Open Access
Add to BibSonomy
Abstract
Germanium (Ge) lateral p-i-n photodetectors with grating and hole-array structures were fabricated on a Ge-on-insulator (GOI) platform. Owing to the low threading dislocation density (TDD) in the transferred Ge layer, a low dark current of 0.279 µA was achieved at −1 V. The grating structure enhances the optical absorption by guiding the lateral propagation of normal incident light, contributing to a 3× improved responsivity at 1,550 nm. Compared with the grating structure, the hole-array structure not only guides the lateral modes but also benefits the vertical resonance modes. A 4.5× higher responsivity of 0.188 A/W at 1,550 nm was achieved on the 260 nm Ge absorptive layer. In addition, both the grating and the hole-array structure attribute to a 2× and a 1.6× enhanced 3dB bandwidth at −5 V due to significantly reduced capacitance. The planar configuration of p-i-n photodiodes is favorable for large-scale monolithic integration. The incorporated surface structures offer promising approaches to reinforce the responsivity and bandwidth simultaneously, paving the way for the development of high-performance Ge photodetectors on silicon substrate.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
As a group IV material, Ge has been widely investigated for its application in electronics and photonics. The investigation of Ge starts from the advent of the first transistor. Owing to its high carrier mobility and low effective mass, Ge has been manifested promising for high-speed metal-oxide-semiconductor field-effect transistors (MOSFETs) [1], [2]. Although Ge is an indirect bandgap material, the energy difference between two conduction band minimums (L valley and Γ valley) is only 136 meV. The incorporation of ∼2% tensile strain or high n-type doping of ∼3×1019/cm3 into Ge could effectively compensate for the energy difference between L valley and Γ valley, turning Ge into a direct bandgap material. The bandgap transition endorses Ge the feasibility as semiconductor lasers operating at short-wave infrared (SWIR) region [3–5]. In addition, the remarkable Franz-Keldysh (FK) effect has been discovered in Ge p-i-n diodes, enabling it another application as electro-absorption (EA) modulators [6–8]. Moreover, the direct bandgap of Ge is ∼0.8 eV at room temperature. Owing to the considerable absorption coefficient from 1,200 to 1,600 nm, Ge photodetectors are capable of photodetection from O-band to L-band, covering the two paramount wavelengths of 1,310 and 1,550 nm in the telecommunication system [9–13].
Compared with III-V InGaAs photodetectors operating at the same wavelength range [14–16], Ge photodetectors have a significant advantage in the cost due to the complementary metal-oxide-semiconductor (CMOS) compatibility. Such superiority of Ge photodetectors is more prominent in the large-scale integration of photodetector arrays. It is convenient for the flexible integration of photodetector arrays if light enters the photodetectors through the surface. Among the surface-illuminated photodetectors, vertical p-i-n configuration, which normally requires in-situ doping, is dominant in the active research. Lateral p-i-n doping offers an alternative solution for the formation of photodiode [17–20]. The planar configuration which involves the ion implantation process is also more favorable in large-scale monolithic integration. More importantly, high-efficiency and high-speed photodetection can be achieved simultaneously in the lateral p-i-n photodetectors. For the vertical p-i-n photodetectors, there is a trade-off between responsivity and bandwidth since the absorption of photons and the collection of carriers are in the same direction. A thick intrinsic layer contributes to the photon absorption while impeding the transit time for the photon-generated carriers. However, the lateral p-i-n configuration avoids the trade-off by separating different paths for photon absorption and carrier collection. The photon absorption in the vertical direction will not interfere with the carrier collection in the lateral direction. The responsivity of lateral p-i-n photodetectors is usually limited by external quantum efficiency. For Ge photodetectors, the reflectance on the top surface is ∼25% for the normal incident light at 1,550 nm. To enhance the responsivity, miscellaneous structures have been incorporated into photodetectors [21], [22]. Resonant-cavity-enhanced structure [23], [24], photon-trapping structure [25–27], and plasmonic structure [28–30] have been demonstrated effective in the reinforcement of optical response.
In this work, grating and hole-array enhanced Ge lateral p-i-n photodetectors were fabricated on an insulator platform. After the layer transfer process, the defective Ge epitaxy layer near the Ge/Si interface was etched away, contributing to a low dark current of 0.279 µA at −1 V. The incorporation of grating structure benefits a 3× improved responsivity of 0.13 A/W at 1,550 nm. The polarization-dependent behavior was also investigated for the grating structure. In terms of the hole-array structure, a 4.5× higher responsivity of 0.188 A/W was achieved at 1,550 nm. Considering the thickness of the Ge absorptive layer is 260 nm, such responsivity is relatively high. The absorption enhancement for the two surface structures was investigated using the finite-difference time-domain (FDTD) method. Both the surface structures show a positive influence on the frequency response. The improvement in 3dB bandwidth for the grating and the hole-array structures are 2.1× and 1.6×, respectively. A 3dB bandwidth of 560 MHz was achieved on the photodetector with grating structure at −5 V.
2. Material characterization and device fabrication
The GOI platform was realized through the direct wafer bonding (DWB) process [31], [32]. Firstly, the Ge film was grown on a 6” Si (100) wafer using reduced pressure chemical vapor deposition (RPCVD). Subsequently, a SiO2 layer was deposited on the Ge/Si donor wafer and Si handle wafer respectively by plasma enhanced chemical vapor deposition (PECVD). It should be noted that an additional SiN layer was also deposited on the donor wafer to suppress the void at the bonding interface [33]. Chemical mechanical polishing (CMP) and annealing in N2 ambient were performed on the Ge/Si donor wafer for densification and planarization of the bonding surface to achieve a root mean square (RMS) surface roughness of < 0.5 nm. Both the wafers were exposed to O2 plasma for 15 s before they were brought into contact and bonded at room temperature. Post bonding annealing was then carried out at 300 °C in N2 ambient for 4 hours to enhance the bonding strength. The Si donor layer was removed away by wet etching in potassium hydroxide (KOH) solution at 80 °C. It is well-known that the threading dislocations mainly distribute near the Ge/Si interface due to the lattice mismatch. The defective Ge layer was removed by the CMP process. The thickness and optical constants of the GOI sample were measured using the ellipsometer. As shown in Fig. 1, the refractive index n and extinction coefficient k of the Ge layer was plotted from 1,000 to 2,000 nm wavelength. The inset illustrates the thickness of each layer for the GOI platform. The remaining Ge layer is ∼260 nm and the extracted absorption coefficient of Ge at 1,550 nm is ∼3242 cm−1.
Fig. 1. The refractive index and extinction coefficient of the Ge layer were obtained through ellipsometer fitting. The inset shows the structure of the GOI sample.
Lateral p-i-n Ge photodetectors with grating and hole-array structures were fabricated based on the GOI platform. The schematics of the proposed photodetectors are shown in Fig. 2(a) and 2(b). The red and blue regions represent P-type and N-type doped Ge respectively. It should be noted that part of the doped regions is covered by the metal electrodes. The interdigitated doping profile rather than the interdigitated electrodes configuration benefits a large illuminated-surface, as well as the convenient incorporation of surface structures. The pitch direction of the grating structure is parallel to the interdigitated doped regions. Meanwhile, the square holes of the hole-array structure are uniformly distributed on the surface regardless of the doping profile.
Fig. 2. (a) Schematic of Ge lateral p-i-n photodetector. (b) Schematic of Ge photodetector with grating structure. (c) Schematic of Ge photodetector with hole-array structure. (d) The key fabrication process for the Ge photodetectors with surface structures.
The key fabrication processes of the photodetectors are summarized in Fig. 2(c). It is noteworthy that the manufacture of the surface structures is the first step, prior to the formation of lateral p-i-n doped configuration. The reason is related to the shallow penetration depth of dopants, which could be etched away in the fabrication of the surface structures. Electron beam lithography (EBL) and reactive ion etching (RIE) were used in the patterning and etching of grating and hole-array structures. The implant dose for boron and phosphorus ion implants are 1 ×1015 and 1 ×1015 cm−2, and the dose energies are 40 and 50 keV, respectively. The doping profile before thermal annealing was simulated using Monte Carlo method. Relatively high doping concentration of > 1×1018/cm3 was achieved on top 140 nm Ge layer for both dopants. After the formation of the doping profile, a 100 nm SiO2 layer was deposited on the sample as the passivation layer. Subsequently, the contact windows were opened on the passivation layer to expose the doped region. Finally, the metal electrodes of Ti/TiN/Al were formed through the lift-off process. Rapid thermal annealing (RTA) was carried out at 450 °C in N2 ambient. The RTA process not only activates the dopants but also enhances the Ohmic contact between Ge and metal electrodes.
The scanning electron microscope (SEM) images of the fabricated photodetectors are shown in Fig. 3. The illuminated surface has an area of 70 µm × 60 µm and the metal electrodes have a ground-signal-ground (GSG) configuration for the measurement of bandwidth. For the interdigital doping profile circled in Fig. 3(a), the distance between adjacent P-type and N-type Ge regions is fixed at 2 µm. Figure 3(b) and 3(c) illustrate the zoomed-in views of grating and hole-array structures, respectively. The surface structures are uniform and well-defined. All the photodetectors were fabricated on the same sample. Since the micro loading effect in the dry etch is negligible, these two structures are assumed to have the same etching depth of ∼170 nm.
Fig. 3. (a) Top-view SEM image of Ge lateral p-i-n photodetector with surface structure. (b) Zoomed-in view of the grating structure. (c) Zoomed-in view of the square hole-array structure.
3. Design and simulation
The parameters for the surface structures were simulated using the FDTD method. The n, k parameters for each layer were extracted from ellipsometer fitting. A plane-wave light source with TE polarization direction was adopted in the simulation. Here, the TE polarization is defined as the electric field direction of the incident light which is perpendicular to the bridge direction of the grating structure. It should be noted that the propagation direction of the incident light is fixed to be normal to the photodetector surface. The smallest repeating structure for the grating and hole-array was treated as the unit cell. According to the polarization direction, symmetric and anti-symmetric boundary conditions were employed on the unit cell to reduce the calculation amount. The reflection from the bottom of the Si substrate was ignored. Perfect matching layer boundary condition was adopted in the propagation direction of the incident light. In order to obtain the optical absorption in the Ge layer, two monitors were placed above the optical source and below the Ge layer respectively. After collecting the output power from the monitors, the absorption A in Ge can be simply calculated as A = 1-R-T, where R is the reflectance and T is the transmittance. Additional top-view and cross-section monitors were placed inside the Ge absorptive layer to obtain the electric field intensity distribution.
For the grating structure, pitch size, duty cycle, and etch depth were optimized separately using parameter sweep to maximize the optical absorption at 1,550 nm. The optimal parameters for pitch size, duty cycle and etch depth are 0.54 µm, 0.5 and 168 nm, respectively. Meanwhile, the corresponding electric field density distribution at 1,550 nm are shown in Fig. 4(a) and 4(c). The color bar stands for the electric field density with a unit of V/m. The strong electric field distributed on two sides of the trench manifests the lateral propagation of normal incident light in the grating structure. The grating structure is well-known as the surface coupler, especially for the thin layer. As mentioned before, the absorptive Ge layer has a thickness of 260 nm and an area of 70 µm × 60 µm. The lateral propagation attributes to adequate absorption, reinforcing the responsivity accordingly. In addition, the grating structure helps reduce the reflectance since the etched parts result in a smaller effective refractive index for the Ge layer. The absorption spectrum for the grating-enhanced photodetectors is shown in Fig. 5. Compared with photodetector without surface structure, the absorption is significantly enhanced across the entire wavelength range. The two absorption curves have a similar trend which decreases as the wavelength increases. The absorption reinforcement at 1,550 nm is ∼ 4×, demonstrating the effectiveness of the grating structure.
In terms of the hole-array structure, the parameters were optimized similarly using parameter sweep for the highest absorption in Ge layer. For the perspective wavelength of 1,550 nm, the optimized side length of the square hole, periodicity of the array and etch depth are 0.54 µm, 1 µm, and 160 nm, respectively. As shown in Fig. 4(b) and 4(d), the corresponding electric field intensity distributions were simulated at 1,550 nm. The top-view image shows similar lateral propagation modes in the Ge layer. Different from the grating structure, the lateral propagation occurs in two directions due to the square hole-array configuration. The optical modes along the x and y directions are different because the light source has a TE polarization direction. Besides the lateral modes, the resonance in the vertical direction can be observed in the cross-sectional figure. The remaining Ge layer is sandwiched by the air and the insulator layer, forming a cavity for the normal incident light. By optimizing the structures, significant resonance can be achieved at 1,550 nm. The lateral propagation and vertical resonance collectively contribute to the enhanced absorption in the photodetector. The absorption spectrum in Fig. 5 shows significant enhancement compared with the reference photodetector. Moreover, the absorption curve has a remarkable peak at 1,550 nm. The peak is assumed to be related to the resonance which is selective to wavelength. The absorption spectrum can be divided into two parts: lateral absorption and resonance absorption. The former is assumed to have a similar trend with the grating structure, while the latter shows prominently wavelength-dependent behavior. The collective reinforcement of the hole-array structure leads to a ∼ 7.4× higher absorption at 1,550 nm, which is superior to the grating structure.
Fig. 4. Cross-section view of simulated electric field density distribution in Ge photodetectors with (a) grating structure and (b) hole-array structure. Top view of simulated electric field density distribution in Ge photodetectors with (c) grating structure and (d) hole-array structure.
Fig. 5. Simulated absorption spectra for the Ge lateral p-i-n photodetectors from 1,500 to 1,600 nm.
4. Characterization and discussion
The current-voltage (I-V) characteristics of the fabricated photodetectors were measured in Fig. 6(a). The high-power laser (T100S-HP) used in the setup has a tunable wavelength ranging from 1,500 to 1,630 nm. In terms of the dark current, the reference photodetector without surface structure has a low dark current of 0.279 µA at −1 V. The on-off ratio at ±1 V is ∼1.2×104, demonstrating the excellent rectifying behavior realized by the ion implantation process. Since the lateral p-i-n photodetector has an interdigitated doping profile, it can be equivalently treated as several smaller p-i-n structures which are connected with each other. In addition, the sidewall area (60 µm × 260 nm) is equivalent to the surface area of the vertical p-i-n photodetectors. Therefore, the corresponding dark current density at −1 V is 17.9 mA/cm2. As mentioned above, the defective Ge layer near the interface has been removed after the DWB process. The TDD in the GOI platform is ∼1×106 cm−2 according to our published works [10]. The main source of the dark current could be attributed to surface leakage. The lateral p-i-n configuration leads to a ∼2 orders larger equivalent sidewall area compared with the vertical p-i-n configuration, rendering surface passivation paramount. The simple passivation using 100 nm SiO2 layer in this work is not sufficiently effective. In the future, advanced Al2O3/GeOx/SiO2 passivation will be adopted to suppress the surface leakage [10,34]. It can be observed that the photodetectors with grating and hole-array structures have higher dark currents of 0.62 and 0.54 µA at −1 V. The etched parts bring additional sidewall area, resulting in the increased dark current.
Fig. 6. (a) The I-V characteristics for the Ge lateral p-i-n photodetectors. The dashed lines represent dark currents, and the solid lines represent photocurrents at 0.9 mW. (b) The I-V characteristics for Ge photodetectors with the grating structure under TE and TM polarized light.
Before the photocurrent measurement, the propagation loss in the fiber and the coupling loss in the connectors were calibrated using the power meter (FPM-8220). The calibrated optical power is fixed at 0.9 mW and the operating wavelength is 1,550 nm. Compared with the illuminated area of the photodetector, the single mode fiber (Corning-SM28) used in the setup has a small core diameter of 8.2 µm. The slightly titled angle of the fiber has negligible influence on the measurement results. The fiber position and the coupling height were manually changed to maximize the output photocurrent. As shown in Fig. 6(a), the photocurrent curves show flat response under reverse bias, manifesting the effective collection of the photon-generated carriers. It also demonstrates the reliability of doping profile formed by ion implantation. In addition, the photocurrents for all the three photodetectors have saturated at −0.3 V. Such characteristic endorses the potential for low-power operation. As expected, the photodetectors with grating and hole-array structures have higher photocurrents at the same optical power. It is well-known that the grating structure is sensitive to the polarization direction. To obtain the polarization-dependent photocurrent for the photodetector with grating structure, a 3-paddle polarization controller was added in the optical path. The polarization direction can be manually changed by rotating the paddle. The optical loss brought by the polarization controller was also calibrated. As shown in Fig. 6(b), the photodetector with the grating structure shows a higher photocurrent under TE polarized illumination. However, the difference between TE and TM polarized illumination is not significant. It will not limit the practical application of the grating structure. In terms of the photodetector with the hole-array structure, the symmetry of the square configuration enables the same optical response under TE and TM polarized illumination.
The optical response for the three photodetectors was measured under varying optical power of 0.90, 4.31, and 8.55 mW, respectively. As shown in Fig. 7, the photocurrents and optical powers are linearly dependent. The extracted responsivity for the reference photodetector is 0.042 A/W at 1,550 nm. The low responsivity originates from the thin absorptive layer of 260 nm. As mentioned above, the measured absorption coefficient of the Ge layer is 3242 cm−1 at 1,550 nm. The reflectance can be simply estimated using the formula: R = (n1-n2)2/(n1 + n2)2, where n1 and n2 are refractive indexes of SiO2 and Ge respectively. Since the core layer of the single mode fiber is SiO2, the reflection at the interface between the fiber and the passivation layer is ignored. Therefore, the calculated reflectance R is ∼25% at 1,550 nm. The absorbance A can be expressed using the formula: A = (1-R)×(1-e-αd), where α is the absorption coefficient of Ge and d is the thickness of Ge. The theoretical absorbance in the Ge layer is ∼6%, explaining the low responsivity for the reference photodetector. The incorporation of the grating structure achieves a ∼3× higher responsivity of 0.130 A/W at 1,550 nm. The responsivities under TE and TM polarized illumination are 0.139 and 0.102 A/W, respectively. A higher responsivity of 0.188 A/W was achieved on the photodetector with hole-array structures. It has a ∼4.5× improvement compared with the reference photodetector. Compared with the responsivity of 0.39 A/W on the 850 nm Ge layer [9], the achieved responsivity of 0.188 A/W on the 260 nm Ge layer is relatively high. The noise equivalent power (NEP) for the photodetector was calculated using the formula: NEP = Irms/R, where Irms is the root-mean-square noise current and R is the responsivity. At 1,550 nm, the calculated NEPs for the Ge photodetector with grating and hole-array structure are 1.43×10−12 and 9.62×10−13 W/Hz1/2, respectively at −0.025 V. The corresponding specific detectivities are 4.52×109 and 6.74×109 cm·Hz1/2/W.
Fig. 7. The relationship between photocurrents and optical power for the Ge lateral p-i-n photodetectors. The responsivities were extracted from the slopes of the linear interpolation.
The responsivity spectra of the three photodetectors from 1,500 to 1,630 nm were shown in Fig. 8. For the reference photodetector, the responsivity decreases with the increasing wavelength as expected. The incorporation of the grating structure benefits an enhanced responsivity across the entire range. It should be noted that the responsivity spectrum has a small peak at 1,540 nm. The grating structure is designed for the optimized responsivity at 1,550 nm. However, the sidewall of the grating ridges is not perfectly straight, and the etching process is also difficult to precisely control. The deviations in duty cycle and etch depth attribute to the mismatch between simulation and experiment results. In terms of the hole-array structure, the responsivity enhancement at 1,550 nm is only ∼4.5, which is much lower than the simulated value of ∼8.4. It is expected the lateral and vertical modes in the photodetector collectively contribute to the enhanced absorption. Nevertheless, the deviation in etch depth weakens the resonance in the vertical direction since the resonance is sensitive to the thickness of the remaining Ge layer. Different from the prominent peak in the simulation curve, the responsivity peak at 1,550 nm is lower than the responsivity at 1,500 nm. It can be predicted that a higher responsivity peak will occur at a wavelength of < 1,500 nm. Due to the setup limit, the higher peak cannot be measured. Even though the experimental enhancement is lower than simulation results, both the grating and hole-array structures have been proved to be effective for the enhanced responsivity. Compared with the conventional fabrication process, only one additional lithography step is needed for the achievement of these structures, costing negligible expense in the practical manufacture.
Fig. 8. Responsivity spectra of Ge lateral p-i-n photodetectors from 1,500 to 1,630 nm. The bias voltage is fixed at −1 V.
The frequency response of the three photodetectors was measured using a lightwave component analyzer (Keysight N4373D) at 1,550 nm. As shown in Fig. 9(a), the 3 dB bandwidth of all photodetectors increases as the reverse bias increases. It is well-known that the 3 dB bandwidth is mainly limited by two factors: carrier transit time and resistance-capacitance (RC) delay. The increase of reverse bias voltage benefits the shorter carrier transit time. In this work, the distance between adjacent N-type and P-type Ge regions is fixed at 2 µm. By assuming the thickness of depletion region is equal to the width of intrinsic Ge region, the electric filed within the depletion region can be roughly calculated to be 25 kV/cm at −5 V. The electrons have reached the saturation velocity of 6×106 cm/s, while the velocity of holes is ∼2×106 cm/s [35]. The carrier-transit-limited bandwidth can be calculated using the formula: f = 0.45v/d, where v is the carrier velocity and d is the thickness of depletion region. The calculated carrier-transit-limited bandwidth is ∼ 4.5 GHz. Since the 3 dB bandwidth of the reference photodetector is 272 MHz at −5 V, it is mainly limited by RC delay which originates from the large surface of the Ge photodetector.
Fig. 9. (a) Frequency response of Ge lateral p-i-n photodetectors at −3 and −5 V. The inset table summarizes the 3dB bandwidth for varying photodetectors. (b) Schematic of reduced capacitance for Ge photodetectors with grating structure.
The incorporation of the surface structures contributes to a higher 3 dB bandwidth. The 3 dB bandwidth for the photodetectors with the grating and hole-array structures are 560 and 434 MHz at −5 V. The transit time is assumed to be the same for the three photodetectors and the enhanced bandwidth is mainly attributed to the reduced RC delay. For the interdigitated doping profile, the capacitance can be roughly estimated using the formula: C = (N−3)CI/2 + 2CICE/(CI+CE), where N is the number of electrodes, CI is the half of the capacitance of one interior electrode relative to the ground potential, and CE is the capacitance of one outer electrode relative to the ground plane [36–38]. The interior capacitor CI can be treated as a parallel plate capacitor and the capacitance is proportional to the surface area. As illustrated in Fig. 9(b), the incorporation of grating structure significantly reduces the surface area, leading to the reduced capacitance. The incorporation of hole-array structure contributes to a decreased capacitance in a similar way. For the uniform distributed hole-array, the interior capacitor CI can be divided into smaller parallel plate capacitors. The small capacitor which contains the hole structure has a decreased capacitance, rendering to a reduced total capacitance. Besides the junction capacitance, the series resistance related to the doping profile could also decreases for the photodetector with surface structures. As a result, the RC-delay-limited bandwidth was significantly enhanced for the Ge lateral p-i-n photodetectors with grating and hole-array structures. In the future, the distance between the adjacent doped regions will be decreased to increase the carrier-transit-limited bandwidth. Moreover, the RC delay limited bandwidth can be further improved by shrinking the size of the photodetector.
5. Conclusion
Ge lateral p-i-n photodetectors with grating and hole-array structures were fabricated on the advanced GOI platform. The grating structure benefits a 3× enhanced responsivity at 1,550 nm and 2× improved 3 dB bandwidth at −5 V. The grating structure changes the direction of normal incident light and guides the propagation in the lateral dimension. For the hole-array structure, 4.5× enhanced responsivity at 1,550 nm and 1.6× improved 3 dB at −5 V were achieved. The enhanced absorption for the hole-array structure is attributed to not only the lateral propagation modes but also the vertical resonance modes. The increased bandwidth for both the structures originates from the significantly reduced RC delay. This paper provides effective approaches to improve responsivity and bandwidth simultaneously. In the future, the optimization of the surface passivation, the thickness of the absorptive Ge layer, and the distance between the adjacent doped regions will improve the device performance further. It is desirable to achieve high-efficiency and high-speed Ge photodetectors for low-cost SWIR cameras and optical interconnects.
Funding
National Research Foundation Singapore (NRF–CRP19–2017–01); Ministry of Education - Singapore (T2EP50121-0001); Ministry of Education - Singapore (2021-T1-002-031 (RG112/21)).
Acknowledgements
The authors acknowledge Ms. Xiaohong Yang, Dr. Gang Yih Chong, and Ms. Ling Ling Ngo in Nanyang NanoFabrication Centre for the assistance in EBL, PECVD, and sputtering.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. M. L. Lee, E. A. Fitzgerald, M. T. Bulsara, M. T. Currie, and A. Lochtefeld, “Strained Si, SiGe, and Ge channels for high-mobility metal-oxide-semiconductor field-effect transistors,” J. Appl. Phys. 97(1) (2005).
2. M. J. H. van Dal, G. Vellianitis, G. Doornbos, C.-H. Hsieh, B.-H. Lee, K.-M. Yin, M. Passlack, and C. H. Diaz, “Germanium p-Channel FinFET Fabricated by Aspect Ratio Trapping,” IEEE Trans. Electron Devices 61(2), 430–436 (2014). [CrossRef]
3. S. Bao, D. Kim, C. Onwukaeme, S. Gupta, K. Saraswat, K. H. Lee, Y. Kim, D. Min, Y. Jung, H. Qiu, H. Wang, E. Fitzgerald, C. S. Tan, and D. Nam, “Low-threshold optically pumped lasing in highly strained germanium nanowires,” Nat Commun 8(1), 1–7 (2017). [CrossRef]
4. J. Liu, X. Sun, R. Camacho-Aguilera, L. C. Kimerling, and J. Michel, “Ge-on-Si laser operating at room temperature,” Opt. Lett. 35(5), 679–681 (2010). [CrossRef]
5. R. Camacho-Aguilera, Z. Han, Y. Cai, L. C. Kimerling, and J. Michel, “Direct band gap narrowing in highly doped Ge,” Appl. Phys. Lett. 102(15), 152106 (2013). [CrossRef]
6. J. Liu, M. Beals, A. Pomerene, S. Bernardis, R. Sun, J. Cheng, L. C. Kimerling, and J. Michel, “Waveguide-integrated, ultralow-energy GeSi electro-absorption modulators,” Nat. Photonics 2(7), 433–437 (2008). [CrossRef]
7. N. Feng, D. Feng, S. Liao, X. Wang, P. Dong, H. Liang, C. Kung, W. Qian, J. Fong, R. Shafiiha, Y. Luo, J. Cunningham, A. V. Krishnamoorthy, and M. Asghari, “30 GHz Ge electro-absorption modulator integrated with 3µm silicon-on-insulator waveguide,” Opt. Express 19(8), 7062–7067 (2011). [CrossRef]
8. S. A. Srinivasan, M. Pantouvaki, S. Gupta, H. T. Chen, P. Verheyen, G. Lepage, G. Roelkens, K. Saraswat, D. V. Thourhout, P. Absil, and J. V. Campenhout, “56 Gb/s Germanium Waveguide Electro-Absorption Modulator,” J. Lightwave Technol. 34(2), 419–424 (2016). [CrossRef]
9. Y. Lin, K. Lee, S. Bao, X. Guo, H. Wang, J. Michel, and C. S. Tan, “High-efficiency normal-incidence vertical p-i-n photodetectors on a germanium-on-insulator platform,” Photonics Res. 5(6), 702–709 (2017). [CrossRef]
10. B. Son, Y. Lin, K. H. Lee, Y. Wang, S. Wu, and C. S. Tan, “High speed and ultra-low dark current Ge vertical pin photodetectors on an oxygen-annealed Ge-on-insulator platform with GeOx surface passivation,” Opt. Express 28(16), 23978–23990 (2020). [CrossRef]
11. T. Yin, R. Cohen, M. M. Morse, G. Sarid, Y. Chetrit, D. Rubin, and M. J. Paniccia, “31 GHz Ge nip waveguide photodetectors on Silicon-on-Insulator substrate,” Opt. Express 15(21), 13965–13971 (2007). [CrossRef]
12. M. Jurgen, J. Liu, and L. C. Kimerling, “High-performance Ge-on-Si photodetectors,” Nat. Photonics 4(8), 527–534 (2010). [CrossRef]
13. J. Liu, D. D. Cannon, K. Wada, Y. Ishikawa, S. Jongthammanurak, D. T. Danielson, J. Michel, and L. C. Kimerling, “Tensile strained Ge p-i-n photodetectors on Si platform for C and L band telecommunications,” Appl. Phys. Lett. 87(1), 011110 (2005). [CrossRef]
14. F. Olbricha, J. Höschele, M. Müller, J. Kettler, S. L. Portalupi, M. Paul, M. Jetter, and P. Michler, “Polarization-entangled photons from an InGaAs-based quantum dot emitting in the telecom C-band,” Appl. Phys. Lett. 111(13), 133106 (2017). [CrossRef]
15. B. Chen, W. Jiang, J. Yuan, A. L. Holmes, and B. M. Onat, “SWIR/MWIR InP-based pin photodiodes with InGaAs/GaAsSb type-II quantum wells,” IEEE J. Quantum Electron. 47(9), 1244–1250 (2011). [CrossRef]
16. M. Verdun, G. Beaudoin, B. Portier, N. Bardou, C. Dupuis, I. Sagnes, R. Haïdar, F. Pardo, and J. Pelouard, “Dark current investigation in thin P-i-N InGaAs photodiodes for nano-resonators,” J. Appl. Phys. 120(8), 084501 (2016). [CrossRef]
17. G. Dehlinger, S. J. Koester, J. D. Schaub, J. O. Chu, Q. C. Ouyang, and A. Grill, “High-speed germanium-on-SOI lateral PIN photodiodes,” IEEE Photonics Technol. Lett. 16(11), 2547–2549 (2004). [CrossRef]
18. C. Cheng, C. Tsai, P. Yeh, S. Hung, S. Bao, K. H. Lee, C. S. Tan, and G. Chang, “Ge-on-insulator lateral pin waveguide photodetectors for optical communication,” Opt. Lett. 45(24), 6683–6686 (2020). [CrossRef]
19. J. Wang, W. Y. Loh, K. T. Chua, H. Zang, Y. Z. Xiong, T. H. Loh, M. B. Yu, S. J. Lee, G. Lo, and D. L. Kwong, “Evanescent-coupled Ge pin photodetectors on Si-waveguide with SEG-Ge and comparative study of lateral and vertical pin configurations,” IEEE Electron Device Lett. 29(5), 445–448 (2008). [CrossRef]
20. S. Xu, Y.-C. Huang, K. H. Lee, W. Wang, Y. Dong, D. Lei, S. Masudy-Panah, C. S. Tan, X. Gong, and Y.-C. Yeo, “GeSn lateral pin photodetector on insulating substrate,” Opt. Express 26(13), 17312–17321 (2018). [CrossRef]
21. C. Li, K. Bao, S. Qin, K. Guan, G. Xu, and J. Su, “Grating-enabled high-speed high-efficiency surface-illuminated silicon photodiodes,” Opt. Express 29(3), 3458–3464 (2021). [CrossRef]
22. S. Wu, H. Zhou, Q. Chen, L. Zhang, K. H. Lee, S. Bao, W. Fan, and C. S. Tan, “Suspended germanium membranes photodetector with tunable biaxial tensile strain and location-determined wavelength-selective photoresponsivity,” Appl. Phys. Lett. 119(19), 191106 (2021). [CrossRef]
23. Q. Chen, S. Wu, L. Zhang, D. Burt, H. Zhou, D. Nam, W. Fan, and C. S. Tan, “GeSn-on-insulator dual-waveband resonant-cavity-enhanced photodetectors at the 2 µm and 1.55 µm optical communication bands,” Opt. Lett. 46(15), 3809–3812 (2021). [CrossRef]
24. C. H. Tsai, B. J. Huang, R. A. Soref, G. Sun, H. H. Cheng, and G. E Chang, “GeSn resonant-cavity-enhanced photodetectors for efficient photodetection at the 2 µm wavelength band,” Opt. Lett. 45(6), 1463–1466 (2020). [CrossRef]
25. H. Cansizoglu, C. Bartolo-Perez, Y. Gao, E. P. Devine, S. Ghandiparsi, K. G. Polat, H. H. Mamtaz, T. Yamada, A. F. Elrefaie, and S.-Y. Wang, “Surface-illuminated photon-trapping high-speed Ge-on-Si photodiodes with improved efficiency up to 1700nm,” Photonics Res. 6(7), 734–742 (2018). [CrossRef]
26. H. Zhou, S. Xu, Y. Lin, Y.-C. Huang, B. Son, Q. Chen, X. Guo, K. H. Lee, S. C.-K. Goh, and X. Gong, “High-efficiency GeSn/Ge multiple-quantum-well photodetectors with photon-trapping microstructures operating at 2 µm,” Opt. Express 28(7), 10280–10293 (2020). [CrossRef]
27. Y. Gao, H. Cansizoglu, K. G. Polat, S. Ghandiparsi, A. Kaya, H. H. Mamtaz, A. S. Mayet, Y. Wang, X. Zhang, and T. Yamada, “Photon-trapping microstructures enable high-speed high-efficiency silicon photodiodes,” Nat. Photonics 11(5), 301–308 (2017). [CrossRef]
28. H. Zhou, L. Zhang, J. Tong, S. Wu, B. Son, Q. Chen, D. H. Zhang, and C. S. Tan, “Surface plasmon enhanced GeSn photodetectors operating at 2 µm,” Opt. Express 29(6), 8498–8509 (2021). [CrossRef]
29. J. Tong, L. Y. Tobing, S. Qiu, D. H. Zhang, and A. Unil Perera, “Room temperature plasmon-enhanced InAs0. 91Sb0.09-based heterojunction nip mid-wave infrared photodetector,” Appl. Phys. Lett. 113(1), 011110 (2018). [CrossRef]
30. A. I. Yakimov, V. V. Kirienko, A. A. Bloshkin, A. V. Dvurechenskii, and D. E. Utkin, “Near-infrared photoresponse in Ge/Si quantum dots enhanced by localized surface plasmons supported by aluminum nanodisks,” J. Appl. Phys. 128(14), 143101 (2020). [CrossRef]
31. K. H. Lee, S. Bao, G. Y. Chong, Y. H. Tan, E. A. Fitzgerld, and C. S. Tan, “Defects reduction of Ge epitaxial film in a germanium-on-insulator wafer by annealing in oxygen ambient,” APL Mater. 3(1), 016102 (2015). [CrossRef]
32. K. H. Lee, S. Bao, B. Wang, C. Wang, S. F. Yoon, J. Michel, E. A. Fitzgerald, and C. S. Tan, “Reduction of threading dislocation density in Ge/Si using a heavily As-doped Ge seed layer,” AIP Adv. 6(2), 025028 (2016). [CrossRef]
33. K. H. Lee, S. Bao, Y. Wang, E. A. Fitzgerald, and C. Seng Tan, “Suppression of interfacial voids formation during silane (SiH4)-based silicon oxide bonding with a thin silicon nitride capping layer,” J. Appl. Phys. 123(1), 015302 (2018). [CrossRef]
34. J. Kang, R. Zhang, M. Takenaka, and S. Takagi, “Suppression of dark current in GeOx-passivated germanium metal-semiconductor-metal photodetector by plasma post-oxidation,” Opt. Express 23(13), 16967–16976 (2015). [CrossRef]
35. G. Du, X. Liu, Z. Xia, J. Yang, and R. Han, “Effect of interface roughness on the carrier transport in germanium MOSFETs investigated by Monte Carlo method,” Chinese Phys. B 19(5), 057304 (2010). [CrossRef]
36. R. Igreja and C. J. Dias, “Analytical evaluation of the interdigital electrodes capacitance for a multi-layered structure,” Sensors and Actuators A: Physical 112(2-3), 291–301 (2004). [CrossRef]
37. G. González, E. S. Kolosovas-Machuca, E. López-Luna, H. Hernández-Arriaga, and F. J. González, “Design and fabrication of interdigital nanocapacitors coated with HfO2,” Sensors 15(1), 1998–2005 (2015). [CrossRef]
38. T. Satitchantrakul, W. Jeamsaksiri, and S. Chalermwisutkul, “Droplet Detection with Interdigitated Capacitor Sensor,” Research, Invention, and Innovation Congress (RI2C). IEEE, 1–4 (2019).
