Open Access
Add to BibSonomyAbstract
We demonstrate the monolithic integration of germanium (Ge) p-i-n photodetector (PDs) with silicon (Si) variable optical attenuator (VOAs) based on submicrometer Si rib waveguide. A PD is connected to a VOA along the waveguide via a tap coupler. The PDs exhibit low dark current of ~60 nA and large responsivity of ~0.8 A/W at the reverse bias of 1 V at room temperature. These characteristics are uniform over the chip scale. The PDs generate photocurrents precisely with respect to DC optical power attenuated by the VOAs. Two devices work synchronously for modulated optical signals as well. 3-dB cut-off frequency of the VOA is ~100 MHz, while that of the PD is ~1 GHz. The synchronous response speed is limited by the VOA response speed. This is the first demonstration, to the best of our knowledge, of monolithic integration of Ge PDs with high-carrier-injection-based optical modulation devices based on Si.
©2010 Optical Society of America
1. Introduction
Over the past decade, time domain multiplexed passive optical networks (TDM-PONs) have been deployed for fiber-to-the-X, where X can be a home, curb, cabinet, or building [1]. For future network systems, wavelength-division multiplexed PONs (WDM-PONs) are promising for higher transmission rates (>10 Gbps) because existing TDM-PONs cannot cope with aggregated bandwidth and power budget [2,3]. Particularly, WDM-PONs implemented with erbium-doped fiber amplifiers (EDFAs) can provide long reach and high splitting ratio. However, optical bursts due to the EDFA gain transient require receivers with an extremely large dynamic range and result in a power penalty. To address these issues, implementation of variable optical attenuators (VOAs) to WDM-PONs has been under intensive investigations for level control of signals [4–6]. Although all-optical feedback loops directly linked to EDFAs can control gain transient tilt [7], signal equalization for individual channels is more effective for compensating complicated gain spectrum tilt. In addition, level equalization of bursty upstream optical packets can enhance the dynamic range of the receivers [6]. For these purposes, a compact component integrating VOAs and photodetectors (PDs) is indispensible. The operation principle is that the PDs monitor the optical power and feed the information back to the VOAs to suppress the gain transient tilt or to equalize different amplitudes of optical packets.
Silicon (Si) photonics enable monolithic integration of modulators, arrayed-waveguide grating (AWG) MUX/DEMUXs, VOAs, and PDs on a Si platform, which provides low-cost and compact solutions for WDM networks [8]. Si p-i-n VOAs fabricated on Si rib waveguides have been demonstrated [9,10]. They operate on the basis of free carrier absorption (FCA) in the intrinsic optical-path by carrier injection [11,12]. In particular, Si VOAs based on submicrometer rib waveguides have many advantages, such as a small footprint less than a square millimeter and fast response on the order of nanoseconds [10]. In addition, convenient electric control of the VOAs enables monolithic integration with PDs and feedback circuits, which is crucial for application to WDM-PONs. Furthermore, PDs integrated with Si waveguides have also been demonstrated for the near-infrared (NIR) optical communications band (1300 - 1550 nm). Specifically, some researchers have employed defect-level formation in the Si band gap by ion-implantation [13,14], while others have utilized germanium (Ge) epitaxially grown on Si as an absorptive material [15,16]. Although a large dark current in Ge PDs on Si is unavoidable due to misfit and threading dislocations caused by the 4.2% lattice mismatch between Si and Ge, Ge PDs are beneficial for application to WDM from several perspectives. First, short device length (a few ten of micrometers) can be achieved by evanescent light coupling because Ge has a large absorption coefficient of ~4,000 cm−1 in the NIR range. Second, Ge absorption can cover not only the C band (1528 −1560 nm) but also the L band (1561 - 1620 nm) when Ge is under tensile strain [17]. Although the integration of Si VOAs and Ge PDs is the perfect candidate for level control in WDM-PONs, there has been no report on the monolithic integration of Ge PDs and carrier-injection-based Si modulation devices i.e., VOAs.
In this work, we fabricated Ge p-i-n PDs monolithically integrated with Si p-i-n VOAs based on submicrometer Si rib waveguides. The PDs are connected to the VOAs via a tap coupler. We used selective epitaxial growth to deposit a Ge mesa on Si slab. The PDs show a low dark current, large responsivity, and high response speed. Multiple Ge PDs on few square centimeter chips show good uniformity in terms of dark current and responsivity. We confirmed synchronous operation between the VOAs and PDs for DC and modulated optical signals.
2. Fabrication
We used a 4-inch silicon-on-insulator (SOI) wafer as a starting substrate. The thicknesses of the top Si layer and buried oxide (BOX) are 200 nm and 3 µm, respectively. First, Si rib waveguides were defined by e-beam lithography and transferred by electron cyclotron resonance (ECR) plasma etching [18]. The core cross-section is 600 nm (width) and 200 nm (thickness), while the slab thickness is 100 nm. To electrically separate the Si VOA and Ge PD, we etched the Si slab down to the BOX to make an isolation groove. Boron and phosphorous were ion-implanted for n + and p + contacts of the VOA and simultaneously for bottom p + contact for the PD. The peak concentration of impurities was intended to be ~1020 cm−3. A VOA with a lateral p-i-n structure has an approximately 3-µm-wide intrinsic region. To recrystallize ion-damaged regions, the wafer was annealed at 1000 °C for 1 hour in nitrogen ambient. A 500-nm-thick SiO2 layer was deposited by plasma enhanced chemical vapor deposition (PE-CVD) and was then wet-etched to open a selective mask on the Si slab where Ge mesa would selectively grow. On the mask, 1-μm-thick Ge mesas were grown by ultrahigh vacuum chemical vapor deposition (UHV-CVD). Diluted GeH4 in argon and Si2H6 were the gas precursors. The mesa size is 8 μm × 50 μm. To suppress 3D nucleation due to the Si-Ge lattice mismatch, we employed two-step growth [19]. The growth temperatures for the buffer (~30 nm) and the rest of the Ge mesa were 370 and 600 °C, respectively. As shown in Fig. 1(a) , the Ge mesa has a flat top surface. The sidewall facets are {311} planes with the Ge mesa aligned along the <110> direction. A 1-µm-thick SiO2 layer was additionally deposited on the entire wafer. Post-growth annealing at 900 °C for 10 min was then employed to reduce threading dislocation density in Ge. After patterning the SiO2 layer on the top of Ge mesa to open a window for n + ion-implantation, phosphorous was implanted with peak concentration of ~1020 cm−3 and the wafer was annealed at 600 °C for 5 min to activate n + top contact. Finally, aluminum was deposited and etched to make electrode pads. The cross-sections of the VOA and the PD are shown in Fig. 1(b). Half of the optical power that has passed through the 1-mm-long VOA is guided to the PD via a tap coupler, which is a 3-dB splitter with a 1 × 2 multi-mode interference (MMI) coupler [20]. The other half of the light is coupled to an output fiber that is monitored with a power meter. Figure 1(c) shows one channel of devices consisting of the VOA, PD, and 3-dB splitter. We fabricated eight channels in an array to examine device uniformity.
Fig. 1 (a) SEM image after Ge selective epitaxial growth on a Si slab. (b) Schematic cross-sections of Si VOA (top) and Ge photodetector (bottom). (c) Optical microscope image of as-fabricated devices.
3. Characterization results
3.1 Characteristics of Ge photodetectors
To investigate the mechanism of reverse current generation in the Ge PD, we carried out current-voltage (I-V) measurements at various temperatures, as shown in Fig. 2(a) . The sample was cooled in a cryostat with a temperature range of 161 - 263 K and heated up on a peltier stage with a temperature range of 278 - 313 K. We obtained an Arrhenius plot i.e., reverse current versus the reciprocal of temperature to determine activation energy. As shown in Fig. 2(b), the activation energy is roughly ~0.3 eV, which corresponds to half of the Ge band-gap energy.
Fig. 2 (a) I-V curves of a Ge PD under dark condition varying temperature from 161 to 313 K. (b) Arrhenius plots of dark current at reverse biases of 0.5, 1, and 2 V.
Figure 3(a) shows I-V curves of the PD under dark and illuminated conditions. Dark current is as low as 60 nA at the reverse bias of 1 V at room temperature (293 K), corresponding to dark current density of 15 mA/cm2. To introduce guided light into Si rib waveguides, we cleaved the wafer into a chip with ~10-mm width. We placed adiabatic tapers at both facets with the facet rib-width of 3 µm. A continuous wave (CW) infrared light source with peak wavelength of 1560 nm was coupled to the chip through lensed fiber. The polarization was set to the TE mode. Under the illumination, photocurrent is generated at the PD. The intensity of the light at the PD needs to be estimated to obtain the responsivity. In the estimation, we took into account the monitored light power from the output fiber and the total insertion loss considering coupling loss (~7 dB/facet), propagation loss (~5 dB/cm), and excess losses at the p-i-n VOA (~0.5 dB) and at the MMI 3-dB splitter (~0.5 dB). These loss values were measured from waveguides with different lengths located on the same chip with the VOA-PD devices. The average responsivity is ~0.8 A/W, which corresponds to internal quantum efficiency of 65%. When dark current exceeds photocurrent, the photocurrent cannot be detected as a signal. In this sense, the minimum detectable optical power under DC illumination condition can be defined as dark current divided by responsivity. As a result, we obtained ~75 nW which corresponds to −41 dBm. Figure 3(b) shows the responsivity and dark current for the eight channels of devices. Other parts of the chip were characterized as well. Dark current and responsivity shows good uniformity over the chip scale of over a square centimeter.
Fig. 3 (a) I-V curves of a Ge PD under dark and illuminated conditions. (b) Measured dark currents and responsivities of Ge PDs on eight separate optical channels.
3.2 Synchronous operation for DC optical input
We varied the forward current injected to the VOA with a DC source as the CW light of 1560 nm was guided through the waveguide. For DC performance of the VOA, we obtained attenuation efficiency of 0.26 dB/mA and power consumption for 10-dB attenuation was 49 mW (1.4 V, 35 mA), which is inferior to our previous result of the VOA without the Ge PD [10]; 0.55 dB/mA and 9 mW (0.9 V, 10 mA), respectively. We believe that such degraded attenuation performance of the VOA in this work is attributed to presence of unknown recombination centers or insufficient passivation. Figure 4(a) shows I-V curves of Ge PD under both dark and illuminated condition for each injection current ranging from 0 to 150 mA. While photocurrent under illumination varies in accordance with the optical power attenuated by the VOA, the dark current remains the same regardless of injection current. This implies that synchronous current injection to the VOA does not affect the dark current of the PD. Figure 4(b) shows photocurrent detected at the Ge PD and optical powers monitored through the output fiber. Since attenuation is linearly proportional to carrier density in the intrinsic region of a VOA [9], the attenuated optical power in decibel linearly decreases with respect to the injection current. The change of the photocurrent corresponds well to the optical power. This indicates that the PD precisely detects the light intensity attenuated by the VOA.
Fig. 4 (a) I-V curves of a Ge PD under dark and illumination conditions as forward current (0 - 150 mA) was injected to Si VOA. (b) Photocurrent at the PD and optical power through output fiber at various injection currents.
3.3 Synchronous operation for modulated optical signal
To observe synchronous operation between the VOA and PD for modulated optical signals, we used the measurement setup shown in Fig. 5(a) . The VOA was modulated by the signal from a pulse pattern generator (PPG) whereas the CW light propagated through the waveguide. The modulated light signal was divided in amplitude at the 3-dB splitter. The signal detected at the PD was directly connected to an oscilloscope (Agilent 86100C DCA-J), while the optical signal from the output fiber was once converted to electric signal at an O-E converter and the electric signal was then fed to the scope. Thus, the optical signal from the output fiber represents the pulse response of the VOA. Here, we applied reveres bias of 1 V to the PD.
Fig. 5 (a) Block diagram for measurement of the pseudo-random beam stream (PRBS) at the optical output and Ge PD as the Si VOA is modulated by pulsed signal from the pulse pattern generator (PPG). (b) Pulse trains measured at the output fiber and Ge PD with a 29-1 PRBS at 100 MHz. (c) 20-ns single pulse response at the output fiber and Ge PD with varying PPG bias.
Figure 5(b) shows signal trains at the output fiber and PD when a 29-1 pseudo-random beam stream (PRBS) modulates the VOA. The results indicate that the pulse responses at the VOA and PD are well synchronized at 100-MHz bandwidth. In fact, we observed that the pulse shape began to collapse when the bandwidth was greater than 200 MHz. This accounts for the response speed of the VOA, which is further discussed in the next section. Figure 5(c) shows temporal pulse responses obtained when a single pulse with 20-ns width was sent to the VOA from the PPG. From the rise and fall times of the pulse, the time constant is approximately a few nanoseconds. Furthermore, the pulse heights and shapes are also well synchronized between the PD and VOA in accordance with injection currents to the VOA. These results suggest that the PD and the VOA operate synchronously for the pulsed signals, simultaneously responding to the signal amplitude.
We measured the 3-dB cut-off frequency individually for the Si VOA and Ge PD as well as the synchronous one between the VOA and PD. For these measurements, we commonly used a CW light source with a peak wavelength of 1560 nm and a network analyzer (Agilent E5071C) to launch sinuous signals at 0-dBm power with 50-Ω impedance matching. DC sources were connected via a bias tee so that injection current and revere bias can be independently applied to the VOA and the PD, respectively. First, the 3-dB cut-off frequency of the VOA was measured by modulating the VOA by the signal from the network analyzer. To match 50-Ω impedance between the VOA and transmission cable, a fixed resistor was inserted to a micro-probe near the VOA. The output optical signals of the device chip were guided to an external O-E converter, and then collected by the network analyzer. Second, the 3-dB cut-off frequency of the PD was measured by introducing a modulated light through a commercial lithium niobate (LN) modulator (Sumitomo Osaka Cement T-MZH1.5-10). The signals from the PD were directly collected by the network analyzer. Lastly, the synchronous 3-dB cut-off frequency was measured with the experimental setup shown in Fig. 6(a) . The PD directly sent the signals modulated at the VOA to the network analyzer as the CW light was guided through the VOA.
Fig. 6 (a) Block diagram for measurement of synchronous frequency response between the Si VOA and Ge PD. (b) Separately measured frequency responses of a Si VOA at injection current of 10 and 50 mA and of a Ge PD at −1 and −10 V. (c) Synchronous frequency responses between the Si VOA and Ge PD at injection current of 10 and 50 mA with the Ge PD at −1 (solid) and −10 V (dotted).
Figure 6(b) shows individual frequency responses. The 3-dB cut-off frequencies of the VOA at injection current of 10 and 50 mA are 130 and 46 MHz, respectively, while those of the PD at reverse bias of 1 and 10 V are 1.3 and 2.6 GHz. Figure 6(c) shows the synchronous frequency response. Regardless of the reverse bias at the PD, the synchronous 3-dB cut-off frequencies do not differ from the VOA ones.
4. Discussion
There are several advantages in using selective epi-growth of a Ge mesa for integration with Si waveguide devices. First, processes for further defining the Ge mesa are not necessary because the mesa area is limited by the SiO2 hard mask and accurate thickness control is possible in UHV-CVD growth. A process such as dry etching or chemical-mechanical polishing (CMP) would cause surface damage, which would act as a center for leakage current generation. On the contrary, the atomically smooth facets of the mesa can actually suppress surface-induced leakage current. Second, threading dislocations in the mesa can be readily annihilated because of the small mesa area, and thus dark current can be lowered. Due to the Si-Ge lattice mismatch, threading dislocations in the mesa are inevitable and they form a deep defect-level in the Ge band gap (~0.3 eV). We confirmed from the Arrhenius plot in Fig. 2(b) that Shockley-Reed-Hall electron-hole generation at such deep-level defects i.e., threading dislocations is the dominant leakage current mechanism. Some observed temperature-independent dark-current-generation of Ge PDs on Si at low temperature (<240 K) under large reverse bias (−2 V) and they claim that that phenomenon is caused by tunneling [21,22], although out Ge PD barely shows such a behavior. We believe that this is due to weak electric field in our selective Ge mesa and therefore, a primary origin for leakage current is threading dislocations which should be reduced to lower dark current. Some threading dislocations are glissile; they glide out of the mesa due to thermal stress. In fact, it has been reported that a low threading-dislocation density of 106 cm−2 can be achieved by annealing at 900 °C [23]. Third, tensile strain can be built up in the mesa by high-temperature annealing and this enables a red-shift of the absorption band up to the L band (~1600 nm). The thermal-expansion coefficient difference builds up tensile strain in Ge on Si, which is not relaxed as long as the mesa width is larger than 1 μm [24]. Large coverage of absorption would be favorable for broad bandwidth optical communication.
Generally, there is a variety of applications in optical communication from implementing integrated VOAs and PDs. In this section, however, we focus on the specification of VOAs and PDs for WDM-PONs as an example of practical applications. In WDM-PONs, the key role of the PD will be reliable power detection of light intensity attenuated by the VOA. We show the DC synchronous operation between the VOA and the PD in Fig. 4. Furthermore, our PD has a large dynamic range of over 30 dB, in which photocurrent is linear to light power. In WDM-PONs, 10 - 20 dB of attenuation range is required for gain tilt compensation and signal equalization [6,25]. Therefore, the PD meets the practical application demands under the DC condition. In addition, the pulse response shows good synchronicity as well. The PD has a higher response speed (~1 GHz) than the VOA (~100 MHz). It appears obvious that the response speed of synchronous operation is limited by the slower device, i.e., the VOA. In fact, 100-MHz response speed is sufficient for level control in 10-Gbps WDM-PONs [6,25]. However, we observed a trade-off relation between attenuation and bandwidth of the VOA in accordance with injection current. 10 - 20 dB of attenuation and 100-MHz synchronous bandwidth cannot be simultaneously achieved by the VOA in this work, although attenuation and response speed can be enhanced by optimizing the VOA length and/or the width between p + and n + contacts. In particular, attenuation efficiency can be significantly increased by eliminating unwanted recombination centers along the intrinsic region of the VOA. Ultimately, both the DC and RF requirements for level control in WDM-PONs can be satisfied by our VOA-PD configuration. There are some challenges in this configuration as well. In Fig. 6(b), we observe noisy frequency response at frequency higher than 500 MHz, indicating that a high-frequency noise generated in the VOA operation leaks to the PD. DC injection current to the VOA has no effect on the dark current of the PD, as shown in Fig. 4(a), because these two devices are electrically separated by an isolation groove. Nevertheless, the high frequency noise would leak to the PD via capacitive coupling. A solution to this problem would be to change the layout of the devices by, for instance, placing two devices further away from each other or changing the in-plane device direction. Another important challenge is polarization dependence. In this work, all the waveguide devices were designed to operate with TE-mode light. However, polarization of the optical signals in WDM-PONs is random and thus, polarization-independent VOA-PD devices are necessary. In fact, development of polarization-independent spot-size converters (SSCs), rib waveguides, bends, and 3-dB splitters is one of our on-going tasks [26].
Lastly, Ge has expanded its territory in Si photonics as a high-performance active device, such as modulator and light emitter [27,28]. There is a growing need to integrate such Ge active devices with other Si photonic devices. Recently, waveguide-integrated Ge PD was fabricated by Ge wafer bonding technique [29], and hybrid integration of Si micro-ring modulator and Ge PD was demonstrated based on that technique [30]. Therefore, we believe that our monolithic integration of Si VOA and Ge PD via selective epi-growth might inspire those who study integrated Si photonics for computing as well as communication.
5. Conclusion
We integrated Si p-i-n VOAs fabricated on submicrometer Si rib waveguides with Ge p-i-n PDs. These two devices are connected by a 3-dB splitter that serves as a tap coupler. The Ge mesa was grown by selective UHV-CVD epi-growth. To the best of our knowledge, this is the first attempt at the monolithic integration of Ge PDs with high-carrier-injection-based optical modulation devices on Si. The PD shows low dark current (~60 nA) and large responsivity (~0.8 A/W). These characteristics are uniform over a chip of a square centimeter in area. The PD generates photocurrent proportionally to the optical power attenuated by the VOA. Furthermore, those two devices show good synchronicity for modulated optical signal, as well. Due to much broader bandwidth of the PD (~1 GHz) than that of the VOA (~100 MHz), the synchronous response speed is limited by the VOA bandwidth. Such synchronous performance under DC and RF conditions is suitable for implementing our VOA-PD device to equalize signal levels in WDM-PONs.
Acknowledgment
This work was partly supported by the SCOPE program of the Ministry of Internal Affairs and Communications, Japan. The authors are grateful to Dr. Yukinori Ono of NTT Basic Research Laboratories for helping with the low-temperature I-V measurement and to Dr. Jiro Osaka of the University of Tokyo for helping with the selective epitaxial growth of Ge on Si.
References and links
1. A. A. de Albuquerque, A. J. N. Houghton, and S. Malmros, “Field trials for fiber access in the EC,” IEEE Commun. Mag. 32(2), 40–48 (1994). [CrossRef]
2. A. Banerjee, Y. Park, F. Clarke, H. Song, S. Yang, G. Kramer, K. Kim, and B. Mukherjee, “Wavelength-division-multiplexed passive optical network (WDM-PON) technologies for broadband access: a review [Invited],” J. Opt. Netw. 4(11), 737 (2005), http://www.opticsinfobase.org/abstract.cfm?URI=JON-4-11-737. [CrossRef]
3. C. Lin, Broadband Optical Access Networks and Fiber-to-the-Home: Systems Technologies and Deployment Strategies (Wiley, 2006), Chap 12.
4. M. Fujiwara, J. Kani, H. Suzuki, and K. Iwatsuki, “Impact of backreflection on upstream transmission in WDM single-fiber loopback access networks,” J. Lightwave Technol. 24(2), 740–746 (2006), http://www.opticsinfobase.org/abstract.cfm?URI=JLT-24-2-740. [CrossRef]
5. J. Moon, K. Choi, S. Mun, and C. Lee, “Effects of back-reflection in WDM-PONs based on seed light injection,” IEEE Photon. Technol. Lett. 19(24), 2045–2047 (2007). [CrossRef]
6. Y. Liu, C. W. Chow, C. H. Kwok, H. K. Tsang, and C. Lin, “Optical burst and transient equalizer for 10Gb/s amplified WDM-PON,” in Proceedings of Optical Fiber Communication and the National Fiber Optic Engineers Conference, (Academic, Anaheim, USA, 2007), OThU7.
7. H. Dai, J. Pan, and C. Lin, “All-optical gain control of in-line erbium-doped fiber amplifiers for hybrid analog/digital WDM systems,” IEEE Photon. Technol. Lett. 9(6), 737–739 (1997). [CrossRef]
8. L. Pavesi, and D. J. Lockwood, eds., Silicon photonics (Springer, 2004).
9. D. W. Zheng, B. T. Smith, and M. Asghari, “Improved efficiency Si-photonic attenuator,” Opt. Express 16(21), 16754–16765 (2008), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-16-21-16754. [CrossRef] [PubMed]
10. K. Yamada, T. Tsuchizawa, T. Watanabe, H. Fukuda, H. Shinojima, and S. Itabashi, “Application of low-loss silicon photonic wire waveguides with carrier injection structures,” in Proceedings of 4th International Conference on Group IV Photonics, (Institute of Electrical and Electronics Engineers, Tokyo, Japan, 2007), pp. 116–118.
11. R. Soref and B. Bennett, “Electrooptical effects in silicon,” IEEE J. Quantum Electron. 23(1), 123–129 (1987). [CrossRef]
12. C. K. Tang, G. T. Reed, A. J. Walton, and A. G. Rickman, “Low-loss, single-model optical phase modulator in SIMOX material,” J. Lightwave Technol. 12(8), 1394–1400 (1994). [CrossRef]
13. Y. Liu, C. W. Cho, W. Y. Cheung, and H. K. Tsang, “In-line channel power monitor based on helium ion implantation in silicon-on-insulator waveguides,” IEEE Photon. Technol. Lett. 18(17), 1882–1884 (2006). [CrossRef]
14. M. W. Geis, S. J. Spector, M. E. Grein, R. T. Schulein, J. U. Yoon, D. M. Lennon, S. Deneault, F. Gan, F. X. Kaertner, and T. M. Lyszczarz, “CMOS-compatible all-Si high-speed waveguide photodiodes with high responsivity in near-infrared communication band,” IEEE Photon. Technol. Lett. 19(3), 152–154 (2007). [CrossRef]
15. D. Ahn, C.-Y. Hong, J. Liu, W. Giziewicz, M. Beals, L. C. Kimerling, J. Michel, J. Chen, and F. X. Kärtner, “High performance, waveguide integrated Ge photodetectors,” Opt. Express 15(7), 3916–3921 (2007), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-15-7-3916. [CrossRef] [PubMed]
16. L. Vivien, M. Rouvière, J.-M. Fédéli, D. Marris-Morini, J. F. Damlencourt, J. Mangeney, P. Crozat, L. El Melhaoui, E. Cassan, X. Le Roux, D. Pascal, and S. Laval, “High speed and high responsivity germanium photodetector integrated in a Silicon-On-Insulator microwaveguide,” Opt. Express 15(15), 9843–9848 (2007), http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-15-15-9843. [CrossRef] [PubMed]
17. J. Liu, J. Michel, W. Giziewicz, D. Pan, K. Wada, D. D. Cannon, S. Jongthammanurak, D. T. Danielson, L. C. Kimerling, J. Chen, F. Ö. Ilday, F. X. Kärtner, and J. Yasaitis, “High-performance, tensile-strained Ge p-i-n photodetectors on a Si platform,” Appl. Phys. Lett. 87(10), 103501 (2005). [CrossRef]
18. T. Tsuchizawa, K. Yamada, H. Fukuda, T. Watanabe, J. Takahashi, M. Takahashi, T. Shoji, E. Tamechika, S. Itabashi, and H. Morita, “Microphotonics devices based on silicon microfabrication technology,” IEEE J. Sel. Top. Quantum Electron. 11(1), 232–240 (2005). [CrossRef]
19. S. Park, Y. Ishikawa, T. Tsuchizawa, T. Watanabe, K. Yamada, S. Itabashi, and K. Wada, ““Effect of post-growth annealing on morphology of Ge mesa selectively grown on Si,” IEICE Trans. Electron,” E 91-C, 181 (2008).
20. K. Yamada, T. Tsuchizawa, T. Watanabe, J. Takahashi, E. Tamechika, M. Takahashi, S. Uchiyama, H. Fukuda, T. Shoji, S. Itabashi, and H. Morita, ““Microphotonics devices Based on silicon wire waveguiding system,” IEICE Trans. Electron,” E 87-C, 351 (2004).
21. S. J. Koester, L. Schares, C. L. Schow, G. Dehlinger, and R. A. John, “Temperature-dependent analysis of Ge-on-SOI photodetectors and receivers,” in Proceedings of 3th International Conference on Group IV Photonics, (Institute of Electrical and Electronics Engineers, Ottawa, Canada, 2006), pp. 179–181.
22. Z. Huang, J. Oh, S. K. Banerjee, and J. C. Campbell, “Effectiveness of SiGe buffer layers in reducing dark currents of Ge-on-Si photodetectors,” IEEE J. Quantum Electron. 43(3), 238–242 (2007). [CrossRef]
23. H. Luan, D. R. Lim, K. K. Lee, K. M. Chen, J. G. Sandland, K. Wada, and L. C. Kimerling, “High-quality Ge epilayers on Si with low threading-dislocation densities,” Appl. Phys. Lett. 75(19), 2909 (1999). [CrossRef]
24. S. Park, Y. Ishikawa, K. Wada, Y. Tsusaka, and J. Matsui, “Strain and absorption coefficient of finite Ge structures on Si,” Jpn. J. Appl. Phys. 48(6), 064501 (2009). [CrossRef]
25. S. Nishihara, M. Nakamura, K. Nishimura, K. Kishine, S. Kimura, and K. Kato, “10.3 Gbit/s burst-mode PIN-TIA module with high sensitivity, wide dynamic range and quick response,” Electron. Lett. 44(3), 222 (2008). [CrossRef]
26. H. Nishi, T. Tsuchizawa, T. Watanabe, H. Shinojima, K. Yamada, and S. Itabashi, “Compact and polarization-independent variable optical attenuator based on a silicon wire waveguide with a carrier injection Structure,” Jpn. J. Appl. Phys. (to be published).
27. 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), http://www.opticsinfobase.org/ol/abstract.cfm?URI=ol-34-8-1198. [CrossRef] [PubMed]
28. 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]
29. 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), http://www.opticsinfobase.org/abstract.cfm?URI=oe-16-15-11513. [CrossRef] [PubMed]
30. L. Chen, K. Preston, S. Manipatruni, and M. Lipson, “Integrated GHz silicon photonic interconnect with micrometer-scale modulators and detectors,” Opt. Express 17(17), 15248–15256 (2009), http://www.opticsinfobase.org/abstract.cfm?URI=oe-17-17-15248. [CrossRef] [PubMed]
