We investigated the optical, electrical, and structural properties of epitaxially grown Ge-on-Si substrates after phosphorous implantation. Ion implantation increases n-type doping in Ge for an on-chip light source. However, its effects on Ge should be carefully studied as implantation may increase the recombination sites, and possibly reduce light-emitting efficiency. We studied the light-emitting efficiency of implanted Ge using various material characterizations. We found that phosphorous implantation increased the doping concentration of in situ doped Ge-on-Si, which boosted the photoluminescence by 12–30%. It is therefore critical to optimize the post-annealing and implantation doses to increase light-emitting efficiency of Ge.
© 2016 Optical Society of America
While transistor scaling improves the performance of complementary metal–oxide–semiconductor integrated circuits, electrical interconnections have begun to encounter serious problems, particularly with respect to signal delay and power consumption [1,2]. To address these issues, optical interconnections that communicate data using light have attracted much attention and are expected to become an alternative to problematic electrical interconnections [3,4]. Although most of the optical components required for on-chip optical interconnections have been successfully demonstrated on a Si platform [5–8]. a practical on-chip light source is still a challenge because group IV semiconductors exhibit indirect bandgaps, leading to an extremely poor efficiency of light emission .
Tensile strain and n-type doping have been used to increase the light-emitting efficiency of Ge [10,11]. Process-induced tensile strain due to thermal expansion mismatch between epitaxial Ge films and Si substrates (Ge-on-Si) reduces the energy difference between indirect and direct bandgaps. N-type doping can further enhance the light-emitting efficiency by increasing the fraction of conduction electrons in the direct energy bandgap . In recent years, n-type doped Ge-on-Si has attracted much attention as an efficient on-chip light source [13–15]. Electrically pumped Ge lasers have been demonstrated on Ge-on-Si using in situ phosphorus doping (4 × 1019 cm−3) and by using multiple delta-doping  during epitaxial growth .
However, high-threshold current density remains a challenging issue for the operation of efficient Ge lasers. A high concentration of n-type dopant is required to lower the threshold current at a given tensile strain. Most commonly, n-type doping has been achieved through in- situ doping during epitaxial growth as well as spin on dopant (SOD) technique using dopant diffusion. Both methods have the advantage that it does not damage the surface. However, the difficulties of diffusion control and removal of oxide layer, limiting the use of SOD . In addition, the activated dopant concentration achieved through in situ doping is generally limited to 4 × 1019 cm−3.
More recently, ion implantation has also been studied and n-type doping greater than 1 × 1020 cm−3 has been demonstrated [19,20]. Ion implantation may enhance the activated dopant concentration, but it is not clear if the light-emitting efficiency is truly improved by ion implantation owing to the possible lattice damage induced by high ion energies and doses. Ion implantation degrades crystal quality much more significantly than does in situ doping. Light-emitting efficiency may be reduced by the increased number of recombination sites arising from implantation damage. Therefore, it is critical to carefully investigate these two competing factors, namely increased dopant concentration and possible material damage for the enhancement of light-emitting efficiency.
In this work, we examine the efficacy of ion implantation in improving light-emitting efficiency by using various characterization techniques: photoluminescence (PL), high-resolution transmission electron microscopy (HRTEM), secondary ion mass spectrometry (SIMS), four-point probe measurements, and Raman spectroscopy. Instead of performing ion implantation on intrinsic Ge, we perform ion implantation into in situ doped Ge-on-Si to compare with in situ doping for improving the light-emitting efficiency.
2. Material preparation
Epitaxial Ge layers were grown on Si using an ultra-high vacuum evaporator with a base pressure of under 5 × 10−9 Torr. P-type Si (100) substrates were chemically cleaned using buffered oxide etchant with a 6:1 HF solution. Surface native oxide layers were thermally removed by heating the samples up to 800°C. After de-oxidation, Si substrates were immediately transferred to the deposition chamber without breaking the vacuum break to minimize re-contamination of the surface. The working pressure was increased up to 5 × 10−8 Torr after the Ge source evaporated. The Ge growth rate ranged from 5 to 7 nm/min, as measured by a quartz crystal oscillator.
To overcome the large lattice mismatch between Ge and Si, we introduced a two-step growth method. Low-temperature (LT) Ge buffer layer of around 100 nm in thickness were grown at 300°C. Subsequently, a 1500-nm-thick high-temperature (HT) Ge layer was grown at 500°C on top of the LT Ge buffer layer. Phosphorus 31P+ ions were implanted into nominally undoped and in situ doped Ge-on-Si (phosphorus doped at 2 × 1018 cm−3) with doses of 2 × 1014–2 × 1015 cm−2 at an energy of 100 keV at an angle of 7° off the crystallization axis to avoid the channeling effect. Post-implantation annealing was carried out using rapid thermal processing (RTP-5000 of SNTEK) in an ambient N2 at temperatures of 400–800°C for durations of 30–300 seconds for dopant activation and recrystallization. To prevent out-diffusion of dopant from the Ge-on-Si, a 100-nm SiO2 cap was deposited using plasma-enhanced chemical vapor deposition at 300°C before annealing.
3. Characterization results and discussion
The PL measurements were carried out at room temperature using two different lasers emitting at wavelengths of 532 nm and 1064 nm. Crystalline phase recovery was measured using bright-field HRTEM (JEM-ARM200F of JEOL). Sheet resistance was measured using a four-point probe station to determine the activated dopant concentration. Strain variation and crystal quality were characterized through Raman spectrometry (Lab Ram ARAMIS of HORIBA). Tensile strain was evaluated from the Raman peak shift and the recovery of damage was evaluated from the full width at half maximum (FWHM) of the Raman peak.
Figure 1 shows bright-field HRTEM images of phosphorous-implanted Ge-on-Si with a dose of 1 × 1015 cm−2 at an energy of 100 keV (a) before thermal annealing and (b) after thermal annealing at 650°C for 100 s. Ion implantation with a high dose and high energy damaged the crystalline structure of the top Ge. In Fig. 1(a), a 180-nm-thick amorphous phase was seen from the surface to the depth of the projection range for P-implanted Ge-on-Si. Selected area electron diffractions (SAED) taken in this region produced multiple ring patterns scattered from amorphous Ge phases. After thermal annealing, amorphous Ge was recrystallized epitaxially on a Si substrate, as shown in Fig. 1(b), which was confirmed by a series of diffraction spots in the SAEDs.
Figure 2 shows SIMS (CAMECA IMS-6f Magnetic Sector SIMS) profiles of P-implantation into (a) undoped Ge-on-Si and (b) in situ doped Ge-on-Si. The primary ion was Cs- with ion energy of 15keV, the beam current of 50 nA, and the raster size of 200 μm × 200 μm. A Gaussian profile of the P concentration showed a projected range of 70 nm, a straggle range of 30 nm, and a peak concentration of 1.8 × 1020 cm−3. After annealing at 500°C–700°C, the high diffusivity of phosphorous cause the dopant profiles to spread out, resulting in a uniform phosphorous concentration of 2.0 × 1019 cm−3 over a range of 350 nm in the Ge-on-Si. Implanted ions were retained without dopant loss since the SiO2 capping layers were deposited on the Ge-on-Si prior to thermal annealing. The implanted phosphorous concentration of 1.4 × 1019–1.8 × 1019 cm−3 exceeded the in situ doping concentration of 2.1 × 1018–2.4 × 1018 cm−3.
Figure 3 compares the PL spectra of Ge-on-Si after P implantation with a dose of 1.0 × 1015 cm−2 at an energy of 100 keV. In situ P-doped Ge-on-Si exhibited a higher PL intensity than undoped Ge-on-Si, which is attributed to the indirect L-valley filling effect with high electron concentration . The PL intensity increased after P implantation in the in situ P-doped Ge-on-Si at 750°C and 800°C. The increase in dopant concentration enhanced the photon emission from the direct bandgap. On the other hand, prior to annealing, the as-implanted Ge-on-Si did not exhibit a PL signal because of the number of nonradiative recombination sites due to implantation damage. The PL intensity of implanted Ge-on-Si increased as the damaged Ge-on-Si layers were recrystallized with thermal annealing. In addition, in the implanted samples, there is red shift compared with no implanted samples, which is explained by band gap narrowing (BGN) occurring at high doping.
Figure 4 shows the PL spectra of Ge-on-Si after implantation at different P doses (4 × 1014 cm−2 and 2 × 1015 cm−2) at an energy of 100 keV. The PL intensity of Ge-on-Si after implantation with a dose of 4 × 1014 cm−2 exceeded that of in situ doped Ge-on-Si after annealing at 700°C. The annealing temperature required to exceed the PL intensity of the in situ doped sample is determined by the dopant concentration and lattice damage for different implantation doses. The PL intensity of the sample with an implantation dose of 2 × 1015 cm−2 did not reach that of the in situ doped sample up to the annealing temperature of 800°C. A high implantation dose increased the dopant concentration in Ge-on-Si, but did not necessarily increase the PL intensity. Heavy implantation inevitably involves nonradiative recombination at crystal defects. Therefore, n-type doping concentration and ion-induced crystal defects should be taken care of simultaneously.
Figure 5 shows the variation of sheet resistance (Rsh) and the PL intensity for P-implanted Ge-on-Si. P-implantation into Ge-on-Si with a dose of 2 × 1014–2 × 1015 cm−2 substantially decreased Rsh compared to undoped Ge-on-Si. The PL intensities of P-implanted and in situ doped Ge-on-Si were higher than that of undoped Ge-on-Si. However, an implantation dose of 2 × 1014 cm−2 was not sufficient to exceed the in situ doping, showing higher Rsh and slightly lower PL. At high doses of 4 × 1014 cm−2 and 1 × 1015 cm−2, the PL of P-implanted Ge-on-Si outperformed the in situ doping by 12% and 30%, respectively, at optimum annealing. Even though the crystal lattice was damaged as the implantation dose increased, the n-type doping concentration increased L-valley filling effect for a high emission efficiency. At an implantation dose of 2 × 1015 cm−2, the PL intensity declined. This suggests that lattice damage due to heavy implantation overrides the contribution to PL intensity of a high n-type concentration or a low Rsh.
Figure 6(a) shows the Raman spectra after activation annealing at different temperatures for 100 s. In Fig. 6(b), in-plane tensile strain (ε//) was calculated from the equation Δω = bε//, where Δω is the difference of the Raman shift between the bulk wafer and the Ge layer and b = −408 cm−1 . Tensile strain arises from the difference in the Si and Ge thermal expansion coefficients . During cooling after Ge epitaxy, restoring the original lattice constant of Ge is suppressed by the lattice constant of Si, which generates residual tensile strain in the Ge layer. From this calculation, we determined the in-plane tensile strain to be ~0.2% ± 0.04% after annealing at various temperatures. This tensile strain is comparable with that before ion implantation, which suggests that the implantation and thermal annealing processes do not affect the tensile strain.
The Raman peak of as-implanted Ge was not detected, owing to the implantation damage; however, it was detected after annealing. This can be explained by the fact that annealing is sufficient to cause a phase transition from amorphous to single–crystal structures for implanted Ge-on-Si. The Raman FWHM decreased in the interval of 400°C to 650°C, indicating that recrystallization occurred in this temperature range. However, there were insignificant changes in the crystalline surface recovery above 650°C. Considering the previous results for PL intensity, the increase of PL intensity above 650°C is more influenced by the dopant activation than by crystalline recovery.
We performed phosphorus implantation into in situ doped Ge-on-Si to examine whether an ion implantation actually improves the light-emitting efficiency without damaging the crystal lattice. We used various characterization techniques, namely PL, SIMS, four-point probe measurements, Raman spectroscopy, and HRTEM to study the optical, electrical, and structural properties of the P-implanted Ge-on-Si. We found that P implantation into Ge-on-Si with doses of 4 × 1014–1 × 1015 cm−2 and annealing at 700°C –800°C indeed enhanced the emission efficiency by 12%–30% compared to in situ doped Ge-on-Si. Our results indicate the importance of optimized annealing conditions for implantation doses in order to achieve maximum light-emitting efficiency . The enhanced light-emitting efficiency demonstrates the feasibility of ion implantation as an alternative to in situ doping for high-efficiency Ge light emission.
This work was supported by the Future Semiconductor Device Technology Development Program (10044735) funded by MOTIE (Ministry of Trade, Industry & Energy) and KSRC (Korea Semiconductor Research Consortium). This research was supported by the MSIP (Ministry of Science, ICT and Future Planning), Korea, under the “IT Consilience Creative Program” (IITP-2015-R0346-15-1008) supervised by the IITP(Institute for Information & Communications Technology Promotion).
References and links
1. K. C. Saraswat and F. Mohammadi, “Effect of scaling of interconnections on the time delay of VLSI circuits,” Solid-State Circuits, IEEE Journal of 17(2), 275–280 (1982). [CrossRef]
2. M. T. Bohr, “Interconnect scaling-the real limiter to high performance ULSI,” in International Electron Devices Meeting, (Institute Of Electrical & Electronic Engineers, Inc (IEEE), 1995), 241–244. [CrossRef]
3. D. A. Miller, “Physical reasons for optical interconnection,” Int. J. Optoelectron. 11, 155–168 (1997).
4. D. A. Miller, “Rationale and challenges for optical interconnects to electronic chips,” Proc. IEEE 88(6), 728–749 (2000). [CrossRef]
6. L. Kimerling, D. Ahn, A. Apsel, M. Beals, D. Carothers, Y.-K. Chen, T. Conway, D. Gill, M. Grove, and C.-Y. Hong, “Electronic-photonic integrated circuits on the CMOS platform,” in Integrated Optoelectronic Devices 2006, (International Society for Optics and Photonics, 2006), 612502.
7. C. Gunn, “10 Gb/s CMOS photonics technology,” in Integrated Optoelectronic Devices 2006, (International Society for Optics and Photonics, 2006), 612501.
8. 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]
9. H. Liu, T. Wang, Q. Jiang, R. Hogg, F. Tutu, F. Pozzi, and A. Seeds, “Long-wavelength InAs/GaAs quantum-dot laser diode monolithically grown on Ge substrate,” Nat. Photonics 5(7), 416–419 (2011). [CrossRef]
10. D. Nam, D. Sukhdeo, S.-L. Cheng, A. Roy, K. C.-Y. 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]
11. Y. Cai, R. E. Camacho-Aguilera, J. T. Bessette, L. C. Kimerling, and J. Michel, “High n-type doped germanium for electrically pumped Ge laser,” in Integrated Photonics Research, Silicon and Nanophotonics, (Optical Society of America, 2012), IM3A. 5.
12. 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), 11272–11277 (2007). [CrossRef] [PubMed]
14. J. Liu, L. C. Kimerling, and J. Michel, “Monolithic Ge-on-Si lasers for large-scale electronic–photonic integration,” Semicond. Sci. Technol. 27(9), 094006 (2012). [CrossRef]
15. X. Xu, K. Nishida, K. Sawano, T. Maruizumi, and Y. Shiraki, “Resonant photoluminescence from Ge microdisks on Ge-on-insulator,” in 2014 7th International Silicon-Germanium Technology and Device Meeting (ISTDM), 2014). [CrossRef]
16. R. E. Camacho-Aguilera, Y. Cai, J. T. Bessette, L. C. Kimerling, and J. Michel, “High active carrier concentration in n-type, thin film Ge using delta-doping,” Opt. Mater. Express 2(11), 1462–1469 (2012). [CrossRef]
17. 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] [PubMed]
18. J. Liu, R. Camacho-Aguilera, J. T. Bessette, X. Sun, X. Wang, Y. Cai, L. C. Kimerling, and J. Michel, “Ge-on-Si optoelectronics,” Thin Solid Films 520(8), 3354–3360 (2012). [CrossRef]
19. J. Kim, S. W. Bedell, S. L. Maurer, R. Loesing, and D. K. Sadana, “Activation of implanted n-type dopants in Ge over the active concentration of 1× 1020 cm− 3 using coimplantation of Sb and P,” Electrochem. Solid-State Lett. 13(1), H12–H15 (2010). [CrossRef]
20. J. Kim, S. W. Bedell, and D. K. Sadana, “Multiple implantation and multiple annealing of phosphorus doped germanium to achieve n-type activation near the theoretical limit,” Appl. Phys. Lett. 101(11), 112107 (2012). [CrossRef]
21. F. Cerdeira, C. Buchenauer, F. H. Pollak, and M. Cardona, “Stress-induced shifts of first-order Raman frequencies of diamond-and zinc-blende-type semiconductors,” Phys. Rev. B 5(2), 580–593 (1972). [CrossRef]
22. Y. Ishikawa, K. Wada, D. D. Cannon, J. Liu, H.-C. Luan, and L. C. Kimerling, “Strain-induced band gap shrinkage in Ge grown on Si substrate,” Appl. Phys. Lett. 82(13), 2044–2046 (2003). [CrossRef]