Epitaxial liftoff (ELO) releases single-crystalline GaAs/AlGaAs films from their growth substrate by selectively etching a buried AlAs release layer in HF, after which the freed film attaches to nearly any smooth substrate by natural intermolecular surface forces, a process the originating work called van der Waals bonding. Cross-sectional transmission electron microscopy of such bonded interfaces shows an amorphous interlayer, about 20 angstrom on GaAs and up to about 100 angstrom on silicon, whose precise chemistry, bonding mechanism, and electrical and thermal transport were left unquantified, along with the origin of its substrate-dependent thickness and the systematic process limits (pressure, thermal stability, reliability, and area) of the technique. The open problem is to determine the microscopic nature of the bonded interlayer, establish the mechanism that holds the films to substrates and produces thicker interlayers on silicon, measure electrical tunneling and thermal conductance across it, and quantify the process limits and reliability needed to generalize adhesive-free ELO integration. The release-etch and scalability sides have since been substantially developed by other routes, and thermal boundary conductance of ELO-transferred GaAs films was measured only in 2024 and remains explicitly interface-limited, while the chemical speciation and bonding mechanism of the interlayer itself remain open.

Background

Epitaxial liftoff (ELO) is a technique for removing a thin, single-crystalline compound-semiconductor film from the wafer on which it was grown. In the original demonstration (Yablonovitch, Hwang, Gmitter, Florez, and Harbison, Applied Physics Letters 56, 2419 (1990)), an AlAs 'release layer' (a thin buried layer of aluminum arsenide) is selectively etched laterally in dilute hydrofluoric acid (HF); because HF etches AlAs roughly ten orders of magnitude faster than the surrounding AlGaAs alloy (with Al mole fraction x, for x below about 0.4), the epitaxial film above it can be separated intact while the growth substrate is preserved for reuse. The freed film, supported by a domed layer of Apiezon W hydrocarbon wax, is then placed on the target substrate with a residual water droplet: if the mean wetting angle on the two surfaces is below 90 degrees, water surface tension pulls the film into contact, applied pressure squeezes the water out to a thickness of roughly one optical fringe (about 1000 angstrom), and the remaining water evaporates over hours, leaving a permanent, adhesive-free bond that the authors attributed to van der Waals (intermolecular) surface forces. Films were bonded this way onto silicon, glass, sapphire, lithium niobate (LiNbO3), indium phosphide (InP), diamond, and onto GaAs itself, and bonded GaAs-on-silicon films survived thermal cycling to 750 degrees C, high enough for subsequent organometallic vapor-phase epitaxy (MOCVD, also called metal-organic chemical vapor deposition).

Cross-sectional transmission electron microscopy (TEM, imaging in which the atomic lattice planes of both film and substrate are resolved simultaneously in a thinned cross-section) revealed that the bond is not atomically intimate: a thin amorphous (non-crystalline) interlayer sits between film and substrate, averaging about 20 angstrom (varying locally from 0 to about 40 angstrom) for GaAs bonded to GaAs, and about 100 angstrom for GaAs bonded to silicon or LiNbO3. Energy-dispersive X-ray fluorescence (EDX, a spectroscopy that identifies elements by their characteristic X-ray emission) limited heavy-element contaminants (atomic number above 11) in the interlayer to no more than a monolayer; the interlayer was interpreted as consisting predominantly of semiconductor oxides. Ammonium sulfide ((NH4)2S) treatments, which remove native GaAs oxide, did not measurably thin the interlayer, though they improved wetting enough to allow bonding to diamond. Trapped dust particles formed micrometer-scale tent-like separations that blistered during heating above about 250 degrees C unless dehydrated by a long vacuum bake.

What the originating work left open is precisely the microscopic and quantitative side of this bond: the chemical speciation of the amorphous interlayer, the actual bonding mechanism, the electrical tunneling resistance and thermal conductance across it, the origin of its substrate-dependent thickness (about 100 angstrom of apparently silicon-rich oxide on Si versus about 20 angstrom on GaAs, a difference the source did not explain mechanistically), the systematic limits of the empirical process parameters (applied pressure, long-term thermal stability, bond strength, reliability statistics), and the scaling of the technique beyond the reported lateral dimensions (stripes wider than 2 cm could not be undercut with the wax-support curvature method).

The question sits between two later literatures. The ELO engineering literature (notably the weight-induced ELO and etch-rate modeling work of the Radboud group and the 2024 review of ELO for III-V solar cells) developed the release-etch side: the lateral etch rate is now described by a diffusion-and-reaction model, and film curvature is deliberately controlled to keep the etch crack open, superseding the 2 cm wax-support limit. Separately, the direct (fusion) wafer-bonding literature for GaAs/GaAs and hydrophilic Si/Si pairs measured interface electrical conductivity, interfacial oxide stability, and interface recrystallization after high-temperature anneals, showing that bonded GaAs/GaAs interfaces contain amorphous oxide inclusions that transform on annealing; and a 2024 study measured the thermal boundary conductance (the heat-transfer coefficient of an interface, in units of W per square meter per kelvin) of van der Waals bonded ELO-transferred GaAs nanomembranes by nanosecond transient thermoreflectance (a pump-probe optical method that tracks surface temperature from its reflectivity change). None of these, however, chemically specifies or mechanistically explains the water-mediated, unannealed interlayer of ELO-bonded GaAs films. Modern 2D-material release layers (graphene-based remote epitaxy, in which a single-crystal film is grown through a one-atom-thick graphene sheet on the substrate and can then be peeled from the graphene; and boron-nitride buffer layers combined with controlled spalling, i.e. fracture-driven film release) achieve large-area transfer by replacing the release mechanism altogether, so they address the scalability limitation without answering the interfacial question.

Problem Statement

For GaAs/AlGaAs epitaxial films released by epitaxial liftoff and attached to foreign smooth substrates by the water-mediated, adhesive-free 'van der Waals' bonding procedure (residual-water placement, capillary-assisted contact, pressure-assisted water expulsion, evaporation), establish the microscopic nature of the bond and the quantitative limits of the technique, specifically:

  1. Determine the chemical composition and structure of the amorphous interlayer observed between bonded film and substrate (approximately 20 angstrom on GaAs, up to approximately 100 angstrom on Si or LiNbO3), including its chemical speciation (native oxide, hydrated oxide, residual adsorbed water, other species) beyond the original energy-dispersive X-ray constraint of at most a monolayer of elements above atomic number 11.

  2. Establish the bonding mechanism: identify the intermolecular forces that actually hold the film to the substrate (van der Waals attraction, hydrogen bonding through hydroxylated or hydrated oxide surfaces, or partially covalent bonding), and relate the bond strength and stability to the measured wetting and surface-energy condition under which the trapped water is expelled.

  3. Explain, experimentally and mechanistically, why the interlayer is several times thicker on silicon substrates (apparently silicon-oxide-rich) than on GaAs, and quantify how deliberate surface chemical treatments change interlayer thickness and bonding quality.

  4. Measure the interfacial transport properties across the bonded interlayer: its electrical tunneling resistance and its thermal boundary conductance, and determine how these scale with interlayer thickness, substrate material, and processing conditions. For the thermal boundary conductance, the single value already reported for ELO-transferred GaAs films in 2024 is a reference point rather than a resolution of this item: what resolves it is how the interlayer's transport scales with interlayer thickness, substrate material, and processing conditions, whether obtained from an interlayer-specific determination or by superseding that reported value with a direct measurement.

  5. Quantify the process and reliability limits needed to generalize the technique: the systematic limits on applied pressure during bonding, long-term thermal stability beyond the demonstrated thermal cycling to 750 degrees C, bond-strength and yield statistics (including dependence on particulate cleanliness), and time-to-failure under thermal and mechanical stress, for the transferred areas now achievable with the process.

The single overarching objective is a predictive, mechanism-level understanding of the adhesive-free ELO film/substrate interface: what the interlayer is, why it forms with its observed substrate-dependent thickness, and what it does to electrical and thermal transport, sufficient to rationally choose surface chemistry and process parameters rather than rely on the purely empirical 1990-era practice values. All five enumerated items above are mandatory and jointly required under this single overarching objective; addressing only a subset does not resolve the task.

References

  1. E. Yablonovitch, D. M. Hwang, T. J. Gmitter, L. T. Florez, J. P. Harbison (1990), 'Van der Waals bonding of GaAs epitaxial liftoff films onto arbitrary substrates', Applied Physics Letters 56, 2419-2421, DOI 10.1063/1.102896, https://doi.org/10.1063/1.102896
  2. E. Yablonovitch, K. Kash, T. J. Gmitter, L. T. Florez, J. P. Harbison, E. Colas (1989), 'Regrowth of GaAs quantum wells on GaAs liftoff films van der Waals bonded to silicon substrates', Electronics Letters 25, 171, DOI 10.1049/el:19890123, https://doi.org/10.1049/el:19890123
  3. A. Ersen, I. Schnitzer, E. Yablonovitch, T. Gmitter (1993), 'Direct bonding of GaAs films on silicon circuits by epitaxial liftoff', Solid-State Electronics 36, 1731-1739, DOI 10.1016/0038-1101(93)90220-k, https://doi.org/10.1016/0038-1101(93)90220-k
  4. X. Y. Lee, M. Goertemiller, M. Boroditsky, R. Ragan, E. Yablonovitch (1997), 'Thin film GaAs solar cells on glass substrates by epitaxial liftoff', AIP Conference Proceedings 394, 719-727, DOI 10.1063/1.52872, https://doi.org/10.1063/1.52872
  5. M. M. A. J. Voncken, J. J. Schermer, G. Maduro, G. J. Bauhuis, P. Mulder, P. K. Larsen (2002), 'Influence of radius of curvature on the lateral etch rate of the weight induced epitaxial lift-off process', Materials Science and Engineering B 95, 242-248, DOI 10.1016/S0921-5107(02)00240-4, https://doi.org/10.1016/S0921-5107(02)00240-4
  6. A. T. J. van Niftrik, J. J. Schermer, G. J. Bauhuis, P. Mulder, P. K. Larsen, J. J. Kelly (2007), 'A Diffusion and Reaction Related Model of the Epitaxial Lift-Off Process', Journal of the Electrochemical Society 154, D629, DOI 10.1149/1.2779968, https://doi.org/10.1149/1.2779968
  7. G. J. Bauhuis, P. Mulder, E. J. Haverkamp, J. C. C. M. Huijben, J. J. Schermer (2009), '26.1% thin-film GaAs solar cell using epitaxial lift-off', Solar Energy Materials and Solar Cells 93, 1488-1491, DOI 10.1016/j.solmat.2009.03.027, https://doi.org/10.1016/j.solmat.2009.03.027
  8. D. van der Woude, L. B. Reboucas, E. Vlieg, J. Smits, J. Schermer (2024), 'A review on epitaxial lift-off for III-V solar cells', Thin Solid Films 808, 140570, DOI 10.1016/j.tsf.2024.140570, https://doi.org/10.1016/j.tsf.2024.140570
  9. H. Takita, N. Hashimoto, C. T. Nguyen, M. Kudo, M. Akabori, T. Suzuki (2010), 'Electron transport properties of InAs ultrathin films obtained by epitaxial lift-off and van der Waals bonding on flexible substrates', Applied Physics Letters 97, 012102, DOI 10.1063/1.3459137, https://doi.org/10.1063/1.3459137
  10. Y. Jeong, M. Shindo, M. Akabori, T. Suzuki (2008), 'Epitaxial Lift-Off of InGaAs/InAlAs Metamorphic High Electron Mobility Heterostructures and Their van der Waals Bonding on AlN Ceramic Substrates', Applied Physics Express 1, 021201, DOI 10.1143/apex.1.021201, https://doi.org/10.1143/apex.1.021201
  11. F. Gucmann, B. Meng, A. Chvala, R. Kudela, C. Yuan, M. Tapajna, M. Florovic, F. Egyenes, P. Elias, F. Hrubisak, J. Kovac, J. Fedor, D. Gregusova (2024), 'Improved Thermal Performance of InGaAs/GaAs Nanomembrane HEMTs Transferred onto Various Substrates by Epitaxial Lift-Off', ACS Applied Electronic Materials 6, DOI 10.1021/acsaelm.4c00659, https://doi.org/10.1021/acsaelm.4c00659
  12. F. F. Shi, K. C. Hsieh (2003), 'Kinetic study of thermally induced electronic and morphological transitions of a wafer-bonded GaAs/GaAs interface', Journal of Applied Physics 94, 2423-2425, DOI 10.1063/1.1592293, https://doi.org/10.1063/1.1592293
  13. K. Yeung, J. Mc Kay, C. Roberts, M. S. Goorsky (2013), 'Electrical Conductivity of Direct Wafer-Bonded GaAs/GaAs Structures for Wafer-Bonded Tandem Solar Cells', ECS Transactions 50, 99-108, DOI 10.1149/05007.0099ecst, https://doi.org/10.1149/05007.0099ecst
  14. K.-Y. Ahn, R. Stengl, T. Y. Tan, U. Gosele, P. Smith (1989), 'Stability of interfacial oxide layers during silicon wafer bonding', Journal of Applied Physics 65, 561-563, DOI 10.1063/1.343141, https://doi.org/10.1063/1.343141
  15. Z. L. Liau, L. R. Danielson, P. M. Fourspring, L. Hu, G. Chen, G. W. Turner (2008), 'Reduction of thermal conductivity in wafer-bonded silicon', Applied Physics Letters 93, 021917, DOI 10.1063/1.2959063, https://doi.org/10.1063/1.2959063
  16. Y. Kim, S. S. Cruz, K. Lee, B. O. Alawode, C. Choi, Y. Song, J. M. Johnson, C. Heidelberger, W. Kong, S. Choi, K. Qiao, I. Almansouri, E. A. Fitzgerald, J. Kong, A. M. Kolpak, J. Hwang, J. Kim (2017), 'Remote epitaxy through graphene enables two-dimensional material-based layer transfer', Nature 544, 340-343, DOI 10.1038/nature22053, https://doi.org/10.1038/nature22053
  17. M. Snure, E. W. Blanton, V. Soukhoveev, T. Vogt, A. Osinsky, T. Prusnick, W. J. Kennedy, N. Glavin (2023), 'Spalling induced van der Waals lift-off and transfer of 4-in. GaN epitaxial films', Journal of Applied Physics 134, 025307, DOI 10.1063/5.0153634, https://doi.org/10.1063/5.0153634