International Radiation Detectors, Inc.

Linearity

Figure 1 shows the linearity of the UVG-20 photodiode and a widely used P on N photodiode of equivalent area when exposed to increasing levels of 430 nm radiation. The standard ac-dc method was used to measure linearity. The P on N photodiode showed a noticeable decrease in responsivity at photocurrents greater than 500 µA while the UVG-20 photodiode showed only 0.02% decrease in responsivity at a photocurrent of 3 mA. Application of a reverse bias will extend the linear range of the UVG photodiodes when measuring UV radiation.

Fig. 1: Linearity of UVG-20 photodiode and a widely used P-on-N photodiode when tested at 430 nm with no reverse bias.

It is believed that the difference in the series resistance of these diodes can explain the large difference in linear range.

Figure 2 shows the linearity of a UVG-20 diode and a P-on-N photodiode with an internal quantum efficiency of 80% at 430 nm. This 80% IQE is a result of photogenerated carrier recombination in the front region. The supralinearity (increased responsivity) in the low IQE device is caused by the filling of trap centers with increasing flux [1]. As the trap centers are filled the minority carrier lifetime increases reducing the photogenerated carrier recombination resulting in increased responsivity. The UVG series photodiodes have 100% internal quantum efficiency (no photogenerated carrier recombination) at 430 nm so they show no supralinearity. Because it is difficult to correct for non-linearity errors, high accuracy applications require linear photodiodes like the UVG series diodes.

Fig. 2: Linearity of P-on-N photodiode with 80% internal quantum efficiency and UVG-20 photodiode with 100% internal quantum efficiency.

Figure 3 compares the linearity of the UVG-20 photodiode and a widely used P on N photodiode with equivalent area when exposed to increasing levels of 980 nm radiation. The P on N photodiode showed noticeable supralinear behavior for photocurrents above 20 µA while no noticeable supralinearity was observed in the UVG-20 diode. At high irradiance levels, the P-on-N photodiode was found to lose responsivity much more rapidly than the UVG series diode.

As the UVG series diode internal quantum efficiency drops rapidly after 700 nm owing to the limited silicon thickness, IRD also provides P-on-N photodiodes (for example: UVG-PN100, UVG-PN20, etc.) which have shown over 97% internal quantum efficiency at 950 nm.

Fig. 3: Linearity of UVG-20 photodiode and widely used P-on-N photodiode when tested at 980 nm with no reverse bias

Interestingly, the linearity of the IRD P-on-N diodes (IRD model # UVG-PN20) was exactly the same as that of the standard UVG-20 diode depicted in Figure 3. Also, the internal quantum efficiencies of the IRD P-on-N and the widely used P-on-N photodiodes are nearly equivalent at 980 nm indicating they will have the same minority carrier lifetime. This suggests that the minority carrier lifetime (diffusion length) is not the only factor which determines the supralinearity as was previously believed [1].

The substrate doping concentration of the UVG-PN20 diode is 2 x 10¹³/cm³ and that of the widely used P-on-N diode is 5 x 10¹²/cm³. Computer modeling has shown that this difference in the starting materials can qualitatively explain why these photodiodes exhibit such different levels of supralinearity.

Figure 4 shows the structure of the UVG-PN20 photodiode that was investigated.

Fig. 4: Structure of the UVG-PN20 photodiode

References:

1] A.R. Schaefer, E.F. Zalewski, and Jon Geist "Silicon detector nonlinearity and related effects"

Applied Optics, Vol. 22, 1232-1235 (1983)