 |
Figure 3: Drawing of amorphous silicon microbolometer pixel |
|
Technology
Advancements of Amorphous Silicon Microbolometers
Since their initial development, amorphous
silicon (a-Si) microbolometers have substantially matured.
Technological breakthroughs have occurred at a constant
pace, resulting in improvements in performance (such as
resolution and sensitivity) as well as in detector packaging.
Pixel uniformity and operability have also continued to
improve as well due to production changes.
Reduced Pixel Geometries
Because amorphous silicon is inherently
stiff, microbolometers manufactured with this material
can be made with very small bridge structures. As a result,
a-Si microbolometers exhibit higher effective thermal
insulation compared with VOx based devices due to thinner
membranes. This reduction in bridge size allows for the
development of small pixels and therefore more pixels
per area, smaller physical sizes and lower costs. As shown
in Figure 3, the amorphous silicon microbolometer pixel
is quite simple in structure, so reducing pixel size does
not require a complex microbridge structure since the
structure scales appropriately. As a result, the number
of fabrication operations in the manufacturing process
is optimized.
 |
Figure 4: Pixel pitch for amorphous silicon microbolometers have continued to decrease due to technological advancements. |
|
As shown in the chart in Figure 4, over
the past ten years, a-Si IR detector arrays have seen
a steady reduction in pixel pitch from 45μm back in
2000 to 17μm pixel pitch in 2008. The next generation
arrays are projected to have a 12μm pixel pitch geometry,
further reducing the size of detectors and optics as well
as the size of infrared cameras.
In conjunction with pixel pitch reductions,
arrays offering higher resolution have been introduced.
Many microbolometer array sizes are available as shown
in Figure 5, from a low resolution 160x120 array to
a large format 1024x768 array. In light of the reduction
in pixel size and the increase in resolution, the active
detection area for these arrays remains nearly constant
for certain comparative array sizes. As shown in Figure
6, the width of a detector having 320x240 pixels with
45µm pixels (14.4mm) is quite similar to the width
of a detector having 640x480 pixels with 25µm
pixels (16.0mm) and also to the width of a detector
having 1024x768 pixels with 17µm pixels (17.4mm).
With similar detection areas, production and packaging
costs are maintained and the size and cost of infrared
camera housings and optics can also be stabilized.
 |
| Figure 5: A wide range of array sizes are now available, from 160x120 pixel arrays, for use in economical infrared cameras, to 1024x768 pixel arrays for very high resolution products. |
|
 |
| Figure 6: As pixel pitch decreases and pixel counts increase, detector active area stays fairly stable. |
|
Thermal Time
Constant and Sensitivity
| For portable 30Hz infrared cameras, it is highly desirable
to have pixel response times shorter than 10 ms. This
is easily achieved with a-Si pixels. |
|
Thermal time constant is one of the most
important parameters in the design of an infrared imaging
detector exceeding the importance of other parameters.
Again, because of the intrinsic properties of silicon,
a-Si pixels can be produced with very thin microstructures
because of their rigid behavior. With these thin microstructures,
pixels exhibit a very low thermal conductance between
the pixel and substrate and low suspended thermal pixel
mass. This results in pixels having a very short time
constant and very fast response time which is an enormous
benefit for any infrared detector. It is commonly accepted
that pixel response times should not exceed one-third
of the reciprocal of the frame time (i.e. 10ms for an
array operating at 30Hz). This is easily achieved with
a-Si pixels.
As pixel pitch decreases, other parameters
need to be appropriately modified in order to keep pixel
sensitivity from decreasing since the active detection
area becomes smaller. Because sensitivity has a great
impact on image quality, certain design changes are important
to assure that detector sensitivity remains at the desired
level. For example, when the pixel pitch of a-Si microbolometers
was reduced from 45μm to 25μm, improvements were
made in the photolithographic resolution without adversely
impacting fill factor. While 1.5µm minimum design
rules were used for 45µm pitch detectors, 25µm
detectors required 0.8 µm minimum design rules,
and 0.5µm for 17µm pitch detectors. However,
improvements in the design of the support legs as well
as a reduction in both the width and thickness of the
patterned legs resulted in higher thermal insulation.
The result was a microbolometer array having the same
sensitivity but with all the advantages of smaller pixels.
In order to further reduce
the pixel pitch, modifications were required to the design
of pixels in order to maintain the pixel time constant
and detector sensitivity. It was necessary to reduce the
thickness of the microbridge films in order to reduce
its thermal mass. In order to maintain good noise immunity,
the microbridge was thinned down at the same rate of decrease
as the detector active area. At the same time (but more
challenging), the thermal conductance also needed to be
reduced at a rate corresponding to the square of the pixel
area. Because of the simplicity of the silicon substrate,
the thickness of the a-Si microbridge could be easily
reduced with very few restrictions.
As an example of the importance of pixel
time constant, compare the results of different a-Si focal
planes as shown in the table below, one standard in design,
and another designed for high sensitivity, narrow intra-scene
dynamic range applications such as would be required in
astronomical missions.
| |
Temporal
NETD |
Pixel
time constant |
Intra-scene
dynamic range |
Standard design |
30mK |
7ms |
200K |
MERTIS design |
11mK |
13.6ms |
19K |
|
Image Uniformity
and Operability
As with other imaging arrays, each microbolometer
pixel has a slightly different response given by the pixel
gain and pixel offset. During the manufacturing process,
each array has a resulting distribution of gain and offset
values from each pixel. Because of the well-established
silicon manufacturing methods and the simplicity of its
structure, amorphous silicon microbolometers exhibit excellent
image uniformity as shown by the very similar behavior
of all of the pixels. Excellent array uniformity means
that the array gain can be increased prior to digitization
in order to improve overall system signal-to-noise performance.
 |
| Figure 7: Amorphous silicon detectors are well known for their excellent array uniformity, resulting in improved image quality, reduced system complexity and lower overall costs. |
|
The operability of an IR detector
array is defined by the number of defective pixels in
the array, or the percentage of defective pixels. Highly
uniform arrays have a smaller number of bad pixels since
pixels are defined as defective if their characteristics
are significantly different than the average. To achieve
the best image, array pixel response should be as uniform
as possible across the surface of the device prior to
any non-uniformity correction. Amorphous silicon detectors
traditionally offer high operability rates partly because
of the simplicity of silicon technology and the benefits
of established processes. Operability of better than
99.5% is standard in amorphous silicon detectors and
rates as high as 99.9% are not uncommon.
| Amorphous silicon detectors traditionally offer
high pixel operability partly because of the simplicity
of silicon technology and the benefits of well-established
manufacturing processes. Operability of better
than 99.5% is typical and rates as high as 99.9%
are not uncommon. |
|
Because of their intrinsic
uniformity, operating a-Si detectors has other important
advantages over other types of microbolometer detectors.
For example, these detectors can be operated quite easily
without a thermoelectric cooler (“TEC-less”).
Since the a-Si pixels can only have one activation energy
that defines the bolometer resistance as a function of
temperature, knowing the pixel resistance at one temperature
is sufficient for predicting the resistance at other temperatures.
Microbolometer arrays manufactured with a composite of
materials (e.g. VOx) have numerous factors that impact
the response of the pixels in the array and are not so
easily characterized. Consequently, they exhibit higher
non-uniformity characteristics and are more difficult
to operate in TEC-less mode since the wide variation in
response cannot be easily predicted. In addition, if not
appropriately compensated, some of the pixels in the array
may appear saturated, having a response outside the detector
digitization range. Because of the high uniformity of
a-Si detector arrays, the raw signal delivered by the
detector is sufficiently uniform to be processed without
any additional electronics, which reduces system complexity,
calibration, yield, power consumption, size and finally
cost of the overall system.
Applications
Amorphous silicon detectors are ideally
suited for many IR applications due to their high performance,
high reliability and lower cost. Applications for microbolometers
exist in both commercial and military sectors. Some of
the top applications are as follows: |
 |
| Figure 8: Infrared cameras are ideal for night-time surveillance applications. |
|
Surveillance
Because of the high contrast between humans
and vehicles, uncooled IR detectors are an ideal choice
for surveillance application. Surveillance applications
encompass many different areas including, night vision,
security, and basic surveillance. Because of the technological
advances including 17µm pixel pitch and large format
detectors, a-Si microbolometers are finding their way
into many more military and commercial night vision applications.
For example, one military application that benefits from
these advances is unmanned aerial vehicles (UAVs) which
require lightweight and reliable IR detectors that can
stand physical punishment including high vibrations and
high G force loads. These requirements make uncooled a-SI
microbolometers ideally suited for this application.
 |
| Figure 9: Because infrared cameras can see through smoke, they are a valuable tool for firefighters. |
|
Firefighting
Another application for IR detectors is
firefighting. Firefighters are often thrust into situations
where visible light cannot be used due to smoky conditions.
In low visibility conditions, finding trapped victims
or downed fire fighters can be near impossible. Thermal
imaging cameras have allowed fire fighters to see through
the smoke and locate trapped victims and firefighters.
Additionally, Thermal imaging allows for the detection
of hot spots after a fire is extinguished, as well as
to find the source of active fires.
 |
| Figure 10: Identifying unexpected heat loss in electrical connections has been found to be key in preventing costly failures to occur. |
|
Industrial Inspection
Certain infrared cameras can be calibrated
so that they provide for the non-contact measurement of
object temperature. These find plentiful applications
in industrial inspection. One of the most common uses
of these radiometric infrared cameras is in industrial
environments having high power usage. Inspectors use IR
cameras for thermal analysis of electrical and mechanical
components for problem prevention and detection. In these
systems, excess heat can be a sign of a potential problem.
Infrared cameras are used to scan for many different issues
including loose or dirty connections, overloaded circuits,
plugged cooling lines in transformers, and much more.
Additionally, the technology is non-invasive and non-destructive
and allows for the testing of many points per day. |
 |
 |
| Figure 11: With rising energy costs, identifying heat loss in homes and buildings has become a valued application for infrared cameras. |
|
Energy Conservation
According to the US Department of Energy “heating
and cooling account for 50 to 70% of the energy used in
the average American home. Inadequate insulation and air
leakage are leading causes of energy waste in most homes” [See
Reference 1]. Energy conservation is a booming industry
fueled by higher energy costs, global warming, and dwindling
natural resources. In an effort to make homes and businesses
more energy efficient, energy audits are becoming more
commonplace. IR detectors and thermal imaging technologies
provide an easy to use tool for conducting energy audits.
Energy audits primarily focus on evaluating the effectiveness
of insulation and seals around exterior doors and windows
in addition to looking for gaps or cracks that are leaking
energy. In addition, thermal imaging cameras can save
time and money in detecting leaks in pipes as well. Both
in wall and buried pipes can be quickly scanned without
the need to destroy a wall or dig up a pipe. Common applications
include: hot water leaks in buried pipe, oil and gas pipeline
leaks, and leak detection in municipal water mains.
 |
| Figure 12: Fever screening cameras are now commonplace at airports and other transit locations to monitor travelers having elevated body temperature. |
|
Medical Monitoring
Infrared cameras have become quite popular
at airports and other transit locations for non-contact “fever
screening” where public health officials can swiftly
scan and measure the skin temperature of people as they
pass. Individuals showing an elevated temperature can
be evaluated in more detail to help prevent the spread
of disease.
 |
| Figure 13: Infrared cameras are now available as an option on many new cars for improved night vision. |
|
Automotive Night Vision
Over the last decade IR detectors have become
more commonplace in civilian applications. In addition
to being used for surveillance, IR detectors and cameras
are now becoming standard in automobiles as automotive
night vision systems. Automotive night vision systems
are used to assist drivers in dark conditions, as well
as in bad weather. Automotive night vision systems increase
a driver’s seeing distance and perception in these
conditions beyond the reach of the vehicle’s headlights. |
| |
|
|
Summary
Significant technological improvements
have occurred in uncooled microbolometers over the last
few years fueling the growth of infrared camera designs
that use them. With improvements in sensitivity and resolution
and because of a wide variety of available pixel sizes,
infrared cameras meeting a variety of application-specific
requirements have been introduced. In particular, amorphous
silicon microbolometers are shown to have specific benefits
in performance (including thermal time constant and uniformity)
and cost due to steady technical advances, silicon-based
manufacturing advantages as well as yield improvements.
Their simplified design, small size and mature silicon-based
manufacturing technology have been leveraged to make everything
from handheld thermal imaging cameras to night vision
systems. Thermal imaging products have found their way
into a variety of different applications, including: surveillance,
firefighting, industrial inspection, energy conservation,
medical monitoring and automotive night vision. |
References
- US Dept. of Energy Insulation Fact Sheet: www.ornl.gov/sci/roofs+walls/insulation/ins_01.html.
- Jean-Jacques Yon, Eric Mottin,
Jean-Luc Tissot, “Latest
amorphous silicon microbolometer developments at LETI-LIR”,
Proc. SPIE Vol. 6940, May 2008.
- C. Trouilleau, B. Fièque,
S. Noblet, “High-performance uncooled amorphous silicon TEC less XGA IRFPA with 17μm pixel-pitch”, Proc. SPIE Vol. 7298, May 2009.
|
|
|
|