Fiber-Optic Sensors - Applied Optics Fundamentals and Device Applications - page 164

As a final consideration we note that since the area of detectors is usually
larger than the fiber core area, it is conceptually a straightforward matter to
couple light from the fiber to the detector. Coupling loss arises mostly from
Fresnel reflections and can be minimized by refractive index matching or by
the use of an antireflection layer on the detector.
4.11 FiberSensorDetectorConsiderations
There exist a large number of different photodetectors, ranging from the
large photoemissive and photomultiplier vacuum tubes to solid-state and
semiconductor devices. The latter are the appropriate choice for fiber sys-
tems, with the most practical operating in the near infrared range. These are
the silicon p-n and PIN diodes operated in reverse-biased photoconductive
mode. These devices provide a change in conductivity proportional to the
optical generation rate. When operated in the third quadrant of the IV char-
acteristic, current is essentially independent of voltage but proportional to
the generation rate of electron-hole pairs. Such a device provides a useful
means of measuring illumination levels or of converting time-varying opti-
cal signals into electrical signals.
When an election-hole pair is created in the depletion region, the electric
field sweeps the electron to the n side and the hole to the p side. It is therefore
desirable to make the depletion width W large enough so that most of the
photons are absorbed within W (depletion-layer photodiode) rather than in
the neutral p and n regions. The appropriate W is chosen as a compromise
between sensitivity and speed of response, since W must not be so wide that
the time required for drift of photogenerated carriers out of the depletion
region is excessive.
Internal gain is produced by applying external voltage sufficient to extend
the high-field region into the two regions adjoining the junction. Accelerated
carriers now have an increased probability of colliding with the semicon-
ductor atoms, freeing even more carriers. The PIN structure includes an
intrinsic region tailored to meet particular drift and multiplication effects.
For detection of very low optical signals, the photodiode is often operated in
the avalanche region (avalanche photodiode). In this mode, each photogen-
erated carrier results in a significant change in the current due to avalanche
multiplication. Significant current gain can be achieved, but at the expense
of the introduction of excess noise. The APD is a more expensive component
than the PD, and requires a greater complexity for the bias circuit and power
supply.
Numerous commercial PIN diodes are available for operation at around
850 nm. This is, in fact, near the peak sensitivity of silicon devices, with a
typical range up to 400-600 μA mW −1 . Minimum detectable signal levels are