We have investigated the dark current of a germanium (Ge) photodetector (PD) with a GeO2 surface passivation layer and a gas-phase-doped n+/p junction. The gas-phase-doped PN diodes exhibited a dark current of approximately two orders of magnitude lower than that of the diodes formed by a conventional ion implantation process, indicating that gas-phase doping is suitable for low-damage PN junction formation. The bulk leakage (Jbulk) and surface leakage (Jsurf) components of the dark current were also investigated. We have found that GeO2 surface passivation can effectively suppress the dark current of a Ge PD in conjunction with gas-phase doping, and we have obtained extremely low values of Jbulk of 0.032 mA/cm2 and Jsurf of 0.27 μA/cm.
©2012 Optical Society of America
Germanium (Ge) photodetectors (PDs) on Si have been intensely investigated over the past ten years for optical fiber communications with a 1.55-μm wavelength range owing to the marked improvements in the epitaxial growth of Ge on Si despite the 4.2% lattice mismatch between Si and Ge [1–4]. Ge PDs on Si have been successfully demonstrated thus far in normal-incidence geometries [3,5–8] and waveguide geometries [9–13]. However, a high dark current is still one of the critical technology issues for Ge PDs as compared with InGaAs-based PDs. The origin of the high dark current has been historically attributed to threading dislocations in Ge, and in many studies conducted to reduce the threading dislocation density in Ge, a dark current density of 0.041 - 0.15 mA/cm2 in normal-incidence Ge PDs on Si with in situ-doped junctions has been achieved [3,5]. However, the dark current density of the Ge waveguide PDs on Si is still typically on the order of 10 mA/cm2 [9–12]. Recently a dark current of approximately 0.2 nA was demonstrated by Ge selective growth in sub-micrometer narrow trenches with amorphous silicon contact , however the dark current density was still approximately 0.67 mA/cm2, which was one order of magnitude higher than that of normal-incidence Ge PDs. Although commercially available normal-incidence Ge PDs with a-few-mm junction diameter fabricated on single crystalline Ge wafers exhibit the dark current density of approximately 0.06 mA/cm2 because bulk Ge wafers does not suffer from defects related to the 4.2% lattice mismatch between Ge and Si , the Ge PDs on the Ge wafers with 100-μm junction diameter exhibited the high dark current density of approximately 0.6 – 1.3 mA/cm2 [16,17], suggesting the surface leakage current is significant when the junction area is small. Thus, we expect that crystal defects related to the lattice mismatch between Ge and Si is not only the source of the dark current of the Ge PDs. One of the origins of the dark current is the junction bulk leakage current (Jbulk) due to the defects induced by an ion implantation process. The rapid diffusion of phosphorus (P) and arsenic (As) in Ge makes it difficult to carry out in situ doping due to the high-temperature growth of Ge or the subsequent high-temperature anneal for reducing dislocation density , which indicates the reason why the ion implantation process has been frequently used after the Ge growth to form an n+/p junction [9–12]. However, it is difficult to annihilate the defects by activation annealing because of the rapid dopant diffusion. Thus, the residual defects in the junction induce the large dark current of the Ge PDs. The low-temperature Ge growth below 600 °C without the subsequent high-temperature anneal has been examined in Refs 6 and 19, however the dark current density is still more than 3 mA/cm2 due to the relatively low crystal quality. Thus, the ex-situ doping method to form high quality n+/p junction after Ge growth instead of the ion implantation process is required. The surface leakage current (Jsurf) is also expected to be another source of the dark current. As is well known, the surface passivation of Ge has been more difficult than that of a SiO2/Si system. Thus, the surface leakage current, such as the trap-assisted tunneling current, is significant, particularly for waveguide devices, because the ratio of the surface area to the volume increases.
In this study, we have investigated the gas-phase doping and GeO2 surface passivation processes in order to suppress the dark current of the Ge PDs. The gas-phase doping of As in Ge allows us to form a high-quality n+/p junction ; thus, the dark current can be reduced by two orders of magnitude as compared with that in the case of the ion implantation process. We have recently found that the GeO2 formed by high-temperature oxidation is effective for the surface passivation of Ge [21,22]. GeO2 surface passivation effectively suppressed the dark current of the Ge detector in conjunction with gas-phase doping, and we have successfully obtained extremely low values of Jbulk of 0.032 mA/cm2 and Jsurf of 0.27 μA/cm. Although all the experiments presented in this paper have been done using single crystalline Ge wafers, the process technologies of gas-phase doping and thermal oxidation for GeO2 formation can also be applied to Ge PDs on Si because both processes are basically complementary metal-oxide-semiconductor (CMOS) compatible. The thermal oxidation process is one of the common processes in the CMOS process, and the gas-phase doping is one of the most promising doping methods for deeply scaled CMOS . We expect that the gas-phase doping and the GeO2 surface passivation can also contribute the reduction in the dark current of the Ge PDs on Si.
2. Gas-phase-doped PN junction
We have investigated the PN diodes formed by the gas-phase doping of As in Ge using a metal-organic source, tertiary-butyl-arsine (TBAs), instead of the diodes formed by ion implantation. The gas-phase doping with TBAs was carried out by using the metal-organic vapor phase epitaxy (MOVPE) system originally designed for III-V compound semiconductor growth . Low-damage TBAs-based gas-phase doping results in the slower diffusion of As in Ge than ion implantation, which is favorable for obtaining low junction leakage. High-performance Ge n-type metal-oxide-semiconductor field-effect transistors (MOSFETs) have been successfully demonstrated by TBAs-based gas phase doping .
PN diodes were fabricated on p-type Ge (100) substrates with a doping concentration of 1 × 1016 cm−3 by gas-phase doping as follows. After cleaning the surfaces of the Ge samples by cyclic etching with HF solution and deionized water, a 100-nm SiO2 mask was deposited on the samples using a plasma-enhanced chemical vapor deposition (PECVD) system. The SiO2 mask was patterned by photolithography in order to define the junction regions. Then, the samples were injected into the MOVPE reactor for gas-phase doping. After heating the reactor up to 600 °C with H2 flow, gas-phase doping was carried out for 60 min by injecting TBAs with hydrogen carrier gas as a source for As diffusion. During gas-phase doping, the pressure of the reactor was maintained at 100 mbar, and the partial pressure of TBAs was set to be 0.18 mbar. After As doping, Ni electrodes were deposited on the front surface of the samples and Al was deposited as the back electrode. For comparison, PN diodes were also fabricated by the ion implantation of P and As. The acceleration energy was set to be 10 keV for P and 15 keV for As. The implantation dose was 1 × 1015 cm−2. Rapid thermal annealing (RTA) was carried out at 400 - 650 °C for 10 s for dopant activation.
Figure 1 shows the As distribution in the gas-phase-doped Ge samples measured by secondary ion mass spectroscopy (SIMS). As shown in the SIMS profile in Fig. 1, arsenic atoms decomposed from the TBAs that diffused into the Ge surfaces, and the surface concentration of As was approximately 4 × 1019 cm−3. Figure 2(a) shows a cross-sectional transmission electron microscopy (TEM) image of the gas-phase-doped n+/p junction, in which no obvious dislocation can be observed. To analyze the two-dimensional doping profile of the junction, the electrical potential distribution was observed using the electron holography technique for the fabricated PN diode. As shown in Fig. 2(b), an abrupt junction with a doping depth of 300 nm was clearly observed; the doping depth agreed well with the SIMS profile shown in Fig. 1. The spherical distribution due to gas-phase doping was also obtained as expected. Since the precise gas control can be available owing to the MOVPE system, the shallow doping depth of less than 300 nm is available by reducing the doping time. When the doping time is 1 min at 500 °C, the doping depth is expected to be less than 20 nm.
To clarify the characteristics of the n+/p junctions, we evaluated the electrical properties of the PN diodes fabricated by gas-phase doping. The I-V characteristics of the gas-phase-doped diode are shown in Fig. 3 . The diameter of PN diodes was 150 μm. The surfaces of the diodes were passivated by SiO2, as shown in the inset of Fig. 3. For comparison, the I-V characteristics of the PN diode fabricated by the ion implantation of P are also shown in Fig. 3. The P-implanted PN diode activated at 600 °C for 10 s exhibited the lowest dark current among the implanted samples. The off current of the n+/p junction formed by gas-phase doping was two orders of magnitude lower than that of the junction formed by ion implantation. It was confirmed that the characteristics of the PN diodes diffused at 500 °C and 600 °C were the same, where the on/off ratio was 105 and the ideality factor n was 1.2. The ideality factor of the gas-phase-doped samples was smaller than 1.8 in the case of the implanted diode.
Figure 4 shows the temperature dependence of the dark current measured at a bias voltage of −0.5 V for evaluating the carrier conduction mechanism. The activation energy (Ea) was estimated on the basis of the slope of the Arrhenius plots. The PN junctions formed by gas-phase doping exhibited an Ea value of 0.36 eV that corresponded to half of the band gap of Ge, suggesting that the generation–recombination current was dominant in the diodes fabricated by gas-phase doping. On the other hand, the Ea values of 0.18 – 0.24 eV for ion implantation were smaller than half of the band gap of Ge, suggesting that the dark current of the ion-implanted diodes was attributed to the defect-assisted tunneling current. These results indicate that gas-phase doping enables low-damage PN junction formation, which is effective for reducing junction leakage current. As mentioned in , the diffusion constant of boron in Ge is quite small as compared with those of arsenic and phosphorus. Thus, ion implantation of boron to form p + Ge is less problematic. We expect that the combination of ion implantation of boron and gas-phase doping of arsenic is one of the promising ways to achieve high-quality Ge PIN diodes through suppressing dopant diffusion, which is suitable for Ge PDs on Si.
We have also investigated the surface leakage current of the SiO2-passivated n+/p junction by preparing samples with different peripheral lengths but with the same junction area. Figure 5(a) shows the I-V characteristics of the samples with a junction area of 22500 μm2. The dark current increased with an increase in peripheral length, indicating that the surface leakage current was not negligible. Figure 5(b) shows the dark current as a function of the peripheral length measured at the bias voltage of −1 V. The dark current was approximately proportional to the peripheral length, and the surface leakage per unit length was estimated to be 8.4 μA/cm. This value was approximately two times smaller than that reported by Loh et al . However, the surface leakage current would still be significant in small-junction-area Ge PDs. Thus, a more effective surface passivation of Ge is required for further reduction in the dark current of Ge PDs.
3. GeO2-passivated Ge PD with gas-phase-doped junction
GeO2 was not previously regarded as a good passivation material for Ge despite many attempts to form high-quality GeO2/Ge interfaces [25–29]. However, we have recently found that the high-temperature oxidation of Ge can form a high-quality GeO2/Ge interface with an interface state density of less than 1011 cm−2 eV−1 . Ge pMOSFETs and nMOSFETs exhibiting superior performance characteristics to Si MOSFETs have successfully been demonstrated using a thermally oxidized GeO2/Ge MOS interface [24,30].
To suppress the surface leakage current of the Ge PDs, we have applied GeO2 passivation to ring-shaped Ge PDs with a gas-phase-doped junction. A schematic of the device is shown in the inset of Fig. 6(a) . First, As was diffused into p-Ge substrates with a SiO2 mask by gas-phase doping at 600 °C for 60 min to form an n+/p junction. After removing the SiO2 mask, the 20-nm-thick GeO2 passivation layer was formed by thermal oxidation at 550 °C. Owing to the small diffusion constant of gas-phase-doped As at 550 °C , the dopant diffusion is expected to be approximately 30 nm after GeO2 formation, which is not significant for the characteristics of the Ge PDs. We found that 5-nm-thick GeO2 was enough for passivating the GeO2 surface . Thus, further suppression of As diffusion is expected by reducing the oxidation time down to 5 min. A 20-nm-thick Al2O3 layer was deposited by atomic layer deposition (ALD) as a capping layer. Although GeO2 itself is water-soluble and unstable if GeO2 is exposed to the atmosphere, the Al2O3 capping layer effectively protects GeO2. By using the Al2O3 capping layer, we can use the standard fabrication processes, and the stable device operation can also be obtained. Then, Ni was deposited in the n+ Ge region. Al was also deposited on the Al2O3/GeO2 passivation layer surrounding the PD region as a gate electrode to control the surface potential of the surrounding region. Finally, Al was deposited on the back surface of the substrates.
The I-V characteristics of the GeO2-passivated PD with a junction diameter of 430 μm are shown in Fig. 6(a). Since we observed the negative flat-band voltage shift of the GeO2/Ge interface due to the contamination after gas-phase doping, the surface of the p-Ge surrounding the PD region was expected to be strongly inverted. To compensate for the influence of the negative flat-band voltage of the GeO2-passivated Ge surrounding the PD region, a gate voltage of −3 V was applied during the measurements. When the gate voltage was −2V to −3V, the surface of the p-Ge was expected to be weakly inverted. Owing to GeO2 surface passivation, the dark current of the Ge PD was reduced to less than 100 nA despite its large junction area of more than 105 μm2. The dark current showed almost no dependence on the applied bias voltage up to −2 V, indicating that the n+/p junction formed by gas-phase doping showed almost no defects. Owing to the low dark current, the on/off ratio of the diode measured at ± 1 V was more than 106, which was one magnitude higher than that of the SiO2-passivated PD shown in Fig. 3. The negative flat-band voltage shift can be reduced to be nearly zero by optimizing the cleaning process after gas-phase doping; thus, the reduction in the dark current will also be expected even without applying a gate voltage after process optimization. A photocurrent produced by 1550-nm-wavelength continuous-wave light with 0-dBm input power is also shown in Fig. 6(a). The light was coupled through a cleaved single-mode fiber with normal incidence. A photocurrent of approximately 550 μA was observed. Figure 6(b) shows responsivity as a function of wavelength. A responsivity of approximately 0.55 A/W was obtained for the C-band wavelength. The 0.55-A/W responsivity at 1550 nm is below 1 A/W due to reflection at the surface and shadowing effect of the electrodes, which can be improved by applying the anti-reflection coating and optimizing the electrode pattern. Since the direct band gap energy of Ge is approximately 0.8 eV, the penetration depth of the incident light at 1630 nm in Ge is larger than the estimated diffusion length of approximately 27 μm . Thus, the responsivity was gradually decreased as the wavelength increased from 1550nm.
Figure 7 shows the dark current measured at −1 V as a function of the junction area of the Ge PD. To extract the junction bulk current density (Jbulk) and the peripheral surface current density (Jsurf) separately, numerical fitting was performed using6 and 7. Thus, the GeO2 surface passivation effectively suppressed the dark current of the Ge PD in conjunction with gas-phase doping.
We have investigated the origins of the dark current of Ge PDs using a GeO2 surface passivation layer and a gas-phase-doped junction. The conventional ion implantation of P and As, which was frequently carried out to form PN junctions of Ge waveguide PDs because the rapid diffusion of P and As made it difficult to perform in situ doping during the high-temperature growth of Ge, induced a large dark current due to the residual crystal defects formed by ion implantation. The gas-phase-doped junction exhibited a dark current of two orders of magnitude lower than that of the ion-implanted junctions because of the low-damage junction formation process. We have also revealed that the SiO2 passivation of Ge surfaces is not sufficient to reduce the surface leakage current. We have successfully demonstrated that a high-temperature thermally oxidized GeO2 layer can effectively passivate Ge surfaces, reducing the surface leakage current of Ge PDs by two orders of magnitude. An ultralow junction bulk leakage of 0.032 mA/cm2 and a surface leakage of 0.27 μA/cm were obtained by GeO2 surface passivation in conjunction with gas-phase doping, which can contribute to the reduction in the dark current of the Ge waveguide PDs.
This work was partly supported by Precursory Research for Embryonic Science and Technology (PRESTO), JST.
References and links
1. J. Michel, J. Liu, and L. C. Kimerling, “High-performance Ge-on-Si photodetectors,” Nat. Photonics 4(8), 527–534 (2010). [CrossRef]
2. S. Luryi, A. Kastalsky, and J. C. Bean, “New infrared detector on a silicon chip,” IEEE Trans. Electron. Dev. 31(9), 1135–1139 (1984). [CrossRef]
3. S. B. Samavedam, M. T. Currie, T. A. Langdo, and E. A. Fitzgerald, “High-quality germanium photodiodes integrated on silicon substrates using optimized relaxed graded buffers,” Appl. Phys. Lett. 73(15), 2125–2127 (1998). [CrossRef]
4. H. C. 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), 2909–2911 (1999). [CrossRef]
5. J. Osmond, G. Isella, D. Chrastina, R. Kaufmann, M. Acciarri, and H. von Kanel, “Ultralow dark current Ge/Si(100) photodiodes with low thermal budget,” Appl. Phys. Lett. 94(20), 201106 (2009). [CrossRef]
6. H. Y. Yu, S. Ren, W. S. Jung, A. K. Okyay, D. A. B. Miller, and K. C. Saraswat, “High-efficiency p-i-n photodetectors on selective-area-grown Ge for monolithic integration,” IEEE Electron Device Lett. 30(11), 1161–1163 (2009). [CrossRef]
7. T. H. Loh, H. S. Nguyen, R. Murthy, M. B. Yu, W. Y. Loh, G. Q. Lo, N. Balasubramanian, D. L. Kwong, J. Wang, and S. J. Lee, “Selective epitaxial germanium on silicon-on-insulator high speed photodetectors using low-temperature ultrathin Si0.8Ge0.2 buffer,” Appl. Phys. Lett. 91(7), 073503 (2007). [CrossRef]
8. L. Colace, P. Ferrara, G. Assanto, D. Fulgoni, and L. Nash, “Low dark-current germanium-on-silicon near-infrared detectors,” IEEE Photon. Technol. Lett. 19(22), 1813–1815 (2007). [CrossRef]
9. 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). [CrossRef] [PubMed]
10. S. Park, T. Tsuchizawa, T. Watanabe, H. Shinojima, H. Nishi, K. Yamada, Y. Ishikawa, K. Wada, and S. Itabashi, “Monolithic integration and synchronous operation of germanium photodetectors and silicon variable optical attenuators,” Opt. Express 18(8), 8412–8421 (2010). [CrossRef] [PubMed]
11. 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] [PubMed]
12. K. W. Ang, T. Y. Liow, M. B. Yu, Q. Fang, J. Song, G. Q. Lo, and D. L. Kwong, “Low thermal budget monolithic integration of evanescent-coupled Ge-on-SOI photodetector on Si CMOS platform,” IEEE J. Sel. Top. Quantum Electron. 16(1), 106–113 (2010). [CrossRef]
14. M. Beals, J. Michel, J. F. Liu, D. H. Ahn, D. Sparacin, R. Sun, C. Y. Hong, L. C. Kimerling, A. Pomerene, D. Carothers, J. Beattie, A. Kopa, A. Apsel, M. S. Rasras, D. M. Gill, S. S. Patel, K. Y. Tu, Y. K. Chen, and A. E. White, “Process flow innovations for photonic device integration in CMOS,” Proc. SPIE 6898, 689804(2008). [CrossRef]
15. Thorlabs Inc, http://www.thorlabs.com.
16. H. Ando, H. Kanbe, T. Kimura, T. Yamaoka, and T. Kaneda, “Characteristics of germanium avalanche photodiodes in the wavelength region of 1-1.6 μm,” IEEE J. Quantum Electron. 14(11), 804–809 (1978). [CrossRef]
17. S. Kagawa, T. Kaneda, T. Mikawa, Y. Banba, and Y. Toyama, “Fully ion-implanted p+ -n germanium avalanche photodiodes,” Appl. Phys. Lett. 38(6), 429–431 (1981). [CrossRef]
18. C. O. Cui, K. Gopalakrishnan, P. B. Griffin, J. D. Plummer, and K. C. Saraswat, “Activation and diffusion studies of ion-implanted p and n dopants in germanium,” Appl. Phys. Lett. 83(16), 3275–3277 (2003). [CrossRef]
19. O. Fidaner, A. K. Okyay, J. E. Roth, R. K. Schaevitz, Y.-H. Kuo, K. C. Saraswat, J. S. Harris, and D. A. B. Miller, “Ge–SiGe quantum-well waveguide photodetectors on silicon for the near-infrared,” IEEE Photon. Technol. Lett. 19(20), 1631–1633 (2007). [CrossRef]
20. M. Takenaka, K. Morii, M. Sugiyama, Y. Nakano, and S. Takagi, “Gas phase doping of arsenic into (100), (110), and (111) germanium substrates using a metal–organic source,” Jpn. J. Appl. Phys. 50, 010105 (2011). [CrossRef]
21. H. Matsubara, T. Sasada, M. Takenaka, and S. Takagi, “Evidence of low interface trap density in GeO2/Ge metal-oxide-semiconductor structures fabricated by thermal oxidation,” Appl. Phys. Lett. 93(3), 032104 (2008). [CrossRef]
22. T. Sasada, Y. Nakakita, M. Takenaka, and S. Takagi, “Surface orientation dependence of interface properties of GeO2/Ge metal-oxide-semiconductor structures fabricated by thermal oxidation,” J. Appl. Phys. 106(7), 073716 (2009). [CrossRef]
23. N. D. Nguyen, E. Rosseel, S. Takeuchi, J. L. Everaert, L. Yang, J. Goossens, A. Moussa, T. Clarysse, O. Richard, H. Bender, S. Zaima, A. Sakai, R. Loo, J. C. Lin, W. Vandervorst, and M. Caymax, “Use of p- and n-type vapor phase doping and sub-melt laser anneal for extension junctions in sub-32 nm CMOS technology,” Thin Solid Films 518(6), S48–S52 (2010). [CrossRef]
24. K. Morii, T. Iwasaki, R. Nakane, M. Takenaka, and S. Takagi, “High-performance GeO2/Ge nMOSFETs with source/drain junctions formed by gas-phase doping,” IEEE Electron Device Lett. 31(10), 1092–1094 (2010). [CrossRef]
25. M. D. Jack and J. Y. M. Lee, “DLTS measurements of a germanium MIS interface,” J. Electron. Mater. 10(3), 571–589 (1981). [CrossRef]
26. E. E. Crisman, J. I. Lee, P. J. Stiles, and O. J. Gregory, “Characterisation of n-channel germanium mosfet with gate insulator formed by high-pressure thermal oxidation,” Electron. Lett. 23(1), 8–10 (1987). [CrossRef]
27. Y. Wang, Y. Z. Hu, and E. A. Irene, “Electron cyclotron resonance plasma and thermal oxidation mechanisms of germanium,” J. Vac. Sci. Technol. A 12(4), 1309–1314 (1994). [CrossRef]
28. V. Craciun, I. W. Boyd, B. Hutton, and D. Williams, “Characteristics of dielectric layers grown on Ge by low temperature vacuum ultraviolet-assisted oxidation,” Appl. Phys. Lett. 75(9), 1261–1263 (1999). [CrossRef]
29. R. S. Johnson, H. Niimi, and G. Lucovsky, “New approach for the fabrication of device-quality Ge/GeO2 /SiO2 interfaces using low temperature remote plasma processing,” J. Vac. Sci. Technol. A 18(4), 1230–1233 (2000). [CrossRef]
30. Y. Nakakita, R. Nakakne, T. Sasada, M. Takenaka, and S. Takagi, “Interface-controlled self-align source/drain Ge p-channel metal–oxide–semiconductor field-effect transistors fabricated using thermally oxidized GeO2 interfacial layers,” Jpn. J. Appl. Phys. 50, 010109 (2011). [CrossRef]
31. S. J. Koester, J. D. Schaub, G. Dehlinger, and J. O. Chu, “Germanium-on-SOI infrared detectors for integrated photonic applications,” IEEE J. Sel. Top. Quantum Electron. 12(6), 1489–1502 (2006). [CrossRef]