Photoelectric gain

For a comparison of different photoconductors, a knowledge of their photoelectric gains is very important. However, it should not be forgotten that G depends, among other factors, on the mean lifetime r. Since the parameters responsible for the mean carrier lifetime depend considerably on structural effects... 

Interestingly, the action spectrum of self-sensitization follows the photoelectric gain curve of the multilayers It was therefore concluded that the same transition... 

Polydiacetylene multilayers were also found to be photoconducting . Due to their low thickness d the condition ad < 1 (a being the absorption coefficient) can be established and the photoelectric gain ought to reflect the absorption spectrum of the ionizing transition directly. 

Photoelectric gain r The number of electrons flowing through the photodetector circuit per an absorbed photon the ratio of lifetime and transit time r = T/r ... 

This is actually the increase of the dark current due to illumination and is a dc value (valid for/= 0 Hz). Here rj denotes the quantum efficiency of the detector, < ) is the incident photon flux density, A is the detector active area. The factor T denotes the photoelectric gain or photogain (the ratio between the number of electrons flowing through the electric circuit and the number of absorbed photons). The fundamental equation of photoconductivity is also valid without changes for the short circuit current of a photovoltaic detector (photodiode operating in photo-conductive mode). In that mode of operation T = 1 in most of the cases. 

Specific detectivity is proportional to photoelectric gain. This is why it is convenient to use e.g., avalanche devices, where the photogain is large. However, the photogain increase must not compromise other factors, for instance response time or noise level. 

We consider further the determination of local values of generation-recombination noise for the case when carrier concentration within detector is position-dependent, i.e., when g-r rate and photoelectric gain are spatially inhomogeneous. We use such spatial distribution to determine total noise current (g-r plus thermal) through the whole detector. For the sake of simphcity, we assume that the gradient exists only along one direction, parallel to the y-axis. We consider a photocon-ductive device. 

We divide the detector stmcture into a large number of layers sufficiently thin to permit us to regard g-r rate and photoelectric gain as constant. To one such layers with a thickness dy and located at a position y we apply the standard expression for the g-r noise of photoconductors (1.88). For a photovoltaic detector (1.89) is used in the same manner. Thus, we obtain the increase of noise current due to the processes taking place in the layer itself... 

In (1.92) F(y) represents the position-dependent photoelectric gain. In a position y it will be proportional to the ratio of the total detector current change and the generation rate change in the considered thin layer, so that we may write [67]... 

Photoelectric gain should be maximized, but without compromising other detector characteristics like response time or noise level. 

We will focus our attention to the last item, noise decrease in photodetectors. We start from (1.95), assuming that the given parameters are temperature (must be as near to the room temperature as possible) and photoelectric gain (must furnish maximum sensitivity and basically is given by the chosen detection mechanism). We conclude that the following two conditions are to be met... 

Figure 3.39 shows spatial distribution of photoelectric gain within a p jin mercury cadmium extraction photodiode for different bias voltages. The detector parameters are given in the Fig. 3.39. Gain is given for the whole stmcture, including not only the depleted region, but also highly doped parts. An important...

F. 3.39 Spatial distribution of photoelectric gain in a p jm" extraction diode for different bias levels... 

The next diagram, Fig. 3.40, shows spatial distribution of noise density factor (product of the sum of absolute values of generation and recombination rates and squared photoelectric gain, ( G + i )r. ... 

Arrays of 64 HgCdTe elements with mesa diodes (20-80) pm were fabricated [362]. The doping level in the ti region was 2-4 x lO cm the region width 5 pm. The n-type dopant was iodine, p-type dopant was arsenic. In these devices an excessive 1/f noise component was noted with a knee at about 10 MHz. The authors noted that the diodes they fabricated had a photoelectric gain T > 1. 

In later generation devices exclusively heterostructures have been used. Skauli et al. [383] reported MBE-fabricated Hgi- CdxTe extraction diodes using silver as dopant in P region and indium in the N region. All regions in their three-layer structures were about 3.5 pm thick, and they worked at = 9 pm for a temperature of 295 K. These authors also noticed photoelectric gain higher than zero within the active areas of their diodes. 

The calculated photoelectric gain of a magnetoconcentration photoconductor increases with the material composition approaching the optimum at a given tem-peramre and tends to unity, except near the surface with high recombination rate. 

Noise current slowly decreases with the position within the photoconductor, reaches a minimum and then sharply increases toward the back surface (with high recombination velocity.) After reaching its maximum, it becomes lower again because of a sharp drop of photoelectric gain, in spite of the fact that the g-r processes are actually largest just in that position. 

---