CHAPTER 11 Digital Microscope Cameras 71
Initially, the easiest way to determine the sensitivity of your camera is to take
a glance at the quantum efficiency curve. This curve tells you how much of
the light hitting the camera from the sample is used to produce the final image.
Silicon-based
sensors (CCD or CMOS) work from around 350 nm up to the range
of 1,000 nm. Peak sensitivity is usually reached around 500 to 550 nm. The peak
quantum efficiency is in the range of 70 % for standard sensors and up to 98 % for
special back-illuminated devices.
Alongside quantum efficiency, the pixel size of the camera also influences sensitivity:
the larger a pixel, the more photons can be caught and subsequently the
more signal will be generated. This parameter is very dominant: a 6.5 μm pixel offers
double the area compared to a 4.5 μm pixel and so can be twice as sensitive.
Sometimes you might rather give up resolution than sacrifice sensitivity. In that
case you would be better off choosing a microscope camera with a large pixel size.
These collect more light so are the better choice for low-light applications. The
high NA of the objective lens is then used to collect light rather than increase resolution
in this combination. And remember: pixel size can always be enlarged by
applying the binning method.
Imagine this: you are standing in your kitchen and the range hood is turned on
and making noise. Someone outside the kitchen is trying to ask you a question.
You know this won’t work because the noise of the fan overlays and distorts the
voice of the person asking the question in such a way that it becomes incomprehensible.
Something similar can happen in an electronic camera. Several noise
sources exist and can obstruct the electron signal you want to record.
Some dynamic examples are: readout noise, dark current noise, photon shot
noise.
radio telegraph noise …
Some static examples are: photo response non-uniformity, dark current non-
uniformity …
Some of these unwanted signals can be compensated for or minimized, some not.
For example, the dark current noise can be minimized by careful sensor design
and production, and by sensor cooling. Cooling reduces this type of noise especially
for longer exposure times. However, it makes a camera more expensive. The
static noise can be removed by doing a black reference image, which will be subtracted
from the final image automatically after exposure. This requires the careful
acquisition of a black image in complete darkness and can be done just once
at the beginning of the work.