## Photoconductors and photodiodes

Photoconductor: This is the simplest application of a semiconductor for detection of photons. A typical photoconductor arrangement is shown in Figure 5.16. Photons are absorbed and create electron-hole pairs. If the material is extrinsic rather than intrinsic, then E{ must be substituted for EG. Also, for extrinsic materials there are limits on solubility of the dopants, and high concentrations introduce unwanted conductivity modes such as "hopping" which involves conduction between neighboring impurity atoms without raising an electron to the conduction band. In the discussion below we assume that the semiconductor has been cooled to eliminate thermally generated charges. In practice, both electrons and holes contribute to the photocurrent, but it is usually the electrons that dominate. The average photocurrent (I) between the terminals that is generated by an incident flux with power P (watts) is given by Figure 5.16. The basic construction and operation of a semiconductor used in photoconduction mode.

In this expression q is the quantum efficiency; and P/hv is just the photon arrival rate. The quantity r is called the mean carrier lifetime and measures how long the photogenerated charge exists before recombination. Values are usually less than to much less than a few milliseconds but depend on doping and temperature. The average charge carrier velocity is v, which is related to the applied electric field across the photoconductor E = V/l by v = ^E, where ^ is called the mobility of the charge carrier. Thus, l/v is the transit time across the device from one terminal to the other, and the quantity G = vr/l is just the ratio of mean carrier lifetime to transit time. It is known as the "photoconductive gain". The response of the detector (in amps per watt or volts per watt) is just I/P or V/RP where V is the bias voltage across the photoconductor, and the resistance R due to the photocurrent is l/aA and the conductivity a = ne^, where n is the average density of carriers. It follows that S = (eqG/hc)\. Finally, the root-mean-square noise for a photoconductor is given by y/(4eGIB) where B is the electrical bandwidth of the measurement.

Photodiodes: Junctions between p-type and n-type regions are used many times in semiconductor structures to produce different devices. One such device is the photodiode. When a p-n junction is formed, electrons from the n region tend to diffuse into the p region near the junction and fill up some of the positively ionized states or holes in the valence band thus making that p-type region more negative than it was. Similarly, the diffusion of holes from the p-side to the n-side leads to an increasingly more positive electrical potential. A narrow region forms on either side of the junction in which the majority charge carriers are "depleted" relative to their concentrations well away from the junction. As the concentration of electrons in the n-type material is usually very much larger than in the p-type material, the flow of electrons would tend to be one way were it not for the fact that the diffusion process itself begins to build up an electrostatic potential barrier which restrains the flow of electrons from the n-type region; the build-up of electrons on the p-side makes it negatively charged which starts to repel further diffusion. The magnitude of this potential barrier (V0) depends on impurity concentrations (i.e., on the number of donor electrons at the junction that are available for transfer to nearby acceptor

Depletion region

Depletion region electrons Figure 5.17. The formation of a p-njunction between p-doped and n-doped materials results in a region depleted of carriers and the creation of a potential barrier.

"rr

VOLTAGE

CHARGE

electrons

Figure 5.17. The formation of a p-njunction between p-doped and n-doped materials results in a region depleted of carriers and the creation of a potential barrier.

levels) and is just equal to the required shift of the energy bands needed to ensure that the Fermi level (Ep) remains constant throughout the crystal. The Fermi level is the energy at which there is a 50/50 chance of the corresponding electron energy state or orbit being occupied by an electron. For an intrinsic semiconductor, Ep lies half-way between the valence and conduction bands whereas for an n-type doped semiconductor the Fermi level moves up toward the conduction band, and conversely p-type doping lowers the Fermi level.

The potential drop across the "depletion layer" is about 0.6 V for silicon and 0.3 V for germanium, half of the forbidden energy gap. Figure 5.17 shows how this condition is represented by an energy band model of voltage or potential vs. distance from the junction. The width of the junction region (Z1 + X2) increases with increasing voltage (F0) and decreasing number density of acceptor atoms.

If a positive voltage is applied to the p-side of the junction it will tend to counteract or reduce the built-in potential barrier and attract more electrons across the junction, whereas a negative voltage on the p-side will enhance the internal barrier and increase the width of the depletion region; these conditions are called "forward" and "reversed" bias, respectively. Therefore, on one side of a p-n junction there is a region which is more negative and on the other side there is a region which is more positive than elsewhere in the crystal. When light of the correct wavelength is absorbed near the junction an electron-hole pair is created and the potential difference across the junction sweeps the pair apart before they can recombine. Electrons are drawn toward the region of greatest positive potential buried in the n-type layer which therefore behaves like a charge storage capacitor. Of course, as more electrons accumulate, the positive potential is progressively weakened. In the photodiode, an electron-hole pair is created within the depletion region by the absorption of a photon, and the charge carriers are immediately separated by the electric field across the junction. The current due to an incident photon flux (signal and background) of power P is just I = e-qP/hv and the root-mean-square noise is given by ^J(2eIB), where B is the electrical frequency bandwidth of the measurement. Comparing this with the photoconductor we see that G = 1 for the photodiode and that the noise is less by a factor of 2 because recombination does not occur in the depletion region. 