U.S. patent application number 10/521427 was filed with the patent office on 2006-01-12 for tfa image sensor with stability-optimized photodiode.
This patent application is currently assigned to STMicroelectronics N.V.. Invention is credited to Arash Mirhamed, Jens Prima, Peter Rieve, Konstantin Seibel, Marcus Walder.
Application Number | 20060006482 10/521427 |
Document ID | / |
Family ID | 30009999 |
Filed Date | 2006-01-12 |
United States Patent
Application |
20060006482 |
Kind Code |
A1 |
Rieve; Peter ; et
al. |
January 12, 2006 |
Tfa image sensor with stability-optimized photodiode
Abstract
The invention relates to a TFA image sensor with
stability-optimized photodiode for converting electromagnetic
radiation into an intensity-dependent photocurrent with an
intermetal dielectric, on which, in the region of the pixel matrix,
a lower barrier layer is situated and a conductive layer is
situated on the barrier layer, and vias being provided for the
contact connection to the ASIC, the vias in metal contacts on the
ASIC. A TFA image sensor having improved electrical properties is
provided. This is achieved in that an intrinsic absorption layer is
provided between the TCO layer and the barrier layer with a layer
thickness of between 300 nm and 600 nm. Before the application of
the photodiodes, the topmost, comparatively thick metal layer of
the ASIC is removed and replaced by a matrix of thin metal
electrodes which form the back electrodes of the photodiodes, the
matrix being patterned in the pixel raster.
Inventors: |
Rieve; Peter;
(Windeck-Dattenfeld, DE) ; Walder; Marcus;
(Wipperfurth, DE) ; Seibel; Konstantin; (Siegen,
DE) ; Prima; Jens; (Gehrde, DE) ; Mirhamed;
Arash; (Vellmar, DE) |
Correspondence
Address: |
STMicroelectronics Inc.;c/o WOLF, GREENFIELD & SACKS, PC
Federal Reserve Plaza
600 Atlantic Avenue
BOSTON
MA
02210-2206
US
|
Assignee: |
STMicroelectronics N.V.
WTC Schiphol Airport, Schiphol Boulevard 265
Schiphol Airport Amsterdam
NL
1118BH
|
Family ID: |
30009999 |
Appl. No.: |
10/521427 |
Filed: |
July 14, 2003 |
PCT Filed: |
July 14, 2003 |
PCT NO: |
PCT/EP03/07593 |
371 Date: |
July 25, 2005 |
Current U.S.
Class: |
257/414 ;
257/437; 257/443; 257/464; 257/E27.141; 257/E31.062 |
Current CPC
Class: |
H01L 31/1055 20130101;
H01L 27/14665 20130101 |
Class at
Publication: |
257/414 ;
257/443; 257/437; 257/464 |
International
Class: |
H01L 27/14 20060101
H01L027/14; H01L 31/06 20060101 H01L031/06; H01L 31/00 20060101
H01L031/00; H01L 31/0232 20060101 H01L031/0232 |
Foreign Application Data
Date |
Code |
Application Number |
Jul 16, 2002 |
DE |
102 32 019.5 |
Claims
1. A TFA image sensor with stability-optimized photodiode for
converting electromagnetic radiation into an intensity-dependent
photocurrent with an intermetal dielectric, on which, in the region
of the pixel matrix, a lower barrier layer is situated and a
conductive layer is situated on said barrier layer, and vias being
provided for the contact connection to the ASIC, said vias ending
in metal contacts on the ASIC, wherein an intrinsic absorption
layer is provided between the TCO layer and the barrier layerwith a
layer thickness of between 300 nm and 600 nm.
2. The TFA image sensor as claimed in claim 1, wherein the layer
thickness of the intrinsic absorption layer is approximately 450
nm.
3. The TFA image sensor as claimed in claim 1, wherein the band gap
of the intrinsic absorption layer of the photodiode is
increased.
4. The TFA image sensor as claimed in claim 1, wherein the increase
in the band gap is realized by using an amorphous silicon-carbon
alloy (a-Sic:H) as absorption layer.
5. The TFA image sensor as claimed in claim 1, wherein, in
particular, the photodiode of reduced layer thickness is arranged
on a surface that is as planar as possible.
6. The TFA image sensor as claimed in claim 1, wherein the
photodiode with small intrinsic layer thickness is deposited on an
ASIC having a flat surface topography.
7. The TFA image sensor as claimed in claim 1, wherein the ASIC is
coated with a passivation.
8. The TFA image sensor as claimed in claim 1, wherein, within the
pixel matrix, firstly the back electrodes of all the pixels are
connected to one another via the topmost CMOS metal plane, which is
made planar in the region of the pixel matrix.
9. The TFA image sensor as claimed in claim 8, wherein the metal
plane is situated on a CMP-planarized surface (CMP=Chemical
Mechanical Polishing) of the topmost intermetal dielectric
layer.
10. A method for fabricating a TFA image sensor as claimed in claim
1, wherein, before the application of the photodiodes, the topmost,
comparatively thick metal layer of the ASIC is removed and replaced
by a matrix of thin metal electrodes which form the back electrodes
of the photodiodes, said matrix being patterned in the pixel
raster.
11. The method as claimed in claim 10, wherein an antireflection
layer that is present and the metal layer are completely removed
above the pixel matrix, so that all that remains is the barrier
layer situated underneath.
12. The method as claimed in claim 10, wherein the lower barrier
layer is completely removed, this then being followed by the
deposition and patterning of the further metal layer in the form of
pixel back electrodes.
13. The method as claimed in claim 10, wherein the ASIC passivation
is opened in the photoactive region of the TFA sensor.
14. The method as claimed in claim 10, wherein the antireflection
layer of the upper metallization layer of the ASIC in the
photoactive region of the TFA sensor is removed.
15. The method as claimed in claim 10, wherein the conductive layer
of the upper metallization layer of the ASIC in the photoactive
region of the TFA sensor is removed.
16. The method as claimed in claim 10, wherein the lower barrier
layer of the upper metallization layer of the ASIC in the
photoactive region of the TFA sensor is patterned or removed.
17. The method as claimed in claim 16, wherein a further metal
layer is deposited and patterned.
18. The method as claimed in claim 17, wherein further layers, such
as color filter layers, are deposited and patterned.
19. A method for fabricating a TFA image sensor as claimed in claim
1, wherein the ASIC passivation in the photoactive region of the
TFA sensor is opened, the antireflection layer of the upper
metallization layer of the ASIC in the photoactive region of the
TFA sensor is removed, the conductive layer of the upper
metallization layer of the ASIC in the photoactive region of the
TFA sensor is removed, the lower barrier layer of the upper
metallization layer of the ASIC in the photoactive region of the
TFA sensor is patterned or removed, a further metal layer is
deposited and patterned, the photodiode layers are deposited and
patterned, and further layers, such as color filter layers, are
deposited and patterned.
Description
[0001] The present patent application relates to a TFA image sensor
with stability-optimized photodiode for converting electromagnetic
radiation into an intensity-dependent photocurrent with an
intermetal dielectric, on which, in the region of the pixel matrix,
a lower barrier layer (metal 2) is situated and a conductive layer
(metal 2) is situated on said barrier layer, and vias being
provided for the contact connection to the ASIC, said vias in metal
contacts on the ASIC.
[0002] Such a TFA sensor (Thin Film on ASIC (TFA) Technology)
comprises a matrix-organized or linear arrangement of pixels. The
electronic circuits for operating the sensor (e.g. pixel
electronics, peripheral electronics, system electronics) are
usually realized using CMOS-based silicon technology and form an
application specific integrated circuit (ASIC).
[0003] Isolated therefrom by an insulating layer and connected
thereto by means of corresponding electrical contacts, there is
situated on the ASIC a multilayer arrangement as photodiode, which
performs the conversion of electromagnetic radiation into an
intensity-dependent photocurrent. Said photocurrent is transferred
at specific contacts--present in each pixel--of the pixel
electronics underneath (B. Schneider, P. Rieve, M. Bohm, Image
Sensors in TFA (Thin Film on ASIC) Technology, ed. B. Jahne, H.
Hausecker, P. Gei.beta.ler, Handbook of Computer Vision and
Applications, pp. 237-270, Academic Press, San Diego, 1999).
[0004] According to the prior art (J. A. Theil, M. Cao, G. Kooi, G.
W. Ray, W. Greene, J. Lin, A J. Budrys, U. Yoon, S. M a, H. Stork,
Hydrogenated Amorphous Silicon Photodiode Technology for Advanced
CMOS Active Pixel Sensor Imagers, MRS Symposium Proceedings, vol.
609, 2000), what is used as photodiode is a pin configuration based
on amorphous silicon, i.e. a sequence comprising a p-conducting, an
intrinsically conducting (intrinsic) and an n-conducting amorphous
silicon layer. The n-type layer usually forms the bottom most layer
facing the ASIC.
[0005] The electrical contacts are formed by a metal layer, for
example, on said side facing the ASIC, while the contact connection
on the side facing the direction of light incidence is generally
effected by a transparent and conductive layer.
[0006] Over and above the pin photodiode mentioned, further
component structures are also possible, e.g. Schottky photodiodes,
in which an intrinsic semiconductor layer is brought into contact
with a suitable metal (for example chromium, titanium, platinum,
palladium, silver), so that the metal-semiconductor junction forms
a Schottky photodiode.
[0007] A typical layer configuration is disclosed in the patent
application TFA image sensor with extremely low dark current (file
reference 10063837.6). Furthermore, detector structures with a
controllable spectral sensitivity are known (P. Rieve, M. Sommer,
M. Wagner, K. Seibel, M. Bohm, a-Si:H Color Imagers and
Colorimetry, Journal of Non-Crystalline Solids, vol. 266-269, pp.
1168-1172, 2000). This basic structure of a TFA image sensor can
furthermore be extended by additional, upstream layers in the
direction of light incidence, for example by color filter layers
(e.g. Bayer pattern, U.S. Pat. No. 3,971,065).
[0008] If amorphous silicon is used as photoactive sensor material,
then the metastability observed in the case of this material
becomes apparent, under certain circumstances. Hydrogenated
amorphous silicon (a-Si:H) comprises a silicon-hydrogen atomic
composite lacking a long-range order as is typical of semiconductor
crystals. Modifications of the atom bonding parameters occur with
respect to the ideal semiconductor crystal. The consequence of this
is that, in the context of the solid-state band model, a state
density that differs from zero exists in the band gap between
conduction band and valence band, which affects the electrical and
optical properties of the material. States in the middle of the
band gap predominantly act as recombination centers, while states
in the vicinity of the band edges function as traps for charge
carriers. On account of light being radiated in or injection of
charge carriers, more precisely through recombination of injected
charge carriers, weak silicon bonds are broken and additional band
gap states arise.
[0009] These band gap states caused by light irradiation represent
additional recombination or trapping centers and influence the
charge carrier transport and the distribution of the electric field
strength in the components fabricated from amorphous silicon. In
pin photodiodes, for example, predominantly positively charged
states are concentrated in traps in that region of the intrinsic
layer (i-type layer) which adjoins the p-type layer, and negatively
charged states in that region of the i-type layer which adjoins the
n-type layer. These stationary charges result in a decrease in the
magnitude of the electric field strength within the i-type layer,
so that the accumulation of photogenerated charge carriers
deteriorates. An efficient charge carrier accumulation in pin
photodiodes made of amorphous silicon is provided when the drift
length (.mu..tau.E) of the charge carriers significantly exceeds
the thickness d of the intrinsic layer: .mu..tau.E>>d (1)
[0010] On account of the increase in the defect density associated
with light being radiated in, on the one hand the lifetime .tau. is
reduced due to intensified recombination of charge carriers and on
the other hand the electric field E is reduced on account of the
charged states in the i-type layer. Both have the consequence that
the ratio of drift length to i-type layer thickness is reduced and
the photocurrent thus decreases. The decrease in the photocurrent
becomes apparent particularly when the photodiode is operated near
the short-circuit point without additional reverse voltage, i.e.
when only the built-in potential difference brought about by the
doped layers is effective.
[0011] When reverse voltage is applied, by contrast, the electric
field intensifies, so that the charge carrier accumulation is
impaired to a less extent. The essential consequence of the light
irradiation of a photodiode made of amorphous silicon with regard
to the photocurrent thus consists in a reduction of the
photocurrent saturation.
[0012] The dark current of an a-Si:H photodiode, i.e. the current
which flows even in the unilluminated state is likewise influenced
by the degradation of the material. On account of the defect states
additionally generated by light being radiated in, the thermal
generation of charge carriers increases in the case of a
reverse-biased photodiode (extraction), which is manifested in an
increase in the dark current.
[0013] The invention is now based on the object of providing a TFA
image sensor with stability-optimized photodiode for converting
electromagnetic radiation into an intensity-dependent photocurrent
having improved electrical properties.
[0014] The formulated object on which the invention is based is
achieved, in the case of a TFA image sensor with
stability-optimized photodiode for converting electromagnetic
radiation into an intensity-dependent photocurrent, in that a layer
thickness of the intrinsic absorption layer of between 300 nm and
600 nm is provided.
[0015] Further refinements of the invention emerge from the
associated subclaims.
[0016] The object on which the invention is based is furthermore
achieved by means of a method which is characterized in that,
before the application of the photodiodes, the topmost,
comparatively thick metal layer of the ASIC is removed and replaced
by a matrix of thin metal electrodes which form the back electrodes
of the photodiodes, said matrix being patterned in the pixel
raster.
[0017] Further refinements of the method according to the invention
emerge from the associated subclaims.
[0018] One particular refinement of the invention is characterized
by opening of the ASIC passivation in the photoactive region of the
TFA sensor, removal of the antireflection layer of the upper
metalization layer of the ASIC in the photoactive region of the TFA
sensor, removal of the conductive layer of the upper metalization
layer of the ASIC in the photoactive region of the TFA sensor,
patterning or removal of the lower barrier layer of the upper
metalization layer of the ASIC in the photoactive region of the TFA
sensor, deposition and patterning of a further metal layer,
deposition and patterning of the photodiode layers, and deposition
and patterning of further layers, such as color filter layers.
[0019] The changes in the dark current and photocurrent brought
about by light being radiated in are reduced, according to the
invention, by reducing the thickness of the intrinsic layer. This
measure brings about an increase in the electric field strength
over the i-type layer, so that the field strength depth caused by
the increase in the defect density on account of light being
radiated in, within the i-type layer, is less sharply
pronounced.
[0020] In this way, the accumulation condition for photogenerated
charge carriers which is given by equation (1) can be met even in
the state of increased defect density (after light has been
radiated in), and a decrease in photosensitivity is avoided. With
regard to the behavior of the photodiode without illumination,
photodiodes with a small i-type layer thickness, in the aged state,
have a lower dark current than those with a thick i-type layer,
which can be attributed to the smaller number of generation centers
present in the band gap.
[0021] The method of improving the stability of photodiodes made of
amorphous silicon by means of a thin absorber layer is known from
the field of photovoltaic technology, where it is employed
successfully in solar cells based on amorphous silicon. Application
to image sensors in TFA technology is novel. The method is suitable
both for photodiodes of the pin or nip type and for Schottky
diodes. A layer thickness of the intrinsic absorption layer of
between 300 nm and 600 nm has proved to be advantageous with regard
to the stability of the photodiode, and it should preferably be
approximately 450 nm.
[0022] One advantageous development consists in increasing the band
gap of the intrinsic absorber layer of the photodiode. The dark
current can be reduced in this way. At the same time, it is
possible to counteract the increase in the diode capacitance which
accompanies the reduction of the i-type layer thickness.
Technologically, it is possible to increase the band gap for
example by using an amorphous silicon-carbon alloy (a-SiC:H) as
absorption layer.
[0023] In photodiodes with a small i-type layer thickness, the
configuration of the surface on which the diode is situated is of
crucial importance for the magnitude of the dark current. Besides
the thermal generation currents already mentioned, inhomogeneities
of the ASIC surface form, caused by the structures (metal tracks,
holes in passivation layer, etc.) situated thereon, a further
source of undesirably high dark currents in TFA image sensors. In
this case, the influence of the surface topography is greater, the
thinner the photodiode situated thereon. In this respect, it is
necessary in particular to deposit the photodiode of reduced layer
thickness on a surface that is as planar as possible.
[0024] One advantageous development of the invention thus consists
in depositing the photodiode with small i-type layer thickness (as
mentioned above) on an ASIC having a flat surface topography. This
is ensured by the fabrication process explained below. The ASIC
can, but need not necessarily, be coated with a passivation.
[0025] Within the pixel matrix, firstly the back electrodes of all
the pixels are connected to one another via the topmost CMOS metal
plane, which is made planar in the region of the pixel matrix. This
metal area is situated on a CMP-planarized surface (CMP=Chemical
Mechanical Polishing) of the topmost intermetal dielectric layer.
Before the application of the photodiodes, this topmost,
comparatively thick metal layer of the ASIC is removed and replaced
by a matrix of thin metal electrodes which form the back electrodes
of the photodiodes, said matrix being patterned in the pixel
raster. The topmost metallization of the ASIC generally comprises a
multilayered arrangement comprising a lower barrier layer, e.g.
titanium nitride or titanium, the actual conductive layer, e.g.
aluminum (alloys) and, if appropriate, an upper antireflection
layer, e.g. titanium nitride. In an expedient manner, the
antireflection layer (if present) and the metal layer are
completely removed above the pixel matrix, so that all that remains
is the lower barrier layer. The latter is then patterned in the
pixel raster and either forms the pixel back electrode directly, or
it is coated with a further metal layer, e.g. chromium, which forms
the matrix of the pixel back electrodes after a further patterning
step. As an alternative, the lower barrier layer is completely
removed, this then being followed by the deposition and patterning
of the further metal layer in the form of pixel back
electrodes.
[0026] The process steps are summarized below as key points:
[0027] a) if appropriate opening of the ASIC passivation in the
photoactive region of the TFA sensor,
[0028] b) if appropriate removal of the antireflection layer of the
upper metallization layer of the ASIC in the photoactive region of
the TFA sensor,
[0029] c) removal of the conductive layer of the upper
metallization layer of the ASIC in the photoactive region of the
TFA sensor,
[0030] d) patterning or removal of the lower barrier layer of the
upper metallization layer of the ASIC in the photoactive region of
the TFA sensor,
[0031] e) if appropriate deposition and patterning of a further
metal layer,
[0032] f) deposition and patterning of the photodiode layers,
[0033] g) if appropriate deposition and patterning of further
layers (e.g. color filter layers).
[0034] In this way, a largely planar surface is ensured in the
region of the active pixel matrix of the sensor because the etching
attack into the topmost intermetal dielectric layer is reduced to a
minimum. It is only during the patterning or removal of the lower
barrier layer of the topmost ASIC metallization layer that the
CMP-planarized dielectric layer is uncovered and is removed locally
by the etching attack, which can be minimized by a suitable choice
of process parameters. Apart from that, the flat surface topography
of the CMP planarization is maintained, thus avoiding any
influencing of the dark current of the photodiodes deposited
thereon.
[0035] The invention is explained below with reference to some
drawings. FIGS. 1 and 2 show a pin and, respectively, a Schottky
photodiode with an intrinsic absorption layer i according to the
invention made of amorphous silicon in the layer thickness range of
between 300 nm and 600 nm. The following figures relate to the
abovementioned fabrication process which ensures a largely planar
surface topography.
[0036] In this case, the illustrations only include the topmost
layers of the ASIC which are relevant to the interface with the TFA
layers.
[0037] FIG. 3 illustrates the initial state before the beginning of
the TFA processing in the form of a passivated ASIC with
passivation that has been opened in the region of the pixel matrix.
The antireflection layer of the topmost metallization layer of the
ASIC is likewise removed in the pixel region.
[0038] In this case, firstly an intermetal dielectric is arranged
on the ASIC and, in the region of the pixel matrix, a lower barrier
layer (metal 2) is situated on said intermetal dielectric and a
conductive layer (metal 2) is situated on said barrier layer. Vias
are provided for the contact connection to the ASIC, said vias
ending in metal contacts on the ASIC. Furthermore, a bond pad
(metal 2) for external contact connection is provided, which is
contact-connected to the ASIC by means of vias and a metal 1.
[0039] The state after the removal of the conductive layer of the
topmost metallization is recorded in FIG. 4.
[0040] FIG. 5 documents the result after the patterning of the
lower barrier layer. This produces the pixel back electrodes, which
are subsequently coated with the multilayer system comprising
amorphous silicon and TCO (FIG. 6). FIGS. 7 and 8 show a process
variant in which, proceeding from the situation according to FIG.
5, the patterned regions of the lower barrier layer are covered by
a further patterned metal layer before the deposition of the
photodiode. Beginning with FIG. 9, a further variant is illustrated
in which, after the situation outlined in FIG. 4, the lower barrier
layer of the topmost metallization layer of the ASIC is completely
removed. The further metal layer is subsequently deposited directly
onto the intermetal dielectric, and forms the pixel back electrodes
after patterning (FIG. 10). The photodiode-forming layers are then
applied thereto (FIG. 11).
* * * * *