Open Access
Add to BibSonomy
Abstract
To enhance the performance of multi-junction photovoltaics, we investigated three different InP-based tunnel junction designs: p++-InGaAs/n++-InP tunnel junction, p++-InGaAs/i-InGaAs-/n++-InP tunnel junction, and p++-InGaAs/i-InGaAs/n++-InGaAs tunnel junction. The p++-InGaAs/i-InGaAs/n++-InGaAs tunnel junction demonstrated a peak tunneling current density of 495 A/cm2 and a resistivity of 9.3 × 10−4 Ωcm2, allowing the tunnel junction device to operate at a concentration over 30000 suns. This was achieved by inserting an undoped InGaAs quantum well at the p++-InGaAs/n++InGaAs junction interfaces, which enhanced its stability within the operating temperature range of multi-junction solar cells. Moreover, the p++-InGaAs/i-InGaAs/n++-InGaAs tunnel junction exhibited the lowest resistance.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
Solar cells are a promising source of renewable energy [1]. Multi-junction solar cells are more efficient than traditional designs in capturing and utilizing light energy [2]. To maximize efficiency, sunlight concentration is key, which requires improving tunnel junction performance to avoid tunnel breakdown [3]. InP-based tunnel junctions have proven to be a more challenging template for developing high-performance transparent tunnel junctions [4]. High-performance tunnel junctions require higher peak tunnel currents to accommodate the photocurrent in multi-junction solar cells, lower external bias resistors to reduce voltage losses during operation, and high transparency to reduce optical loss (It is necessary to balance the optical absorption, tunneling probability caused by bandgap width, and material defects caused by lattice mismatch.) [5].
Tunnel junction materials that match the InP substrate lattice include InP, InGaAs, InAlAs, InGaAsP, and InGaAlAs. Typically, a combination of these materials is used to form the anode and cathode of the tunnel junction [6–9]. Kenneth J et al. fabricated an InAlAs: C/InP: Si tunnel junction with a peak tunnel current of 3.5 A/cm2, an InGaAs: Zn/InP: Si tunnel junction with a peak tunnel current of 69 A/cm2[4]. Matthew P. Lumb grew an InAlGaAs tunnel junction on InP with a band gap of 1.18 eV. After adding two InGaAs quantum wells to the structure, the peak tunnel current density reached 113 A/cm2, higher than the bulk InAlGaAs tunnel junction [10]. This shows that the performance of InP-based tunnel junctions with quantum wells is significantly better than conventional bulk material tunnel junctions with equivalent doping amounts.
In this study, we investigated tunnel junctions with three different materials and structures: p++-InGaAs/n++-InP tunnel junction, p++-InGaAs/i-InGaAs-/n++-InP tunnel junction, and p++-InGaAs/i- InGaAs/n++-InGaAs tunnel junction. We first compared the performance of p++-InGaAs/n++-InP tunnel junction and p++-InGaAs/i-InGaAs-/n++-InP tunnel junction. The results show that inserting an undoped quantum well layer in the tunnel junction increases the peak current density and reduces the resistance. Further, we replaced n++-InP with n++-InGaAs. demonstrating a peak tunneling current density of 495 A/cm2 and a resistivity of 9.3 × 10−4 Ω cm2. To evaluate the impact of tunnel junctions on multi-junction solar cell performance under actual operating conditions, we investigated the peak tunnel current density and resistance of tunnel junctions over the operating temperature range of multi-junction solar cells.
2. Experimental
A series of tunnel junctions were grown on n-type InP substrates using commercial production Aixtron metal-organic chemical vapor deposition (MOCVD) at a temperature of 630°C and a pressure of 50 mbar, the growth rate is less than 10 Å/s, interrupt for 3 s when growing n++/p++ interface [11]. Zn (DMZn source) and Si (Si2H6 source) were employed as the p-type and n-type dopants, respectively. The doping levels were evaluated using Secondary Ion Mass Spectrometry (SIMS). The tunnel junction device structure schematic is shown in Fig. 1 and Table 1. Tunnel junction structures were composed of a 15 nm p++ layer and a 15 nm n++ layer. The p++/n++ junction is surrounded by a 30 nm Si-doped n-cladding layer and a 30 nm Zn-doped p-cladding layer, the p++InGaAs and n++InGaAs layer with In composition of 0.53, the i-InGaAs layer with In composition of 0.45. A less than 5 nm undoped InGaAs quantum well was inserted at the tunnel junction interface of structures B and C. To accomplish ohmic contact with the surface, all three structures contain a Zn-doped InGaAs cap layer, and the tunnel junctions were annealed at 600°C for 15 minutes.
Fig. 1. (a) The device structure of the p++-lnGaAs/n++-InP tunnel junction. (b) The device structure of the p++-lnGaAs/i-lnGaAs/n++-InP tunnel junction. (c) The device structure of the p++-lnGaAs/i-lnGaAs/n++-InGaAs tunnel junction.
Table 1. Tunnel junction architecture
The grown tunnel junction devices underwent conventional photolithography and wet etching [12]. SiOx was deposited using plasma-enhanced chemical vapor deposition (PECVD) to isolate the mesa sidewall. The top contact metal of Ti/Pt/Au was deposited by thermal evaporation, while the bottom of the tunnel junction was deposited with AuGe/Ni/Au contact metal. The area of the tunnel junction test structures is 0.01 cm2 and 0.02 cm2. Each device’s current density versus voltage (J-V) characteristics were measured using a four-probe technique [13].
3. Result and discussion
The doping concentration distributions of the three tunnel junction structures were measured with SIMS (the SIMS resolution is about 6 nm),as shown in Fig. 2. The p-type doping part of the three structures is identical. The doping concentration of the p-cladding (Zn-InGaAsP) layer is more than 1 × 1019 cm−3, and the p++ layer (Zn-InGaAs) doping concentration is more than 1 × 1019 cm−3. In the n-type doping part, the n++ layer material has a doping concentration of more than 3 × 1019 cm−3 for structure A (Si-InP) and structure B (Si-InP), and of more than 1 × 1019 cm−3 for structure C (Si-InGaAs). The n-cladding has a doping concentration of more than 3 × 1018 cm−3 for structure A (Si-InP), structure B (Si-InP), and structure C (Si-InGaAsP). In particular, an undoped quantum well layer (i-InGaAs) is inserted at the p++/n++ interface of both structure B and structure C. Comparing structure A and structure B, it can be found that the quantum well insertion layer significantly inhibits the diffusion of high-concentration Zn dopant.
The peak tunneling current density and resistance of structures A and B are shown in Fig. 3. The peak tunneling current density (Jpeak) of Structure A is 7.3 A/cm2, and the resistance (Rp) is 1.4 × 10−2 Ω cm2. The insertion of the undoped lnGaAs quantum well layer leads to a 10-fold increase in Jpeak and a 2-fold reduction in resistance. Structure B obtains a peak current density of 70.2 A/cm2 and resistance of 5.7 × 10−3 Ω cm2. Assuming a one-sun shortcircuit current density of 15 mA/cm2 [14]. This measurement value significantly exceeds the requirements for ultra-high concentration operations, enabling the device to operate within a concentration of 4680 suns. The voltage drop across the tunnel junction is 85 mV at a concentration of 1000 suns. Figure 3(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. 3. (a) Current density versus voltage plots for the structure A and structure B tunnel junctions. (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.
Inserting an undoped quantum well layer into InGaAs/lnP tunnel junction improves performance significantly. The reasons are as follows: (1) The non-doped layer InGaAs quantum well creates an isolation layer that prevents Zn dopants from diffusing to other regions of the tunnel junction, effectively reducing the occurrence of self-compensation effects by limiting the interaction between dopants and surrounding materials. Comparison between structure A and B in Fig. 2 shows that structure B has a steeper doping interface. (2) The InGaAs quantum well layer is inserted into the p++/n++ interface, significantly shortening the tunneling distance from the energy band structure, as shown in Fig. 4. The undoped quantum well layer transfers the quantum well's valence band into the tunneling region, thereby increasing the tunneling current [13]. The peak tunneling current is expressed as
Fig. 4. Band structure of p++InGaAs/i-InGaAs/n++InP (red solid line) and p++InGaAs /n++InP (blue dashed line) tunnel junctions. The differences in band bending shorten the tunneling distance for the undoped InGaAs quantum well inserted in the p++InlGaAs /n++InP tunnel junction structure.
The insertion of the undoped quantum well has a significant effect. The n++-lnP layer was replaced with an n++-lnGaAs layer in Structure C, which was designed based on Ref. [15]. Figure 5 depicts the peak tunneling current density and resistance of Structure C with 495A/cm2 and 9.3 × 10−4Ω cm2, respectively. This measurement value significantly exceeds the requirements for ultra-high concentration operations, enabling the device to operate within a concentration over 30000 suns. The voltage drop is 0.1 V at a concentration of 10000 suns. Besides, the box and whisker chart of the peak tunnel current density shows a higher device reproducibility than structures A and B. The performance of Structure C is better than that of Structure B because n++-InGaAs (Si doping) in Structure C has a higher electronic activation rate than n++-InP (Si doping) in Structure B.
Fig. 5. Current density versus voltage plots for the p++-lnGaAs: Zn/i-lnGaAs/n++-InGaAs: Si (structure C) tunnel junctions. Adjacent to curve is a box and whisker chart of the peak tunnel current density for the entire sample set.
To investigate the impact of the tunneling junction on the performance of multi-junction solar cells under practical working conditions, we investigated the peak tunneling current density and resistance of the tunneling junction at typical operating temperatures (25-60°C). Figure 6(a) lacks a distinct pattern, potentially due to measurement errors and other factors. However, it can be concluded that resistance tends to decrease as the temperature rises. Figures 6(b) and 6(c) depict box and whisker diagrams of the tunneling peak current density and resistance results for various samples within the multi-junction solar cells’ operating temperature range. The frames indicate the 25th and 75th percentiles and the median, the hollow squares represent the mean values, and the whiskers indicate the range. Table 2 summarizes the experimental results at different temperatures. The results indicate that Structure B has greater device reproducibility and stability than Structure A. The tunnel junction of p++-InGaAs/i-InGaAs-/n++-InGaAs has the highest peak tunneling current density, lowest resistance, and highest reproducibility among the three structures. The peak tunneling current value is around 0.45 V, indicating that the resistance of the p++-InGaAs/i-InGaAs-/n++-InGaAs tunnel junction needs to be further improved.
Fig. 6. (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.
Table 2. Experimental results of tunnel-junction devices under various temperatures
4. Conclusion
We grew three InP-based tunnel junctions using MOCVD and measured their J-V performance. In structure B (p++-InGaAs/i-InGaAs-/n++-InP tunnel junction), the peak current density is 10 times higher than that of structure A (p++-InGaAs/n++-InP tunnel junction). The undoped InGaAs quantum well insertion layer results in a significant performance improvement. Structure C (p++-InGaAs/i-InGaAs/n++-InGaAs tunnel junction) replaces n++InP with n++InGaAs based on structure B. Its peak current density reaches 495 A/cm2, and its resistivity reaches 9.3 × 10−4 Ωcm2, allowing the device to operate within a concentration over 30000 suns. Comparing the performance of the three structures within the operating temperature range of multi-junction solar cells, it was found that structure C has the best performance and the highest reproducibility.
Funding
Fundamental Research Funds for the Central Universities (2023SCU12010); China Postdoctoral Science Foundation (2022M722243); National Natural Science Foundation of China (62301347).
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. E. Kabir, P. Kumar, S. Kumar, et al., “Solar energy: Potential and future prospects,” Renewable Sustainable Energy Rev. 82, 894–900 (2018). [CrossRef]
2. J. F. Geisz, R. M. France, K. L. Schulte, et al., “Six-junction III–V solar cells with 47.1% conversion efficiency under 143 Suns concentration,” Nat. Energy 5(4), 326–335 (2020). [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. K. J. Schmieder, T. C. Mood, E. A. Armour, et al., “InP-Based Tunnel Junctions for Microconcentrator Photovoltaics,” IEEE J. Photovoltaics 13(6), 819–824 (2023). [CrossRef]
5. D. Masson, F. Proulx, and S. Fafard, “Pushing the limits of concentrated photovoltaic solar cell tunnel junctions in novel high-efficiency GaAs phototransducers based on a vertical epitaxial heterostructure architecture,” Prog. Photovoltaics 23(12), 1687–1696 (2015). [CrossRef]
6. J. Yin, Y. Sun, S. Yu, et al., “1064 nm InGaAsP multi-junction laser power converters,” J. Semicond. 41(6), 062303 (2020). [CrossRef]
7. J. Yin, Y. Sun, A. Wang, et al., “High-voltage 1064 nm InGaAsP multijunction laser power converters,” IEEE Electron Device Lett. 43(8), 1291–1294 (2022). [CrossRef]
8. M. Beattie, C. Valdivia, M. Wilkins, et al., “High current density tunnel diodes for multi-junction photovoltaic devices on InP substrates,” Appl. Phys. Lett. 118(6), 062101 (2021). [CrossRef]
9. S. Fafard and D. P. Masson, “High-efficiency and high-power multijunction InGaAs/InP photovoltaic laser power converters for 1470 nm,” in Photonics, (MDPI, 2022), 438.
10. M. P. Lumb, M. K. Yakes, M. González, et al., “Double quantum-well tunnel junctions with high peak tunnel currents and low absorption for InP multi-junction solar cells,” Appl. Phys. Lett. 100(21), 213907 (2012). [CrossRef]
11. Y. Gou, J. Wang, Y. Cheng, et al., “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]
12. Y. Gou, J. Wang, Y. Cheng, et al., “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]
13. Y. Gou, H. Wang, J. Wang, et al., “High performance p++-AlGaAs/n++-InGaP tunnel junctions for ultra-high concentration photovoltaics,” Opt. Express 30(13), 23763–23770 (2022). [CrossRef]
14. E. Barrigón, I. García, L. Barrutia, et al., “Highly conductive p++-AlGaAs/n++-GaInP tunnel junctions for ultra-high concentrator solar cells,” Prog. Photovoltaics 22(4), 399–404 (2014). [CrossRef]
15. K. Louarn, Y. Claveau, C. Fontaine, et al., “Thickness limitation of band-to-band tunneling process in GaAsSb/InGaAs type-II tunnel junctions designed for multi-junction solar cells,” ACS Appl. Energy Mater. 2(2), 1149–1154 (2019). [CrossRef]
