Open Access
Add to BibSonomy
Abstract
Germanium (Ge) is an attractive material for monolithic light sources and photodetectors, but it is not easy to integrate Ge light sources and photodetectors because their optimum device structures differ. In this study, we developed a monolithically integrated Ge light emitting diode (LED) that enables current injection at high density and a Ge photodiode (PD) having low dark current, and we fabricated an on-chip optical interconnection system consisting of the Ge LED, Ge PD, and Si waveguide. We investigated the properties of the fabricated Ge LED and PD and demonstrated on-chip optical interconnection.
© 2021 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Data centers and information and communication technology (ICT) are reaching their performance limits in terms of power consumption, miniaturization, and speed because of the transmission delay and heat generation due to metal wiring. To solve this problem, silicon (Si) photonics, which enables high-speed optical interconnects in electronic devices, have attracted much attention [1–3]. As for Si photonics, since optical devices are manufactured by a complementary metal-oxide-semiconductor (CMOS) process, they can be monolithically integrated with Si electronic components. There have been many reports on photonic devices that can be monolithically integrated on a Si chip, such as waveguides [4–8], optical modulators [9–13], germanium (Ge) photodiodes (PDs) [14–43] and so on. However, the biggest challenge facing Si photonics is monolithic integration of light emitters [44]. This is because Si is an indirect-bandgap material with extremely low light-emission efficiency, whereas conventional lasers are based on III-V compound semiconductors and are not compatible with CMOS processes. III-V semiconductor lasers bonded to Si waveguides have been reported [44–46], but it is difficult to achieve large-scale integration with them. Monolithic integrated light sources compatible with the CMOS process are needed to take full advantage of Si photonics. Ge has attracted much attention as a promising candidate for not only PDs but also monolithic light sources due to its CMOS compatibility and pseudo-direct-bandgap characteristics resulting from the small difference between the L valley and the Γ valley of the conduction band [47]. It is necessary to excite enough electrons in the Γ valley of the conduction band to increase the light emission efficiency of Ge because most electrons are excited in the indirect L valley of the conduction band because of the lower energy of the L valley than that of the Γ valley. An effective approach is filling the L valley by heavy n-type doping [47–53]. The effect of L-valley filling is to increase the proportion of electrons excited in the Γ valley. Another approach is introduction of tensile strain into Ge [47,48,53–55]. The tensile strain reduces the bandgap energy of Ge more in the Γ valley than in the L valley. The number of electrons injected into the Γ valley is therefore increased by the tensile strain. There have been reports on achieving lasing by using above approaches [56–61]. However, there is a problem in terms of the manufacturing processes used to monolithically integrate Ge light emitting diodes (LEDs) and Ge PDs because the most appropriate diode structure for Ge LEDs is different from that for PDs. For LEDs, formation of a pn junction by connecting p- and n- type heavily doped electrodes to Ge active layers is effective for injecting current at high density [50,60–64]. On the other hand, for PDs, a pin junction structure is used in order to have a light receiving region. Also, dark current should be small because a higher dark current needs higher optical input power to maintain the on/off ratio. Applying a high electric field to Ge causes a high dark current because Ge layers grown on Si substrates have a large number of dislocations due to the lattice mismatch between these materials [65–67]. Dark current increases with the electric field at the Ge/Si interface, which in vertical Ge/Si diodes is caused by increasing the doping concentration in Si (e.g., p-Ge/i-Ge/n-Si) [42]. Hence, vertical Ge pin diodes on Si substrates have been used in many reports on Ge PDs [14,15,26,37–39] because the junctions of the diodes should be separated from the Ge/Si interfaces. It is difficult to monolithically integrate vertical diodes with different structures by depositing multiple layers. To overcome this problem, we previously reported lateral Ge LEDs to which an external stressor can be easily applied [68]. It is easy to adjust the designs of lateral diodes because the junction can be defined by photolithography.
In this study, we devised Ge LEDs and PDs that can be monolithically integrated by using simple fabrication processes. We investigate the properties of the fabricated Ge LEDs and PDs and demonstrated on-chip optical interconnection using monolithically integrated Ge LEDs, Ge PDs, and Si waveguides.
2. Device structure and fabrication process
Figure 1(a) shows a schematic top view of an on-chip optical interconnection system consisting of a monolithically integrated Ge LED, PD, Si waveguide, and electrical isolation part, and Figs. 1(b)–(d) show schematic cross sections of (b) a Ge LED (between A and A’ in Fig. 1(a)), (c) a Si waveguide (between B and B’ in Fig. 1(a)), and (d) a Ge PD (between C and C’ in Fig. 1(a)). The Ge LED (Fig. 1(b)) consists of a lateral pn diode in a 200-nm-thick SOI layer and a Ge waveguide with a thickness of 500 nm, width of 500 nm, and length of 500 μm on the pn junction of the lateral SOI pn diode. The LED structure in this study is suitable to apply tensile strain to the Ge waveguide using a SiNx stressor because a device structure with p- and n-type electrodes underneath the Ge waveguide (a Ge waveguide on SOI diode) can prevent the electrodes from interfering with stress distribution in the Ge waveguide and SiNx stressor [68]. In addition, a lateral Ge pn junction formed by ion implantation technique is not suitable to light source because crystallinity of Ge deteriorates by ion implantation. Therefore, an LED structure using Si for the p- and n-type electrodes was adopted in this study. In addition, as pointed out in reports on PD with a similar structure [23–25,27–29], by using Si instead of Ge for the p- and n-type electrodes, the contact resistance and the free carrier absorption loss can be reduced. On the other hand, the LED structure in this study is disadvantageous in terms of injected current density compared to the p-Si/n-Ge junction such as the first Ge laser structure [60]. However, as reported in our previous work [68], forming a p-Si/n-Ge junction using an n-type Ge waveguides on SOI pn diodes improves current injection efficiency, while reduced carrier lifetime due to high-concentration doping causes non-uniform carrier distribution in the Ge waveguide. In this study, Ge was not doped with priority given to carrier lifetime and integration with the Ge PD. The Ge PD (Fig. 1(d)) consists of a lateral pin diode with a 2-μm-width i-region in a 200-nm-thick SOI layer and a Ge waveguide with a thickness of 500 nm, width of 1000 nm, and length of 100 μm on the i-region of a lateral SOI pin diode. The LED has a 10-μm-long Ge taper to help guide the light down, but the PD does not have a Ge taper. The reason why the PD was not tapered is that the Ge taper on the Si waveguide is far from the p- and n-type electrodes, and the photocurrent generated in the Ge taper causes a decrease in the speed of the PD. On the other hand, In the LED, since Ge is not doped, and the free carrier absorption loss is low, so a taper was provided. To help guide light up to the PD from the Si waveguide, it is promising to provide a taper with other materials such as poly-Si [23]. The Si waveguide consists of a SOI layer with a thickness of 200 nm and a width of 2 μm. The lengths of the Si waveguides were varied (10, 100, and 1000 μm) to investigate the waveguide loss. To enhance the optical coupling to the Ge LED and PD, the Si waveguide was not separated from the SOI layers of the Ge LED and Ge PD. The electrical isolation part (Fig. 1(c)) is located at the center of the Si waveguide for electrical isolation and maintaining optical coupling. The electrical isolation part has a cladding layer that consists of a SiNx layer with a thickness of 500 nm and width of 4 μm. The Si waveguide was separated to ensure electrical isolation, and the end of the Si waveguide was tapered [69–72]. Si tapers have a length of 0, 1, 10 or 20 μm facing each other and a 1-μm-long gap at the center.
Fig. 1. (a) Schematic top view of on-chip optical interconnection system consisting of monolithically integrated Ge LED, PD, Si waveguide, and electrical isolation part. Schematic cross sections of (b) Ge LED (between A and A’), (c) Si waveguide (between B and B’), and (d) Ge PD (between C and C’).
Cross-sectional schematic views at each step of the fabrication process of the monolithic integrated Ge LEDs, PDs, Si waveguides, and electrical isolation parts are shown in Figs. 2(a-1)–(a-4), (b-1)–(b-4), and (c-1)–(c-4). An 8-inch SOI substrate with a 2000-nm-thick BOX layer and a 200-nm-thick SOI layer having a plane orientation of <100> was used (Figs. 2(a-1), (b-1), and (c-1)). First, the SOI layer was oxidized to form a 20-nm-thick SiO2 layer. Subsequently, dopants were implanted to form p-type and n-type diffusion regions, and dopants were activated by N2 annealing to form lateral SOI diodes. The SiO2 damaged by implantation was removed by a hydrofluoric acid (HF) solution, and a 500-nm-thick Ge layer was epitaxially grown by two-step epitaxial growth technique, which includes post-growth annealing and high-temperature regrowth, using a cold-wall rapid thermal chemical vapor deposition (CVD) system [73] (Figs. 2(a-2), (b-2), and (c-2)). Although in-situ doping was not performed, trap levels induced by defects caused the Ge layer to have p-type conductivity with a carrier concentration of 1016 cm−3 [74–76]. Ge waveguides were formed by patterning using electron beam lithography and dry etching. Subsequently, Si waveguides and mesa isolation of Ge LEDs and PDs were made by patterning of the SOI layer using photolithography and dry etching (Figs. 2(a-3), (b-3), and (c-3)). The surfaces of the Ge waveguides were passivated with a 5-nm-thick GeO2 layer formed by low-temperature oxidization to suppress nonradiative recombination due to interfacial traps [77–79]. Then a 20-nm-thick SiO2 layer was deposited by low-temperature CVD to prevent dissolution of the GeO2 layer, which is water soluble, during the cleaning process. Thereafter, contact holes were opened on the SOI layer by wet etching, and metal electrodes were formed. Then a 500-nm-thick SiNx layer under 100 MPa tensile stress was deposited and patterned by photolithography and dry etching to form clad layers of electrical isolation part (Figs. 2(a-4), (b-4), and (c-4)). Figure 2(d) shows an optical microscope image of the fabricated on-chip optical interconnection system. Figures 2(a-4), (b-4), and (c-4) show schematic cross sections between A and A’, B and B’, and C and C’ in Fig. 2(d).
Fig. 2. Cross-sectional schematic views at each step of fabrication process of monolithic integrated (a-1)–(a-4) Ge LED, (b-1)–(b-4) electrical isolation part, and (c-1)–(c-4) Ge PD. (d) Optical microscope image of fabricated on-chip optical interconnection system.
3. Simulation study of electrical properties of Ge LEDs and PDs
The difference between the Ge LED and Ge PD is the width of the Ge waveguide and the junction structure of the SOI diode beneath the Ge waveguides. We investigated the effects of different device designs on the electrical properties. First, to investigate the properties of the PDs, we simulated the electric field in the devices under application of reverse bias by using technology computer aided design (TCAD). Figures 3(a) and (b) show the strength of the electric fields in Ge waveguides with a width of 1 μm on a SOI pn diode (Fig. 3(a)) and on a pin diode (Fig. 3(b)) under a reverse bias of 1 V. The doping concentration was set to 5×1019cm−3 in p+-Si and n+-Si, 1015 cm−3 in p−-Si, and 1016 cm−3 in p−-Ge. In Fig. 3(a), the electric field in the p−-Ge is high near the n+-Si because the pn junction there is formed from p−-Ge and n+-Si. On the other hand, in Fig. 3(b), the electric field is high in the p−-Si near the n+-Si because the pn junction is formed from p−-Si and n+-Si. A large number of defects exist in a Ge layer grown on a Si substrate, especially near the Ge/Si interface, because of the lattice mismatch between these materials [65]. Such high-density defects at the Ge/Si interface cause the high dark current of Ge PDs [66,67]. A Ge waveguide on a SOI pin diode (Fig. 3(b)) is expected to have a dark current lower compared with a waveguide on a SOI pn diode because the electric field near the Ge/Si interface is low. The band diagrams of the Ge PD under a reverse bias of 1 V are shown in Fig. 3(c) and (d). In particular, Fig. 3(c) shows profiles of the electrostatic potential at the bottom of the conduction band (Ec) and top of the valence band (Ev) along the X direction at A–A’ and B–B’ in Fig. 3(b). The electric field along the X direction in the Ge waveguide is lower than that in the SOI layer beneath the Ge because the carrier density in p−-Ge is higher than that in p−-Si. Therefore, the difference in electrostatic potential between p−-Ge and p−-Si at the Ge/Si interface is large at the edge of the Ge waveguide. Figure 3(d) shows profiles of the electrostatic potential at the bottom of the Ec and the top of the Ev along the Z direction at C–C’ and D–D’ in Fig. 3(b). The electric field moves holes from the Ge waveguide into the SOI layer at the Ge/Si interface at C–C’ in Fig. 3(b) for the reason described above. The electrostatic potential is higher in the Ge waveguide than in the SOI layer at the Ge/Si interface at D–D’ in Fig. 3(b), so that the electric field moves electrons from the Ge waveguide to the SOI layer at the Ge/Si interface. These results show that photo-excited carriers can be moved from both edges of the Ge waveguide to the SOI diode, i.e., holes are moved from the left side of the Ge/Si interface (C–C’ in Fig. 3(b)) and electrons are moved from the right side (D–D’ in Fig. 3(b)).
Fig. 3. Electric field strength in 1-μm-wide Ge waveguides (a) on SOI pn diode and (b) on pin diode under reverse bias of 1 V. Band diagrams of Ge PD at (c) A–A’ and B–B’, and (d) C–C’ and D–D’ in Fig. 3(b).
Next, to investigate the properties of the LEDs, we simulated the electron density under a forward bias. Ge has pseudo-direct-bandgap characteristics, where the Γ valley is 140 meV higher than the L valley in the conduction band [47]. Current injection at high density is needed to excite enough electrons in the Γ valley of the conduction band and increase the electroluminescence (EL) intensity [47,48,53]. We investigated the effect of the differences between the Ge LED and the Ge PD, i.e., the width of the Ge waveguide and the junction structure of the SOI diode beneath the Ge waveguides, on the carrier density. Figures 4(a)–(c) show the electron density in the devices under a forward bias of 2 V. The black solid lines in the devices indicate equipotential lines. The doping concentrations in Ge and Si were set as above and the carrier lifetime in Ge was set to 1 ns [80]. To investigate the effect of the junction structure on the injected current, we simulated 1-μm-wide Ge waveguide on a SOI pin diode (Fig. 4(a)) and on a SOI pn diode (Fig. 4(b)). The electrostatic potential in the Ge waveguide on the SOI pin diode (Fig. 4(a)) is lower than that of Si at the Ge/Si interface near the n+-Si, so that electrons are injected from the SOI layer to the Ge waveguide. However, the electron density in the Ge waveguide is limited to 1018 cm−3 because of the low electron density in the p−-Si layer. On the other hand, electrons are directly injected from the n+-SOI layer to the Ge waveguide on the SOI pn diode (Fig. 4(b)) so that the electron density in the Ge waveguide on a SOI pn diode can be increased to about 1019 cm−3. These results indicate that the electron density in the Ge waveguide can be increased about tenfold simply by changing the junction structure. From these results, the device structure for current injection should differ from the device structure for photo detection, and a Ge waveguide on a SOI pn diode, in which Ge is directly connected to highly doped Si, is suitable for current injection at high density. Next, the suitable width of the waveguide for current injection was investigated by comparing Ge waveguides with widths of 1 μm (Fig. 4(b)) and 500 nm (Fig. 4(c)). The electron density in the waveguide far from the p−-Ge/n+-Si interface decreases because the carrier lifetime in Ge grown on Si substrate is very short [80]. Thus, the region in which carrier density is low is large in the wide Ge waveguide. Also, self-heating is higher in the wide waveguide than it is in the narrow one because the bias voltage is higher in the wide waveguide. Therefore, we decided to use the narrower 500-nm-wide Ge waveguide in the Ge LED because it is more suitable for high-density current injection.
Fig. 4. Distribution of electrons under forward bias of 2 V in (a) 1-μm-wide Ge waveguide on SOI pin diode, (b) 1-μm-wide Ge waveguide on SOI pn diode, and (c) 500-nm-wide Ge waveguide on SOI pn diode.
Simulated current-voltage characteristics of the Ge LED, PD, and a lateral SOI diode for reference under forward bias are shown in Fig. 5. The Ge LED has a lower forward voltage, defined at a current of 1 μA (i.e., 0.36 V), than that of the SOI diode (i.e., 0.86 V) because the top of the valence band of Ge is higher than that of Si [68,81]. Also, the electrical resistance of the Ge LED is lower than that of the Ge PD because of the higher electron density in the Ge waveguide of the Ge LED. These results indicate that Ge LEDs that enable high injection current density and Ge PDs that enable low dark current density can be monolithically integrated by using photolithography to change the junction structure and width of the Ge waveguide.
4. Electrical and optical characterization of Ge LEDs and Ge PDs
Next, the electrical and optical properties of fabricated Ge LEDs and PDs in an on-chip optical interconnection system (Fig. 1) were investigated. Current-voltage characteristics of the fabricated Ge LED, Ge PD and SOI pin diode are shown in Fig. 6. The dashed lines are displayed on a logarithmic scale, and the solid lines are displayed on a linear scale. Dark current under reverse bias is two orders of magnitude lower in the Ge PD than in the Ge LED owing to the low electric field applied to the Ge/Si interface. Dark currents in literature were summarized in Fig. 7. The dark current of 12 nA at −2 V is comparable to those of state-of-the-art Ge photodetectors, such as vertical Ge pin diodes [15,20–22], vertical Ge/Si pin diodes [17,19,34–36], lateral Ge pin diodes [26,30–33,39], lateral Si/Ge/Si diodes [23–25,27–29], and metal-semiconductor-metal (MSM) detectors [40]. The minimum record of dark current in each Ge area is not proportional to the Ge area, and the larger the area, the lower the dark current density tends to be. It is presumed that this is because the contribution of parasitic components such as surface leakage is large in devices having small sizes. Regarding the forward characteristics, the forward voltages of the Ge LED and PD are lower than that of the SOI pin diode due to current injected into the Ge waveguide [68]. Also, the differential resistances of the LED and PD under a forward bias of 1 V are 4.7 Ω and 53 Ω, respectively, so high-density current can be injected into the waveguide of the Ge LED. These experimental results show that monolithic integration of Ge PDs with low dark current and Ge LEDs that enable carrier injection at high density can be fabricated simply by changing the junction design and width of the Ge waveguide.
Next, we investigated the optical properties of the fabricated Ge PDs and LEDs. First, the emission properties of the LEDs were investigated. The EL spectrum was detected by an InGaAs photodetector that has a cut-off wavelength of 1620 nm and was set above the LED. Figure 8(a) shows an EL spectrum of a typical Ge LED at an injected current of 100 mA and the sensitivity of the InGaAs photodetector. The peak wavelength of the EL spectrum was 1589 nm, and it is shorter than the cut-off wavelength of the photodetector. So, it is considered that the actual peak of the EL was detected, and it corresponds to the direct-bandgap energy of Ge under a biaxial tensile strain of 0.14%; these results confirm direct-bandgap EL from the Ge LED. The tensile strain in the Ge waveguide was due to the thermal expansion mismatch between Ge and Si during Ge epitaxial growth [73]. The output power of the LED was measured to estimate the quantum efficiency of the Ge LED. Since the spontaneous emission from the Ge LED may have low coupling efficiency with the waveguide, the output power was measured from the top surface of the LED using an InGaAs photodetector with a large detecting area. Since the detecting area of the InGaAs photodetector was large, it is considered that almost all the light emitted to the upper part of the Ge LED could be detected. The output power was calculated by converting the photocurrent at 1.29 A/W. Figure 8(b) plots the output power of the LED against injection current. The output power increased with increasing injection current, and an output power of 265 nW was obtained at an injection current of 300 mA. Although not all the emitted light was detected, the external quantum efficiency could be estimated to be more than 1.14×10−4% from output power of 265 nW at current injection of 300 mA.
Fig. 8. (a) EL spectrum of Ge LED under current injection of 100 mA and (b) output power of the Ge LED as a function of injection current.
The frequency response of the Ge LED was measured using the optical network analyzer and a voltage amplifier with a bias-T for driving the LED. The emitted light was focused by the lens coupled to a multimode fiber. Figure 9(a) shows the frequency response of the LED and the voltage amplifier. The 3-dB bandwidth of the LED was 65 MHz, the same as that of the voltage amplifier. This suggests that the frequency response of the LED is limited by that of the voltage amplifier. From this result, it is estimated that the 3-dB bandwidth is above 65 MHz. Waveforms of the Ge LED was evaluated using an oscilloscope (Tektronix 3052) to confirm the modulation operation of the Ge LED. Input signal to the Ge LED was a sinusoidal current with an amplitude of 500 mA, offset of 250 mA and frequency of 50 MHz. The emitted light from the Ge LED was received by a photo receiver (New Focus 1811). Figure 9(b) shows the input and output waveforms of the Ge LED. The output signal followed the input signal, and the 50 MHz-modulation operation of the Ge LED was confirmed.
Fig. 9. (a) Frequency response of Ge LED and voltage amplifier. (b) Waveforms of input signal and output signal of the Ge LED.
Next, the properties of the Ge PD were investigated. The characteristics of the PD in the on-chip optical interconnection system, the characteristics of the Ge PD were measured by irradiating the upper surface. Figure 10(a) plots the photocurrent of PDs illuminated by light of various wavelengths through the lensed fiber. The lens curvature and the mode field diameter of the lens fiber were 10 μm and 1.4 μm, respectively. The photocurrent decreases as the wavelength of the input light shifts to the red. Because this plot corresponds to the energy dependence of the absorption coefficient of Ge [43], it shows that PD worked by excitation across the direct bandgap of Ge. It can be seen in Fig. 10(a) that the larger the fiber output power, the slightly higher the increasing rate of photocurrent. One possible reason of this is thought to be temperature rising due to laser irradiation. Energy dependence of the absorption coefficient of Ge is high near the direct bandgap of Ge, and the direct bandgap of Ge shrinks by temperature rising [68]. Therefore, the absorption coefficient in a wavelength range from 1550 nm to 1600 nm is temperature sensitive. The coupling loss between the fiber and the PD was estimated to be 5 dB from a beam profile of the fiber measured using knife-edge method (shown in inset of Fig. 10(a)). Photo responsivity by vertically coupling was 40 mA/W at a wavelength of 1550 nm and 5.5 mA/W at a wavelength of 1600 nm. Considering the 500-nm-thick Ge layer and the 5-dB coupling loss, the absorption coefficient at a wavelength of 1550 nm was estimated to be 3000 cm−1 and it corresponds to that reported in a previous study [43]. The absorption coefficient at a wavelength of 1600 nm, which is the peak wavelength of the Ge LED, was estimated to be 280 cm−1. Wavelength dependence of the absorption coefficient is large around these wavelengths because they are near direct band edge of Ge. The absorption coefficient is low at the wavelength for on-chip optical interconnection because the LED and the PD are based on the same material, i.e. Ge. These results show that a waveguide-coupling PD is essential to compensate for the reduced photo responsivity due to the low absorption coefficient in the Ge-based optical interconnection. In this study, the Ge PD was placed in the on-chip interconnection system, so the coupling efficiency between the Si waveguide and the Ge PD cannot be evaluated. However, in the literature reporting similar structures, high photo responsivity has been reported at 1550 nm, even with a short Ge waveguide, for example, more than 0.5 A/W at 10-μm-long Ge waveguide [25] and more than 0.9 A/W at 20-μm-long Ge waveguide [24]. In this study, the emission wavelength of the Ge LED is about 1600 nm and the absorption coefficient is small about 0.1 times of that at a wavelength of 1550 nm, but since the length of the Ge waveguides is 100 μm, photo responsivity close to 0.5 A/W can be expected. That is, the photo responsivity at a wavelength of 1600 nm can be further improved by further increasing the Ge waveguide length of the Ge PD. Figure 10(b) shows the frequency response of the PD under various reverse bias voltages. The measurement was performed using the optical network analyzer where the upper surface of the PD was irradiated with light with a wavelength of 1550 nm. The response increased with increasing bias applied to the PD. The 3-dB bandwidth was 4.6 GHz at a reverse bias of 3 V. Although the bandwidth of 4.6 GHz is lower than those of state-of-the-art devices [15,17,19,26,37,38,41], it is sufficiently high to receive the modulated signal of the Ge LED having a low bandwidth. There are two possible factors that limit the bandwidth: the time constant of the circuit and the drift velocity of carriers. Since the differential resistance at a bias of 1 V is 53 Ω and the capacitance is 38 fF, the bandwidth limited by the time constant of resistance including the 50-Ω termination of the measurement system and capacitance is 40 GHz. Hence, the frequency response of the Ge PD is limited not by the time constant due to the resistance and capacitance of the device but by the carrier-transit time. As shown in Fig. 3(b), the electric field concentrates in the SOI layer and in part of the Ge waveguide. In most of the Ge waveguide, the electric field is less than 100 V/cm. Hence, the carrier-transit time that determines the bandwidth can be divided into a drift process in the SOI layer and a diffusion process in the Ge waveguide with a weak electric field. The experimental values can be explained by the drift time in the SOI layer. Since the carrier lifetime in the Ge waveguide is short, the contribution of the diffusion current to the signal is considered to be small.
Fig. 10. (a) Photocurrent of Ge PDs whose upper surfaces were illuminated by light of various wavelengths through a lens fiber. Inset shows beam profile of the fiber. (b) Frequency response of Ge PD under various reverse bias voltages.
5. Demonstration of the on-chip optical interconnection
We demonstrated on-chip optical interconnection systems incorporating the Ge LEDs and PDs described above. Figure 11 plots the current of the Ge PD caused by light from the Ge LED, which was connected to the PD via a 1-mm-long Si waveguide and an electrical isolation part without a Si taper. The current of the Ge PD increased as the current injected into the Ge LED increased, indicating that the photocurrent was generated by the light emitted from the Ge LED. The standard deviation of the dark current in the range of 2 V to 1.5 V is 17 pA, which is sufficiently low with respect to the measured photo current. Therefore, on-chip optical interconnection consisting of a monolithically integrated Ge LED, Ge PD, and Si waveguide was confirmed from this result. It is necessary to apply a reverse bias larger than 0.5 V to obtain high photo responsivity, as reported in [25], and we have confirmed the operation with a reverse bias up to 3 V. In addition, it is possible to inject a current of up to 800 mA into the Ge LED, but if a larger current is applied, the LED will be broken.
Fig. 11. Current of Ge PD under reverse bias when various currents were injected to Ge LED connected to the Ge PD via a 1-mm-long Si waveguide and an electrical isolation part without a Si taper.
Next, the propagation loss of the Si waveguides was investigated by evaluating the dependence of the photocurrent of the Ge PDs on the waveguide length. Figure 12(a) plots the photocurrent of the PDs under a 2 V reverse bias for various currents injected to the Ge LEDs and various Si waveguide lengths. Photocurrent increases super-linearly against injected current because the intensity of light emitted from the LED increases super-linearly against injected current. The increase in the ratio of electrons in the Γ valley to electrons in the L valley of the conduction band with increasing Fermi level and shrinkage of the direct bandgap of Ge due to self-heating by current injection are considered to be the reasons for the super-linear increase in the intensity of light emitted from the Ge LED [62,68]. Photocurrent decreased as the length of the Si waveguide was increased. Figure 12(b) shows the dependence of the photocurrent on the length of the Si waveguide in the case of a 2-V reverse bias applied to the PDs and a current of 500 mA injected to the LEDs. Dots indicate experimental data, while the solid lines indicate the calculated components due to waveguide propagation, the dashed lines indicate the calculated components due to free space propagation, and the dotted lines show the sum of the calculated components due to waveguide propagation and free space propagation. The fitting parameter of the calculation were waveguide-propagation loss and ratio of the components due to waveguide propagation and free space propagation. We assumed the free space propagation to be of light emitted radially from a 500-μm-long Ge waveguide and the calculation was performed with the waveguide-propagation loss varied from −3 to −30 dB cm−1. In Fig. 12(b) red lines indicate calculation with −30 dB cm−1, green lines indicate calculation with −10 dB cm−1, and blue lines indicate calculation with −3 dB cm−1. Under all waveguide-propagation-loss conditions, the contribution of the free space propagation is larger than that of the waveguide propagation in the waveguides with length of 10 μm and 100 μm, whereas waveguide propagation is dominant in the 1-mm-long waveguide, i.e. the waveguide-propagation component is 1.6 nA in the total photocurrent of 1.9 nA. The free space propagation component of the light emission from the Ge LED causes low coupling efficiency and crosstalk in high-density integrated multichannel communications, so that the coupling efficiency to the waveguide and light directivity should be improved for practical use. The power efficiency was 3.8×10−7% in a 1-mm-long optical interconnection. Considering the estimated external quantum efficiency of the Ge LED, which is more than 1.14×10−4%, the coupling efficiency of the light emission from the Ge LED to the Ge PD via 1-mm-long Si waveguide was estimated to be less than 0.33%. In addition, the link speed of the optical interconnection is considered to be limited by the signal intensity of the Ge PD in this study. The photocurrent in the 1-mm-long interconnection is about two order of magnitude lower than that in the measurement of the Ge LED alone (Fig. 8(b)). Therefore, it is difficult to demonstrate the dynamic operation of the on-chip optical interconnection system in this study, since the signal intensity of the Ge PD is low. However, since the noise in dynamic operation is proportional to the square root of the frequency, operation speed can be expected on the order of kHz, which is four order of magnitude lower than the 50 MHz operation confirmed in Fig. 9(b). The quantum efficiency of the Ge light source and the coupling efficiency must be improved for higher link speeds and more energy efficient interconnections. In addition, improvement of the photo responsivity of the PD at a wavelength of 1600 nm by increasing the Ge waveguide length of the PD can also help to increase signal intensity of the PD. The most effective way to improve the quantum efficiency and the coupling efficiency is to replace the light source with a laser diode [60,61]. Application of strong tensile strain [57–59,68] to the Ge waveguide is a promising way to increase optical gain and achieve lasing.
Fig. 12. (a) Photocurrent of Ge PDs under 2 V reverse bias for various currents injected into Ge LEDs and various lengths of the Si waveguide between the Ge LEDs and PDs. (b) Si-waveguide-length dependence of photocurrent with 2 V reverse bias applied to the Ge PDs and current of 500 mA injected into the Ge LEDs.
The effect of varying the length of the Si taper of the electrical isolation parts was evaluated to investigate the optical coupling loss in the electrical isolation parts. Since waveguide propagation was found to be dominant in the optical interconnection with the 1000-μm-long Si waveguide, on-chip optical interconnection systems having 1000-μm-long waveguides were used in this measurement. Photocurrent of the Ge PD under a 2-V reverse bias is plotted in Fig. 13 against LED injection currents for various lengths of Si tapers of the electrical isolation part. It is clear that varying the taper length had little effect on the photocurrent. In this study, the 2-μm-wide Si waveguide was tapered using an i-line photolithography, so the tip width is considered to be about 500 nm, which is the resolution limit. Since the tip width of the Si waveguide is as wide as about 500 nm, it is considered that the mode conversion is not sufficient and a loss due to reflection occurs at the tip of the Si waveguide. In fact, it has been reported that a transmission of 90% or more can be obtained when the SOI waveguide width is converted from 2 μm to 500 nm by tapering [69]. This is consistent with the fact that the taper length dependence was small in this study, and it is considered that the loss mainly occurs at the tip of the taper. Thus, in order to investigate the coupling loss at the electrical isolation part, it is considered necessary to pattern the tapers by using electron beam lithography.
Fig. 13. Photocurrent of Ge PD under 2 V reverse bias as a function of injected current for various lengths of Si tapers of electrical isolation parts.
6. Conclusions
We developed a Ge LED that enables current injection at high density and a Ge PD that has a low dark current and fabricated an on-chip optical interconnection system consisting of the Ge LED, Ge PD, and a Si waveguide. Operation of the Ge LED and PD was confirmed. The dark current of the Ge PD was less than one-hundredth that of the Ge LED, and the forward resistance of the Ge LED was less than one-tenth that of the Ge PD. It was shown that monolithic integration of these Ge LEDs and PDs is easy. Furthermore, the on-chip optical interconnection system was demonstrated. The dependence of the photocurrent on the waveguide length indicated that waveguide propagation is dominant when the Si waveguide of the interconnection system is 1000 μm long. From the above results, it can be concluded that the Ge waveguide on lateral SOI diode is a device structure appropriate for realizing monolithic integrated Ge light sources and PDs for on-chip optical interconnections.
Funding
Japan Society for the Promotion of Science; Funding Program for World-Leading Innovative R&D on Science and Technology; the Council for Science and Technology Policy (CSTP).
Acknowledgments
We thank Professor Yasuhiko Arakawa, Professor Satoshi Iwamoto, Dr. Satoshi Kako, and Professor Yukihiro Shimogaki for their enlightening discussions. This research was granted by the Japan Society for the Promotion of Science (JSPS) through the “Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST Program)”, initiated by the Council for Science and Technology Policy (CSTP).
Disclosures
The authors declare no conflicts of interest.
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. R. Soref, “The past, present, and future of silicon photonics,” IEEE J. Sel. Top. Quantum Electron. 12(6), 1678–1687 (2006). [CrossRef]
2. B. Jalali and S. Fathpour, “Silicon photonics,” J. Light. Technol. 24(12), 4600–4615 (2006). [CrossRef]
3. Y. Arakawa, T. Nakamura, Y. Urino, and T. Fujita, “Silicon photonics for next generation system integration platform,” IEEE Commun. Mag. 51(3), 72–77 (2013). [CrossRef]
4. R. Soref and J. Larenzo, “All-silicon active and passive guided-wave components for λ = 1.3 and 1.6 µm,” IEEE J. Quantum Electron. 22(6), 873–879 (1986). [CrossRef]
5. A. M. Agarwal, L. Liao, J. S. Foresi, M. R. Black, X. Duan, and L. C. Kimerling, “Low-loss polycrystalline silicon waveguides for silicon photonics,” J. Appl. Phys. 80(11), 6120–6123 (1996). [CrossRef]
6. T. Tsuchizawa, K. Yamada, H. Fukuda, T. Watanabe, J. I. Takahashi, M. Takahashi, T. Shoji, E. Tamechika, S. I. Itabashi, and H. Morita, “Microphotonics devices based on silicon microfabrication technology,” IEEE J. Sel. Top. Quantum Electron. 11(1), 232–240 (2005). [CrossRef]
7. M. Melchiorri, N. Daldosso, F. Sbrana, L. Pavesi, G. Pucker, C. Kompocholis, P. Bellutti, and A. Lui, “Propagation losses of silicon nitride waveguides in the near-infrared range,” Appl. Phys. Lett. 86(12), 121111 (2005). [CrossRef]
8. R. Sun, K. McComber, J. Cheng, D. K. Sparacin, M. Beals, J. Michel, and L. C. Kimerling, “Transparent amorphous silicon channel waveguides with silicon nitride intercladding layer,” Appl. Phys. Lett. 94(14), 141108 (2009). [CrossRef]
9. A. Liu, R. Jones, L. Liao, D. Samara-Rubio, D. Rubin, O. Cohen, R. Nicolaescu, and M. Paniccia, “A high-speed silicon optical modulator based on a metal–oxide–semiconductor capacitor,” Nature 427(6975), 615 (2004). [CrossRef]
10. Y. H. Kuo, Y. K. Lee, Y. Ge, S. Ren, J. E. Roth, T. I. Kamins, D. A. B. Miller, and J. S. Harris, “Quantum-confined stark effect in Ge/SiGe quantum wells on Si for optical modulators,” IEEE J. Sel. Top. Quantum Electron. 12(6), 1503–1513 (2006). [CrossRef]
11. 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]
12. G. T. Reed, G. Mashanovich, F. Y. Gardes, and D. J. Thomson, “Silicon optical modulators,” Nat. Photonics 4(8), 518–526 (2010). [CrossRef]
13. S. Akiyama, T. Baba, M. Imai, T. Akagawa, M. Takahashi, N. Hirayama, H. Takahashi, Y. Noguchi, H. Okayama, T. Horikawa, and T. Usuki, “125-Gb/s operation with 029-V·cm V_πL using silicon Mach-Zehnder modulator based-on forward-biased pin diode,” Opt. Express 20(3), 2911 (2012). [CrossRef]
14. S. Luryi, A. Kastalsky, and J. C. Bean, “New infrared detector on a silicon chip,” IEEE Trans. Electron Devices 31(9), 1135–1139 (1984). [CrossRef]
15. M. Jutzi, M. Berroth, G. Wöhl, M. Oehme, and E. Kasper, “Ge-on-Si vertical incidence photodiodes with 39-GHz bandwidth,” IEEE Photonics Technol. Lett. 17(7), 1510–1512 (2005). [CrossRef]
16. L. Colace, M. Balbi, G. Masini, G. Assanto, H. C. Luan, and L. C. Kimerling, “Ge on Si p-i-n photodiodes operating at 10 Gbits,” Appl. Phys. Lett. 88(10), 101111 (2006). [CrossRef]
17. T. Yin, R. Cohen, M. M. Morse, G. Sarid, Y. Chetrit, D. Rubin, and M. J. Paniccia, “31 GHz Ge n-i-p waveguide photodetectors on Silicon-on-Insulator substrate,” Opt. Express 15(21), 13965–13971 (2007). [CrossRef]
18. J. Osmond, G. Isella, D. Chrastina, R. Kaufmann, M. Acciarri, and H. Von Känel, “Ultralow dark current Ge/Si(100) photodiodes with low thermal budget,” Appl. Phys. Lett. 94(20), 201106 (2009). [CrossRef]
19. L. Vivien, J. Osmond, J.-M. Fédéli, D. Marris-Morini, P. Crozat, J.-F. Damlencourt, E. Cassan, Y. Lecunff, and S. Laval, “42 GHz pin Germanium photodetector integrated in a silicon-on-insulator waveguide,” Opt. Express 17(8), 6252–6257 (2009). [CrossRef]
20. Y. Lin, K. H. 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 (2017). [CrossRef]
21. Y. Lin, B. Son, K. H. Lee, J. Michel, and C. S. Tan, “Sub-mA/cm2 Dark Current Density, Buffer-Less Germanium (Ge) Photodiodes on a 200-mm Ge-on-Insulator Substrate,” IEEE Trans. Electron Devices 68(4), 1730 (2021). [CrossRef]
22. Y. Lin, K. H. Lee, B. Son, and C. S. Tan, “Low-power and high-detectivity Ge photodiodes by in-situ heavy As doping during Ge-on-Si seed layer growth,” Opt. Express 29(3), 2940 (2021). [CrossRef]
23. H. Chen, P. Verheyen, P. De Heyn, G. Lepage, J. De Coster, S. Balakrishnan, P. Absil, W. Yao, L. Shen, G. Roelkens, and J. Van Campenhout, “−1 V bias 67 GHz bandwidth Si-contacted germanium waveguide p-i-n photodetector for optical links at 56 Gbps and beyond,” Opt. Express 24(5), 4622 (2016). [CrossRef]
24. H. Chen, M. Galili, P. Verheyen, P. De Heyn, G. Lepage, J. De Coster, S. Balakrishnan, P. Absil, L. Oxenlowe, J. Van Campenhout, and G. Roelkens, “100-Gbps RZ Data Reception in 67-GHz Si-Contacted Germanium Waveguide p-i-n Photodetectors,” J. Light. Technol. 35(4), 722 (2017). [CrossRef]
25. D. Benedikovic, L. Virot, G. Aubin, F. Amar, B. Szelag, B. Karakus, J.-M. Hartmann, C. Alonso-Ramos, X. Le Roux, P. Crozat, E. Cassan, D. Marris-Morini, C. Baudot, F. Boeuf, J.-M. Fédéli, C. Kopp, and L. Vivien, “25 Gbps low-voltage hetero-structured silicon-germanium waveguide pin photodetectors for monolithic on-chip nanophotonic architectures,” Photonics Res. 7(4), 437 (2019). [CrossRef]
26. S. J. Koester, C. L. Schow, L. Schares, G. Dehlinger, J. D. Schaub, F. E. Doany, and R. A. John, “Ge-on-SOI-detector/Si-CMOS-amplifier receivers for high-performance optical-communication applications,” J. Lightwave Technol.25(1), 46–57 (2007) Institute of Electrical and Electronics Engineers Inc.
27. L. Virot, D. Benedikovic, B. Szelag, C. Alonso-Ramos, B. Karakus, J.-M. Hartmann, X. Le Roux, P. Crozat, E. Cassan, D. Marris-Morini, C. Baudot, F. Boeuf, J.-M. Fédéli, C. Kopp, and L. Vivien, “Integrated waveguide PIN photodiodes exploiting lateral Si/Ge/Si heterojunction,” Opt. Express 25(16), 19487 (2017). [CrossRef]
28. Y. Zhang, S. Yang, Y. Yang, M. Gould, N. Ophir, A. E.-J. Lim, G.-Q. Lo, P. Magill, K. Bergman, T. Baehr-Jones, and M. Hochberg, “A high-responsivity photodetector absent metal-germanium direct contact,” Opt. Express 22(9), 11367 (2014). [CrossRef]
29. H. T. Chen, P. Verheyen, P. De Heyn, G. Lepage, J. De Coster, P. Absil, G. Roelkens, and J. Van Campenhout, “High-Responsivity Low-Voltage 28-Gb/s Ge p-i-n Photodetector With Silicon Contacts,” J. Light. Technol. 33(4), 820 (2015). [CrossRef]
30. L. Vivien, A. Polzer, D. Marris-Morini, J. Osmond, J. M. Hartmann, P. Crozat, E. Cassan, C. Kopp, H. Zimmermann, and J. M. Fédéli, “Zero-bias 40Gbit/s germanium waveguide photodetector on silicon,” Opt. Express 20(2), 1096 (2012). [CrossRef]
31. G. Li, Y. Luo, X. Zheng, G. Masini, A. Mekis, S. Sahni, H. Thacker, J. Yao, I. Shubin, K. Raj, J. E. Cunningham, and A. V. Krishnamoorthy, “Improving CMOS-compatible Germanium photodetectors,” Opt. Express 20(24), 26345 (2012). [CrossRef]
32. L. Virot, L. Vivien, J.-M. Fédéli, Y. Bogumilowicz, J.-M. Hartmann, F. Bœuf, P. Crozat, D. Marris-Morini, and E. Cassan, “High-performance waveguide-integrated germanium PIN photodiodes for optical communication applications [Invited],” Photonics Res. 1(3), 140 (2013). [CrossRef]
33. S. Lischke, D. Knoll, C. Mai, L. Zimmermann, A. Peczek, M. Kroh, A. Trusch, E. Krune, K. Voigt, and A. Mai, “High bandwidth, high responsivity waveguide-coupled germanium p-i-n photodiode,” Opt. Express 23(21), 27213 (2015). [CrossRef]
34. H. Chen, P. Verheyen, P. De Heyn, G. Lepage, J. De Coster, S. Balakrishnan, P. Absil, G. Roelkens, and J. Van Campenhout, “Dark current analysis in high-speed germanium p-i-n waveguide photodetectors,” J. Appl. Phys. 119(21), 213105 (2016). [CrossRef]
35. C. T. DeRose, D. C. Trotter, W. A. Zortman, A. L. Starbuck, M. Fisher, M. R. Watts, and P. S. Davids, “Ultra compact 45 GHz CMOS compatible Germanium waveguide photodiode with low dark current,” Opt. Express 19(25), 24897 (2011). [CrossRef]
36. H. Zhou and Y. Sun, “Size reduction of Ge-on-Si photodetectors via a photonic bandgap,” Appl. Opt. 57(12), 2962 (2018). [CrossRef]
37. D. Feng, S. Liao, P. Dong, N. N. Feng, H. Liang, D. Zheng, C. C. Kung, J. Fong, R. Shafiiha, J. Cunningham, A. V. Krishnamoorthy, and M. Asghari, “High-speed Ge photodetector monolithically integrated with large cross-section silicon-on-insulator waveguide,” Appl. Phys. Lett. 95(26), 261105 (2009). [CrossRef]
38. S. Klinger, M. Berroth, M. Kaschel, M. Oehme, and E. Kasper, “Ge-on-Si p-i-n photodiodes with a 3-dB bandwidth of 49 GHz,” IEEE Photonics Technol. Lett. 21(13), 920–922 (2009). [CrossRef]
39. T. Y. Liow, K. W. Ang, Q. Fang, J. F. Song, Y. Z. Xiong, M. Bin Yu, G. Q. Lo, and D. L. Kwong, “Silicon modulators and germanium photodetectors on SOI: Monolithic integration, compatibility, and performance optimization,” IEEE J. Sel. Top. Quantum Electron. 16(1), 307–315 (2010). [CrossRef]
40. L. Chen, P. Dong, and M. Lipson, “High performance germanium photodetectors integrated on submicron silicon waveguides by low temperature wafer bonding,” Opt. Express 16(15), 11513–11518 (2008). [CrossRef]
41. S. Assefa, F. Xia, S. W. Bedell, Y. Zhang, T. Topuria, P. M. Rice, and Y. A. Vlasov, “CMOS-integrated high-speed MSM germanium waveguide photodetector,” Opt. Express 18(5), 4986 (2010). [CrossRef]
42. C. Masini, L. Calace, G. Assanto, H.-C. Luan, and L. C. Kimerling, “High-performance p-i-n Ge on Si photodetectors for the near infrared: from model to demonstration,” IEEE Trans. Electron Devices 48(6), 1092–1096 (2001). [CrossRef]
43. 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]
44. D. Liang and J. E. Bowers, “Recent progress in lasers on silicon,” Nat. Photonics 4(8), 511–517 (2010). [CrossRef]
45. A. W. Fang, H. Park, O. Cohen, R. Jones, M. J. Paniccia, and J. E. Bowers, “Electrically pumped hybrid AlGaInAs-silicon evanescent laser,” Opt. Express 14(20), 9203 (2006). [CrossRef]
46. B. Corbett, C. Bower, A. Fecioru, M. Mooney, M. Gubbins, and J. Justice, “Strategies for integration of lasers on silicon,” Semicond. Sci. Technol. 28(9), 094001 (2013). [CrossRef]
47. J. Liu, X. Sun, D. Pan, X. Wang, L. C. Kimerling, T. L. Koch, and J. Michel, “Tensile-strained, n-type Ge as a gain medium for monolithic laser integration on Si,” Opt. Express 15(18), 11272 (2007). [CrossRef]
48. M. Virgilio, C. L. Manganelli, G. Grosso, G. Pizzi, and G. Capellini, “Radiative recombination and optical gain spectra in biaxially strained n-type germanium,” Phys. Rev. B 87(23), 235313 (2013). [CrossRef]
49. M. El Kurdi, T. Kociniewski, T. P. Ngo, J. Boulmer, D. Dbarre, P. Boucaud, J. F. Damlencourt, O. Kermarrec, and D. Bensahel, “Enhanced photoluminescence of heavily n -doped germanium,” Appl. Phys. Lett. 94(19), 191107 (2009). [CrossRef]
50. M. Oehme, M. Gollhofer, D. Widmann, M. Schmid, M. Kaschel, E. Kasper, and J. Schulze, “Direct bandgap narrowing in Ge LED’s on Si substrates,” Opt. Express 21(2), 2206 (2013). [CrossRef]
51. X. Sun, J. Liu, L. C. Kimerling, and J. Michel, “Direct gap photoluminescence of n -type tensile-strained Ge-on-Si,” Appl. Phys. Lett. 95(1), 011911 (2009). [CrossRef]
52. T. Okumura, K. Oda, J. Kasai, M. Sagawa, and Y. Suwa, “Optical net gain measurement in n-type doped germanium waveguides under optical pumping for silicon monolithic laser,” Opt. Express 24(9), 9132 (2016). [CrossRef]
53. Y. Suwa, “First-principles study of light emission from strained germanium,” in 10th International Conference on Group IV Photonics (IEEE, 2013).
54. M. El Kurdi, G. Fishman, Ś. Sauvage, and P. Boucaud, “Band structure and optical gain of tensile-strained germanium based on a 30 band kp formalism,” J. Appl. Phys. 107(1), 013710 (2010). [CrossRef]
55. K. Tani, K. Oda, and T. Ido, “Analysis of stress distribution in microfabricated germanium with external stressors for enhancement of light emission,” Opt. Express 28(25), 38267 (2020). [CrossRef]
56. 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]
57. S. Bao, D. Kim, C. Onwukaeme, S. Gupta, K. Saraswat, K. H. Lee, Y. Kim, D. Min, Y. Jung, H. Qiu, H. Wang, E. A. Fitzgerald, C. S. Tan, and D. Nam, “Low-threshold optically pumped lasing in highly strained germanium nanowires,” Nat. Commun. 8(1), 1845 (2017). [CrossRef]
58. F. T. Armand Pilon, A. Lyasota, Y. M. Niquet, V. Reboud, V. Calvo, N. Pauc, J. Widiez, C. Bonzon, J. M. Hartmann, A. Chelnokov, J. Faist, and H. Sigg, “Lasing in strained germanium microbridges,” Nat. Commun. 10(1), 2724 (2019). [CrossRef]
59. A. Elbaz, M. El Kurdi, A. Aassime, S. Sauvage, X. Checoury, I. Sagnes, C. Baudot, F. Boeuf, and P. Boucaud, “Germanium microlasers on metallic pedestals,” APL Photonics 3(10), 106102 (2018). [CrossRef]
60. R. E. Camacho-Aguilera, Y. Cai, N. Patel, J. T. Bessette, M. Romagnoli, L. C. Kimerling, and J. Michel, “An electrically pumped germanium laser,” Opt. Express 20(10), 11316–11320 (2012). [CrossRef]
61. R. Koerner, M. Oehme, M. Gollhofer, M. Schmid, K. Kostecki, S. Bechler, D. Widmann, E. Kasper, and J. Schulze, “Electrically pumped lasing from Ge Fabry-Perot resonators on Si,” Opt. Express 23(11), 14815–14822 (2015). [CrossRef]
62. X. Sun, J. Liu, L. C. Kimerling, and J. Michel, “Room-temperature direct bandgap electroluminesence from Ge-on-Si light-emitting diodes,” Opt. Lett. 34(8), 1198–1200 (2009). [CrossRef]
63. S.-L. Cheng, J. Lu, G. Shambat, H.-Y. Yu, K. Saraswat, J. Vuckovic, and Y. Nishi, “Room temperature 16 μm electroluminescence from Ge light emitting diode on Si substrate,” Opt. Express 17(12), 10019–10024 (2009). [CrossRef]
64. D. Nam, D. Sukhdeo, S. L. Cheng, A. Roy, K. Chih-Yao Huang, M. Brongersma, Y. Nishi, and K. Saraswat, “Electroluminescence from strained germanium membranes and implications for an efficient Si-compatible laser,” Appl. Phys. Lett. 100(13), 131112 (2012). [CrossRef]
65. R. People and J. C. Bean, “Calculation of critical layer thickness versus lattice mismatch for Ge x Si 1− x /Si strained-layer heterostructures,” Appl. Phys. Lett. 47(3), 322–324 (1985). [CrossRef]
66. L. M. Giovane, H. C. Luan, A. M. Agarwal, and L. C. Kimerling, “Correlation between leakage current density and threading dislocation density in SiGe p-i-n diodes grown on relaxed graded buffer layers,” Appl. Phys. Lett. 78(4), 541–543 (2001). [CrossRef]
67. G. Wang, R. Loo, E. Simoen, L. Souriau, M. Caymax, M. M. Heyns, and B. Blanpain, “A model of threading dislocation density in strain-relaxed Ge and GaAs epitaxial films on Si (100),” Appl. Phys. Lett. 94(10), 102115 (2009). [CrossRef]
68. K. Tani, K. Oda, M. Deura, and T. Ido, “Enhanced room-temperature electroluminescence from a germanium waveguide on a silicon-on-insulator diode with a silicon nitride stressor,” Opt. Express 29(3), 3584 (2021). [CrossRef]
69. Y. Fu, T. Ye, W. Tang, and T. Chu, “Efficient adiabatic silicon-on-insulator waveguide taper,” Photonics Res. 2(3), A41 (2014). [CrossRef]
70. T. Ye, Y. Fu, L. Qiao, and T. Chu, “Low-crosstalk Si arrayed waveguide grating with parabolic tapers,” Opt. Express 22(26), 31899 (2014). [CrossRef]
71. P. Sethi, A. Haldar, and S. K. Selvaraja, “Ultra-compact low-loss broadband waveguide taper in silicon-on-insulator,” Opt. Express 25(9), 10196 (2017). [CrossRef]
72. J. Zhang, J. Yang, H. Xin, J. Huang, D. Chen, and Z. Zhaojian, “Ultrashort and efficient adiabatic waveguide taper based on thin flat focusing lenses,” Opt. Express 25(17), 19894 (2017). [CrossRef]
73. K. Oda, K. Tani, S. Saito, and T. Ido, “Improvement of crystallinity by post-annealing and regrowth of Ge layers on Si substrates,” Thin Solid Films 550, 509 (2014). [CrossRef]
74. A. Dimoulas, P. Tsipas, A. Sotiropoulos, and E. K. Evangelou, “Fermi-level pinning and charge neutrality level in germanium,” Appl. Phys. Lett. 89(25), 252110 (2006). [CrossRef]
75. P. Tsipas and A. Dimoulas, “Modeling of negatively charged states at the Ge surface and interfaces,” Appl. Phys. Lett. 94(1), 012114 (2009). [CrossRef]
76. Y. Zhou, W. Han, Y. Wang, F. Xiu, J. Zou, R. K. Kawakami, and K. L. Wang, “Investigating the origin of Fermi level pinning in Ge Schottky junctions using epitaxially grown ultrathin MgO films,” Appl. Phys. Lett. 96(10), 102103 (2010). [CrossRef]
77. A. Delabie, F. Bellenger, M. Houssa, T. Conard, S. Van Elshocht, M. Caymax, M. Heyns, and M. Meuris, “Effective electrical passivation of Ge(100) for high- k gate dielectric layers using germanium oxide,” Appl. Phys. Lett. 91(8), 082904 (2007). [CrossRef]
78. D. Kuzum, T. Krishnamohan, A. J. Pethe, A. K. Okyay, Y. Oshima, Y. Sun, J. P. McVittie, P. A. Pianetta, P. C. McIntyre, and K. C. Saraswat, “Ge-interface engineering with ozone oxidation for low interface-state density,” IEEE Electron Device Lett. 29(4), 328–330 (2008). [CrossRef]
79. H. Matsubara, T. Sasada, M. Takenaka, and S. Takagi, “Evidence of low interface trap density in Ge O2 Ge metal-oxide- semiconductor structures fabricated by thermal oxidation,” Appl. Phys. Lett. 93(3), 032104 (2008). [CrossRef]
80. S. Kako, K. Oda, T. Ido, and Y. Arakawa, “Excess carrier lifetime in epitaxially grown layers of germanium on silicon,” in 2015 IEEE 12th International Conference on Group IV Photonics (GFP) (IEEE, 2015).
81. C. G. Van de Walle and R. M. Martin, “Theoretical calculations of heterojunction discontinuities in the Si/Ge system,” Phys. Rev. B 34(8), 5621–5634 (1986). [CrossRef]
