Dry Transfer Based on PMMA and Thermal Release Tape for Heterogeneous Integration of 2D-TMDC Layers

Agata Piacentini; Amir Ghiami; Andrei Vescan; Eva Mayer; Hleb Fiadziushkin; Holger Kalisch; Jochen M. Schneider; Max C. Lemme; Michael Heuken; Songyao Tang

REVIEW 3 major objections 5 minor 3 references

Dry Transfer Based on PMMA and Thermal Release Tape for Heterogeneous Integration of 2D-TMDC Layers

T0 review · 3 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper claims a dry transfer method using PMMA and thermal release tape that peels MOCVD-grown 2D-TMDCs off sapphire reproducibly, keeps them on Si/SiO2 during polymer removal, and yields FETs with more than 50 times higher average ON…

desk verdict A clearly described dry-transfer process with one genuinely useful plasma step, undermined by a single un-replicated dry-vs-wet comparison that cannot support the headline electrical claims. read the letter →

arxiv 2412.02407 v1 pith:UJHUZT5F submitted 2024-12-03 physics.app-ph cond-mat.mtrl-sci

Amir Ghiami , Hleb Fiadziushkin , Tianyishan Sun , Songyao Tang , Yibing Wang , Eva Mayer , Jochen M. Schneider , Agata Piacentini

show 4 more authors

Max C. Lemme Michael Heuken Holger Kalisch Andrei Vescan

This is my paper · ORCID

classification physics.app-phcond-mat.mtrl-sci

keywords 2D-TMDCdrytransferPMMAthermalreleasetapeMOCVDWSe2field-effecttransistorplasmacross-linking

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper sets out to solve two known problems in transferring wafer-scale 2D-TMDC films from their growth sapphire to device substrates: getting the film cleanly off the growth substrate, and keeping it attached to the target while the polymer support is dissolved. It proposes a dry process using a PMMA support and a thermal release tape, with a pre-initiated edge crack to start the peel, a water-assisted bonding step on the target side, and a plasma treatment that cross-links the PMMA surface so that acetone removal is slowed. The authors argue that this combination gives reproducible 100 percent peel-off and preserves the film on Si/SiO2, and that the resulting layers outperform a standard KOH-based wet transfer in Raman, XPS, AFM and SEM characterisation. Transistors made from the dry-transferred WSe2 show over 50 times higher average ON current, about 1.5 times lower subthreshold slope, and lower device-to-device variability than wet-transferred devices, which matters because scalable integration of 2D materials depends on transferring films without wrinkles, cracks or polymer residue.

What carries the argument

The load-bearing mechanism is a CASING-type plasma treatment: exposing the PMMA surface to Ar/BCl3 plasma induces cross-linking and chlorine incorporation, increasing the molecular weight of a thin surface layer so that it acts as a controlled barrier against acetone diffusion and swelling during polymer removal. The other two mechanism elements are a KOH-initiated edge crack, which uses fracture mechanics so the peel-off proceeds from a single initiation point, and a water-assisted bonding step inspired by direct wafer bonding, which helps the 2D film conform to and adhere to the SiO2 target. Together they carry the claim that peel-off is reproducible and that film preservation during the final solvent step no longer depends on strong adhesion to prepatterned metal contacts.

What would settle it

Run repeated dry and wet transfers of identically grown WSe2 layers from the same wafer, matched in size (for example six 3x3 cm2 pieces each), through identical device fabrication, and compare the distributions of ON current, subthreshold slope and threshold voltage. If the wet-transferred transistors match or exceed the dry-transferred ones in average current and variability, the claimed 50-fold improvement is not caused by the transfer method.

Watch

Extended reading notes

Core claim

The central discovery is a dry transfer sequence that avoids the two failure modes of polymer-assisted dry transfer. First, exposing the edge of the as-grown 2D-TMDC/PMMA stack to KOH solution creates a crack that lowers the force needed for the mechanical peel-off, so the entire 3x3 cm2 layer detaches from the sapphire growth substrate. Second, wetting the Si/SiO2 target with deionized water and letting it diffuse out over several hours before tape release improves adhesion, and treating the PMMA with Ar/BCl3 plasma before acetone removal creates a densified, cross-linked surface layer that slows solvent penetration and stops the 2D layer from detaching. The paper reports that with both measures 100 percent of the transferred area survives PMMA removal with no visible pinholes, that the dry-transferred WSe2 has less PMMA residue, fewer wrinkles and cracks than the wet-transferred reference, and that FETs made from it have significantly better and more uniform electrical metrics.

Load-bearing premise

The whole comparison rests on assuming that one dry-transferred 3x3 cm2 layer and one wet-transferred 1x1 cm2 layer fairly represent the two transfer methods, with no replicate transfers or statistical tests reported.

Editorial extensions

If this is right

If the method is right, MOCVD-grown TMDC films can be transferred to plain Si/SiO2 at 3x3 cm2 scale with 100 percent yield, without needing Au or Bi interfacial layers and their acid-etch cleanup.
The FET results imply that dry-transferred films are clean and uniform enough for large-area integration: 100 devices across 3x3 cm2 show narrow distributions of ON current, subthreshold slope and threshold voltage.
The transfer also works on prepatterned substrates with 50 nm Au contacts, beyond the 30 nm contact-height limit reported for earlier PMMA/tape transfers, which widens bottom-contact device options.
Because the method preserves the as-grown structural quality, it should translate directly to other MOCVD TMDC films such as MoS2, which the authors demonstrate, and to stacking of 2D heterostructures.
The remaining PMMA residue means that cleaner polymers or a post-transfer cleaning step are still needed before the interface can be considered device-grade.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The performance comparison in the paper rests on one 3x3 cm2 dry-transferred sample against one 1x1 cm2 wet-transferred sample; a fair reader should wait for replicate transfers of the same growth batch at matched areas before treating the 50-fold ON-current gain as a property of the transfer method rather than of sample area or batch.
The authors attribute the electrical improvements to fewer wrinkles, cracks and KOH-related damage, but they do not separate these contributions; testing dry transfer with the same PMMA/tape flow but no plasma treatment, or wet transfer followed by the same plasma treatment, could isolate which step carries the benefit.
The CASING barrier idea is generic: any polymer support that swells in its solvent could benefit from a cross-linked skin, so the plasma step may transfer to other materials and solvents beyond TMDCs on SiO2.
Introducing KOH only at the film edge is a deliberate compromise; the paper claims the exposed edge is small, but edge contamination and its effect on device yield over many transfers is a natural follow-up measure.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: solid process paper, one clean idea, one overreach. The clean idea is the Ar/BCl3 plasma step that cross-links the PMMA surface so acetone removal doesn't peel the TMDC off the target. The overreach is the headline electrical comparison.

What's actually new: KOH edge crack initiation, DI-water wetting of the target, and the plasma cross-linking step. The first two are incremental; the plasma step is the real contribution. They show XPS evidence of Cl incorporation and a densified surface, control experiments showing the plasma doesn't change the TMDC/PMMA interface, and failure images for each omitted step. That's good mechanistic work. The process is described in enough detail to reproduce, which is more than many methods papers. The demonstration on a 3x3 cm2 WSe2 layer and on a prepatterned substrate with 50 nm Au contacts goes beyond Kwon's 30 nm contact limit for this class of transfer.

Where it gets soft: the FET comparison is a single dry-transferred 3x3 cm2 sample versus a single wet-transferred 1x1 cm2 sample. No replicates, no error bars on the transfer itself, no statement that the two layers came from the same growth batch or had comparable thickness. The >50x ION and lower variability could be batch or area effects, not the transfer method. The authors are careful in the mechanism sections, but they make a causal claim that this experiment cannot support. The '100% transfer yield' claim is too strong: it's optical inspection of one sample, with no denominator. And they acknowledge PMMA residues remain a major issue; their own XPS shows dry transfer leaves a considerable amount, just ~23% less than wet. That tempers 'better preservation of as-grown properties.' The citation pattern is fine; they engage with Kwon and Nakatani, and the self-citations are standard methodology papers, not circular.

Who gets value: anyone working on BEOL integration of MOCVD-grown TMDCs, or on bottom-contact FET fabrication. The process itself is likely useful even if the quantitative claims aren't settled. I'd send this to peer review, but insist on either replicate transfers with matched samples or a revision that clearly frames the electrical data as a single-case demonstration.

Referee Report

3 major / 5 minor

Summary. The manuscript reports a dry-transfer process for wafer-scale 2D-TMDC layers grown by MOCVD on sapphire, targeting Si/SiO2 and prepatterned substrates. The process combines PMMA as a sacrificial support, thermal release tape as a carrier, a KOH edge-crack step to initiate peel-off, DI-water wetting of the target substrate, and Ar/BCl3 plasma treatment of PMMA before removal in hot acetone. The authors claim that this process gives reproducible peel-off, avoids the detachment and pinhole formation seen in their control transfers, preserves the as-grown layer quality, and yields field-effect transistors with over 50x higher average ON current, about 1.5x lower subthreshold slope, and lower device-to-device variability than FETs based on a standard KOH-assisted wet transfer. The evidence includes XPS, Raman, PL, AFM, SEM, optical microscopy, and FET statistics on 100 dry-transferred devices and 20 wet-transferred devices.

Significance. If the central claims are confirmed, the process is a useful contribution to scalable heterogeneous integration of 2D-TMDCs because it uses commercially available materials and standard processing steps. The paper is careful to include multiple characterization channels (XPS, Raman, PL, AFM, SEM, electrical data), and it demonstrates transfer onto a prepatterned substrate with Au contacts, going beyond a bare SiO2 target. There is no mathematical derivation or fitting, so the usual circularity concerns do not apply; the paper's claims are empirical and falsifiable. However, the headline quantitative advantages over wet transfer currently rest on a single realization of each transfer method with unmatched sample areas and no replicates, so the significance and the strength of the claims are not yet commensurate.

major comments (3)

[§2, Fig. 5 / 'Field-effect transistors...' paragraph] The electrical comparison that supports the paper's main claim is based on one dry-transferred layer (3x3 cm2, 100 FETs) versus one wet-transferred layer (1x1 cm2, 20 FETs), with the wet baseline referenced to a prior protocol (Schneider et al.) rather than prepared and characterized on matched material in this work. The dry layer is explicitly described as a coalesced monolayer with ~30% bilayer coverage, while no equivalent coverage or thickness information is given for the wet layer. With no replicate transfers and no statistical significance test, the reported >50x ION improvement, ~1.5x SS improvement, and lower CV values cannot be attributed specifically to the dry-transfer process; they could reflect growth-batch differences, layer-thickness differences, sample-area effects, or processing drift. This is the load-bearing point of the manuscript and needs to be addressed with replicated, matched transfers and appropriate statistical reporting.
[§2, Fig. 3b and Supporting Information 'Transfer onto prepatterned substrate'] The claim of '100% transfer yield with no visible pinholes or detachment' rests on optical inspection of one 3x3 cm2 WSe2 transfer and one MoS2 prepatterned sample. Optical microscopy cannot exclude sub-micrometer pinholes or partial delamination, and no success denominator is reported for the 'numerous experiments' cited in the peel-off discussion. Please report a yield statistic over repeated transfers and complement optical inspection with higher-resolution imaging (SEM/AFM) over representative areas, including edge regions.
[§2, Fig. 4d and XPS paragraph after Fig. 4] The Raman intensity comparison is used as evidence of better structural preservation of the dry-transferred layer, but the displayed spectra appear to be single-point measurements without normalization, error bars, or thickness/bilayer-fraction information. Lower Raman intensity in the wet-transferred sample could also reflect a different bilayer fraction or layer thickness rather than transfer-induced damage. Similarly, the statement that dry-transferred layers contain 'up to 23% less PMMA' is presented without replicate analysis or uncertainty estimates. Quantitative claims of this kind need multiple measurement points, replicate samples, and error bars.

minor comments (5)

[General] Figure numbering is inconsistent: two different figures are labeled Figure 2 and two different figures are labeled Figure 3 in the main text. Please renumber all figures sequentially.
[§2, Raman paragraph] The word 'respecticely' should be 'respectively'.
[§4, MOCVD paragraph] The units for precursor flow are inconsistent: 'flow rates of W and Se precursors were 25 nm/min and 110 µmol/min'. Clarify whether 25 nm/min is a growth rate or a precursor flow expressed in different units, and unify the units.
[§2, Fig. 3b caption] In the caption, 'sever 2D-TMDC detachment' should read 'severe 2D-TMDC detachment'.
[§2, peel-off paragraph] The phrase 'numerous experiments have demonstrated' is not a quantitative yield statement; if it is intended to support the reproducibility claim, it should be replaced with a count and success rate.

Circularity Check

0 steps flagged · score 1.0 of 10

No material circularity: self-citations are methodological, and the device/transfer comparisons are direct measurements rather than derived predictions.

full rationale

This paper is an empirical process-development study and contains no mathematical derivation chain, fitted parameter, or uniqueness theorem whose output is equivalent to its input. The central claims—reproducible peel-off, preservation of the 2D-TMDC during PMMA removal, and improved FET metrics—are supported by directly measured SEM, AFM, Raman, XPS, and electrical data. The self-citations present in the manuscript are methodological rather than load-bearing: references 6–9 document the MOCVD growth procedures, reference 38 describes the wet-transfer protocol used for the comparison, reference 25 gives fabrication details, and references 39/50 provide XPS analysis conventions. None of these is used to define the outcome being claimed. The wet-transfer baseline is an experimental comparison point, not a fitted input, and the dry-transfer result does not reduce to the prior wet-transfer protocol. The comparison in Figure 5 is a single dry-transferred 3x3 cm2 layer versus a single wet-transferred 1x1 cm2 layer, which raises statistical and confound-control concerns, but that is an experimental-design limitation, not a circularity. The '100% transfer yield' statement is an empirical observation under optical microscopy and is not a predicted quantity obtained from the method itself. Overall, the derivation chain is self-contained and the self-citations are not used to force the central conclusion; any weaknesses are evidentiary, not circular.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The central claim rests on five unverified process assumptions: the KOH edge crack is benign, water wetting fully expels from the interface, plasma cross-linking creates an acetone barrier, optical microscopy catches all defects, and the single wet-transfer sample is a fair baseline. No free parameters or invented physical entities are introduced; all characterization uses standard commercial tools.

assumptions (5)

domain assumption A 30-60 s exposure to 1 M KOH at the stack perimeter creates a crack initiation point that enables complete peel-off without damaging the 2D-TMDC layer.
The authors state KOH wets only a small section of the film edge and therefore does not impair quality, but no direct measurement of edge damage, KOH spreading, or peel-force is provided. Section 2 (crack introduction, Fig. 1(iv)).
domain assumption Wetting the Si/SiO2 target with DI water drops before transfer, followed by 6 h lateral water diffusion, improves conformal bonding without leaving trapped water or residues at the interface.
The approach is borrowed from direct wafer bonding, but the paper provides no interfacial characterization or adhesion quantification to confirm water is fully expelled. Section 2 (Fig. 1(vi)).
domain assumption Ar/BCl3 ICP treatment cross-links the PMMA surface into a densified barrier that slows acetone penetration, preventing 2D-TMDC detachment, while leaving the 2D layer and interface unaffected.
XPS shows chlorine incorporation and binding-energy shifts consistent with cross-linking, and Raman/PL show no damage, but the barrier mechanism is inferred from success/failure optical images rather than directly measured. Section 2 and Fig. 2a/b.
domain assumption Optical microscopy is sufficient to establish 100% transfer yield with no pinholes or detachment.
The yield claim is based on visible-light inspection of a 3x3 cm2 layer; no quantitative defect counting, dark-field imaging, or larger-area automated inspection is reported. Section 2, Fig. 3b(iv).
domain assumption One wet-transferred 1x1 cm2 sample and one dry-transferred 3x3 cm2 sample are representative of the two transfer methods for electrical comparison.
No replicate transfers or significance tests are reported; differences in area and history could confound the 50x ION comparison. Section 5, Fig. 5.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Dry Transfer Based on PMMA and Thermal Release Tape for Heterogeneous Integration of 2D-TMDC Layers." pith.science (2026). https://pith.science/paper/UJHUZT5F

@misc{pith2026241202407,
  author       = {Pith},
  title        = {Pith review of: Dry Transfer Based on PMMA and Thermal Release Tape for Heterogeneous Integration of 2D-TMDC Layers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UJHUZT5F}},
  note         = {Machine review of arXiv:2412.02407}
}

abstract

A reliable and scalable transfer of 2D-TMDCs (two-dimensional transition metal dichalcogenides) from the growth substrate to a target substrate with high reproducibility and yield is a crucial step for device integration. In this work, we have introduced a scalable dry-transfer approach for 2D-TMDCs grown by MOCVD (metal-organic chemical vapor deposition) on sapphire. Transfer to a silicon/silicon dioxide (Si/SiO$_2$) substrate is performed using PMMA (poly(methyl methacrylate)) and TRT (thermal release tape) as sacrificial layer and carrier, respectively. Our proposed method ensures a reproducible peel-off from the growth substrate and better preservation of the 2D-TMDC during PMMA removal in solvent, without compromising its adhesion to the target substrate. A comprehensive comparison between the dry method introduced in this work and a standard wet transfer based on potassium hydroxide (KOH) solution shows improvement in terms of cleanliness and structural integrity for dry-transferred layer, as evidenced by X-ray photoemission and Raman spectroscopy, respectively. Moreover, fabricated field-effect transistors (FETs) demonstrate improvements in subthreshold slope, maximum drain current and device-to-device variability. The dry-transfer method developed in this work enables large-area integration of 2D-TMDC layers into (opto)electronic components with high reproducibility, while better preserving the as-grown properties of the layers.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

3 extracted references · 2 canonical work pages

[1]

(1) Zhou, Y.; Tong, L.; Chen, Z.; Tao, L.; Pang, Y.; Xu, J. -B. Contact -Engineered Reconfigurable Two -Dimensional Schottky Junction Field -Effect Tr ansistor with Low Leakage Currents. Nat. Commun. 2023, 14 (1),

work page 2023
[4270]

(2) Huh, W.; Lee, D.; Lee, C

https://doi.org/10.1038/s41467-023- 39705-w. (2) Huh, W.; Lee, D.; Lee, C. Memristors Based on 2D Materials as an Artificial Synapse for Neuromorphic Electronics. Adv. Mater. 2020, 32 (51), 1 –16. https://doi.org/10.1002/adma.202002092. (3) Kwon, J.; Seol, M.; Yoo, J.; Ryu, H.; Ko, D. -S.; Lee, M.-H.; Lee, E. K.; Yoo, M. S.; Lee, G.-H.; Shin, H.-J.; Kim, ...

doi:10.1038/s41467-023- 2020
[6740]

(6) Shallenberger, J

https://doi.org/10.1021/acs.jpcc.0c00733. (6) Shallenberger, J. R. 2D Tungsten Diselenide Analyzed by XPS. Surf. Sci. Spectra 2018, 25 (1), 014001. https://doi.org/10.1116/1.5016189. (7) Grundmann, A.; Beckmann, Y.; Ghiami, A.; Bui, M.; Kardynal, B.; Patterer, L.; Schneider, J.; Kümmell, T.; Bacher, G.; Heuken, M.; Kalisch, H.; Vescan, A. Impact of Synthe...

work page doi:10.1021/acs.jpcc.0c00733 2018

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.

[1] [1]

(1) Zhou, Y.; Tong, L.; Chen, Z.; Tao, L.; Pang, Y.; Xu, J. -B. Contact -Engineered Reconfigurable Two -Dimensional Schottky Junction Field -Effect Tr ansistor with Low Leakage Currents. Nat. Commun. 2023, 14 (1),

work page 2023

[2] [4270]

(2) Huh, W.; Lee, D.; Lee, C

https://doi.org/10.1038/s41467-023- 39705-w. (2) Huh, W.; Lee, D.; Lee, C. Memristors Based on 2D Materials as an Artificial Synapse for Neuromorphic Electronics. Adv. Mater. 2020, 32 (51), 1 –16. https://doi.org/10.1002/adma.202002092. (3) Kwon, J.; Seol, M.; Yoo, J.; Ryu, H.; Ko, D. -S.; Lee, M.-H.; Lee, E. K.; Yoo, M. S.; Lee, G.-H.; Shin, H.-J.; Kim, ...

doi:10.1038/s41467-023- 2020

[3] [6740]

(6) Shallenberger, J

https://doi.org/10.1021/acs.jpcc.0c00733. (6) Shallenberger, J. R. 2D Tungsten Diselenide Analyzed by XPS. Surf. Sci. Spectra 2018, 25 (1), 014001. https://doi.org/10.1116/1.5016189. (7) Grundmann, A.; Beckmann, Y.; Ghiami, A.; Bui, M.; Kardynal, B.; Patterer, L.; Schneider, J.; Kümmell, T.; Bacher, G.; Heuken, M.; Kalisch, H.; Vescan, A. Impact of Synthe...

work page doi:10.1021/acs.jpcc.0c00733 2018