News
Structure of InGaAs photodetector
2024-10-29 10:06:40 | Company News Page views：47
Structure of InGaAs photodetector
Since the 1980s, researchers at home and abroad have studied the structure of InGaAs photodetectors, which are mainly divided into three types. They are InGaAs metal-Semiconductor-metal photodetector (MSM-PD), InGaAs PIN Photodetector (PIN-PD), and InGaAs Avalanche Photodetector (APD-PD). There are significant differences in the fabrication process and cost of InGaAs photodetectors with different structures, and there are also great differences in device performance.
The InGaAs metal-semiconductor-metal photodetector, shown in Figure (a), is a special structure based on the Schottky junction. In 1992, Shi et al. used low pressure metal-organic vapor phase epitaxy technology (LP-MOVPE) to grow epitaxy layers and prepared InGaAs MSM photodetector, which has A high responsiveness of 0.42 A/W at a wavelength of 1.3 μm and a dark current lower than 5.6 pA/ μm² at 1.5 V. In 1996, zhang et al. used gas phase molecular beam epitaxy (GSMBE) to grow the InAlAs-InGaAs-InP epitaxy layer. The InAlAs layer showed high resistivity characteristics, and the growth conditions were optimized by X-ray diffraction measurement, so that the lattice mismatch between InGaAs and InAlAs layers was within the range of 1×10⁻³. This results in optimized device performance with dark current below 0.75 pA/μm² at 10 V and fast transient response up to 16 ps at 5 V. On the whole, the MSM structure photodetector is simple and easy to integrate, showing low dark current (pA order), but the metal electrode will reduce the effective light absorption area of the device, so the response is lower than other structures.
The InGaAs PIN photodetector inserts an intrinsic layer between the P-type contact layer and the N-type contact layer, as shown in Figure (b), which increases the width of the depletion region, thus radiating more electron-hole pairs and forming a larger photocurrent, so it has excellent electron conduction performance. In 2007, A.Poloczek et al. used MBE to grow a low-temperature buffer layer to improve the surface roughness and overcome the lattice mismatch between Si and InP. MOCVD was used to integrate InGaAs PIN structure on the InP substrate, and the responsiveness of the device was about 0.57A /W. In 2011, the Army Research Laboratory (ALR) used PIN photodetectors to study a liDAR imager for navigation, obstacle/collision avoidance, and short-range target detection/identification for small unmanned ground vehicles, integrated with a low-cost microwave amplifier chip that significantly improved the signal-to-noise ratio of the InGaAs PIN photodetector. On this basis, in 2012, ALR used this liDAR imager for robots, with a detection range of more than 50 m and a resolution of 256 × 128.
The InGaAs avalanche photodetector is a kind of photodetector with gain, the structure of which is shown in Figure (c). The electron-hole pair obtains enough energy under the action of the electric field inside the doubling region, so as to collide with the atom, generate new electron-hole pairs, form an avalanche effect, and multiply the non-equilibrium carriers in the material. In 2013, George M used MBE to grow lattice matched InGaAs and InAlAs alloys on an InP substrate, using changes in alloy composition, epitaxial layer thickness, and doping to modulated carrier energy to maximize electroshock ionization while minimizing hole ionization. At the equivalent output signal gain, APD shows lower noise and lower dark current. In 2016, Sun Jianfeng et al. built a set of 1570 nm laser active imaging experimental platform based on the InGaAs avalanche photodetector. The internal circuit of APD photodetector received echoes and directly output digital signals, making the whole device compact. The experimental results are shown in FIG. (d) and (e). Figure (d) is a physical photo of the imaging target, and Figure (e) is a three-dimensional distance image. It can be clearly seen that the window area of area c has a certain depth distance with area A and b. The platform realizes pulse width less than 10 ns, single pulse energy (1 ~ 3) mJ adjustable, receiving lens field Angle of 2°, repetition frequency of 1 kHz, detector duty ratio of about 60%. Thanks to APD’s internal photocurrent gain, fast response, compact size, durability and low cost, APD photodetectors can be an order of magnitude higher in detection rate than PIN photodetectors, so the current mainstream liDAR is mainly dominated by avalanche photodetectors.
Since the 1980s, researchers at home and abroad have studied the structure of InGaAs photodetectors, which are mainly divided into three types. They are InGaAs metal-Semiconductor-metal photodetector (MSM-PD), InGaAs PIN Photodetector (PIN-PD), and InGaAs Avalanche Photodetector (APD-PD). There are significant differences in the fabrication process and cost of InGaAs photodetectors with different structures, and there are also great differences in device performance.
The InGaAs metal-semiconductor-metal photodetector, shown in Figure (a), is a special structure based on the Schottky junction. In 1992, Shi et al. used low pressure metal-organic vapor phase epitaxy technology (LP-MOVPE) to grow epitaxy layers and prepared InGaAs MSM photodetector, which has A high responsiveness of 0.42 A/W at a wavelength of 1.3 μm and a dark current lower than 5.6 pA/ μm² at 1.5 V. In 1996, zhang et al. used gas phase molecular beam epitaxy (GSMBE) to grow the InAlAs-InGaAs-InP epitaxy layer. The InAlAs layer showed high resistivity characteristics, and the growth conditions were optimized by X-ray diffraction measurement, so that the lattice mismatch between InGaAs and InAlAs layers was within the range of 1×10⁻³. This results in optimized device performance with dark current below 0.75 pA/μm² at 10 V and fast transient response up to 16 ps at 5 V. On the whole, the MSM structure photodetector is simple and easy to integrate, showing low dark current (pA order), but the metal electrode will reduce the effective light absorption area of the device, so the response is lower than other structures.
The InGaAs PIN photodetector inserts an intrinsic layer between the P-type contact layer and the N-type contact layer, as shown in Figure (b), which increases the width of the depletion region, thus radiating more electron-hole pairs and forming a larger photocurrent, so it has excellent electron conduction performance. In 2007, A.Poloczek et al. used MBE to grow a low-temperature buffer layer to improve the surface roughness and overcome the lattice mismatch between Si and InP. MOCVD was used to integrate InGaAs PIN structure on the InP substrate, and the responsiveness of the device was about 0.57A /W. In 2011, the Army Research Laboratory (ALR) used PIN photodetectors to study a liDAR imager for navigation, obstacle/collision avoidance, and short-range target detection/identification for small unmanned ground vehicles, integrated with a low-cost microwave amplifier chip that significantly improved the signal-to-noise ratio of the InGaAs PIN photodetector. On this basis, in 2012, ALR used this liDAR imager for robots, with a detection range of more than 50 m and a resolution of 256 × 128.
The InGaAs avalanche photodetector is a kind of photodetector with gain, the structure of which is shown in Figure (c). The electron-hole pair obtains enough energy under the action of the electric field inside the doubling region, so as to collide with the atom, generate new electron-hole pairs, form an avalanche effect, and multiply the non-equilibrium carriers in the material. In 2013, George M used MBE to grow lattice matched InGaAs and InAlAs alloys on an InP substrate, using changes in alloy composition, epitaxial layer thickness, and doping to modulated carrier energy to maximize electroshock ionization while minimizing hole ionization. At the equivalent output signal gain, APD shows lower noise and lower dark current. In 2016, Sun Jianfeng et al. built a set of 1570 nm laser active imaging experimental platform based on the InGaAs avalanche photodetector. The internal circuit of APD photodetector received echoes and directly output digital signals, making the whole device compact. The experimental results are shown in FIG. (d) and (e). Figure (d) is a physical photo of the imaging target, and Figure (e) is a three-dimensional distance image. It can be clearly seen that the window area of area c has a certain depth distance with area A and b. The platform realizes pulse width less than 10 ns, single pulse energy (1 ~ 3) mJ adjustable, receiving lens field Angle of 2°, repetition frequency of 1 kHz, detector duty ratio of about 60%. Thanks to APD’s internal photocurrent gain, fast response, compact size, durability and low cost, APD photodetectors can be an order of magnitude higher in detection rate than PIN photodetectors, so the current mainstream liDAR is mainly dominated by avalanche photodetectors.
Overall, with the rapid development of InGaAs preparation technology at home and abroad, we can skillfully use MBE, MOCVD, LPE and other technologies to prepare large-area high-quality InGaAs epitaxial layer on InP substrate. InGaAs photodetectors exhibit low dark current and high responsiveness, the lowest dark current is lower than 0.75 pA/μm², the maximum responsiveness is up to 0.57 A/W, and has a fast transient response (ps order). The future development of InGaAs photodetectors will focus on the following two aspects: (1) InGaAs epitaxial layer is directly grown on Si substrate. At present, most of the microelectronic devices in the market are Si based, and the subsequent integrated development of InGaAs and Si based is the general trend. Solving problems such as lattice mismatch and thermal expansion coefficient difference is crucial for the study of InGaAs/Si; (2) The 1550 nm wavelength technology has been mature, and the extended wavelength (2.0 ~ 2.5) μm is the future research direction. With the increase of In components, the lattice mismatch between InP substrate and InGaAs epitaxial layer will lead to more serious dislocation and defects, so it is necessary to optimize the device process parameters, reduce the lattice defects, and reduce the device dark current.
