Computationally-tractable synthesis of an MXene metamaterial absorber with a 3D-printable spatially variable substrate by a local approach

Haozhe Wang; Maria-Thaleia Passia; Steven Cummer; Yilin Zhao

REVIEW 3 major objections 5 minor 15 references

Computationally-tractable synthesis of an MXene metamaterial absorber with a 3D-printable spatially variable substrate by a local approach

T0 review · 3 major / 5 minor · reviewed 2026-08-01 · deepseek-v4-flash

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

Pith's one-line read LOCABINACONN3D converts a non-manufacturable inverse-designed MXene absorber into a 3D-printable air/resin structure that keeps nearly the same absorbance, by matching the S-parameters of small local components rather than simulating the wh

desk verdict LOCABINACONN3D is a genuine and useful speedup for converting inverse-designed absorbers into 3D-printable structures, but the single demonstrated example and untested locality assumption keep the central claim conditional. read the letter →

arxiv 2607.17413 v1 pith:FXV36YWE submitted 2026-07-19 physics.app-ph physics.comp-ph

Maria-Thaleia Passia , Yilin Zhao , Haozhe Wang , Steven Cummer This is my paper · ORCID

classification physics.app-phphysics.comp-ph

keywords MXenemetamaterialabsorber3D-printablesubstratespatiallyvariabledielectricmethodoflinesS-parametermatchingtopologyoptimizationbroadbandabsorptionmanufacturabilityconstraints

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 introduces LOCABINACONN3D, a methodology that turns an inverse-designed MXene metamaterial absorber with a continuous, non-manufacturable dielectric substrate into a manufacturable version made only of printer resin and air. Instead of running full-wave simulations of the entire device, it works on small graph components of each material layer. For each component, it generates candidate air/resin configurations, computes their S-parameters with a fast semi-analytical method-of-lines solver inside a small conducting box, and picks the candidate whose response best matches the original. Recombining these locally chosen components gives a full manufacturable absorber with average absorbance 0.9301 over 8–25 GHz, close to 0.9439 for the idealized non-manufacturable design. This matters because spatially variable substrates broaden absorption bandwidth but are hard to 3D-print; the new method makes the conversion computationally tractable and scalable to larger devices.

What carries the argument

The central mechanism is component-level S-parameter matching: each material level forms a graph, and each connected component is treated as an independent design unit. For every component, candidate manufacturable configurations are generated by randomly assigning cells to air (while preserving connectivity and mandatory resin cells), each is simulated in a small box with PEC side walls using a method-of-lines solver, and the configuration whose |S11| and |S21| best match the original component's is selected. The chosen components are then combined to form the full device. This replaces full-device simulation with simulations of much smaller subareas, which is what makes the approach comput

What would settle it

Simulate a small full device containing at least two adjacent components or layers, replace one component with the locally chosen manufacturable candidate, and compare the full-device absorbance to the response predicted from the isolated component's S-parameters. If the deviation exceeds the roughly 1.4-point average-absorbance drop seen here, or if a global optimization of the full device finds a manufacturable configuration with meaningfully higher absorbance, the local-match assumption fails.

Watch

Extended reading notes

Core claim

The central claim is that a non-manufacturable, continuously varying dielectric substrate in a multilayer MXene absorber can be rendered manufacturable by making the air/resin choice locally, component by component, rather than by optimizing or simulating the full device. The local choice is guided by matching the scattering parameters of each graph component inside a small PEC-walled computational box, and a semi-analytical method-of-lines solver makes the search fast. The paper demonstrates that the assembled manufacturable absorber retains nearly the same absorbance as the ideal non-manufacturable one, and compares favorably with a device-level method that requires expensive full-wave sim

Load-bearing premise

The load-bearing premise is that an air/resin substitute chosen to match the S-parameters of one graph component inside a small conducting box will still match when all components are recombined with MXene layers between them; the paper does not test this against a global optimization or on a larger device.

Editorial extensions

If this is right

The method scales to larger MMAs because simulation cost depends on component size rather than full device size.
The manufacturable absorber's average absorbance sits within about 1.4 percentage points of the ideal non-manufacturable design (0.9301 vs 0.9439), and close to the device-level method (0.9344).
Connectivity constraints across consecutive layers are built into the local selection, facilitating multilayer fabrication and assembly.
The semi-analytical MoL solver cuts per-configuration simulation time dramatically, making the search over candidate configurations feasible.
The same local approach could be applied to other inverse-designed all-dielectric devices with spatially variable substrates, not just MXene absorbers.

Reading between the lines

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

The paper does not directly verify that a component chosen in isolation, inside a small PEC-walled box, still performs optimally when recombined with MXene layers and neighboring components; coupling between components and layers could accumulate in larger devices.
Generating only eight random candidate configurations per component may not find the globally best manufacturable replacement; a more systematic search could push fidelity even closer to the ideal.
If the local-match assumption holds generally, the method might be adapted to other printed microwave devices where local electromagnetic response dominates, such as reflectarrays or frequency-selective surfaces.
The claimed scaling advantage is only demonstrated on a proof-of-concept device; a larger multiscale test would confirm that the local approach remains accurate as device size grows.

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

This paper extends the authors' earlier 2D approach into a full 3D multilayer design tool. That is a real contribution: it adds connectivity constraints across consecutive layers and uses a method-of-lines (MoL) solver to evaluate component-sized structures instead of running full-wave FEM on the entire device. The speedup is concrete and meaningful — 0.36 s per component frequency versus 44 s per full-device frequency. The final FEM check on the assembled manufacturable MMA gives absorbance 0.9301, close to the 0.9439 of the non-manufacturable design, so the method clearly works for this particular device.

The soft spots are real but addressable. The component-selection step is based on visual matching of |S11| and |S21| curves from eight random air/resin configurations; there is no quantitative matching metric, no sensitivity analysis, and no error bars. The load-bearing locality assumption is this: a component chosen inside a small PEC-walled box with air elsewhere will behave the same when embedded in the full multilayer stack with MXene layers and other dielectric materials. That assumption is asserted in Section III-D ('we combine all these smaller manufacturable structures to form the entire manufacturable MMA') but never tested. There is no control — for example, comparing the final absorbance against randomly chosen manufacturable configurations — and no demonstration on a larger or second device. The scalability claim for 'even larger-scale MMAs' therefore rests on the same untested principle.

The circularity burden is low: the final result is an independent FEM simulation, not a fitted outcome. The MoL solver comes from the authors' own prior work, which is fine — self-citation is not a flaw — but it means the speedup is not independently benchmarked here. No code or data are released, and there is no fabrication, so the numerical claims are not independently checkable.

These are proportionate concerns. The paper is honest about the local approach being a heuristic, and one FEM-validated proof-of-concept is a reasonable starting point. But the title's 'enables' and the conclusion's 'paves the way' go a bit beyond what is shown. This paper deserves a serious referee. The referee should push for a quantitative selection metric, a locality test with a baseline comparison, and ideally a second device or experimental validation. I would accept it for peer review rather than desk reject.

Referee Report

3 major / 5 minor

Summary. The paper presents LOCABINACONN3D, a methodology for converting an inverse-designed MXene metamaterial absorber (MMA) with a continuously varying dielectric substrate into a manufacturable air/resin structure while preserving absorption performance. The key idea is to perform the binarization locally: each connected component of each discrete material level is replaced by an air/resin configuration whose isolated |S11| and |S21| match those of the original non-manufacturable component, computed with a fast method-of-lines (MoL) solver in a small computational box. The chosen components are then recombined to form the full MMA. The method is demonstrated on a seven-layer MXene MMA with seven discrete material levels. The final FEM-simulated average absorbance (8–25 GHz) is 0.9301 for the LOCABINACONN3D-based device, versus 0.9439 for the non-manufacturable seven-material device and 0.9344 for the previously introduced BINACONN3D approach.

Significance. If the local-global transfer is reliable, LOCABINACONN3D would offer a substantial computational speedup for inverse-designed 3D-printable metamaterials: the paper reports 0.36 s per frequency per configuration with MoL versus 44 s per frequency for a full FEM simulation. The idea of component-level selection with connectivity constraints is a reasonable and potentially impactful extension of the authors' earlier BINACONN3D method. However, the central claim rests on an untested locality assumption: that components optimized in isolation, in an air-filled PEC-walled box, will preserve the performance of the full metal-backed multilayer device after recombination. The paper provides only a single example and no controls. The strength of the contribution therefore depends on the validity of this assumption, which is not established.

major comments (3)

[Sec. III-D (recombination step)] The load-bearing claim is that locally selected manufacturable components can be combined into a full MMA with preserved absorbance. However, the selection is made in a surrogate environment: a small box with PEC side walls and all non-component cells set to air (Sec. III-D). In the assembled device, each component is surrounded by other dielectrics (epsilon_r 1.2–2.7), MXene layers, and the PEC backplane. Matching isolated |S11| and |S21| does not guarantee preservation of the metal-backed multilayer absorbance because inter-component and inter-layer coupling can shift resonances. The paper reports one successful outcome (A=0.9301 vs 0.9439) but provides no control: no random selection, no sensitivity to component choice, and no larger-device test. The scalability claim in the Abstract and Conclusions rests on this untested principle. Please add a validation study, e.g., compare against
[Sec. III-D (selection criterion)] The method selects 'the manufacturable structure that provides the closest match' to the non-manufacturable component, but the matching criterion is not defined quantitatively. Fig. 6 shows a visual comparison, but no error metric (e.g., mean squared error over the band, maximum deviation) or a threshold is specified. This makes the procedure non-reproducible and leaves open whether the chosen candidate is truly the best in any well-defined sense. Please define the objective function used for selection (e.g., minimize a frequency-averaged difference of |S11| and |S21|) and report the selected score. Without this, the method cannot be applied by others.
[Sec. III-D (MoL solver)] The MoL solver is used to compute the S-parameters for all candidate components, and the final selection depends entirely on these predictions. The paper does not validate MoL against a full-wave solver for the dielectric component geometries considered here. If MoL has a systematic error for certain component shapes, the selection could be suboptimal. Please include a validation of MoL versus FEM (or another full-wave reference) for a representative set of components, e.g., the eight candidates in Fig. 6, reporting the S-parameter deviation across the 8–25 GHz band.

minor comments (5)

[Throughout] The method name is spelled inconsistently: LOCABINACONN3D (in the title/abstract) and LOCABINNACONN3D (in most of the body), with a double 'N'. Also 'BINNACONN3D' appears once in Sec. IV. Please standardize.
[Sec. V] Typo: 'absobance' should be 'absorbance'.
[Sec. II] In Fig. 2 caption, '0r = 1.5(1-j0.03)' should be 'epsilon_r = 1.5(1-j0.03)'. Also the mathematical symbols are not rendered consistently (e.g., 'epsilon_r,d' formatting).
[Sec. III-D] The number of candidate configurations is fixed to eight with no justification or convergence study. Please explain how this number was chosen and whether the selection is sensitive to it.
[Sec. IV] The computational comparison reports only the time per frequency for the full FEM device and the MoL component simulations. It would be helpful to state the total number of MoL simulations needed for the complete synthesis and the total time savings, including the overhead of generating the candidate configurations.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the final absorbance claim is independently FEM-validated; local S-parameter matching is a synthesis step, not a definitional prediction.

full rationale

The claimed derivation chain is: (III-A) discretize the optimized continuous substrate into seven material levels; (III-B) assign each level a resin percentage by matching the propagation constant of an air/resin configuration; (III-C) decompose each sub-device into connected graph components; (III-D) generate candidate manufacturable air/resin structures for each component and select the one whose |S11| and |S21| computed in a small MoL box most closely match those of the non-manufacturable component; then combine the components and simulate the full MMA. The one step that could look like fitting-called-prediction is the per-component S-parameter match, but the paper does not claim the final absorbance follows from that match by construction. The final claim—that the manufacturable MMA maintains absorbance (A=0.9301 vs 0.9439)—is obtained from an independent full-device COMSOL FEM simulation (Sec. IV), not from the local S-parameters. The local matching is a synthesis objective; the full-device simulation is an external check that could in principle have disagreed if the locality assumption failed. Citations to the authors' earlier BINACONN3D and MoL work ([9],[11],[12],[15]) provide the baseline and solver, but they are not invoked as a uniqueness theorem or as a substitute for the validation performed here. The untested locality assumption is a correctness risk, not a circularity: it concerns whether an approximation is valid, not whether the output is equivalent to the input by definition.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The method rests on several unvalidated modeling simplifications: locality of component responses, MoL accuracy, effective-medium resin percentages, manual constraint cells, and the 7-level discretization. The only external check is a single FEM simulation of the assembled device; there is no experimental validation and no code release.

free parameters (4)

number of discrete material levels = 7
Section III-A: 'We choose to preserve 7 material levels, as this is sufficient for retaining the absorbance.' A hand-tuned discretization; the central result depends on this choice.
number of candidate manufacturable configurations per component = 8
Section III-D: Fig. 6 shows eight manufacturable structures per component; no stated rule for this number. Selection among these candidates drives final performance.
resin percentage per material level = not stated as a table
Section III-B: resin percentage is assigned by matching the propagation constant of the air/resin configuration to the continuous material level. This is an effective-medium mapping rather than a fit to final absorbance, but it is a modeling choice with no independent calibration.
cells constrained to remain resin (black nodes) = manual subset
Section III-C: 'we specify certain cells that must remain as resin, i.e., the black nodes in Fig. 4.' This manual constraint affects structural connectivity and performance and is not optimized or justified quantitatively.

assumptions (5)

domain assumption Local component S-parameter matching in PEC-walled boxes is sufficient for final assembled-device performance.
Section III-D: the component box is set to air outside the component with PEC side walls, then components are combined. This locality assumption is never tested on the assembled device.
domain assumption The MoL semi-analytical solver is accurate enough for component selection.
Section III-D uses MoL [13]-[15] instead of full-wave FEM; no direct MoL-vs-FEM error comparison is reported for the same component.
domain assumption Propagation-constant matching assigns a valid resin percentage for each material level.
Section III-B: resin percentage is chosen to match the propagation constant of the supported mode; this is an effective-medium/homogenization assumption.
domain assumption The TopOpt continuous design can be discretized to seven materials without significant performance loss, and this generalizes.
Section III-A and Fig. 3 show this for this specific device; no theory guarantees it for other devices.
domain assumption PEC side boundaries and TE10 excitation approximate the waveguide-fed operation of the real device.
Section II: 'We consider PEC side boundaries and a TE10 mode excitation, to resemble a waveguide-fed excitation.'

how reviews work

0 comments

Cite this review

Pith. "Pith review of Computationally-tractable synthesis of an MXene metamaterial absorber with a 3D-printable spatially variable substrate by a local approach." pith.science (2026). https://pith.science/paper/FXV36YWE

@misc{pith2026260717413,
  author       = {Pith},
  title        = {Pith review of: Computationally-tractable synthesis of an MXene metamaterial absorber with a 3D-printable spatially variable substrate by a local approach},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FXV36YWE}},
  note         = {Machine review of arXiv:2607.17413}
}

read the original abstract

We introduce the LOCABINACONN3D methodology to enable the computationally tractable synthesis of MXene metamaterial absorbers (MMA) based on spatially variable 3D-printable substrates. Spatially variable substrates offer enhanced absorption bandwidth compared to constant ones. Such MXene MMAs are typically synthesized by inverse design, which may lead to non-manufacturable optimized dielectric substrates. To transform non-manufacturable dielectric substrates into manufacturable ones, existing methodologies either add manufacturing constraints to the optimization, which may lead to less optimized MMAs, or are computationally expensive, as they require full-wave simulations of the entire manufacturable MMA. We develop a computationally tractable methodology, LOCABINACONN3D, to render optimized MMAs manufacturable, preserving performance. Our methodology (i) accommodates detailed connectivity constraints across consecutive layers, thus facilitating multilayer fabrication, and (ii) enables scaling to larger MMAs, by requiring simulations only of smaller manufacturable MMA subareas and by using a semi-analytical method-of-lines (MoL) solver instead of full-wave methods to determine suitable manufacturable configurations. This work paves the way for synthesizing optimized larger-scale 3D-printable MMAs more efficiently.

Figures

Figures reproduced from arXiv: 2607.17413 by the authors.

**Figure 1.** (a) MXene MMA of constant dielectric substrates. (b) Opti [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗

**Figure 2.** Absorbance of the MXene MMA with (i) constant and (ii) [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Graph of a sub-device. Each cell of the sub-device is shown as [PITH_FULL_IMAGE:figures/full_fig_p003_4.png]

**Figure 7.** Figure 7: Absorbance of the manufacturable MMA by LOCABINA [PITH_FULL_IMAGE:figures/full_fig_p004_7.png]

**Figure 6.** Figure 6: (a) |S11| and (b) |S21| of the non-manufacturable structure of 7 materials of “Sub-device 2 - ϵr=1.2 - Component=1” and 8 manufacturable structures. shown in [PITH_FULL_IMAGE:figures/full_fig_p004_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references

[1]

Constructing metastruc- tures with broadband electromagnetic functionality,

R.-H. Fan, B. Xiong, R.-W. Peng, M. Wang, “Constructing metastruc- tures with broadband electromagnetic functionality,”Adv. Mater ., vol. 32, no. 1904646, pp. 1–22, 2020

2020
[2]

Ultra-thin broadband metamaterial absorber,

Y . Liu and S. Gu and C. Luo and X.Zhao, “Ultra-thin broadband metamaterial absorber,”Appl. Phys. A, vol. 108, p. 19–24, 2012

2012
[3]

Ultra-broadband microwave metamaterial absorber,

F. Ding, Y . Cui, X. Ge, F. Zhang, Y . Jin and S. He, “Ultra-broadband microwave metamaterial absorber,”Appl. Phys. Lett., vol. 100, no. 103506, pp. 1–4, 2012

2012
[4]

Perforated lightweight broadband metamaterial absorber based on 3-d printed honeycomb,

S. Ghosh and S. Lim, “Perforated lightweight broadband metamaterial absorber based on 3-d printed honeycomb,”IEEE Wirel. Antennas Propag., vol. 17, no. 12, pp. 2379–2383, 2018

2018
[5]

Design and full characterization of a 3-d-printed hyperbolic pyramidal wideband microwave absorber,

X. Lleshi and T.Q.V . Hoang and B.Loiseaux and D.Lippens, “Design and full characterization of a 3-d-printed hyperbolic pyramidal wideband microwave absorber,”IEEE Wirel. Antennas Propag., vol. 20, no. 1, pp. 28–32, 2021

2021
[6]

Ultrabroadband transparent meta- material absorbers designed by anomalous brewster effect and gradient impedance matching optimized with deep neural network,

J. Chen, H. Duan, and G. Huang, “Ultrabroadband transparent meta- material absorbers designed by anomalous brewster effect and gradient impedance matching optimized with deep neural network,”Phys. Rev. Appl., vol. 23, no. L051002, 2025

2025
[7]

Tri-band terahertz metamaterial absorber based on structural ti3c2tx mxene for enhanced sensing application,

S. Feng and L. Yang and B.Cai and W. Yang and L. Wu and Y . Cheng, “Tri-band terahertz metamaterial absorber based on structural ti3c2tx mxene for enhanced sensing application,”IEEE Sens. J., vol. 24, no. 18, pp. 28 889 – 28 896, 2024

2024
[8]

Inverse-designed mxene metamaterial absorber for broadband solar energy harvesting,

H.Zhang and J. Yin, “Inverse-designed mxene metamaterial absorber for broadband solar energy harvesting,”Photonics Nanostructures: Fundam. Appl., vol. 63, p. 101362, 2025

2025

Show all 15 references

[9]

Synthesis of broadband multilayer metamaterial absorbers based on spatially variable 3d-printed structures and mxenes,

M.-T. Passia and Y . Zhao and H. Wang and S. A. Cummer, “Synthesis of broadband multilayer metamaterial absorbers based on spatially variable 3d-printed structures and mxenes,”Phys. Rev. Appl., vol. 24, no. 064070, 2025

2025
[10]

Inverse design in photonics by topology optimization: tutorial,

R. E. Christiansen and O. Sigmund, “Inverse design in photonics by topology optimization: tutorial,”J. Opt. Soc. Am. B, vol. 38, no. 2, pp. 496–509, 2021

2021
[11]

A systematic approach for synthesizing 3d-printable all-dielectric devices,

M. T. Passia and S.A. Cummer, “A systematic approach for synthesizing 3d-printable all-dielectric devices,”J. Appl. Phys., vol. 136, no. 213101, pp. 1–10, 2024

2024
[12]

Computationally-efficient synthesis of inversely-designed 3d-printable all-dielectric devices,

M.T. Passia and S. Cummer, “Computationally-efficient synthesis of inversely-designed 3d-printable all-dielectric devices,”IEEE AP-S/URSI, 2025

2025
[13]

The extended method of lines for the characterization of dielectric metasurfaces for ultraviolet chiral sensing,

M. Z. Bideskan, Z. Atlasbaf, and A. V . Lavrinenko, “The extended method of lines for the characterization of dielectric metasurfaces for ultraviolet chiral sensing,”IEEE Trans. Antennas Propag., vol. 71, no. 12, 2023

2023
[14]

Method of lines numerical analysis of conformal antennas,

A. Alu, F. Bilotti, and L. Vegni, “Method of lines numerical analysis of conformal antennas,”IEEE Trans. Antennas Propag., vol. 52, no. 6, 2004

2004
[15]

Analysis and design of periodic multi-element mmWave metasurfaces by a fast extended method-of-lines formulation,

M. Passia, K. Ntokos, and T. Yioultsis, “Analysis and design of periodic multi-element mmWave metasurfaces by a fast extended method-of-lines formulation,”IEEE Trans. Magn.,Early Access, 2025

2025

Pith tools

Reviewed August 1, 2026 · model on record in the stance chip above.

[1] [1]

Constructing metastruc- tures with broadband electromagnetic functionality,

R.-H. Fan, B. Xiong, R.-W. Peng, M. Wang, “Constructing metastruc- tures with broadband electromagnetic functionality,”Adv. Mater ., vol. 32, no. 1904646, pp. 1–22, 2020

2020

[2] [2]

Ultra-thin broadband metamaterial absorber,

Y . Liu and S. Gu and C. Luo and X.Zhao, “Ultra-thin broadband metamaterial absorber,”Appl. Phys. A, vol. 108, p. 19–24, 2012

2012

[3] [3]

Ultra-broadband microwave metamaterial absorber,

F. Ding, Y . Cui, X. Ge, F. Zhang, Y . Jin and S. He, “Ultra-broadband microwave metamaterial absorber,”Appl. Phys. Lett., vol. 100, no. 103506, pp. 1–4, 2012

2012

[4] [4]

Perforated lightweight broadband metamaterial absorber based on 3-d printed honeycomb,

S. Ghosh and S. Lim, “Perforated lightweight broadband metamaterial absorber based on 3-d printed honeycomb,”IEEE Wirel. Antennas Propag., vol. 17, no. 12, pp. 2379–2383, 2018

2018

[5] [5]

Design and full characterization of a 3-d-printed hyperbolic pyramidal wideband microwave absorber,

X. Lleshi and T.Q.V . Hoang and B.Loiseaux and D.Lippens, “Design and full characterization of a 3-d-printed hyperbolic pyramidal wideband microwave absorber,”IEEE Wirel. Antennas Propag., vol. 20, no. 1, pp. 28–32, 2021

2021

[6] [6]

Ultrabroadband transparent meta- material absorbers designed by anomalous brewster effect and gradient impedance matching optimized with deep neural network,

J. Chen, H. Duan, and G. Huang, “Ultrabroadband transparent meta- material absorbers designed by anomalous brewster effect and gradient impedance matching optimized with deep neural network,”Phys. Rev. Appl., vol. 23, no. L051002, 2025

2025

[7] [7]

Tri-band terahertz metamaterial absorber based on structural ti3c2tx mxene for enhanced sensing application,

S. Feng and L. Yang and B.Cai and W. Yang and L. Wu and Y . Cheng, “Tri-band terahertz metamaterial absorber based on structural ti3c2tx mxene for enhanced sensing application,”IEEE Sens. J., vol. 24, no. 18, pp. 28 889 – 28 896, 2024

2024

[8] [8]

Inverse-designed mxene metamaterial absorber for broadband solar energy harvesting,

H.Zhang and J. Yin, “Inverse-designed mxene metamaterial absorber for broadband solar energy harvesting,”Photonics Nanostructures: Fundam. Appl., vol. 63, p. 101362, 2025

2025