Graphene Remote Epitaxy for Solar Cells – A-PROPOS

Graphene is synthesized using a method mastered in our laboratory: chemical vapor deposition (CVD) on a germanium substrate. Unlike conventional "wet" transfer methods—which involve etching the original substrate in solution and using a PMMA layer to support the graphene—we rely on a dry transfer process. This method avoids immersing the target substrate in solution before graphene contact, enabling better control of the interface quality and pre-deoxidation of the substrate.

To optimize the selective growth of GaAs within graphene openings, the graphene is patterned into strips of varying widths and periods (typically 50 to 300 nm for both the strips and openings) using electron beam lithography.

GaAs growth is performed via molecular beam epitaxy (MBE), with systematic optimization of key parameters: deoxidation under ultra-high vacuum, growth temperatures, and gallium (Ga) and arsenic (As) fluxes.

For exfoliation, we employ a literature-inspired method: a 30 nm titanium adhesion layer and a 250 nm stressed nickel layer are deposited, followed by the application of a thermal release tape for manual exfoliation.

To ensure process control, we use a range of characterization techniques: Raman spectroscopy to assess graphene quality before and after transfer, Scanning electron microscopy (SEM) and atomic force microscopy (AFM) to visualize surface morphology and layer structure, Transmission electron microscopy (TEM), used selectively, to observe the III-V/graphene/III-V interface.

Photoluminescence measurements allow us to predict the electrical voltage that a solar cell made from this material could generate. Cathodoluminescence provides sub-micrometer spatial resolution, offering greater precision than photoluminescence.

The work began with the development of a dry transfer method for graphene without degradation. Initial molecular beam epitaxy (MBE) growth tests resulted in polycrystalline layers, suggesting the absence of long-range interaction effects. This hypothesis was supported by consistent observations from SEM, TEM, and AFM analyses.

We then focused on optimizing selective growth within graphene openings, followed by coalescence. We found that precise control of the orientation of these openings relative to the monocrystalline substrate is critical for achieving a high-quality optoelectronic coalesced layer. As a proof of concept, we demonstrated a coalesced material with: low roughness (< 1 nm RMS), no visible dislocations in TEM (density < 10⁷ cm⁻²). Furthermore, the obtained photoluminescence is reduced by a factor of 10 compared to the reference sample grown on an epi-ready substrate without graphene, corresponding to a predicted voltage loss of 60 mV for solar cell fabrication. This performance loss is minimal considering the extensive surface preparation steps before growth and is highly encouraging.

Following exfoliation tests, we observed that structures with 95% graphene coverage were more easily detached compared to those with 50% coverage. Additionally, we noted a dependence between the ability to exfoliate the epitaxial material and the orientation of the graphene openings. We determined that exfoliation must be performed parallel to the openings, attempting to peel perpendicularly prevents successful exfoliation.

The results obtained through this project open several promising perspectives:

The underlying physical principle that enables such high material quality on a structured substrate remains an open question. Clarifying this mechanism is crucial, as it will determine the ultimate material quality we can achieve, as well as the fabrication parameters, including graphene coverage, patterning designs, and substrate orientations.

Luminescence measurements indicate that the obtained material is compatible with the fabrication of high-performance devices. This paves the way for the development of demonstrators, an essential step in validating the applicative potential of this approach.

We have also demonstrated that the method is compatible with exfoliation. To complete the process and enable substrate recycling, two key areas require further investigation: optimizing the exfoliation step and re-preparing the substrate post-exfoliation. These studies demand multidisciplinary expertise, including mechanics for exfoliation, chemistry for surface characterization and re-preparation, in addition to the materials science and device engineering skills employed thus far.

Record efficiency solar cells are made of III-V materials, but their usage is limited to niche applications due to their high cost. More than 80% of this one is made up by costly substrates, so that a method to recycle them for several consecutive growths would constitute a breakthrough for high-efficiency low-cost devices. As an appealing solution to answer this technological problem, this project aims at developing the remote epitaxy. It consists in the epitaxy on a crystalline substrate covered by a monolayer of graphene and was shown to allow the growth of transferable epilayers. While providing convincing results, the method raises fundamental questions regarding the particle interactions during growth. This project provides with a methodology to clarify those phenomena, as well as original developments for robust and controllable fabrication processes and ambitious objectives in terms of device performances. Beyond photovoltaics, this project also opens perspectives in fields such as silicon photonics or flexible opto-electronic devices.