Image Processing Reference
In-Depth Information
1.0
SmartFSI
0.9
BSI
0.8
Lightpipe FSI
0.7
0.6
0.5
0.4
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
Height from Si surface (µm)
FIGURE 5.72
Simulated light energy loss in light propagating path for green light. Arrows indicate energy loss at color filter.
(Reprinted with permission from Watanabe, H., Hirai, J., Katsuno, M., Tachikawa, K., Tsuji, S., Kataoka, M.,
Kawagishi, S. et al., Proceedings of the IEEE International Electron Devices Meeting, Technical Digest , 8.3, pp. 179-182,
Washington, DC, 2011.)
of 46.3%, 72.0%, and 63.1% are obtained in the red, green, and blue channels, respectively,
as well as low optical and electrical cross talk. The stacked lightpipe structures bring the
benefit of higher acceptance angles for incident light, and the maximum incidence angle
was 40 (±20)º at 20% down level of signal quantity of normal light incidence. The situation
is the same for smaller pixel pitch of 1.12 μm. 60
While techniques (1) and (3) seem applicable to BSIs, because they are countermeasures
for common issues for both FSIs and BSIs and cost-effectiveness is very important in this
industry. Future developments will attract much attention.
5.3.3.4 Organic Sensors
Although organic sensors should not necessarily be categorized as CMOS sensors, they
are tentatively classified this way here.
There is a new trend using organic photoconductive material instead of silicon as the
sensor part.
Before describing organic sensors, silicon image sensors are reviewed. In Section 2.2, it
is pointed out that silicon is not necessarily the best material for the photoelectric conver-
sion part of integrated circuits, although it is a natural choice because it senses visual light.
As silicon needs to be 3-4 μm thick to absorb sufficient red light, PDs that are tall or have
a high aspect ratio are required along with pixel pitch shrinkage. This structure is apt to
bring about cross talk, which should be avoided as much as possible in SCCCs, which is
the largest application.
In Section 5.3.2.2, 3-Tr and 4-Tr pixel configurations are compared, and it is pointed out
that the CMOS sensor most comparable to CCDs is the 4-Tr type with a combination of
a buried/pinned PD. To be pinned, PDs have to be depleted just after signal charges are
read out. This means the maximum signal charge, electron, number that can be integrated
in a pinned PD is the same as the maximum impurity atom, donor, number in the n -type
region of the PD. This leads to limitation of the saturation level or dynamic range and
inadequate SNR for high-illumination scenes along with pixel pitch shrinkage.
Search WWH ::




Custom Search