InGaAs/InP double heterostructures on InP/Si templates fabricated by wafer bonding and hydrogen-induced exfoliation

Abstract

Hydrogen-induced exfoliation combined with wafer bonding has been used to transfer ϳ600-nm-thick films of ͑100͒ InP to Si substrates. Cross-section transmission electron microscopy ͑TEM͒ shows a transferred crystalline InP layer with no observable defects in the region near the bonded interface and an intimately bonded interface. InP and Si are covalently bonded as inferred by the fact that InP/Si pairs survived both TEM preparation and thermal cycles up to 620°C necessary for metalorganic chemical vapor deposition growth. The InP transferred layers were used as epitaxial templates for the growth of InP/In 0.53 Ga 0.47 As/InP double heterostructures. Photoluminescence measurements of the In 0.53 Ga 0.47 As layer show that it is optically active and under tensile strain, due to differences in the thermal expansion between InP and Si. These are promising results in terms of a future integration of Si electronics with optical devices based on InP-lattice-matched materials. © 2003 American Institute of Physics. ͓DOI: 10.1063/1.1637429͔Applications of InP-based materials are numerous, and thus integration of InP on Si may enable realization of powerful integrated III-V-on-Si systems. InP and its lattice matched quaternary counterpart In 1Ϫx Ga x As y P 1Ϫy are direct gap semiconductors, which have high carrier mobilities, therefore finding applications in lasers, multijunction solar cells 1 and high-speed devices. Additionally, they cover the low dispersion and minimum loss wavelengths for optical fiber communication at 1.3 and 1.5 m, respectively, making them attractive materials for fabricating semiconductor lasers and detectors for telecommunications applications. However, InP is mechanically fragile, is not available in large substrates, and is expensive. Integrating InP thin films on Si substrates improves its mechanical strength and may also allow InP integration on large substrates by a process of tiling transferred thin films. Most importantly, a viable approach to InP/Si may enable cost-effective integration of infrared optoelectronic devices with well-established silicon electronics.Because InP and Si have a large lattice mismatch ͑8.1%͒, heteroepitaxial growth on Si has not yet been able to produce the high quality electronic material needed in optoelectronic devices, since the dislocation density is typically 10 7 cm 2 . 2In some cases densities of 10 5 cm Ϫ2 can be reached by conformal growth, but this is only possible on very small areas. 3 Other methods have been attempted to circumvent the restrictions of heteroepitaxial growth. However, most of these processes require an entire substrate to be consumed. 4 Direct wafer bonding of III-V semiconductors to Si has been previously demonstrated. 5,6 However, it would be more desirable to have a method for InP/Si integration in which the InP substrate can be repeatedly reused, rather than consuming one InP substrate per active InP-based device layer. Hydrogen-induced exfoliation and layer transfer has shown to be a successful method for Si film fabrication, ...

Role of hydrogen in hydrogen-induced layer exfoliation of germanium

“…Wafer bonding and layer transfer has now been applied to semiconductors other than silicon such as Ge, InP, GaAs, and GaN. [2][3][4][5][6] The mechanism of H-induced exfoliation has been extensively studied in the case of silicon, 7,8 and, more recently, in the case of InP. 9 The early work of Weldon et al provided important conclusions: ͑i͒ implantation induced defects serve to trap H within the Si substrate.…”

Spectroscopic studies of the mechanism for hydrogen-induced exfoliation of InP

“…Indeed, it was only recently shown that the concept of hydrogen-induced exfoliation can be applied to the transfer of thin films of InP and GaAs to foreign substrates. [6][7][8] Moreover, a detailed understanding of the role of H in these materials is still lacking.…”

Direct-bonded GaAs∕InGaAs tandem solar cell

“…2,3 An alternative approach is to employ direct-bonded interconnects between subcells of a multijunction cell, which enables dislocationfree active regions by confining the defect network needed for lattice mismatch accommodation to tunnel junction interfaces. [4][5][6][7][8] We report here a direct-bond interconnected multijunction solar cell, a two-terminal monolithic GaAs/ InGaAs two-junction cell, to demonstrate a proof of principle for the viability of direct wafer bonding for solar cell applications.…”

“…The wafers were diced into ϳ1 cm 2 area and bonded following the procedure described elsewhere. 8 Special care was taken to keep the surface of the wafers clean of organic contaminations and particles. After degreasing the surface, the native oxide was removed by dipping the GaAs and InP pieces in 7 vol % HCl ͑aq͒ and 10 vol % HF ͑aq͒, respectively, for 30 s. Then the wafers were brought into contact with the ͑011͒ edges aligned.…”