In this paper, deep level transient spectroscopy (DLTS) characterization was performed on Beryllium compensation doping of InGaAs/GaAsSb type-II superlattice photodiode. Three electron traps with the energy levels located at Ec-0.11 eV (E1), Ec-0.28 eV (E2), Ec-0.17 eV (E3), and a hole trap situated at Ev + 0.25 eV (H1) were revealed. The position distribution and depth concentration of these traps in SL absorption region was also explored. Furthermore, the bandlike states (E2) and localized states (E1 and H1) of extended defects were confirmed by DLTS measurements as a function of the filling-pulse time, these traps as generation-recombination centers are responsible for dominant dark current.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Type-II superlattice (T2SL) photodiodes based on the InGaAs/GaAsSb material system have attracted a lot of attention for short-infrared and mid-infrared (2 μm to 5 μm) detection, which have potential applications in infrared imaging, environmental monitoring, medical diagnostics, gas sensing, free-space communications, etc [1,2]. The InGaAs and GaAsSb are both lattice matched to InP substrates, and the detection wavelength of photodiodes can be arbitrarily extended by changing the layer thickness, composition and structure of each material in the SL region to meet practical application requirements . So far, in order to further extend the detection wavelength, the effective approach is to use the strain-compensated absorption region [4–6]. Chen et al. have been successfully achieved optical response of 3 µm using a novel strain-compensated type-II InGaAs/GaAsSb quantum well active region, and the superior detectivity of 7.73 × 109 cmHz1/2W−1 at 290 K for 2.7 μm was obtained . Thus, T2SL photodiodes have the advantage of operating without deep cooling for the application of short-infrared and mid-infrared wavelength .
Nevertheless, the quantum efficiency (QE) of InGaAs/GaAsSb T2SL photodiodes at bias close to zero was less than 30% in the reported literatures . As our previous reported in , it has been proposed and experimentally demonstrated that Beryllium (Be) compensation doping in the active region can significantly improve the InGaAs/GaAsSb type-II SL photodiodes’ QE up to 48.2%. By doping Be in the absorption region, the residually n-type SL is compensated to become slightly p-type. The photo generated minority carriers become electrons and they have longer diffusion length owing to the higher electron mobility . Thus, the improvement of photodiodes’ QE was successfully realized. In addition, T2SL photodiode performance is often limited by excess dark current due to a large numbers of traps in the absorption layer. In these traps, the electronically active defects act as generation-recombination (G-R) centers, resulting in the increase of the dark current and poor carrier lifetime, and then the sensitivity performance of photodiodes is seriously degraded . Therefore, further effort is urgently required to identify the defects in the SL region and evaluate their role in affecting the performance of photodiodes. It is important for optimizing material quality, heterojunctions of SL and detector performance for future-generation infrared detectors.
In this work, deep level transient spectroscopy (DLTS) characterization was performed on Be doped InGaAs/GaAsSb type-II SL photodiode to obtain the activation energies, capture cross sections and trap concentrations of all the traps. Pushing the DLTS probing region from bulk to near the surface of SL, the position distribution and depth concentration of these traps in SL absorption region was also explored. Furthermore, the physical properties of the revealed defects were determined by the DLTS measurements as a function of the filling-pulse time.
2. Experimental procedures
The schematic structure of InGaAs/GaAsSb T2SL photodiode is shown in Fig. 1. In this experiment, the lattice-matched In0.53Ga0.47As/GaAs0.49Sb0.51 type-II SL photodiodes were grown by solid source molecular beam epitaxy (MBE) on (100) InP substrates. The growth rates were 1 ML/s for InGaAs and 0.5 ML/s for GaAsSb, respectively. Silicon (Si) and Beryllium (Be) were used as n-type and p-type dopants, respectively. The SL photodiode consists of a 500 nm thick n-type InAl0.52As0.48 bottom contact layer with a doping level of 1 × 1018 cm−3, 300 pairs of 7 nm-thick In0.53Ga0.47As and 5 nm-thick GaAs0.49Sb0.51 SL p-type absorption region with Be doping concentration of 9.5 × 1015 cm−3, and a 500 nm thick p-type InAl0.52As0.48 layer with a doping concentration of 1 × 1018 cm−3. The SL photodiodes were mesa isolated, using wet etched with mixed solution of phosphoric acid, citric acid and hydrogen peroxide, and Au n-type and p-type ohmic contacts were deposited by electron-beam evaporation on both the bottom and the top contact layers. It should be pointed out that there is no dielectric layer passivation on the detector surface and sidewalls. A more detailed process have been published elsewhere . Deep level characterization of the photodiodes were carried out using DLS-83D Deep Level Transient Spectroscopy test system. Temperature ranging from 80~300 K was selected by a temperature controller at a heating rate of 0.5 Ks−1.
3. Results and discussion
Figure 2(a) shows the current-voltage (I–V) characteristics of Be doped InGaAs/GaAsSb T2SL photodiode measured in the temperature range from 77 K to 300 K. It can be seen that the dark current has strong temperature dependence. It is commonly known that the dark current consists of the following four mechanisms in the inset of Fig. 2(b): (1) diffusion; (2) generation-recombination (G-R); (3) defect assisted tunneling (TAT) and (4) band-to-band tunneling (BTBT) [13,14]. To illustrate the dominant mechanism of dark current, the dark current density as a function of inverse temperature is given in Fig. 2(b). It is noteworthy that the calculated activation energy Ea is 0.24 eV under an operating bias of −50 mV, extracted from the slope of the linear fitting curve in the temperature range of 200~300 K, based on . It means the dark current of the photodiode is mainly dominated by the G-R current because the activation energy is about 1/2 Eg for the InGaAs/GaAsSb SL (an effective band gap Eg of 0.49 eV) . The result indicates that carriers through near-mid-gap traps make a greater contribution to the dark current. Moreover, the extracted activation energy always signifies the average energy of several traps rather than a single defect level. Providing that a large number of deep level defects in the photodiodes, these trap centers may release electrons and participate in the carrier transport process, which lead to the larger G-R current.
A better explanation for the origin of the dark current in the photodiode needs identification of the defects by DLTS measurements. Figure 3 shows the DLTS spectra for Beryllium compensation doping of InGaAs/GaAsSb T2SL photodiode in the temperature range from 77 K to 300 K. The reverse bias voltage VR was tuned from −2 V to −0.2 V. The applied pulse VP of 0.5 V and a pulse width tp of 20 μs were chosen in order to ensure saturation trap filling. The lock-in frequencies f0 was set as 680 Hz for each temperature scan. The amplitude of peak detected at the lock-in frequencies (T = 1/f0) can be expressed 15]. Keeping pulse voltage fixed while increasing the reverse voltage modulus, the maximum peak of E2 and H1 shift towards lower temperatures in DLTS spectra, which are attributed to bulk trap related peaks. Owing to the bulk trap peaks with strong emission rate dependence on the electric field, the increase of reverse voltage modulus leads to an effective addition of active traps in a stronger electric field area, and then allows the maximum peak shift towards lower temperature . A .V. P. Coelho et al. proposed and verified this method to distinguish the bulk traps and interface states related peaks in DLTS spectra, and identical phenomenon related to the bulk traps was obtained in studies of GaAs diodes . It is worthwhile mentioning that the depth concentration profile of the observed traps was also reflected by adjusting the variation of depletion region. When the reverse voltage VR was changed from −2 V to −0.2 V, pushing the DLTS probing region from bulk to near the surface of SL, the depletion region edge sweeps approximately from 728 nm to 472 nm below the surface according to . It is found that the electron trap density of E1 and E2 slightly increases and then decreases, while the hole trap concentration of H1 increases rapidly with the drop of reverse voltage modulus. Therefore, it is concluded that the electron traps E1 and E2 reside in the bulk of SL and the hole trap H1 is localized near the interface of SL.
To clearly determine the trap levels of these observed traps, DLTS was repeated at different lock-in frequencies f0 range from 80 Hz to 2200 Hz and fixed other parameters. Figures 4(a)-4(c) show the DLTS spectra of the traps E1 and E2, E3 and H1 at the different reverse bias (VR = −2 V, VR = −0.7 V and VR = −0.5 V), respectively. The activation energy Ea, capture cross-section σ and trap concentration Nt was calculated using the following expressionFig. 4(d). And the positions of all detected traps in the band gap are shown in the Fig. 4(e). The corresponding capture cross sections of 2.63 × 10−16 cm2, 4.93 × 10−17 cm2, 1.01 × 10−18 cm2 and 1.03 × 10−18 cm2 were calculated from the y-intercept of Arrhenius plots, respectively. The high trap densities are 10+13~10+14 cm−3 proportional to DLTS signal. All the defect parameters are calculated from the DLTS scans and are listed in Table 1. It is worth noting that W. Chen etc. have found three deep levels of 0.14 eV, 0.34 eV and 0.43 eV in lattice-matched InGaAs/GaAsSb and strain-compensated InGaAs/GaAsSb using both low-frequency noise spectroscopy (LFNS) and DLTS [12,18,19]. However, these energy levels are different from that in our sample (Be doped InGaAs/GaAsSb), which may be attributed to the factors of Be doping and different thickness etc. It is important to highlight that the Ea extracted from DLTS are consistent with the estimated value obtained from the I-V characteristics, indicating that the Ea (0.24 eV) extracted from the temperature dependence of the dark current denotes an average of multiple trap levels. Hence, it is confirmed that the excess dark current is dominated by G-R current due to the existence of defects in the SL, and the large trap densities are significant contributors to deteriorate photodiode performance.
The next step of our investigations was to distinguish the point or extended character of the observed deep level traps. Introducing defect states within the band gap can be divided into two categories, point defects (such as vacancies, interstitial atoms, etc.) and extended defects (such as dislocations, stacking faults, grain boundaries, point defect clusters, etc.) . In DLTS, the logarithmic capture law is exploited as a principal argument for distinguishing between point defects and extended defects, which can be expressed as a linear dependence of the DLTS-peak amplitude on the logarithm of the filling-pulse time . In contrast, isolated point defects or impurities typically reveal exponential capture kinetics (the exponential capture law) . In Figs. 5(a) and 5(b), the DLTS spectra are shown for different filling-pulse times ranging from 10 μs to 20 ms. It can be seen for the trap E1 and H1, (i) the position of DLTS-peak stays constant, and (ii) the maximum of DLTS signal increases consistently with increasing filling-pulse time and are related to by △C~ ln (). As it is shown in Figs. 5(c) and 5(d), a linear trend is observed in semi-logarithmic plot of DLTS-peak vs. filling-pulse time, indicating two traps all present a typical characteristic of extended traps. This coincides with the observation for dislocations themselves or dislocations decorated with point defects in various semiconductors, such as Si, InGaAs/GaAs, and GaAs, etc [23–25]. The extended defects can be classified as the bandlike states caused by the atomic structures of the defects themselves and the localized states formed by their reaction with point defects. W. Schröter proposed a model by computer simulation and experimental verification, revealing the basic rules of DLTS spectrum of extended defects (bandlike states and localized states) under different filling-pulse time . On the basis of features (i) and (ii), the trap E1 and H1 are determined as localized states of extended defects.
However, bandlike states of extended defects exhibit basic differences from features (i) and (ii). It is noticeable that the observed DLTS peak E2 is a broad asymmetric peak shape, and the low-temperature edge of the DLTS peak shifts to lower temperature with the increase of filling-pulse time. That is, as the pulse time increases, the DLTS peak broadens on its low-temperature side due to strain fields or defect interaction, and the temperature of DLTS-peak decreases. The same phenomenon was also observed and reported in GeSi/Si samples due to the bandlike states of extended defects . On the other hand, for the trap E2, (i) as the filling-pulse time increases, DLTS-peak moves to the low temperature side, and (ii) even if the normalization is not carried out, the peak shape on the high temperature side coincide. It is illustrated that the trap E2 could be a typical characteristic of bandlike states. On the basis of these results, these traps may be associated with dislocation due to the presence of interacting trap levels characteristic for extended defects, causing excess dark current by emission of electrons from these traps in the bulk or interface of SL.
In summary, the G-R dominated dark current observed in Be doped InGaAs/GaAsSb T2SL photodiodes grown by MBE is attributed to the existence of the deep level defects. DLTS measurements were performed by changing probing region from bulk to near the surface of SL in the temperature range of 77~300 K, three electron traps E1 (Ec-0.11 eV), E2 (Ec-0.28 eV), E3 (Ec-0.17 eV) and a hole trap H1 (Ev + 0.25 eV) were explored, respectively. It is also found that the traps E1 and E2 reside in the bulk of SL and the trap H1 is localized near the interface of SL. Besides, the measurements of capture kinetics for the related trap E1 and H1 indicated a linear increase of the DLTS-peak amplitude on increasing of the filling-pulse time in a semi-logarithmic plot, these traps can be attributed to localized states of extended defects, and the revealed features of the trap E2 is related to bandlike states of extended defects. It is concluded that the excess dark currents are caused by the extended defects related to dislocation and it is helpful to optimize the growth of SL and detector performance.
This work was supported by the National Key Research and Development Program of China (No. 2016YFB0402403 and No. 2016YFGX020080), the China Academy of Space Technology Foundation (No. CAST 08201601).
References and links
1. B. Chen, W. Jiang, J. Yuan, A. L. Holmes, and B. M. Onat, “SWIR/MWIR InP-based p-i-n photodiodes with InGaAs/GaAsSb Type-II quantum wells,” IEEE J. Quantum Electron. 47(9), 1244–1250 (2011). [CrossRef]
2. A. Yamamoto, Y. Kawamura, H. Naito, and N. Inoue, “Optical properties of GaAs0.5Sb0.5 and In0.53Ga0.47As/GaAs0.5Sb0.5 type II single hetero-structures lattice-matched to InP substrates grown by molecular beam epitaxy,” J. Cryst. Growth 201(3), 872–876 (1999). [CrossRef]
3. R. Sidhu, N. Duan, J. C. Campbell, and A. L. Holmes, “A long-wavelength photodiode on InP using lattice-matched GaInAs-GaAsSb type-II quantum wells,” IEEE Photonics Technol. Lett. 17(12), 2715–2717 (2005). [CrossRef]
4. B. Chen, “Active region design and gain characteristics of InP-based dilute Bismide type-II quantum wells for Mid-IR lasers,” IEEE Electron Dev. 64(4), 1–6 (2017). [CrossRef]
5. B. Chen, A. L. Holmes, W. Y. Jiang, and J. Yuan, “Design of strain compensated InGaAs/GaAsSb type-II quantum well structures for mid-infrared photodiodes,” Opt. Quantum Electron. 44(3–5), 103–109 (2012). [CrossRef]
6. B. Chen, W. Y. Jiang, J. Yuan, A. L. Holmes, and B. M. Onat, “Demonstration of a room-temperature InP-based photodetector operating beyond 3 μm,” IEEE Photonics Technol. Lett. 23(4), 218–220 (2011). [CrossRef]
7. B. Chen and A. L. Holmes Jr., “InP-based short-wave infrared and midwave infrared photodiodes using a novel type-II strain-compensated quantum well absorption region,” Opt. Lett. 38(15), 2750–2753 (2013). [CrossRef] [PubMed]
8. T. Kawahara, K. Machinaga, B. Sundararajan, K. Miura, M. Migita, H. Obi, T. Fuyuki, K. Fujii, T. Ishizuka, H. Inada, and Y. Iguchi, “InGaAs/GaAsSb type-II quantum-well focal plane array with cutoff-wavelength of 2.5 μm,” Proc. SPIE 10111, 1011115 (2017). [CrossRef]
9. G. F. Fulop, “Extended wavelength SWIR detectors with reduced dark current,” Proc. SPIE 9451, 945106 (2015). [CrossRef]
10. C. Jin, F. Wang, Q. Xu, C. Yu, J. Chen, and L. He, “Beryllium compensation doped InGaAs/GaAsSb superlattice photodiodes,” J. Cryst. Growth 477, 100–103 (2017). [CrossRef]
11. D. Jiang, W. Xiang, F. Guo, H. Hao, X. Han, X. Li, G. Wang, Y. Xu, Q. Yu, and Z. Niu, “Very high quantum efficiency in InAs/GaSb superlattice for very long wavelength detection with cutoff of 21 μm,” Appl. Phys. Lett. 108(12), 61–161 (2016). [CrossRef]
12. W. Chen, B. Chen, J. Yuan, A. Holmes, and P. Fay, “Bulk and interfacial deep levels observed in In0.53Ga0.47As/GaAs0.5Sb0.5 multiple quantum well photodiode,” Appl. Phys. Lett. 101(5), 59–390 (2012). [CrossRef]
13. D. Hoffman, B. M. Nguyen, P. Y. Delaunay, A. Hood, M. Razeghi, and J. Pellegrino, “Beryllium compensation doping of InAs/GaSb infrared superlattice photodiodes,” Appl. Phys. Lett. 91(14), 085316 (2007). [CrossRef]
14. B. Chen, J. Yuan, and A. L. Holmes, “Dark current modeling of InP based SWIR and MWIR InGaAs/GaAsSb type-II MQW photodiodes,” Opt. Quantum Electron. 45(3), 271–277 (2013). [CrossRef]
15. G. L. Miller, D. V. L. And, and L. C. Kimerling, “Capacitance transient spectroscopy,” Annu. Rev. Mater. Res. 7(7), 377–448 (1977).
16. K. Yamasaki, M. Yoshida, and T. Sugano, “Deep level transient spectroscopy of bulk traps and interface states in Si MOS diodes,” Jpn. J. Appl. Phys. 18(1), 113–122 (1979). [CrossRef]
17. A. V. P. Coelho, M. C. Adam, and H. Boudinov, “Distinguishing bulk traps and interface states in deep-level transient spectroscopy,” J. Phys. D Appl. Phys. 44(30), 416–426 (2011). [CrossRef]
18. W. Chen, B. Chen, A. Holmes, and P. Fayet, “Investigation of traps in strained-well InGaAs/GaAsSb quantum well photodiodes,” Electron. Lett. 51(18), 1439–1440 (2015). [CrossRef]
19. W. Chen, B. Chen, J. Yuan, A. Holmes, and P. Fay, “Characterization and impact of traps in lattice-matched and strain-compensated In1−xGaxAs/GaAs1−ySby multiple quantum well photodiodes,” in IEEE Device Research Conference (IEEE, 2012), pp. 251–252.
20. A. V. P. Coelho and H. Boudinov, “Emission rate dependence on the electric field for two trap levels in proton-irradiated n-type GaAs,” Phys. Rev. B 77(23), 235210 (2008). [CrossRef]
21. P. Omling, E. R. Weber, L. Montelius, H. Alexander, and J. Michel, “Electrical properties of dislocations and point defects in plastically deformed silicon,” Phys. Rev. B Condens. Matter 32(10), 6571–6581 (1985). [CrossRef] [PubMed]
22. Ł. Gelczuk, M. Dąbrowska-Szata, P. Kamyczek, E. Płaczek-Popko, K. Kopalko, B. Ściana, D. Pucicki, D. Radziewicz, and M. Tłaczała, “Investigation of deep-level defects in InGaAsN/GaAs 3xQWs structures grown by AP-MOVPE,” Solid-State Electron. 8902(21), 306–309 (2013).
23. T. Wosiński, “Evidence for the electron traps at dislocations in GaAs crystals,” J. Appl. Phys. 65(4), 1566–1570 (1989). [CrossRef]
24. D. L. Dexter and F. Seitz, “Effects of dislocations on mobilities in semiconductors,” Phys. Rev. 86(6), 964–965 (1952). [CrossRef]
25. A. Y. Du, M. F. Li, T. C. Chong, K. L. Teo, W. S. Lau, and Z. Zhang, “Dislocations and related traps in p-InGaAs/GaAs lattice mismatched heterostructures,” Appl. Phys. Lett. 69(19), 2849–2851 (1996). [CrossRef]
26. P. N. Grillot, S. A. Ringel, E. A. Fitzgerald, G. P. Watson, and Y. H. Xie, “Electron trapping kinetics at dislocations in relaxed Ge0.3Si0.7/Si heterostructures,” J. Appl. Phys. 77(7), 3248–3256 (1995).
27. W. Schröter, J. Kronewitz, U. Gnauert, F. Riedel, and M. Seibt, “Bandlike and localized states at extended defects in silicon,” Phys. Rev. B Condens. Matter 52(19), 13726–13729 (1995). [CrossRef] [PubMed]