InN: A material with photovoltaic promise and challenges
Introduction
The revised bandgap value of InN [1], [2], [3] has attracted global attention due to new technology opportunities for the implementation of high-efficiency InN photovoltaic (PV) devices. However, there remains debate about the bandgap, specifically the origin of the larger bandgap [4], [5] and its relation to oxygen-compound and bandfilling effects [6], [7]. Despite this debate, the development of InN solar cells remains a significant focus.
The current state-of-the art solar cell has an efficiency greater than 35% and was fabricated from lattice-matched III–V semiconductors on a germanium (Ge) substrate [8]. Ge substrates provide adequate lattice matching to InN compared to traditional substrates such as sapphire [3]. Ge also allows for a vertical conduction device design that can reduce shadow losses leading to higher efficiencies compared with top-connected solar cells. However, Ge substrates have not been explored by group III-nitride materials experts because of impediments to growth that have been technically difficult to overcome. Georgia Institute of Technology has demonstrated the successful growth of InN on Ge substrates [3]. This demonstration may allow for increased flexibility in the design of novel InN solar cells.
A current challenge to InN PV devices is strong band bending [9], [10], [11] at the surface/heterointerface, which so far has resulted in the inability to form a rectifying solid-state junction. Specifically, the strong band bending at the collecting junction prevents the collection of minority carriers, leading to ohmic junctions through the tunneling mechanism. If the background doping in InN could be substantially lowered, such band bending near the surfaces could be beneficially used as an efficient means of reducing surface recombination via re-direction of minority carriers away from the defective surfaces.
Another challenge to InN PV device development is that p-type doping of InN has yet to be demonstrated. p-type doping is needed to create the p–n sub-cell collecting junction and for the degenerately doped interconnect regions. The p–n collecting junction facilitates current collection through the electrostatic field created by the spatially separated ionized donors and acceptors. Without such p–n junctions, photogenerated electron hole pairs cannot be separated, therefore no photocurrent is produced. Additionally, traditional tandem solar cells use degenerately doped materials at the sub-cell junctions to facilitate tunnel junctions. In this region the conduction and valence band of the materials overlap, allowing the electrons collected at the sub-cell p–n junctions to tunnel from the n-type emitter of one sub-cell into the neighboring holes in the adjacent sub-cell base. Without p-InN, an InN PV tandem device will need a substantially different means of current collection, such as an internal Schottky barrier or heterojunction for current collection. InN PV tandem devices will also need an alternative means of sub-cell interconnect, independent of the tunneling between degenerate layers. Otherwise, the photogenerated electron hole pairs will not result in current in the external circuit.
Herein, this article details a few challenges and possible solutions for InN solar cell devices; and presents data to aid in the discussion on the discrepancy between the bandgap values of InN. The growth and characterization of InN on (1 1 1) Ge and c-plane sapphire substrates by RF plasma-assisted molecular beam epitaxy (MBE) are investigated to improve the understanding of InN materials for InN solar cells. n-type InN epitaxial layers were grown on p-Ge and n-Ge substrates to investigate the electrical properties. An epitaxial aluminum (Al) buffer layer was grown on Ge to prevent In–Ge eutectic formation and to replace the tunnel junctions in the tandem solar stacks. We also explore crystalline oxygen in InN/AlN/sapphire films and the possibility of an indium oxynitride compound (InONx). In the following, we specify indium oxynitride as a general non-stoichiometric and/or phase separated material as InONx, in analogy to the nomenclature used for non-stoichiometric silicon nitride (SiNx).
Section snippets
Experimental procedure
InN epitaxial layers were grown by RF plasma-assisted MBE. InN/Ge growth conditions have been published elsewhere [3]. InN materials were grown on (1 1 1) epi-ready Ge and on c-plane sapphire substrates that were solvent cleaned. The Ge and sapphire samples were outgassed in the introduction chamber for 1 h at 320 and 500 °C, respectively. The Ge substrate was then loaded into the growth chamber, where a substrate temperature of 360–475 °C was maintained with an In flux of 0.2–1.4×10−7 Torr beam
Results and discussions
One approach to eliminate the need for degenerate p-type doping, and exploit the ohmic behavior of InN/Ge is with the use of an epitaxial layer of Al as a sub-cell interconnect. This epitaxial Al layer also isolates the Ge from the InN, therefore avoiding the eutectic reaction of In–Ge [3]. The Ge–In eutectic forms at approximately 156 °C [13], well below the growth temperature used in these experiments. The In–Ge eutectic may have been the source of the InN–Ge interfacial layers observed by TEM
Conclusion
Several key aspects of InN for use in photovoltaic applications are described. While offering potential for solar cells, several key issues including p-type doping and rectifying solid state junctions have yet to be demonstrated. An investigation of InN grown on Ge and sapphire substrates via plasma assisted MBE was carried out to study the material characteristics of InN for its use in PV devices. The bandgap of InN was determined to be approximately 0.7 eV. However, an additional,
Acknowledgments
This work was supported by the Office of Navel Research monitored by Dr. Colin Wood and the Department of Energy–National Renewable Energy Lab monitored by Dr. Bob McConnell. The authors would like to thank Dr. William Schaff from Cornell University for his unique insight and helpful comments on the growth of InN.
References (18)
- et al.
J. Crystal Growth
(2005) - et al.
J. Crystal Growth
(2003) - et al.
J. Crystal Growth
(2004) - et al.
Nucl. Instrum. Methods Phys. Res.
(1998) - et al.
Phys. Status Solidi
(2002) - et al.
Appl. Phys. Lett.
(2002) - et al.
J. Vac. Sci. Technol.
(1992) - et al.
J. Appl. Phys.
(1986) Phys. Stat. Sol.
(2002)
Cited by (82)
InN nanowire solar cells on Si with amorphous Si interlayer deposited by sputtering
2024, Materials Science in Semiconductor ProcessingProbing the charge state of threading dislocations in indium nitride through advanced atomic force microscopy
2023, Materials CharacterizationAmmonia-free epitaxy of single-crystal InN using a plasma-integrated gas-injection module
2022, Applied Materials TodayCitation Excerpt :For example, InN films have near-infrared photonic behavior without any temperature dependence [8,12,13]. Moreover, the combination of GaN, InGaN, and InN covers a wide range of wavelengths, enabling the development of high-efficiency solar cells with long-term stability and high durability [14,15]. Moreover, the electron mobility of InN is ideally expected to be 14,000 cm2 V–1 s–1 (much higher than that of silicon semiconductors), making this compound advantageous for use in high-speed electronic devices, such as memory and central processing units [16–18].
Growth by MOCVD and photoluminescence of semipolar (202¯1¯) InN quantum dashes
2021, Journal of Crystal GrowthUnderstanding the electronic properties, contact types and optical performances in graphene/InN heterostructure: Role of electric gating
2020, Diamond and Related MaterialsRF-MBE and MOVPE In<inf>x</inf>Ga<inf>1-x</inf>N films over the entire composition range: A study on growth method dependence
2020, Superlattices and Microstructures