The APD used was a Goodrich SU055-GM-APD-FO. Using a thermo-electric cooler the APD was cooled to a temperature of approximatly -60C. The temperature controller that was to be used to stabalize the temperature did not work, so the current to the TEC had to be controlled by hand. This made the temperature somewhat unstable, but with temperature fluctuations less than 3C.
The figure below shows the electronic schematic of the APD circuit. I used a volt meter with an internal resistance of ~10M Ohm to measure the voltage drop across a 68 Ohm resistor, and from this derived the current.
The next figure shows a plot of the output current of the APD vs. bias voltage for several intensities of light of wavelength 1324nm. It also shows how sensitive the dark current is to temperature fluctuations. I measured the dark current three times with the APD cooled to between -60C and -63C. The small change in temperature produces a noticeable change in the magnitude of the dark current.
The next two figures shows the responsivity of the APD as a function of bias voltage. The responsivity was obtained by dividing the output current by the intensity of light from the above figure.
The smaller intensities of light have a large error due to the small current produced relative to the dark current. If the temperature of the APD were more stable this uncertainty could be greatly reduced.
The APD is meant to be run near 90%-99% of the break down voltage, which corresponds to a reverse bias of between 45 and 49.5 volts. At higher bias voltages the APD has a large gain, with a responsivity of near 20 at v_bias = 49.5. The error due to the dark current and temperature fluctuations make the signal to noise ratio very poor, especially for low light intensities and high reverse biases.
I replaced the volt-meter with a oscilloscope of internal resistance of ~1M Ohm. The oscilloscope was unable to measure to the small voltages produced by the APD for the intensities of light used in the previous setup (<200pW). In order to get a signal on the oscilloscope I needed to use an intensity of light that saturated the callibration detector.
The 1324nm laser was pulsed for 1u second, with a rise time of approximatly 5ns. The following figure shows the signal produced by the APD for 4 different reverse bias voltages. At bias voltages of 47 volts or less the output current is approximatly constant. At higher bias voltages the output current seems to decay with time. I am not sure what causes this, but I suspect it is due to some slight capacitance possesed by the APD.
A small bounce-back effect can be seen after the laser is puled off. This seems to be a characteristic that shows up in other photodiodes as well when they are being used with a large intensity of light.
I believe that if lower intensities of light are used the decay in voltage with time and bounce back effects would diminish. However, this would mean that the signal from the APD would have to be amplified in order to be seen on the oscilloscope.
The APD seems fast enough to see the 5ns rise time of the laser. The fall time of the APD is not as fast as the rise time, about 7ns.
