The dark current of a transferred-electron photocathode with an InGaAs absorber, responsive over the 0.9-to-1.7-μm range, must be reduced to an ultralow level suitable for low signal spectral astrophysical measurements by lowering the temperature of the sensor incorporating the cathode. However, photocathode quantum efficiency (QE) is known to reduce to zero at such low temperatures. Moreover, it has not been demonstrated that the target dark current can be reached at any temperature using existing photocathodes.
Changes in the transferred-electron photocathode epistructure (with an InGaAs absorber lattice-matched to InP and exhibiting responsivity over the 0.9-to-1.7-μm range) and fabrication processes were developed and implemented that resulted in a demonstrated >13× reduction in dark current at –40 ºC while retaining >95% of the ≈25% saturated room-temperature QE. Further testing at lower temperature is needed to confirm a >25 ºC predicted reduction in cooling required to achieve an ultralow dark-current target suitable for faint spectral astronomical observations that are not otherwise possible. This reduction in dark current makes it possible to increase the integration time of the imaging sensor, thus enabling a much higher nearinfrared (NIR) sensitivity than is possible with current technology. As a result, extremely faint phenomena and NIR signals emitted from distant celestial objects can be now observed and imaged (such as the dynamics of red-shifting galaxies, and spectral measurements on extra-solar planets in search of water and bio-markers) that were not previously possible. In addition, the enhanced NIR sensitivity also directly benefits other NIR imaging applications, including drug and bomb detection, stand-off detection of improvised explosive devices (IED’s), Raman spectroscopy and microscopy for life/physical science applications, and semiconductor product defect detection.
A set of methods was developed for implementing an InGaAs photocathode whereby the dark current can be reduced by lowering the temperature to the ultralow target level, while at the same time, exhibiting QE that is high enough to perform the astrophysical measurements.
This innovation features a thin, n-type InP cap layer that is etched during final cleaning between the grid lines. Along with an n-type InP layer at the heterointerface and a p-type InP emitting surface layer, the extra degree-of-freedom provided by the n-type InP cap layer enables independent tailoring of the electric field at 3 key locations in the device: beneath the grid lines, at the emitting InP surface between grid lines, and at the p-type InGaAs absorber/n-type InP heterointerface. This enables minimization of the field beneath the grid lines while the emitting surface and heterointerface fields are balanced such that the onset of high escape probability and turn-on completion of the heterointerface occur at the same reduced device bias. The resulting effect is that dark current components are minimized, including those due to undue extension of the depletion region into the low bandgap absorber and premature emitting surface field development with bias, while maintaining high QE and minimal grid line leakage.
The innovation features an InP:Zn emitting surface layer doped below the onset of Zn diffusion (thus minimizing epitaxy and process variability), absence of an undoped InP drift layer (along with the avalanche-current-inducing voltage drop across it), and an InGaAsP step grade layer introduced at the InGaAs absorber/InP:Si layer heterointerface (further reducing dark current components associated with the depleted low bandgap absorber). Employment of a SiON dielectric beneath the grid line promotes device stability and the absence of fixed mobile charge in the metal/dielectric/InP stack.
This work was done by Michael Jurkovic of Intevac Photonics for Goddard Space Flight Center. GSC-16044-1\