Open Access
Add to BibSonomy
Abstract
A p++-AlGaAs: C/n++-InGaP: Te tunnel junction with a record peak tunneling current density of 5518 A/cm2 was developed. This was achieved by inserting a 6.6 Å undoped GaAs quantum well at the junction interface, and the numerical model demonstrated that trap-assisted tunneling contributes to the high peak tunneling current. Furthermore, we found that the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions have lower resistance and better stability than p++-AlGaAs: C/n++-InGaP: Te tunnel junctions in the operating temperature range of the multijunction solar cells, and the peak tunneling current density of the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions excess 3000 A/cm2 with a voltage drop of 7.5 mV at 10000 suns.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Solar energy is abundant and environmentally friendly. Efforts to extract power from solar energy have benefited from the high efficiency of solar cell technology [1,2]. Multijunction solar cells attract more attention than traditional structures because of their higher photoelectric conversion efficiency [1]. Tunnel junctions connect the subcells of the multijunction solar cells vertically, and each subcell is designed to absorb a specific wavelength range of the solar spectrum [3]. It implies that high performance tunnel junctions are essential for multijunction solar cells [4].
The high bandgap tunnel junction transmits the shorter wavelength solar spectrum to the subcells to reach the current match of each cell. Consider the InGaP/GaAs/Ge triple-junction solar cell, where the tunnel junction between the InGaP and GaAs subcells could be p++-AlGaAs/n++-AlGaAs, p++-AlGaAs/n++-InGaP, p++-AlGaAs/n++-GaAs or p++-GaAs/n++-GaAs [4], but the n++-GaAs layer will absorb the spectrum intended for the subsequent subcell, resulting in a 1% efficiency reduction [5]. The p++-AlGaAs/n++-InGaP structure was selected for this study. This tunnel junction must provide sufficiently high peak current and low resistance without introducing parasitic absorption or impeding current flow between adjacent junctions. The peak tunneling current is expressed as [4]
Due to the memory effects, surface segregations, hysteresis, and solubility limitations of the Te dopant in the InGaP layer, obtaining heavily Te doped at the n++-InGaP layer is difficult [6], resulting in an insufficient abrupt doping profile at the junction interface [7]. Furthermore, the composition abruptness is difficult to achieve when As/P needs to switch on and off, requiring the overpressure of group V precursors during growth. A promising way to solve these issues is to insert a GaAs quantum well at the interface of the p++AlGaAs/n++InGaP tunnel junction. Joshua et al. proposed inserting 3 nm n++-GaAs at the junction interface to obtain a peak current of 1000 A/cm2 [7]. Enrique et al. proposed the insertion of 3 nm p++-GaAs to realize a peak tunneling current density of 996 A/cm2 [8]. However, This paper proposed an undoped GaAs quantum well inserted at the tunnel junction interface due to the undoped quantum well has many advantages: 1) The undoped quantum well layer suppresses the self-compensation effect caused by the diffusion of the dopants to achieve an abort doping profile; 2) The undoped quantum well layer moves the valance band of the quantum well into the tunneling region, as shown in Fig. 1, shortening the tunneling distance and increasing the tunneling current. The depletion region of the p++AlGaAs/i-GaAs/n++InGaP tunnel junction is much wider than that of p++AlGaAs /n++InGaP tunnel junction.
Fig. 1. Band structure of (1) p++AlGaAs/i-GaAs/n++InGaP (red solid line) and (2) p++AlGaAs /n++InGaP (blue dashed line) tunnel junctions. The differences in band bending shorten the tunneling distance for the GaAs quantum well inserted in the p++AlGaAs /n++InGaP tunnel junction structure.
In this paper, we first investigated the p++-AlGaAs: C/n++-InGaP: Te tunnel junctions at various GaAs quantum well thickness, obtaining a record peak tunneling current density of 5518 A/cm2. Furthermore, the performances of the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions were investigated in the operating temperature range of the multijunction solar cells, demonstrating that the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions have lower resistance and higher reliability than p++-AlGaAs: C/n++-InGaP: Te tunnel junctions, and the peak tunneling current density of this tunnel junctions excess 3000 A/cm2 with a voltage drop of 7.5 mV at 10000 suns.
2. Experimental
A series of tunnel junctions were grown on (100) n-type GaAs substrates misoriented 6° toward the <111> A direction by metal-organic chemical vapor deposition (MOCVD) [9]. The C (CBr4 source), Si (Si2H6 source) and (or) Te (DeTe source) were used as p-type and n-type dopants, respectively. The ECV (electrical capacitance-voltage) profile and SIMS (secondary ion mass spectroscopy) were used to assess the doping levels. A typical tunnel junction structure included a 30-nm-thick n++-InGaP layer and a 20-nm-thick p++-AlGaAs layer, as well as an undoped GaAs quantum well layer inserted at the p++/n++ junction interface, and the p++/n++ junction was surrounded by a 150 nm n-type (2×1018 cm−3) GaAs buffer layer and a 100 nm p-type (5×1019 cm−3) GaAs cap layer [10]. The growth rate of the p++-AlGaAs layer is 3.0 Å/s, the growth rate of the GaAs layer is 1.0 Å/s, the growth rate of the n++-InGaP layer is 2.5 Å/s, and the tunnel junction was annealed at 600°C for 15 minutes. The experimental conditions are summarized in Table 1, The dopants of the n++-InGaP layer for D0365, D0366 and D0366 are the same, while the thicknesses of the undoped GaAs quantum well are different. D0365 and D0369 have the same thickness of GaAs quantum wells and different dopants of n++-InGaP layer.
Table 1. Abbreviation of experiment
Conventional photolithographic and wet-etching techniques were used to fabricate the tunnel junction devices, and the device structure is schematically depicted in Fig. 2. Tunnel junction devices of various sizes were patterned to simulate resistance [11]. Figure 3 depicts a cross-section of the tunnel junction devices, where the SiOx was deposited using plasma-enhanced chemical vapor deposition (PECVD) to isolate the mesa sidewalls, and the Ti/Pt/Au and AuGe/Ni/Au contact metals were deposited onto the front and backside of the tunnel junctions, respectively. The J-V curve was measured using the four-probe technique to ensure accurate resistance measurements.
Fig. 3. (a) SEM image of the tunnel junction devices. (b) The magnified image shows the cross-section of the tunnel junction structure.
3. Result and discussion
We investigated several processes to obtain an abrupt dopant profile in the InGaP layer due to the well-known issue of tellurium turning on and off in InGaP. Tellurium dopant exhibits a “soft” turn-on and turn–off properties, making it unsuitable for tunnel junctions unless specific procedures are followed [6]. As shown in Fig. 4, pre-doping with DeTe prior to growing the InGaP layer ensures the abrupt doping profile of the front interface. Furthermore, Te atoms tend to attach at the step edges of the junction interface, resulting in Te dopants incorporation into the first few GaAs layers, increasing the depletion width of the tunneling region and thereby lowering the peak tunneling current density values [12]. In order to solve this problem, we turn off DeTe prior to the end of InGaP growth. Following the end of the InGaP layer growth, a 4–5 minutes growth pause was added to raise the wafer temperature by 30°C in a phosphine environment. As a result of the growth pause and the elevated wafer temperature, tellurium adsorbed on the wafer surface sublimated and formed an abrupt back junction interface doping profile. The undoped GaAs quantum well layer then began to grow after a 4-5 minutes AsH3 purge.
Fig. 4. Growth details of the InGaP/GaAs. 1: Te pre-doping; 2: Te early shut-off; 3: Elevated temperature growth pause under phosphine; 4: Pause before GaAs growth under arsine.
The peak tunneling current density results of the undoped GaAs quantum well layer with different thicknesses are shown in Fig. 5 (a). The peak tunneling current density increases from 4686 A/cm2 to 5518 A/cm2 as the thickness of the undoped GaAs quantum well increases from 3.3 Å to 10 Å, with the highest peak tunneling current density achieved at 6.6 Å. According to energy band theory, the tunneling distance decreases as the thickness of the undoped GaAs quantum well decreases, increasing tunneling peak current density [5]. However, the experimental results indicate that the maximum peak tunnel current density occurs at a quantum well thickness of 6.6 Å. At this thickness, a balance is achieved between the p++/n++ dopants’ self-compensation in the thinner quantum well and the decrease in tunneling characteristics in the thicker quantum well. Figure 5 (b) shows a box and whisker chart of the peak tunnel current density for the entire sample set, indicating that the device reproducibility is sensitive to material growth and manufacturing process uniformity.
Fig. 5. (a) Current density versus voltage plots of tunnel junctions with different GaAs quantum well thickness. (b) The box and whisker charts show the peak tunnel current density for the entire sample set. The box defines the 25th and 75th percentiles and the median value of the set, the hollow square is the mean value, and the whiskers show the range.
However, the peak tunneling current density of 5875 A/cm2 in this study is higher than the previous record of 2000 A/cm2 [12], which cannot only be explained as the direct band to band tunneling. The trap-assisted tunneling (TAT) effect must be considered. A simulation model including the band to band tunneling effect, bandgap narrow effect, and trap-assisted tunneling effect is proposed to evaluate the contribution of the TAT [10]. Figure 6 compares the simulated and experimental results, demonstrating that the TAT effect enhances the peak tunneling current density. Since the tellurium has a larger atomic size than phosphorus, high concentrations of Te doped in the InGaP layer cause strain relaxation and defects, which act as traps during the TAT process [13].
In this study, we also investigated the Si + Te co-doped n++-InGaP layer. Because Te is likely to act as a surfactant that aids Si incorporation, thus improves the material quality [14], Si compensates for the delay time between the injection of DeTe into the reactor and the onset of Te incorporation into the epitaxial layer [10,15]. As shown in Fig. 7, we compared the performances of p++-AlGaAs: C/n++-InGaP: Te (D0365) and p++-AlGaAs: C/n++-InGaP: Si + Te (D0369) tunnel junctions. D0365 has a higher peak tunnel current density than D0369. This result can be explained that the Si dopant in the InGaP layer has a lower saturation doping level, resulting in a lower Si + Te co-doping concentration than Te doping concentration, lowering the peak tunneling current density (see Eq. (1)). Besides, the box and whisker chart of the peak tunnel current density shows that the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions have higher device reproducibility than p++-AlGaAs: C/n++-InGaP: Te tunnel junctions.
Fig. 7. Current density versus voltage plots for the p++-AlGaAs: C/n++-InGaP: Te (D0365) and p++-AlGaAs: C/n++-InGaP: Si + Te (D0369) tunnel junctions. Adjacent to each curve is a box and whisker chart of the peak tunnel current density for the entire sample set.
In order to evaluate the influence of the tunnel junction properties on the performance of the multijunction solar cells under actual operating conditions, the peak tunneling current density and resistance of the tunnel junction under typical multi-junction solar cells operation temperatures (10-60°C) were investigated [16]. Figure 8 (a) shows that the resistance decreases with increasing temperature. Figure 8 (b) and (c) show the box and whisker chart of the peak tunnel current density and resistance results for different samples in the operating temperature range of the multijunction solar cells. It indicates that the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions are more stable than p++-AlGaAs: C/n++-InGaP: Te tunnel junctions because the silicon dopant in the Si + Te co-doping condition reduces the trap density and improves the stability of the tunnel junction devices. Figure 8 (c) also shows that p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions have lower resistance than p++-AlGaAs: C/n++-InGaP: Te tunnel junctions due to Si + Te co-doped layer can improve the mobility [14].
Fig. 8. (a) Experimental results of tunnel-junction devices under various temperatures. (b) Box and whisker chart of the peak tunneling current density results under various temperatures. (c) Box and whisker chart of the resistance results under various temperatures.
The voltage drop of the p++-AlGaAs: C/n++-InGaP: Te tunnel junctions and p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions for the current density equivalent to the operation of the multijunction solar cells up to 10000 suns is less than 10.6 and 7.5 mV (assuming 1 sun short-circuit current density of 15 mA/cm2 [8]), respectively, which can be used as a low-loss interconnection in the multi-junction solar cells that operate at ultra-high concentrations.
4. Conclusion
A set of tunnel junctions were investigated based on a p++-AlGaAs/n++-InGaP heterojunction grown by MOCVD. The best device incorporates a 6.6 Å undoped GaAs quantum well at the p++-AlGaAs: C/n++-InGaP: Te tunnel junction interface, achieving a record peak current density of 5518 A/cm2. Furthermore, the performances of the p++-AlGaAs: C/n++-InGaP: Si + Te tunnel junctions were compared to p++-AlGaAs: C/n++-InGaP: Te tunnel junctions in the operating temperature range of the multijunction solar cells. The results demonstrated that the Si + Te co-doped tunnel junctions have lower resistance and higher reliability than Te doped tunnel junctions, and the Si + Te co-doped tunnel junctions exhibit a peak tunneling current density of more than 3000 A/cm2 with a voltage drop at 10000 suns of 7.5 mV.
Disclosures
The authors declare that there are no conflicts of interest related to this article.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. J. F. Geisz, R. M. France, K. L. Schulte, M. A. Steiner, A. G. Norman, H. L. Guthrey, M. R. Young, T. Song, and T. Moriarty, “Six-junction III–V solar cells with 47.1% conversion efficiency under 143 Suns concentration,” Nat. Energy 5(4), 326–335 (2020). [CrossRef]
2. E. Kabir, P. Kumar, S. Kumar, A. A. Adelodun, and K.-H. Kim, “Solar energy: Potential and future prospects,” Renewable Sustainable Energy Rev. 82, 894–900 (2018). [CrossRef]
3. P. Colter, B. Hagar, and S. Bedair, “Tunnel junctions for III-V multijunction solar cells review,” Crystals 8(12), 445 (2018). [CrossRef]
4. J. F. Wheeldon, C. E. Valdivia, A. W. Walker, G. Kolhatkar, A. Jaouad, A. Turala, B. Riel, D. Masson, N. Puetz, and S. Fafard, “Performance comparison of AlGaAs, GaAs and InGaP tunnel junctions for concentrated multijunction solar cells,” Prog. Photovolt: Res. Appl. 19(4), 442–452 (2011). [CrossRef]
5. S. M. Bedair, J. L. Harmon, C. Z. Carlin, I. E. H. Sayed, and P. Colter, “Annealed high band gap tunnel junctions with peak current densities above 800 A/cm 2,” in 2016 IEEE 43rd Photovoltaic Specialists Conference (PVSC), (IEEE, 2016), 2320–2322.
6. C. Ebert, Z. Pulwin, D. Byrnes, A. Paranjpe, and W. Zhang, “Tellurium doping of InGaP for tunnel junction applications in triple junction solar cells,” J. Cryst. Growth 315(1), 61–63 (2011). [CrossRef]
7. J. P. Samberg, C. Zachary Carlin, G. K. Bradshaw, P. C. Colter, J. L. Harmon, J. Allen, J. R. Hauser, and S. Bedair, “Effect of GaAs interfacial layer on the performance of high bandgap tunnel junctions for multijunction solar cells,” Appl. Phys. Lett. 103(10), 103503 (2013). [CrossRef]
8. E. Barrigón, I. García, L. Barrutia, I. Rey-Stolle, and C. Algora, “Highly conductive p++-AlGaAs/n++-GaInP tunnel junctions for ultra-high concentrator solar cells,” Progress in Photovoltaics: Res. Appl. 22(4), 399–404 (2014). [CrossRef]
9. Y. Gou, J. Wang, Y. Cheng, Y. Guo, X. Xiao, and S. Zhou, “A Modeling and Experimental Study on the Growth of VCSEL Materials Using an 8× 6 Inch Planetary MOCVD Reactor,” Coatings 10(8), 797 (2020). [CrossRef]
10. Y. Gou, J. Wang, Y. Cheng, Y. Guo, X. Xiao, H. Liu, S. Tan, L. Zhou, H. Yang, and G. Deng, “Experimental and modeling study on the high-performance p++-GaAs/n++-GaAs tunnel junctions with silicon and tellurium co-doped InGaAs quantum well inserted,” Crystals 10(12), 1092 (2020). [CrossRef]
11. M. Beattie, C. Valdivia, M. Wilkins, M. Zamiri, K. Kaller, M. Tam, H. Kim, J. Krich, Z. Wasilewski, and K. Hinzer, “High current density tunnel diodes for multi-junction photovoltaic devices on InP substrates,” Appl. Phys. Lett. 118(6), 062101 (2021). [CrossRef]
12. S. Bedair, J. L. Harmon, C. Z. Carlin, I. E. Hashem Sayed, and P. Colter, “High performance as-grown and annealed high band gap tunnel junctions: Te behavior at the interface,” Appl. Phys. Lett. 108(20), 203903 (2016). [CrossRef]
13. I. García, I. Rey-Stolle, and C. Algora, “Performance analysis of AlGaAs/GaAs tunnel junctions for ultra-high concentration photovoltaics,” J. Phys. D: Appl. Phys. 45(4), 045101 (2012). [CrossRef]
14. J. Law, A. Carter, S. Lee, C.-Y. Huang, H. Lu, M. Rodwell, and A. Gossard, “Co-doping of InxGa1− xAs with silicon and tellurium for improved ultra-low contact resistance,” J. Cryst. Growth 378, 92–95 (2013). [CrossRef]
15. X. S. Nguyen, S. Yadav, K. H. Lee, D. Kohen, A. Kumar, R. I. Made, K. E. K. Lee, S. J. Chua, X. Gong, and E. A. Fitzgerald, “MOCVD growth of high quality InGaAs HEMT layers on large scale Si wafers for heterogeneous integration with Si CMOS,” IEEE Trans. Semicond. Manufact. 30(4), 456–461 (2017). [CrossRef]
16. J. F. Wheeldon, C. E. Valdivia, A. Walker, G. Kolhatkar, T. J. Hall, K. Hinzer, D. Masson, S. Fafard, A. Jaouad, and A. Turala, “AlGaAs tunnel junction for high efficiency multi-junction solar cells: Simulation and measurement of temperature-dependent operation,” in 2009 34th IEEE Photovoltaic Specialists Conference (PVSC), (IEEE, 2009), 000106–000111.
