Page 45 - Chip Scale Review_March April_2023-digital
P. 45
plasma clean. A protection layer on top of
graphene could also be envisioned to avoid
direct graphene/TBM contact.
Summary
To achieve a reliable transfer process
and to maintain the intrinsic properties
of graphene, it is known that device
performance can improve when SLG
is transferred on smooth target wafers.
To avoid high surface roughness when
integrating graphene in the BEOL (e.g., on
a readout integrated circuit [ROIC] wafer),
a planarization step of the top dielectric will
have to be implemented before graphene
transfer. After transferring graphene on
such a wafer, it will have to be capped with
a dielectric followed by graphene patterning
and contact fabrication. Following this
approach, a BEOL operational graphene
transistor can be achieved that can serve a
multitude of applications [3].
Acknowledgments
This work is supported by the imec
IIAP optical I/O program and received
funding from the European Union’s
Graphene Flagship grant agreement CORE
3 (No 881603) and 2D-EPL (No. 952792).
Figure 3: a) Glass carrier system with laser release and TBM; b) SLG on a glass carrier system obtained after
Cu etch; c) Bonded glass carrier system with graphene on a 200mm target wafer; and d) Graphene on a target
wafer after the full 200mm transfer process. References
1. C. H. Wu, et al., “Graphene electro-
TBM layer are coated on the glass carrier will be important to improve the graphene absorption modulators integrated at
(Figure 3a). Next, an edge bead removal transfer result. Next, the glass carrier that wafer-scale in a CMOS fab,” 2021
step is implemented to avoid excessive contains the graphene layer is bonded on Symp. on VLSI Circuits, Kyoto,
TBM at the bevel of the glass carrier. After top of a target wafer and the temporary Japan, 2021, pp. 1-2, doi: 10.23919/
the temporary bonding step of the glass carrier is subsequently removed using VLSICircuits52068.2021.9492495.
carrier on top of a Cu foil, C residues at the a laser release process (Figure 3c). The 2. K. S. Novoselov, A. K. Geim, et al.,
foil backside are removed and the Cu foil is whole transfer approach ends with a solvent Science Vol. 306, 5696, pp. 666-
selectively etched. After etching the Cu foil, clean that removes the bulk TBM from 669, Oct. 2004.
graphene is exposed on the glass carrier. the 2D material (Figure 3d). Typically, 3. C. Huyghebaert, T. Scharm, et
As can be seen in the microscopy picture in polymer residues are still observed after al., IEEE Inter. Electron Devices
Figure 3b, the roughness of the Cu foil gets the solvent strip process, but these residues Meeting (IEDM), San Francisco,
imprinted in the TBM material. As a result, on top of graphene can be further removed CA, USA, 2018, pp. 22.1.1-22.1.4,
the TBM layer still has considerable TTV with an additional annealing step or even doi: 10.1109/IEDM.2018.8614679.
variation, and improving the foil roughness the implementation of a remote hydrogen
Biographies
Souvik Ghosh is a Research Scientist at imec, Leuven, Belgium. He specializes in 2D materials transfer. Prior
to imec he was a research engineer on chip packaging technologies for the Components Research division at Intel
(Oregon, USA). His background is in flexible electronics, and he holds a PhD in Chemical Engineering from Case
Western Reserve U. (Ohio, USA).
Amaia Zurutuza is the Scientific Director of Graphenea, San Sebastian, Spain, where she leads the R&D
activities on graphene-based materials. Since joining Graphenea in 2010, she has filed for fifteen patents and
published more than 85 publications in peer-reviewed journals, including Nature and Science. She received her PhD in Polymer
Chemistry from the U. of Strathclyde (Glasgow, UK) in 2002.
Contact author: Alice Guerrero/Brewer Science; email aguerrero@brewerscience.com
43
Chip Scale Review March • April • 2023 [ChipScaleReview.com] 43