Page 25 - ChipScale_Nov-Dec_2020-digital

power) can no longer follow the pace boundary conditions make clear that the microscopy image of a monolayer WS 2
of Moore’s Law [5]. To enable further laser debonding that makes use of a rigid sheet transferred to a 50nm-thick SiO 2
physical scaling (“more Moore”), and to glass carrier combined with a laser release layer. Photoluminescence measurements
incorporate new functionalities to devices and adhesive layer is an ideal carrier for a acquired after the different steps of the
that do not necessarily follow Moore’s 2D material transfer approach. transfer process show that the maxima of
Law (“more than Moore”), a high number The transfer of a monolayer WS 2 the photoluminescence peak before and
of novel materials are being investigated. grown on a 300mm Si wafer with a thick after the transfer did not change after the
A promising new material class is 2D SiO 2 (1.6µm) has been demonstrated whole 2D wafer, which indicates that the
materials (e.g., graphene, h-BN, WS 2 , using laser debonding [9]. Figure 6a material strain and doping are controlled
MoS 2 ). Some of the sensor applications shows the 2D WS 2 transfer scheme, when transferring on a blanket target
based on 2D materials have already made and Figure 6b depicts a stitched wafer using laser debonding (Figure 6c).
it to the market (e.g., graphene-based Recently, a full WS 2 transistor integration
visible short-wave infrared [VIS-SWIR] using 300mm processes has been
imagers), while others are still in a demonstrated with a transferred WS2
research phase (e.g., graphene-based light layer that was enabled by the laser release
modulators, WS 2 CMOS transistors). technology [10].
An important reason for the lack of Several challenges remain to be
industrial 2D applications remains the solved when developing a 2D material
absence of a reliable integration process transfer module, but laser debonding is
of high-quality 2D materials in the an enabler for the development of such
semiconductor industry. Synthesizing a transfer module. A first challenge that
epitaxial 2D materials needs to be done needs to be addressed is the increased
on an epitaxial surface (e.g., sapphire adhesion between epitaxial-grown 2D
wafer for transition metal dichalcogenides materials (e.g., WS2 on sapphire or
[6] and Cu(111) for graphene [7]). graphene on sapphire/Cu(111)) and its
Furthermore, a high growth temperature growth template. This makes a reliable
seems preferred to obtain epitaxial 2D dry delamination from the growth
materials on sapphire. The combination substrate more challenging. Adjustments
of these two boundary conditions t o t h e 2D g r o w t h p r o c e s s a n d
(epitaxial template and high growth improvements in intercalation methods
temperature) makes it necessary to and/or modifications to the debonding
develop a 2D material transfer. Advances process are likely needed to improve the
in bonding and debonding technologies reliability of the 2D debonding from the
are clearly needed to enable such a layer growth wafer. Furthermore, the electrical
transfer approach. Almost all 2D material characteristics of these 2D materials are
transfer developments discussed in hampered by their surroundings, which
literature make use of a flexible support implies that the target wafer preparation
layer. This support layer can be a thin (planarization and chemical passivation)
polymer layer, which is often combined are critical before transferring a 2D
with a flexible tape (e.g., thermal or layer. Finally, the transfer flow makes
pressure release tape) [8]. These support use of an adhesive layer that needs to
tapes are not compatible with front-end- be cleaned from the 2D layer after the
of-line (FEOL) semiconductor processing full transfer process. Therefore, cleaning
and make it very difficult to control developments – probably together with the
strain in 2D materials on microscopic use of a protection layer (e.g., h-BN) – are
and macroscopic levels. Furthermore, needed to fully enable a semiconductor-
bonding of the 2D material to the target compatible 2D material transfer module
wafer needs to be performed in a vacuum enabled by laser debonding.
environment to be able to control the 2D/
target wafer interface. To that end, the Summary
support (i.e., carrier) needs to be removed In this brief review, we have introduced
with a very low mechanical force after laser debonding and demonstrated
2D material transfer because of the the advantages that it can bring with
intrinsically low adhesion between a 2D Figure 6: a) Scheme of a fully automated 2D material respect to more complicated 2D-3D
material and the target wafer. transfer process based on laser debonding technology; heterogeneous applications. In all these
Because the temporary and permanent b) A stitched microscopy image of a monolayer WS 2 applications, a debond is required under
bonding steps in a 2D transfer module transferred on a 300mm Si/SiO 2 (50nm) blanket certain mild conditions in order to not
make use of pressure at an elevated target wafer; c) Excitation energies derived from damage the systems. Either the selectivity
photoluminescence measurements after growth and
temperature, a carrier release step without the different stages of a 300mm 2D material transfer or competition of several TBMs is
heat and pressure is preferred. All these process. Excitation peak positions after growth and making the target flow complicated or
after the complete transfer are comparable.

23
Chip Scale Review November • December • 2020 [ChipScaleReview.com] 23