Open Access
Add to BibSonomy
Abstract
The multiple-quantum-well diode (MQW-diode) inherently exhibits simultaneous behavior because of the overlap between the emission spectra and spectral responsivity of the MQW-diode. This feature makes it feasible to form a full-duplex light communication system when two identical MQW-diodes separately function as a transmitter and a receiver at the same time. To verify spatial full-duplex light communication, we fabricated and characterized a monolithic multicomponent system by integrating two InGaN waveguide-based MQW-diodes into a single chip. A 5-μm-wide air gap between two MQW-diodes was manufactured for precise alignment, which could yield spatial light transmission and coupling. Spatial co-time co-frequency full-duplex (CCFD) light communication was experimentally demonstrated using the monolithic multicomponent system, a self-interference cancellation scheme was used to extract the superimposed signals, and a full-duplex audio transmission experiment was performed, opening a promising route toward parallel information processing via free space based on the simultaneous light-emitting and light-detecting phenomenon of the MQW-diode.
© 2019 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Transceivers transmit and receive signals simultaneously in the same band in co-time co-frequency full-duplex (CCFD) communication, which can double the spectral efficiency compared with that of conventional half-duplex communication [1–5]. However, for radio frequency (RF) communication, realizing CCFD communication is an enormous challenge because significant diffraction and reflection in electromagnetic wave propagation generate overwhelming self-interference, which is billions of times stronger (100 dB+) than the useful signals [6,7]. Several studies have adopted various ways to consider the problem of self-interference cancellation. Elsayed Ahmed et al. proposed a novel digital self-interference cancellation technique for full-duplex systems that can significantly mitigate the self-interference signals and the associated transmitter and receiver impairments [8]. Taneli Riihonen et al. reported new baseband signal processing techniques to effectively mitigate loopback self-interference in full-duplex multiple-input multiple-output (MIMO) relays [9]. Dani Korpi et al. proposed a novel method for compensating the image component of the self-interference signal in a direct-conversion full-duplex transceiver [10]. However, expensive and complex devices, such as high-precision analog-to-digital (AD) converters and low-noise analog units, which may be unrealistic for cost-sensitive applications, are required in such systems.
The use of a visible-light communication (VLC) transceiver is another promising way to achieve a CCFD communication system. In our previous work, in-plane VLC systems operating in half-duplex [11] and full-duplex [12, 13] mode were demonstrated, in which suspended membrane p-n junction InGaN/GaN multiple-quantum-wells (MQWs) devices act as transceivers and suspended GaN waveguides serve as communication channels. In this study, we proposed and fabricated a monolithic multicomponent system consisting of two non-suspended waveguide-based multiple-quantum-well diodes (MQW-diodes) on a 2 in. GaN-on-Si platform using a wafer-level procedure [14,15]. Due to the light confinement of the waveguide structure, silicon removal and back wafer etching [16,17] were avoided, and a substantial proportion of photons are trapped and propagated inside the devices which is beneficial for the application of an on-chip spatial VLC. Thus, non-suspended MQW-diodes can be used as light signal transceivers and communicate with each other via free space. The CCFD communication function of the monolithic multicomponent system was experimentally demonstrated.
2. Fabrication of the monolithic multicomponent system
Figure 1(a) shows a cross-sectional high-angle annular dark-field (HAADF) scanning transmission electron microscopy (TEM) image of a GaN-on-Si MQW-diode wafer. The top GaN epitaxy layers, including a 38-nm-thick highly Mg-doped p-type GaN layer, a 600-nm-thick Mg-doped AlGaN cladding layer, an 80-nm-thick GaN waveguide layer, a 40-nm-thick InGaN MQW active region, a 90-nm-thick Si-doped GaN waveguide layer, a 1.5-μm-thick Si-doped AlGaN cladding layer, a 3-μm-thick Si-doped n-type GaN layer, 490-nm-thick Al-composition step-graded AlGaN multilayers consisting of a 180-nm-thick Al0.35Ga0.65N layer and a 310-nm-thick Al0.17Ga0.83N layer, and a 280-nm-thick AlN nucleation layer were grown directly on a Si(111) substrate by metal-organic chemical vapor phase deposition [18]. Figure 1(b) shows an enlarged TEM image of the InGaN MQW active region marked by a blue rectangle in Fig. 1 (a). The InGaN MQW active region with sharp interfaces is sandwiched by two GaN waveguides and two AlGaN cladding layers, as demonstrated in Figs. 1(a) and 1(b).
Fig. 1 (a) Cross-sectional high-angle annular dark-field (HAADF) scanning TEM image of an MQW-diode grown on Si. (b) Enlarged TEM image of the InGaN MQW active region marked by a blue rectangle in (a).
Figure 2(a) schematically illustrates the detailed fabrication process flow of the proposed monolithic multicomponent system. First, the top layer was defined by photolithography, and two isolation mesas for diodes were etched down to the GaN waveguide layer using inductively coupled plasma reactive ion etching (ICP-RIE) with Cl2 and BCl3 hybrid gases. Second, p-electrode contact layers with 20 nm Ni/90 nm Au bilayers were deposited onto the isolation mesas, and then lift-off and rapid thermal annealing processes were conducted. Third, the diodes were defined using photolithography and etched down to the Si-doped n-type GaN layer to a depth of 3 μm. Fourth, a 200-nm-thick SiO2 layer was grown for electrical isolation by plasma-enhanced chemical vapor deposition (PECVD). Finally, p-type and n-type metal electrodes with 50/100/500 nm Ti/Pt/Au metallization stacks were formed using a combination of electron beam evaporation and lift-off process. The absorption of downward-emitting light by the silicon substrate is avoided because most emission photons are confined by the cladding and the waveguide layers. Thus, there is no need to remove the silicon substrate for MQW-diodes, unlike in the fabrication of GaN-on-Si light-emitting diodes (LEDs) [19–21]. Figure 2(b) demonstrates a scanning electron microscopy (SEM) image of the monolithic multicomponent system. The rectangle p-type mesas of MQW-diode A and MQW-diode B are 800 μm × 36 μm and 200 μm μm × 36 μm, respectively, which feature 10-μm gaps to the n-electrode. Two circle bonding mesas measuring 85 μm in diameter are connected to the rectangle p-type mesas. The p-electrodes and the p-type GaN layer are isolated by the SiO2 layer, except for a 4-μm-wide metal line. Figure 2(c) shows a three-dimensional AFM image of the gap between devices. The ridge is 6 μm wide. The p-contact located on the ridge is 4 μm wide. The p-electrode which covers the ridge and p-contact is 19 μm wide. Two MQW-diodes are separated by a gap of 5 μm. The etching depth of the gap is approximately 3.6 μm. Different mesa sizes were adopted to make it easy to distinguish the two MQW-diodes.
Fig. 2 (a) Schematic fabrication process flow of the monolithic multicomponent system. (b) SEM image of the monolithic multicomponent system. (c) AFM image of the gap between two MQW-diodes.
3. Performance of the MQW-diodes
Figure 3 illustrates the electrical and optical characteristics of the proposed MQW-diodes. An Agilent B1500A semiconductor device analyzer was used to measure the current-voltage (I–V) characteristics. Rectifying I–V behavior of MQW-diode A with a turn-on voltage of 2.5 V was observed with forward-bias conditions corresponding to the expected bias polarity, as shown in Fig. 3(a). A leakage current of 1.29 nA was measured at −2 V, and the current increased rapidly with increasing applied bias voltage when the applied bias voltage was higher than the turn-on voltage. The inset in Fig. 3(a) demonstrates MQW-diode A’s electroluminescence spectra in the 400–520 nm range. The light emitted from MQW-diode A’s facet was directly captured by a custom-designed fiber that can be controlled with a three-dimensional stage. When MQW-diode A emitted light, some emitted light was coupled into a 300-μm-diameter multimode fiber and then detected by an Ocean Optics HR4000 spectrometer at the other end. The light intensity increased monotonically with the current injection level. The dominant emission peak, as a function of the excitation of the InGaN MQW active layers, was measured at 458 nm and remained stable as the injection current increased from 1.0 mA to 3.0 mA. The electrical and optical characteristics of MQW-diode B are similar to those of MQW-diode A, as shown in Fig. 3(b). Due to the difference in size, MQW-diode A exhibits higher current and light intensity.
Fig. 3 I–V characteristics of the proposed MQW-diodes measured at room temperature: (a) MQW-diode A. (b) MQW-diode B. The two insets show measured EL spectra of MQW-diode A and MQW-diode B under different current injections.
4. Response of the MQW-diodes
GaN-based MQW-diodes can emit light under current injection and induce photocurrent under light irradiation, which means they are capable of providing multiple functions [22,23]. Due to the intrinsic characteristics of the InGaN MQW active layer, MQW-diodes can exhibit broadband emission and sense light with a short wavelength spanning half of its emission spectrum [24]. When the applied voltage of MQW-diode is higher than its turn-on voltage, the MQW-diode operates in emitter mode. And when the applied voltage is zero bias or negative bias, the MQW-diode operates in detector mode. As a light detector, the MQW-diode usually detects incident light under zero bias or negative bias. Interestingly, MQW-diodes can still sense light when they operate in emitter mode. Four-terminal measurements were conducted to characterize the photoresponse ability of the proposed MQW-diode using a combination of an Agilent B1500A semiconductor device analyzer and a Cascade PM5 probe station. When MQW-diode A emits light, some photons reach MQW-diode B via free space, and MQW-diode B can detect them. These photons are absorbed, and photogenerated electron-hole pairs are formed. The electrons move to the n-electrode, while the holes flow to the p-electrode, which leads to light-power-dependent photocurrent in MQW-diode B. The log-scaled induced photocurrent of MQW-diode B arises in both detector mode and emitter mode, with MQW-diode A operating in emitter mode at different injection current levels ranging from 2 mA to 4 mA, as illustrated in Fig. 4(a). The measured current of MQW-diode B is the sum of the driven current caused by the applied voltage and the photocurrent induced by light irradiation from MQW-diode A. The direction of the driven current is opposite that of the photocurrent. The photocurrent can be calculated by subtracting the measured current values of MQW-diode B at a forward current of 0 mA for MQW-diode A from the other measured current values. The light-emission images of the fabricated devices are shown in the two insets of Fig. 4(a), from which it can be seen that MQW-diode B can sense the light emitted from MQW-diode A when it operates in both detector and emitter mode. The induced photocurrent of MQW-diode B improves with increasing injection current for MQW-diode A because MQW-diode B can absorb more photons supplied by MQW-diode A. Below the turn-on voltage of MQW-diode B, the induced photocurrent gradually decreases with increasing applied voltage. Near the turn-on voltage, the driven current is established due to the applied voltage, leading to a sharp change in the induced photocurrent. Above the turn-on voltage, the induced photocurrent increases rapidly with the increase in the applied voltage over the measurement range. The photocurrent measured under a forward voltage of 4 V is more than 10000 times stronger than the photocurrent at −2 V, indicating that the barrier is managed to form a channel that is beneficial for photon-electron conversion when MQW-diode B operates in emitter mode. The experimental results obtained for MQW-diode A are similar to those obtained for MQW-diode B, which shows that the MQW-diode can simultaneously emit and detect light [24]. In addition, the coupling efficiency between two non-suspended MQW-diodes can be effectively improved and communication distance can be increased when GaN-based laser diodes are employed. Great progresses have been made in GaN-based laser diodes on silicon [18, 25, 26], which opens a promising way towards monolithic laser communication system.
Fig. 4 (a) Log-scaled photocurrent plots for MQW-diode B with different MQW-diode A injection current levels. The insets show the light-emission images of the fabricated devices. (b) Induced photocurrent versus laser power. The inset shows the zoom-in figure of photocurrent measured when MQW-diode operating in detector mode.
To further investigate the photoelectric response characteristics of the MQW-diodes operating in detector and emitter mode, the induced photocurrent was measured using an Agilent B1500A semiconductor device analyzer when the MQW-diodes were illuminated by a 410-nm laser beam, as demonstrated in Fig. 4(b). The 410-nm laser beam was generated by a commercial low noise violet diode laser whose maximum power reaches 300 mW. Photocurrents of 0.127 mA, 0.958 mA and 1.78 mA were measured at laser powers of 13.14 mW, 73.8 mW and 105.6 mW when MQW-diode A at a forward voltage of 0 V, respectively. The inset of Fig. 4(b) shows the zoom-in figure of MQW-diode A’s photocurrent. At the same laser power, the photocurrent increased to 8.68 mA, 16 mA and 23.9 mA when the MQW-diode operated at 3 V. The photocurrents induced by 410-nm laser beam of MQW-diode B are similar to MQW-diode A. These results also show that the MQW-diodes are capable of simultaneously realizing light emission and light detection when bias voltage loading and light irradiation concur. In addition, the MQW-diodes exhibit improved light detection when they operate in emitter mode.
The spatial light communication functions of the proposed photonic communication system were investigated. MQW-diode A was directly driven to emit modulated light at a rate of 100 kHz with an offset voltage of 2 V and a peak-to-peak voltage of 4 V by an Agilent 33522A arbitrary waveform generator (AWG). MQW-diode B detected light signals and generated electron-hole pairs, leading to an induced voltage that was directly characterized by a digital storage oscilloscope (DSO) without any amplification circuits. The amplitude of the received signals, which was due to the bias voltage of the MQW-diode B, was relatively weak when MQW-diode B operated at a forward voltage of 0 V, as demonstrated in Fig. 5(a). In this case, the peaks and troughs are not easy to distinguish. With an increase in applied voltage to 3 V, MQW-diode B emitted and detected light at the same time, and the amplitudes of the received signals reached 4.83 mV, as shown in Fig. 5(b). Peaks and troughs are sufficiently clear despite some attenuation, indicating better spatial communication performance using visible light. When the voltage was increased to 4.5 V, the spatial communication performance of the monolithic multicomponent system was further improved. The amplitudes of the received signals were increased to 6.78 mV, as illustrated in Fig. 5(c). Figure 5(d) shows the amplitudes of the received signals versus the different applied voltages of MQW-diode B from 0 V to 5.4 V. Below the turn-on voltage of 2.5 V, the increase in the applied voltage slightly influences the amplitudes of the received signals. Near the turn-on voltage, with increasing applied voltage, the amplitudes of the received signals increases sharply before reaching a relatively stable region. The same experiments were conducted on MQW-diode A, and similar results were observed. These results show that the MQW-diode can sense light in either detector or emitter mode, and the detection sensitivity of the MQW-diode operating in emitter mode is higher than that in detector mode, indicating the possibility of spatial full-duplex communication using visible light on a single chip.
Fig. 5 The amplitude of the received signals versus the applied voltage of MQW-diode B: (a) 0 V; (b) 3 V; and (c) 4.5 V. (d) The amplitude of the received signals versus the different applied voltages of MQW-diode B from 0 V to 5.4 V.
5. Full-duplex light communication and signal extraction
Furthermore, spatial full-duplex communication of the proposed monolithic multicomponent system using the same frequency as a function of the simultaneous light-emitting and light-detecting ability of the MQW-diodes was investigated. Figure 6 demonstrates that the monolithic photonic communication system operates in full-duplex mode at the same frequency. The two MQW-diodes were simultaneously driven to emit modulated light signals at 100 kHz by an Agilent 33522A arbitrary waveform generator (AWG). MQW-diode A was modulated with an offset voltage of 2 V and a peak-to-peak voltage of 4 V, and MQW-diode B was modulated with an offset voltage of 4 V and a peak-to-peak voltage of 50 mV, as illustrated in Figs. 6(a) and 6(b). The fill factors of the two light signals were both 0.5. When the MQW-diode emits modulated light signals, it can also detect the light signals from another MQW-diode. Thus, the superimposed signals that are mixed together by the transmitted and received light signals can be observed, as demonstrated in Fig. 6(c).
Fig. 6 (a) The transmitted signals from MQW-diode A. (b) The transmitted signals from MQW-diode B. (c) The superimposed signals from MQW-diode B.
The signals received by MQW-diode B can be extracted from the superimposed signals by the self-interference cancellation method (SICM) [27–29]. As shown in Fig. 7, at side A, the received superimposed signals SAB contain the transmitted signals SA, the transmitted signals of side B SB and noise. The SICM at side A estimates the signals SB by removing its own signals SA from SAB. And side B does the same operation to estimate the transmitted signals of side A. Due to signal attenuation and the presence of noise, the superimposed signals are distorted, which cause the rising edges and the trailing edges of the signals to be less steep than those of the transmitted signals in Fig. 8(a). In addition, the bandwidth slightly reduced because device capacitance increased with increasing the forward bias voltage, which also caused the less steeper edge transitions to some extent. Accordingly, the extracted signals are not reliable at the edges, and impulse noise is observed, as shown in Fig. 8(b). To solve this problem, the received signals extracted by the SICM were further denoised and filtered using the median filter. The Median Filter is a nonlinear digital filtering technique which can be used to remove noise from a signal. It run through the signal entry by entry and replaces each entry with the median of neighboring entries. In this way, noises can be reduced and results of later processing can be improved. Figure 8(c) demonstrates the received signals extracted by the optimized SICM, where the instability and impulse noise are effectively mitigated.
Fig. 8 (a) The transmitted signals from MQW-diode A. (b) The extracted signals from MQW-diode B using the SICM. (c) The extracted signals from MQW-diode B using the median filter.
6. Full-duplex audio transmission
To further investigate the spatial full-duplex VLC performance of the monolithic multicomponent system, we performed an audio transmission experiment (please email the author for the experiment video via caiw@njit.edu.cn). A schematic of the spatial full-duplex audio transmission is shown in Fig. 9(a). The amplified audio signals are loaded by a bias-tee circuit to drive the MQW-diode A to emit the modulated light. MQW-diode B detects modulated light signals via free space and completes photo-electron conversion. The received signals are then amplified, and a filter is used to recover the audio signals. At the same time, MQW-diode B is driven by an Agilent 33522A AWG to emit square wave signals. The square wave signals can also be sensed by the MQW-diode A. In addition, spatial full-duplex audio transmission performance of the system was further investigated when the two MQW-diodes were driven by audio signal and sine signal at the same time, respectively. Figure 9(b) demonstrates the transmitted audio signals from MQW-diode A and the superimposed signals of the received and transmitted audio signals from MQW-diode B. Peaks and troughs of the superimposed signals are clear, indicating good duplex audio communication performance of the spatial full-duplex VLC system. The experimental results indicate that the monolithic multicomponent system is capable of realizing parallel information processes via the simultaneous light-emitting and light-detecting ability of the MQW-diode.
Fig. 9 (a) Schematic of spatial full-duplex audio transmission using visible light. (b) Audio signals of spatial full-duplex VLC system.
7. Conclusion
A monolithic multicomponent system consisting of two MQW-diodes was implemented on a GaN-on-Si platform. The two MQW-diodes share the same InGaN/GaN MQW structure and communicate with each other via free space. Due to the light confinement of the cladding and the waveguide layers, silicon removal is avoided, and the substrate can remain intact. Based on the simultaneous light-emitting and light-detecting ability of the MQW-diodes, spatial CCFD communication using visible light was demonstrated, and useful signals were extracted using a combination of the SICM and the median filter. Furthermore, full-duplex audio signal transmission was achieved, which indicates that the monolithic multicomponent system can contribute to the development of CCFD VLC systems, particularly the development of portable optical communication systems.
Funding
Ministry of Science and Technology of the People’s Republic of China Special Project for Inter-government Collaboration of State Key Research and Development Program (2016YFE0118400); Natural Science Foundation of Jiangsu Province (BE2016186); National Natural Science Foundation of China (61322112, 61403188, 61531166004); Nanjing Institute of Technology Research Project (CKJA201705, CKJA201708), Ministry of Science and Technology of the People’s Republic of China (KYZZ16_0258); Ministry of Education of the People’s Republic of China (“111” project).
References
1. Z. Tong and M. Haenggi, “Throughput analysis for full-duplex wireless networks with imperfect self-interference cancellation,” IEEE Transactions on Commun. 63, 4490–4500 (2015). [CrossRef]
2. Z. Zhang, X. Chai, K. Long, and A. V. Vasilakos, “Full duplex techniques for 5g networks: self-interference cancellation, protocol design, and relay selection,” IEEE Commun. Mag. 53, 128–137 (2015). [CrossRef]
3. X. Y. Wu, Y. Shen, Y. X. Tang, and S. C. Xiao, “The power delay profile of the single antenna full-duplex self-interference channel in indoor environments at 2.4 ghz,” in Applied Mechanics and Materials, (Trans Tech Publ, 2015).
4. E. Everett, M. Duarte, C. Dick, and A. Sabharwal, “Empowering full-duplex wireless communication by exploiting directional diversity,” in Signals, Systems and Computers (ASILOMAR), 2011 Conference Record of the Forty Fifth Asilomar Conference on, (IEEE, 2011), pp. 2002–2006. [CrossRef]
5. E. Everett, A. Sahai, and A. Sabharwal, “Passive self-interference suppression for full-duplex infrastructure nodes,” IEEE Transactions on Wirel. Commun. 13, 680–694 (2014). [CrossRef]
6. A. Goldsmith, Wireless Communications (Cambridge University, 2005). [CrossRef]
7. S. Hong, J. Brand, J. Choi, M. Jain, J. Mehlman, S. Katti, and P. Levis, “Applications of self-interference cancellation in 5g and beyond,” IEEE Commun. Mag. 52, 114–121 (2014). [CrossRef]
8. E. Ahmed and A. M. Eltawil, “All-digital self-interference cancellation technique for full-duplex systems,” IEEE Transactions on Wirel. Commun. 14, 3519–3532 (2015). [CrossRef]
9. T. Riihonen, S. Werner, and R. Wichman, “Mitigation of loopback self-interference in full-duplex mimo relays,” IEEE Transactions on Signal Process. 59, 5983–5993 (2011). [CrossRef]
10. D. Korpi, L. Anttila, V. Syrjala, and M. Valkama, “Widely linear digital self-interference cancellation in direct-conversion full-duplex transceiver,” IEEE J. Sel. Areas Commun. 32, 1674–1687 (2014). [CrossRef]
11. W. Cai, Y. Yang, X. Gao, J. Yuan, W. Yuan, H. Zhu, and Y. Wang, “On-chip integration of suspended InGaN/GaN multiple-quantum-well devices with versatile functionalities,” Opt. Express 24, 6004–6010 (2016). [CrossRef] [PubMed]
12. Y. Wang, Y. Xu, Y. Yang, X. Gao, B. Zhu, W. Cai, J. Yuan, R. Zhang, and H. Zhu, “Simultaneous light emission and detection of InGaN/GaN multiple quantum well diodes for in-plane visible light communication,” Opt. Commun. 387, 440–445 (2017). [CrossRef]
13. Y. Wang, X. Wang, B. Zhu, Z. Shi, J. Yuan, X. Gao, Y. Liu, X. Sun, D. Li, and H. Amano, “Full-duplex light communication with a monolithic multicomponent system,” Light. Sci. Appl. 7, 83 (2018). [CrossRef] [PubMed]
14. X. Gao, J. Yuan, Y. Yang, S. Zhang, Z. Shi, X. Li, and Y. Wang, “InGaN directional coupler made with a one-step etching technique,” Semicond. Sci. Technol. 32, 065002 (2017). [CrossRef]
15. J. Yuan, W. Cai, X. Gao, G. Zhu, D. Bai, H. Zhu, and Y. Wang, “Monolithic integration of a suspended light-emitting diode with a y-branch structure,” Appl. Phys. Express 9, 032202 (2016). [CrossRef]
16. Y. Wang, G. Zhu, W. Cai, X. Gao, Y. Yang, J. Yuan, Z. Shi, and H. Zhu, “On-chip photonic system using suspended p-n junction InGaN/GaN multiple quantum wells device and multiple waveguides,” Appl. Phys. Lett. 108, 966–968 (2016). [CrossRef]
17. W. Cai, X. Gao, W. Yuan, Y. Yang, J. Yuan, H. Zhu, and Y. Wang, “Integrated p-n junction InGaN/GaN multiple-quantum-well devices with diverse functionalities,” Appl. Phys. Express 9, 052204 (2016). [CrossRef]
18. Y. Sun, K. Zhou, Q. Sun, J. Liu, M. Feng, Z. Li, Y. Zhou, L. Zhang, D. Li, and S. Zhang, “Room-temperature continuous-wave electrically injected InGaN-based laser directly grown on Si,” Nat. Photonics 10, 595 (2016). [CrossRef]
19. X. Li, Z. Shi, G. Zhu, M. Zhang, H. Zhu, and Y. Wang, “High efficiency membrane light emitting diode fabricated by back wafer thinning technique,” Appl. Phys. Lett. 105, 2211–2213 (2014).
20. S. Qian, Y. Wei, M. Feng, Z. Li, F. Bo, H. Zhao, and Y. Hui, “GaN-on-Si blue/white leds: epitaxy, chip, and package,” J. Semicond. 37, 61–68 (2016).
21. D. Zhu, C. Mcaleese, M. Häberlen, C. Salcianu, T. Thrush, M. Kappers, A. Phillips, P. Lane, M. Kane, and D. Wallis, “Efficiency measurement of GaN-based quantum well and light-emitting diode structures grown on silicon substrates,” J. Appl. Phys. 109, L492 (2011). [CrossRef]
22. S. Nakamura, “The roles of structural imperfections in InGaN-Based blue light-emitting diodes and laser diodes,” Science 281, 955–961 (1998). [CrossRef] [PubMed]
23. Y. Nanishi, “Nobel prize in physics: The birth of the blue led,” Nat. Photonics 8, 884–886 (2014). [CrossRef]
24. Y. Jiang, Z. Shi, S. Zhang, J. Yuan, Z. Hu, X. Shen, B. Zhu, and Y. Wang, “Simultaneous light-emitting light-detecting functionality of InGaN/GaN multiple quantum well diodes,” IEEE Electron Device Lett. 38, 1684–1687 (2017). [CrossRef]
25. M. Feng, Z. Li, J. Wang, R. Zhou, Q. Sun, X. Sun, D. Li, H Gao, Y. Zhou, S. Zhang, D. Li, L. Zhang, J. Liu, H. Wang, M. Ikeda, X. Zheng, and H. Yang, “Room-temperature electrically injected AlGaN-based near-ultraviolet laser grown on Si,” ACS Photonics 5, 699–704 (2018).
26. M. Feng, J. Wang, R. Zhou, Q. Sun, H. Gao, Y. Zhou, J. Liu, Y. Huang, S. Zhang, M. Ikeda, H. Wang, Y. Zhang, Y. Wang, and Y. Hui, “On-chip integration of GaN-based laser, modulator, and photodetector grown on Si,” IEEE J. Sel. Top. Quantum Electron. 24, 1–5 (2018).
27. Z. Zhang, X. Chai, K. Long, and A. V. Vasilakos, “Full duplex techniques for 5g networks: self-interference cancellation, protocol design, and relay selection,” IEEE Commun. Mag. 53, 128–137 (2015). [CrossRef]
28. S. Hong, J. Brand, J. Choi, M. Jain, J. Mehlman, S. Katti, and P. Levis, “Applications of self-interference cancellation in 5g and beyond,” IEEE Commun. Mag. 52, 114–121 (2014). [CrossRef]
29. D. Bharadia, E. Mcmilin, and S. Katti, “Full duplex radios,” Acm Sigcomm Comput. Commun. Rev. 43, 375–386 (2013). [CrossRef]
