|
|
| |
|
| ||
|
HIGH PERFORMANCE NEAR-INFRARED SEMICONDUCTOR PHOTODETECTORS and LASERS InP-BASED LOW DARK-CURRENT 1.55μm PHOTODIODES The evolution of long-haul optical fiber communications systems has created the need for high-sensitivity, low-noise, high quantum efficiency, low-cost photodetectors. We design, fabricate, and characterize InP/InAlAs/InGaAs based p-i-n type low-noise photodiodes. The thickness of the active layer is increased to 15000 Å to decrease the dark current and to increase the quantum efficiency. On the surface, a thin InGaAs layer is employed for p-ohmic contact formation. In the simulations, the quantum efficiency of the designed structure is calculated as 70%, and the corresponding responsivity is 0.88 A/W at 1550 nm wavelength. Since the fabrication steps on an N-type substrate are different than previously performed fabrications on semi-isolating substrates, a new photomask is designed. Figure 1 shows a view of the designed mask. The samples are fabricated by using an eight-step microwave- _compatible fabrication process in a class-100 clean room environment. First, the Ti/Au (100 Å /1000Å) p-ohmic contacts are deposited on the _surface. The next step is the definition of circularly shaped p-mesa regions which defines the detector active area around the p-ohmic contacts by means of a wet etch. This step is followed by a similar yet larger mesa etch that mainly serves to recess the sample surface within a few thousand Å of the n-contact layer. N-type contact areas are first etched to the N+ InP layer, and afterwards a second mask layer is used to deposit Au/Ge/Ni/Au contact metallization. Device mesas are etched away around the n-contacts. The contacts are annealed at 400oC for 60 s in a rapid thermal annealing system. Then, a 200 nm thick Si3N4 is deposited via plasma enhanced chemical vapour deposition (PECVD) for passivation and isolation. The thickness of this layer is also chosen to form an anti-reflection coating. Finally, a ~ 0.6 μm thick Ti/Au interconnect metal is deposited and lifted off to connect the device contacts to the coplanar waveguide transmission line pads. Figure 2 shows the schematic cross-section of a fabricated photodetector.
Fig.1 Photodetector Mask
Fig.2 Schematic cross-section of a fabricated photodiode.
Fig. 3 (a) Photomicrograph showing a part of fabricated sample
Fig. 3 (b) Photomicrograph showing 50 µm active area device. Figure 3 (a) and (b) show photomicrographs of photodetectors after the fabrication. Figure 3 (a) is the general view of a few photodetectors, whereas Fig. 3 (b) is the photomicrograph of a 50 µm active area device. After the completion of fabrication, the devices are tested for dark current and photoresponse. First, dark current is measured on a probe station by means of a semiconductor parametric analyzer. The devices had breakdown voltages higher than 50V. Figure 4 shows the dark I-V graph obtained from a 100 mm diameter device. Up to 10 V reverse bias, the dark current was less than 7nA. For photoresponse measurements, 240 mm diameter photodiodes are tested by using a tunable laser tuned to 1550 nm wavelength. Photodiodes were illuminated by the laser light with the help of a fiber probe on a probe station. Dark current and photocurrent are measured in the 0 to -5V interval for different incident optic power values which was calibrated by a calibrated powermeter. The results of these measurements are shown in Fig. 5. Photocurrent increases linearly with increasing incident optic power and a constant responsivity of 0.48 A/W is obtained for reverse bias values larger than 0.8 V. This responsivity value corresponds to a quantum efficiency of %38.
Fig. 4 Dark current of a 100 μm diameter device.
Fig. 5 Photoresponse of a 240 μm diameter device.
|
||
|
||
|
© Copyright - com.web.tr. This page is best viewed with IE v.5.0 or higher, at minimum 1024x768 resolution. |
||