The Effect of Depletion Layer Thickness in Avalanche Photodiodes for Measurement of Low-energy Electrons

We have tested APDs (Type spl 3989 and Z7966, Hamamatsu Photonics K.K.) using an electron beam. The Z7966, which has a depletion layer of 10μm, is firstly focused on and tested in our former paper [K. Ogasawara, K. Asamura, T. Mukai, Y. Saito, Nucl. Instr. and Meth. A 545(3) (2005) 744] for the energy range of 5-20 keV. The result shows that the pulse height distribution of the APD signal exhibits a significant peak for electrons with energies above 8 keV, and positions of their peaks show a good linearity. The condition of the peak production at energies below 8 keV was attributed to the thickness of the dead layer on the surface of APDs. Now we have tuned up our electron acceleration system up to 40 keV, and tested Z7966 by electrons of higher energies. The result shows that the output pulse height distributions of this Z7966 were distorted over 30 keV. In order to examine the distortion of pulse height distributions, we have made a Monte Carlo numerical simulation of particle transport inside the APD. The result shows that the highest energy limit is expected to be determined by the thickness of the depletion layer inside the APD. Therefore, we have tried an APD type spl 3989, which has a thicker depletion layer (30μm) and a thinner dead layer. As is expected, the spl 3989 responded to 2-40 keV electrons with fine peaks in the output pulse height distributions. The energy resolution was lower than 1 keV for 2-20 keV electrons, and 5 keV for 40 keV electrons. The linearity of the response was also good. According to the Monte Carlo simulation, electrons up to about 60 keV are expected to be well detectable.