, 422b may be
filled with air or a gas. The gaps 422a, 422b may have a depth of
0.6 um or greater between the adjacent color filters. The gap with
the dimensional limitations cited above causes light within the gap
to be diverted into the adjacent color filters and to be
subsequently guided by light guide(s) to the photoelectric
conversion units 102a or 102b. Thus the percentage loss of light
impinging on the pixel due to the light passing through and
penetrating down below the gap 422a, 422b (henceforth "pixel loss")
is substantially reduced.
[0046] Together, the color filter 114a (or 114b) and upper 130 and
lower 316a (or 316b) light guides constitute a "cascaded light
guide" that guides light to the photoelectric conversion unit 102a
(or 102b) by means of total internal reflection at interfaces with
external media such as the insulator 110. (Alternatively, one or
both of the upper and lower light guides may have metal sidewalls
to reflect light inwards.) In FIG. 1B, a ray tracing diagram, rays
a, b, e and f are shown to experience reflections at sidewalls of
the color filters and/or the upper light guides and the lower light
guides. Rays c and d that fall into the wider gap between the color
filters of the second and third pixels are diverted into the color
filters of the second and third pixels respectively and arrive at
the respective photoelectric conversion units.
[0047] FIG. 3 shows an alternative embodiment of the image sensor
10 in which the gaps 422a, 422b between adjacent color filters are
covered under a transparent film 500, and a support film 134 fills
in between adjacent upper light guides 134. The support film 134
shall have a lower refractive index than the upper light guide 130
if the upper light guide 130 is of the total internal reflection
type. A ceiling 510 of the gap 422a, 422b may be concave with
respect to the transparent film 500 (i.e. convex with respect to
the gap) such as being in a dome-shape. The gap 422a, 422b may
contain air or a gas. Lights that enter the gap from above,
crossing the convex ceiling, are diverged towards the adjacent
color filters.
[0048] FIG. 4 shows an alternative embodiment in which the
embodiment in FIG. 3 is further modified so that the color filters
114a, 114b have sidewalls that incline inwards and the support film
134 has a interfaces with the color filters' sidewalls. As in the
second embodiment, across four pixels from left to right, the gap
between consecutive color filters becomes wider, then narrower,
then narrower again. Although the gap has different width at
different height within the gap, for comparison between a wider gap
422b' and a narrower gap 422a', it suffices to measure the gap
width at a horizontal level (i.e. one that is parallel to the plane
of the photoelectric conversion units) that slices through the
color filters and the gaps between them, as shown in FIG. 4.
Likewise, for comparison between a wider pitch between a pair of
consecutive gaps and a narrower pitch between another pair of
consecutive gaps, it suffices to measure at the same horizontal
level. FIG. 7 is a ray tracing diagram that shows a trajectory of a
light ray that enters the gap between two color filters 114a, 114b.
The convexity of the ceiling serves to diverge the light ray
towards one of the color filters. A height from a bottom of the
color filter to a top of an adjacent ceiling shall be 0.6 um or
more. This provides enough depth for light that enters the gap
(below the ceiling and laterally adjacent to the color filter) from
the ceilings to be diverged into the adjacent color filters. For
example, the height is labeled H.sub.a, measuring from the bottom
of the first (from the left) color filter 114a to the top of the
ceiling 510a between the first 114a and second 114b color filters.
Likewise, the height is labeled H.sub.b, measuring from the bottom
of the second 114b color filter to the top of the ceiling 510b
between the second 114b and third 114a color filters.
[0049] FIG. 5 shows an alternative embodiment in which the upper
light guide 130 is dispensed with, and instead microlenses 318 are
placed above the lower light guides 316a, 316b, with a transparent
planarization layer 320 between the microlenses 318 and the light
guides 316a, 316b. The microlenses 318 focus lights into the upper
apertures of the light guides 316a, 316b, which in turn transmit
the lights down to their respective photoelectric conversion units
102a, 102b. For a color image sensor, the light guides 316a, 316b
may comprise colorants such as dyes or an organic or inorganic or
organometallic pigments to give different colors to the light
guides 316a, 361b according to the colors of the color pattern of
the pixel array 12, for example the Bayer pattern.
[0050] Alternatively, the gaps 422a, 422b in the embodiments shown
in FIGS. 1A, 3, 4 and 5 may contain a transparent (liquid or solid)
medium as long as the transparent medium has a refractive index
that is lower than the color filters by at least 20%. For example,
the transparent medium may be a resin having a refractive index
between 1.4 and 1.5 whereas the color filters comprise particles of
silicon nitride having their density adjusted such that its
refractive index becomes 1.7 or above.
[0051] A light guide embedded in the insulator layer 110 and
between wires 108 to transmit lights of shorter wavelengths only
may have a smaller width at its bottom ("bottom width") than
another light guide to transmit lights of longer wavelengths,
regardless of whether the light guide uses total internal
reflection to keep light from exiting or uses metal coating on its
sidewalls. Referring to FIG. 1A, color filters 114a may be blue
color filters that have higher transmittances for wavelengths in
air between 400 nm to 500 nm than for other wavelengths (hence
their pixels are blue pixels), and color filters 114b may be green
color filters that have better transmittances for wavelengths in
air between 500 nm to 600 nm than for other wavelengths (hence
their pixels are green pixels). Accordingly, the lower light guides
316a (for the blue pixels) that transmit blue lights has a smaller
bottom width W.sub.a of its bottom 318a than the bottom width
W.sub.b of a bottom 318b of the lower light guides 316b (for the
green pixels) that transmit green lights only (W.sub.a<W.sub.b).
Generally, the lower light guides 316a is narrower than the lower
light guides 316b. Alternatively, color filters 114b shown in FIG.
1A may be red color filters that have higher transmittances for
wavelengths in air between 600 nm and 700 nm than for other
wavelengths (hence their pixels are red pixels). Alternatively,
color filters 114a may be green color filters whereas color filters
114b being red color filters.
[0052] FIG. 1A also shows that the bottom width varies
non-monotonically across pixels that are arrayed side-by-side in a
direction. In FIG. 1A, from left to right, the sequence of bottom
width is W.sub.a, W.sub.b, W.sub.a, W.sub.b, where
W.sub.a<W.sub.b, exhibiting a sequence of variation of
{increase, decrease, increase}. In an alternative embodiment, a
different sequence of variation of the bottom width is possible
while being non-monotonic, i.e. an increase in the bottom width is
followed by a decrease and further followed by another increase,
and/or a decrease is followed by an increase and further followed
by another decrease. In particular, the sequence of variation may
be a non-monotonic sequence that is repeated. For example, the
variation of bottom width may follow a pattern of {increase, no
change, decrease, increase, decrease} that is repeated.
[0053] In FIG. 1A, color filter 114a may be a blue color filter and
a lower light guide 316a has a bottom width W.sub.a at its bottom.
Further to the right in FIG. 1A, color filter 114b may be a green
color filter and a lower light guide 316b has a bottom width
W.sub.b at its bottom, where the bottom width W.sub.b is larger
than the bottom width W.sub.a (W.sub.a<W.sub.b). For a lower
light guide embedded in the insulating layer 110 to transmit lights
of wavelengths up to 500 nm, its bottom width is preferably between
0.2 um to 0.35 um, more preferably 0.27 um +/-10%. For a lower
light guide to transmit lights having wavelengths up to 600 nm, its
bottom width should preferably be between 0.25 um to 0.4 um, more
preferably within 0.33 um +/-10%. For a lower light guide to
transmit lights having wavelengths up to 700 nm, its bottom width
should preferably be between 0.3 um to 0.5 um, more preferably
within 0.4 um +/-10%. Having smaller widths at the bottom permits
higher packing densities of integrated circuit features (such as
gate electrodes 104a, 104b and 104c, polysilicon contacts and
diffusion contacts) under the insulator 110.
[0054] The gaps 422a, 422b may have the width of gap between the
color filters of side-by-side pixels ("gap width") vary
non-monotonically among themselves. Across a first, a second and a
third pixels, arrayed side-by-side in this order in a direction,
the gap may be wider between the first and the second pixels than
between the second and the third pixels. For example, FIG. 1A shows
two different gap widths V.sub.a, V.sub.b between four adjacent
pixels arrayed side-by-side in a particular direction.
[0055] A distance P ("gap pitch") between centerlines of
consecutive gaps 422a, 422b may be maintained essentially constant
(i.e. within 5% of its maximal value) across five or more pixels
arrayed side-by-side in a direction while the gap width varies
non-monotonically.
[0056] Alternatively, the gap pitch P may be allowed non-monotonic
variation having a maximum-to-minimum difference of at most 20% of
its maximum value across a predetermined number of pixels arrayed
side-by-side in a direction. The predetermined number may be less
than 16, more preferably not more than 8. More particularly, the
gap pitch P may alternately increase and decrease across pixels
arrayed in a direction. The gap pitch P may vary by as much as 0.1
um, more preferably 0.2 um, for an average pixel pitch of 1 um.
Letting the gap pitch P vary non-monotonically in conjunction with
the bottom spacing (or more generally a horizontal spacing between
consecutive lower light guides) gives the lower light guides more
freedom to place themselves more optimally while retaining good
capturing of light. In this way, a lower light guide can shift to
one side to make the spacing on this side narrower while making the
spacing on the opposite side wider enough to accommodate an
additional integrated circuit feature such as a gate electrode or a
diffusion contact, thus helping to support a denser pixel
array.
[0057] A distance between vertical centerlines of consecutive lower
light guides ("bottom pitch") in a vertical cross section across
five or more pixels arrayed side-by-side in a particular direction
may have a non-monotonic variation. For example, FIG. 1A shows two
different bottom pitches X.sub.a, X.sub.b between vertical
centerlines (shown as vertical dashed lines in FIG. 1A) of three
lower light guides 316a, 316b on the left. The larger bottom pitch
X.sub.b is below the larger gap width v.sub.b of gap 422b, whereas
the smaller bottom pitch X.sub.a is below the smaller gap width
v.sub.a of first gap 422a. The larger bottom pitch X.sub.b helps to
accommodate more integrated circuit features under the insulating
layer 110 than the smaller bottom pitch X.sub.a does.
[0058] As FIG. 1A shows, the bottom spacing S.sub.a between the
first pixel (counting from the left) and the second pixel is same
as the bottom spacing S.sub.a between the third pixel and the
fourth pixel, and is smaller than the bottom spacing S.sub.b
between the second and third pixels. The larger bottom spacing
S.sub.b between the second and third pixels accommodates more
integrated circuit features under the insulator or above and
adjacent to the substrate 106, such features including gate
electrodes (such as gate electrodes 104b, 104c as shown in FIG.
1A), polysilicon contact (not shown), diffusion contact (not
shown), and wires. Conventional pixel array, with uniform spacing
between light guides, would have to have all spacings as wide as
S.sub.b, even though such wider spacing is not needed between the
first and second pixels nor between the third and fourth pixels,
resulting in a problem of having less than optimal pixel
density.
[0059] FIG. 2A is a top view looking down at the pixel array 12
from above, showing fifteen pixels 14 in three rows and five
columns of the first, second and third embodiments of the pixel
array 12 where pixels are arranged side-by-side in left-to-right
and top-to-bottom directions. Centerlines of the gaps are drawn in
thick, gray lines between each pair of side-by-side pixels. The
area B represents a top surface area of the upper light guide 130
and the area C represents a bottom surface area of the lower light
guide 316a, 316b. The area A minus the area B is the area of the
gap 422a, 422b between color filters, where A represents a pixel
area. FIG. 2A shows four different pairs of areas B and C: B.sub.1
and C.sub.1, B.sub.2 and C.sub.2, B.sub.3 and C.sub.3, and B.sub.4
and C.sub.4. These four different pairs are repeated in a regular
pattern.
[0060] FIG. 1A can be seen as a vertical section of the first to
fourth pixels, counting from the left, of the middle row in FIG.
2A. In this context, v.sub.a=t.sub.3, v.sub.b=t.sub.4,
S.sub.a=s.sub.3, and S.sub.b=s.sub.4.
[0061] As FIG. 2A shows, the gaps between color filters of
side-by-side pixels have widths t.sub.1, t.sub.2, t.sub.3, t.sub.4,
u.sub.1, u.sub.2, u.sub.3 and u.sub.4 that are repeated. Along any
left-to-right or top-to-bottom cut in FIG. 2A, the gap widths
t.sub.1 to t.sub.4 or u.sub.1 to u.sub.4 are varied among
themselves while the pitch P from a centerline of a first gap to a
centerline of the next first gap remains constant. The gap width
varies across plural pixels arrayed side-by-side in a direction
perpendicular to the gap, which in FIG. 2A is from left to right
(or from right to left) for a gap on the left or right (as shown in
the top view looking down at the pixel array in FIG. 2A) of a pixel
or is from top to bottom (or from bottom to top) for a gap above or
below (again, as shown in the top view looking down at the pixel
array in FIG. 2A) a pixel. The variation is non-monotonic, i.e. an
increase is followed by a decrease that in turn is followed by
another increase, and/or a decrease is followed by an increase
further followed by another decrease. For example, in the top row,
from left to right, the gap width varies in a sequence of t.sub.2,
t.sub.1, t.sub.2, t.sub.1, where t.sub.2<t.sub.1, which exhibits
a sequence of variation in the pattern of {increase, decrease,
increase}; in the middle row, a sequence of t.sub.3, t.sub.4,
t.sub.3, t.sub.4, where t.sub.3<t.sub.4, which exhibits a
sequence of variation in the pattern of {increase, decrease,
increase}; and the bottom row, same sequence as the top row. More
particularly, the gap width alternates between wider gap widths and
narrower gap widths across side-by-side pixels arrayed in a
direction. Although in each of these two sequences of gap widths,
the variation of gap width is shown to alternate between increasing
and decreasing, in an alternative embodiment a different sequence
of variation of the gap width is possible as long as it is
non-monotonic. In particular, the sequence of variation may be a
non-monotonic sequence that is repeated. For example, the variation
of gap width may follow a pattern of {increase, unchanged,
decrease, increase, decrease} that is repeated.
[0062] FIG. 2A also shows that, likewise, the bottom spacing can
vary across pixels that are arrayed side-by-side in a direction.
For example, in the middle row, from left to right, the sequence of
bottom spacing alternate between s.sub.3 and s.sub.4, where
s.sub.3<s.sub.4, exhibiting a sequence of variation of
{increase, decrease, increase}. More particularly, the bottom
spacing alternates between larger spacing and smaller spacing
across pixels arrayed side-by-side in a direction. Conventional
image sensor pixel array that uses light guides places the light
guides uniformly displaced from each other, thus is unable to
utilize a small space between light guide too small to accommodate
an integrated circuit feature such as a gate electrode. The
embodiment shown in FIG. 2A is able to bring the two light guides
on the left of the second row closer to each other and also the
next two light guides of the second row closer to each other in
order to create a wider space in the middle (i.e. between the
second and third light guides) so that an additional gate electrode
104c can fit in. Conventional pixel array, with uniform spacing
between light guides, would have to have all spacings being wider
even though the wider spacing is only needed between certain pairs
of laterally adjacent light guides. In an alternative embodiment,
any sequence of variation of bottom spacing is possible as long as
the sequence of variation is non-monotonic, i.e. an increase in the
bottom spacing is followed by a decrease and further followed by
another increase, and/or a decrease is followed by an increase and
further followed by another decrease. In particular, the sequence
of variation may be a non-monotonic sequence that is repeated. For
example, the variation of bottom spacing may follow a pattern of
{increase, increase, decrease, increase, decrease} that is
repeated. Preferably, the bottom spacing varies by 0.2 um or more
within a group of five consecutive pixels.
[0063] FIG. 2A also shows that the bottom width varies
non-monotonically from left to right in the second row in a
bottom-width sequence of W.sub.a, W.sub.b, W.sub.a, W.sub.b,
W.sub.a, where W.sub.a<W.sub.b, exhibiting a variation pattern
of {increase, decrease, increase, decrease}, which is
non-monotonic. In particular, for the four pixels in the top two
rows and leftmost two columns, it is shown that the bottom width is
smallest for the bottom area C.sub.3, largest for the bottom area
C.sub.1, and intermediate for the bottom areas C.sub.2 and C.sub.4.
In an embodiment where the primary-color Bayer pattern is used for
the color filters over the pixel array 12, the pixels that have the
bottom areas C.sub.2 and C.sub.4 may be the green pixels, the pixel
that has the bottom area C.sub.1 may be the red pixel, and the
pixel that has the bottom area C.sub.3 may be the blue pixel. In a
direction in which alternate pixels alternate between two different
colors, the width of the lower light guides should similarly
alternate so that the lower light guides and their spacings are of
the optimum size for guiding the light to be detected whilst also
optimizing the space for integrated circuit features. Thus, in this
example where bottom areas C.sub.3 corresponds to blue pixels and
bottom areas C.sub.4 corresponds to green pixels, in the second
row, where in the left-to-right direction alternate pixels
alternate between blue and green pixels, the width of light guides
are optimized accordingly, alternating between narrower widths (for
blue pixels) and wider widths (for green pixel), saving space
between the light guides for laying integrated circuit features
such as gate electrodes and contacts. Likewise, in the first row,
where alternate pixels alternate between green and red pixels, the
width of light guides alternate between narrower widths (for green
pixels) and wider widths (for red pixels).
[0064] More generally, for purposes of comparing widths of the
light guides embedded in the insulating layer 110 and of comparing
horizontal spacings between pairs of consecutives ones among these
light guides, the widths and horizontal spacings may be measured at
a horizontal level (i.e. one that is parallel to the plane of the
photoelectric conversion units) between the gate electrode 104a,
104b and the wires 108. Within this range of heights, the
horizontal spacing (and hence the width) is relevant, for it
affects how densely the integrated circuit features in and/or under
the insulating layer 110 can be packed together, such integrated
circuit features including gate electrodes, polysilicon routing
wire (not shown), polysilicon contact (not shown), diffusion
contact (not shown), and metal wires. This horizontal spacing
should be varied non-monotonically across five or more pixels
arrayed side-by-side in a direction, as shown in FIG. 1A, to permit
a light guide to shift to one side to occupy a small space too
small to accommodate an integrated circuit feature so that more
space is aggregated on the other side to become large enough to
accommodate an integrated circuit feature. Preferably, this
variation exhibits itself across eight or fewer pixels arrayed
side-by-side in a direction. Further preferably, this horizontal
spacing exhibits a variation of 0.1 um or larger between the widest
and the narrowest horizontal spacings between consecutive light
guides across the eight or fewer pixels, more preferably within a
group of five consecutive pixels. Still further preferably, the
variation is 0.2 um or larger. Making this width of the light guide
as small as possible (without impeding the transmission of light)
contributes to widening the horizontal spacing, thus benefiting
denser packing of integrated circuit features in and under the
insulating layer 110 and denser packing of pixels in the pixel
array 12. Preferably, this width is varied non-monotonically across
five or more pixels array side-by-side in a direction as shown in
FIG. 1A to take advantage of the fact that this width can be
smaller for a light guide that transmits lights of shorter
wavelengths only than for a light guide that transmits lights of
longer wavelengths. For a light guide embedded in the insulating
layer 110 to transmit lights of wavelengths-in-air between up to
500 nm, this width is preferably between 0.2 um to 0.35 um, more
preferably 0.27 um +/-10%. For a light guide to transmit lights
having wavelengths-in-air up to 600 nm, this width should
preferably be between 0.25 um to 0.4 um, more preferably within
0.33 um +/-10%. For a light guide to transmit lights having
wavelengths-in-air up to 700 nm, this width should preferably be
between 0.3 um to 0.5 um, more preferably within 0.4 um +/-10%.
[0065] Non-monotonically varying the horizontal spacing between
consecutive light guides at a horizontal level (i.e. one that is
parallel to the plane of the photoelectric conversion units)
between a wire 108 and a gate electrode 104a (or in particular the
bottom spacing) and/or a horizontal pitch between light guides at a
height between a wire 108 and a gate electrode 104a, inclusive, (or
in particular the bottom pitch) permits a light guide to shift to
one side to occupy a small space too small to accommodate an
integrated circuit feature so that more space is aggregated on the
other side to become large enough to accommodate an integrated
circuit feature. This can be seen in FIG. 1A where an increased
horizontal spacing between consecutive light guides at any
horizontal level (i.e. one that is parallel to the plane of the
photoelectric conversion units) between wires 108 and the gate
electrodes 104a, 140b, 104c, inclusive, (or in particular the
bottom spacing S.sub.b) between the second and third (counting from
the left) lower light guides accommodates an extra gate electrode
104c than would be possible if all lower light guides were equally
spaced.
[0066] Non-monotonically varying the gap width and/or the gap pitch
between color filters permits the color filters to shift with the
respective lower light guides in such a manner that the respective
lower light guides remain coupled to receive light from the
respective color filters.
[0067] Where the gap is filled with air, or gas(es), or a liquid or
solid medium having a lower refractive index than the color filter,
or where a convex ceiling is over the gap in case where a
refractive index above the ceiling is higher than a refractive
index under the ceiling, light remains able to diverge from the gap
into the color filters even when the gap width changes. Therefore,
in a sequence of consecutive first, second, third and fourth color
filters, the second color filter can be shifted closer to the first
color filter and the third color filter can be shifted closer to
the fourth color filter while all four remain effective in the
gathering of light. As shown in FIG. 1B, ray c and ray d that fall
into the wider gap between color filters of the second and third
pixels are still captured by the respective color filters and
transmitted to the respective photoelectric conversion units.
Having gaps between the color filters and being able to reduce the
gap widths for some gaps while widening a gap between those gaps
that become narrower permits the lower light guide below the color
filters to have similarly reduced spacings below the narrower gaps
and widened spacing below the widened gap while retaining good
light capture of the pixels involved.
[0068] FIG. 2B shows the same top view of the fifteen pixels in
FIG. 2A. FIG. 1A can be seen as a vertical section of the first to
fourth pixels counting from the left along the cut line ZZ' in FIG.
2B. The thick "+" marks within the bottom areas and on the cut line
ZZ' represent the vertical centerlines (shown in FIG. 1A as dashed
lines) of the lower light guides of these pixels in the plane that
is perpendicular to the substrate 106 and that contains the cut
line ZZ'. Bottom pitches X.sub.a, X.sub.b between the four pixels
along the cutline ZZ' are shown to be non-monotonically changing.
In particular, the sequence of the bottom pitch {X.sub.a, X.sub.b,
X.sub.a, X.sub.b} along the cutline ZZ' exhibits a pattern of
{increase, decrease, increase}.
[0069] The description above has shown how packing density of
integrated circuit features can be enhanced by non-monotonic
variations in a horizontal spacing between consecutive light guides
of pixels arrayed in a direction (in particular, the bottom
spacing). Non-monotonic variation in the horizontal spacing (in
particular, the bottom spacing) is helped by non-monotonic
variations in any one or more of the following: a horizontal pitch
between consecutive light guides of pixels arrayed in a direction
(in particular, the bottom pitch), gap width, and gap pitch. For
example, the gap pitch may be allowed to vary by up to Preferably,
the non-monotonicity of the variation happens within a range of a
small number of pixels arrayed side-by-side in a direction, for
example 32 pixels, or for example 16 pixels, or more preferably
within 8 pixels. Within such range, the variation(s) exhibits an
increase followed by a decrease further followed by another
increase, or a decrease followed by an increase followed by another
decrease.
[0070] The description above also shows how the packing density can
be enhanced by optimizing the width of light guides embedded in the
insulating layers for the different colors of light they transmit
so as to take up the least space possible. Light guides that
transmit lights of shorter wavelengths only should be narrower than
light guides that transmit lights of longer wavelengths.
[0071] While certain exemplary embodiments have been described and
shown in the accompanying drawings, it is to be understood that
such embodiments are merely illustrative of and not restrictive on
the broad invention, and that this invention not be limited to the
specific constructions and arrangements shown and described, since
various other modifications may occur to those ordinarily skilled
in the art.
[0072] For example, pixels in a pixel array may be arranged
side-by-side in directions that make 45 degrees with the
left-to-right and top-to-bottom directions (parallel to the plane
of the photoelectric conversion units).
[0073] For example, the top/bottom opening of the upper/lower light
guide may take a different shape than a rectangle, such as an
octagon or a rectangle that has rounded corners.
* * * * *