Page 44 - Chip Scale Review_March April_2023-digital

boundaries and its low surface roughness. adoption and hesitation aside, existing polymer has been applied in a successful
A thermal decomposition route can also layer transfer technology has already and reproducible manner. It is imperative
produce high-quality graphene on SiC, been demonstrated to enable new device that alternate TBMs be identified to enable
but is often considered to be less scalable possibilities. Furthermore, a layer transfer quality-controlled 2D transfer that meets
compared to CVD, due to the need of large – can potentially bring an epitaxial 2D layer industrial-scale requirements because
and as a result – expensive SiC wafers. on an amorphous dielectric surface that uniform bonding of a temporary carrier
Nowadays, CVD synthesis of SLG is is otherwise impossible to achieve via system with a PMMA layer will be very
mostly done on Cu foils at a manufacturing direct growth. difficult to achieve in a reproducible
scale. However, SLG grown on Cu foils The transfer of graphene from a growth manner. Furthermore, such TBMs will need
typically has high topography variations substrate to a target device wafer typically to be specifically selected to be compatible
over a macroscopic scale because of the requires an intermediate carrier to support with a release process (e.g., laser debond
flexible nature of the foils and the high fragile 2D materials. As often reported in approach). To enable this laser release
graphene growth temperature (typically literature, R&D-scale transfer approaches mechanism, the TBM can be complemented
well above 800°C). This foil roughness can use poly-(methyl methacrylate) (PMMA) as with a second layer that specifically is
likely be further improved by incorporating a support layer. This is generally perceived compliant with the release mechanism (e.g.,
electropolishing and annealing before as the golden standard aiding in transfer laser-absorbent materials). Finally, the last
the actual graphene growth process. of 2D flakes and even CVD materials. key element governing the selection of an
However, because of the polycrystalline Here, PMMA is dissolved in a solvent (e.g., appropriate TBM is its adhesion with the
nature of the foils themselves it will be anisole), spin-coated on the 2D material, graphene layer and that with the carrier
difficult to completely avoid graphene grain and possibly adhered to a thermal release system. The adhesion between the TBM and
boundaries. Nevertheless, high graphene tape (TRT). Next, graphene is released from the 2D of interest needs to be high enough
mobilities have been reported for CVD its growth substrate via Cu etching or even such that during mechanical, chemical, or
graphene grown on Cu foil, which could intercalation-based (i.e., electrochemical) electrochemically assisted debonding, the
make this technique ideally suited for release methods. The released graphene 2D is spalled or discharged from the growth
several graphene-based sensor applications. layer is then laminated or bonded to a target substrate while keeping the TBM-2D
To enable high-end graphene applications, a wafer. Finally, the adhesive is stripped interface intact.
controlled and oriented growth on epitaxial to expose the 2D material. While a TRT-
catalyst template wafers is likely needed. based transfer method is viable for proof- A manufacturable graphene
However, this growth approach complicates of-concept device demonstration, it remains transfer route from Cu foil
transfer because a lateral wet-chemical challenging to implement this process in a The highest quality graphene is grown
etch-based transfer is more difficult to production environment. Further, manual on epitaxial template wafers, but debonding
implement and recycling of the growth bonding and debonding steps introduce large-scale graphene from these wafers
wafer is preferred from a cost perspective. user-level variance that manifest in the form has not yet been demonstrated in a fab
of wrinkles, surface-potential variation, environment. Because graphene quality
Layer transfer technologies and macroscopic cracks. A viable and is sufficient for several applications when
applicable to 2D materials scaled industrial approach to demonstrate it is grown on a Cu foil, and the graphene
Because the most mature graphene transfer of graphene on 200mm, or even release can be easily achieved via etching
growth technique is CVD on a transition 300mm target wafers, likely requires the processes, it is likely that this foil approach
metal template, most graphene applications use of a rigid substrate as a temporary combined with an etch-based release step
require the development of a layer transfer carrier instead of TRT. A rigid carrier is the preferred choice for introducing
technology. Such a technology, especially will prevent nonuniform and excessive graphene in the BEOL. To achieve a reliable
for front-end-of-line (FEOL) applications, expansion typically observed for TRT-based transfer, the use of a rigid temporary
is the “new kid on the block.” It is untested transfers. Furthermore, a rigid temporary glass carrier is an option as it optimizes
and historically unqualified to compete carrier is compatible with existing 200mm the bonding step to the target wafer and
with mainstream fab integration processes and 300mm (de)bonding and cleaning facilitates the etching and cleaning steps.
such as selected-area growth, targeted equipment, making it the preferred Figure 2 shows the different steps of
etch, etc. While bonding and debonding temporary carrier to transfer 2D materials. a glass carrier-based graphene transfer
are prevalent in back-end-of-line (BEOL) In terms of temporary bonding materials when graphene is grown via CVD on a Cu
packaging techniques, proposing their (TBMs), PMMA has been the material of foil. The different steps during a 200mm
usage for FEOL applications to transfer choice. In most cases, especially for small- graphene transfer process are visualized
2D layers invites skepticism. Industrial scale demonstration of lab devices, this in Figure 3. First, a laser release and

Figure 2: Schematic of a manufacturable graphene transfer process using a rigid glass carrier and a laser release approach.

42 Chip Scale Review March • April • 2023 [ChipScaleReview.com]
42