Pith. sign in

REVIEW 4 major objections 5 minor 65 references

Sulfur-vacancy diffusion in MoS2 is not a simple constant: it stays zero below about 9% vacancy concentration, rises steeply from 9% to 14%, and plateaus above 14%.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · deepseek-v4-flash

2026-08-02 00:37 UTC pith:R7NOTUAU

load-bearing objection A genuine KMC prediction of a sharp concentration-dependent vacancy diffusivity in MoS2, but the single-slow-rate assumption needs stronger support before the 9%/14% boundaries are taken literally. the 4 major comments →

arxiv 2607.14951 v1 pith:R7NOTUAU submitted 2026-07-16 cond-mat.mtrl-sci physics.comp-ph

Self-organized defect clustering and concentration-dependent vacancy diffusion in MoS₂

Aaron Fl\"ototto , Benjamin Spetzler , Martin Ziegler , Erich Runge , Christian Dre{\ss}ler This is my paper
classification cond-mat.mtrl-sci physics.comp-ph
keywords sulfur vacanciesMoS2kinetic Monte Carlovacancy diffusionself-organized clusteringmemristive devicescooperative jumpsdiffusion coefficient
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper argues that sulfur vacancies in MoS2 diffuse only when they are dense enough to self-organize into clusters, producing a sharply concentration-dependent diffusion coefficient with three distinct regimes. Below roughly 9% vacancy concentration, vacancies are immobile or trapped in small localized clusters; between 9% and 14% they form a connected, fluctuating network and the diffusion coefficient rises steeply; above about 14% the network spans the material and diffusion becomes constant. The authors connect this switch-like behavior to the memristive switching observed in MoS2 devices, suggesting that collective defect dynamics, not single-vacancy hops, control device operation.

Core claim

The central claim is that cooperative, vacancy-assisted sulfur jumps lead to self-organized vacancy clustering, and that this clustering produces a concentration-dependent diffusion coefficient with three regimes. Using kinetic Monte-Carlo simulations parameterized with rates from machine-learning interatomic-potential molecular dynamics, the authors show that the diffusion coefficient is essentially zero below ~9% vacancy concentration, increases monotonically between ~9% and ~14%, and saturates at a plateau above ~14%. They attribute the low-concentration immobility to small clusters confined to triangular mobile regions, and the plateau to extended, anisotropically shaped clusters whose m

What carries the argument

The key machinery is a two-rate kinetic Monte-Carlo model in which every cooperative vacancy jump is classified either as a fast association step (k1←2 = 101.9 ns⁻¹, merging two clusters into one) or a slow non-association step (k2←1 = 1.095 ns⁻¹, cluster dissociation or cluster-count-preserving kinks). These rates were extracted from first-passage-time distributions in MLIP MD simulations of clusters with 3–8 vacancies at 1000 K. The model then simulates large supercells (70×70 sulfur sites) with random initial vacancy distributions, tracking the number of connected lattice-site clusters and the vacancy mean squared displacement to obtain diffusion coefficients via the 2D Einstein–Smoluchow

Load-bearing premise

The entire result rests on assuming that the single slow jump rate measured from small clusters (3–8 vacancies) at 1000 K applies to every cooperative jump that does not merge clusters, regardless of cluster size, shape, or temperature.

What would settle it

Run MLIP MD on a single cluster of 30–50 vacancies and measure the rate of cluster-count-preserving jumps (kink formation); if that rate differs from 1.095 ns⁻¹ by more than about a factor of two, the KMC regime boundaries and plateau value would shift. Alternatively, experimentally measure vacancy diffusion in a MoS2 device as a function of known vacancy concentration and check whether diffusion indeed remains zero below ~9% and plateaus above ~14%.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • If the threshold behavior is correct, MoS2 memristors operate in a regime where small changes in vacancy concentration near 9–14% cause large changes in ionic mobility, offering a natural mechanism for analog switching.
  • Above 14% vacancy concentration, diffusion becomes fast and nearly constant, implying that heavily damaged MoS2 will exhibit strong vacancy migration and potentially rapid degradation, while low-concentration devices remain configurationally frozen.
  • The three-regime picture provides a concrete experimental target: quantifying the sulfur-vacancy concentration in the channel of a MoS2 device should correlate with the onset of memristive switching.
  • The plateau at high concentration suggests that further increasing defect density does not accelerate vacancy transport, so device optimization should focus on the intermediate regime where mobility is most tunable.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same cooperative-jump mechanism may produce analogous concentration thresholds in other transition-metal dichalcogenides, so the three-regime behavior could be a general feature of defect transport in 2D materials with vacancy-assisted migration.
  • Because all rates were derived at 1000 K and temperature independence was assumed rather than demonstrated, the specific threshold concentrations (9% and 14%) may shift at room temperature; a reparameterized simulation at lower temperatures would test this.
  • The sharp onset of diffusion resembles a percolation transition, and the broad cluster-size distribution seen in the intermediate regime hints that device-to-device variability in memristors may reflect stochastic cluster-network formation rather than simple vacancy count.
  • A testable extension would be to apply an electric field or mechanical strain in the simulation to see whether the clustering and plateau boundaries shift, which could inform how to engineer the switching window in real devices.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper constructs a lattice kinetic Monte-Carlo (KMC) model for sulfur-vacancy migration in monolayer MoS2, with transition rates extracted from MLIP-based molecular-dynamics first-passage-time simulations at 1000 K. Two rates are used: a slow dissociation rate k2←1 = 1.095 ns−1 and a fast association rate k1←2 = 101.9 ns−1. The KMC simulations, run for vacancy concentrations from 5% to 20%, identify three regimes in both the density of connected vacancy clusters and the vacancy diffusion coefficient D(c): a localized regime at c ≲ 9% with no long-range diffusion, an intermediate regime from roughly 9% to 14% where D rises with concentration, and an extended regime above about 14% where D plateaus. The clustering is attributed to cooperative vacancy-assisted jumps, and the results are connected to memristive switching in MoS2-based devices.

Significance. If the central D(c) result is robust, the paper provides a concrete, physically motivated mechanism by which defect transport in MoS2 switches from localized to network-mediated as the vacancy concentration crosses a sharp threshold. This is potentially important for interpreting memristive and memtransistor behavior. The paper has notable strengths: the KMC rates are not fitted to the target D(c) curve but derived from an independent MLIP-MD benchmark; the KMC model is checked against MD mean-squared displacements for small clusters; supercell-size convergence is tested; and the manuscript makes code, data, and trajectories openly available. The main weakness is the transfer of a single slow rate measured for one class of events to all non-association cooperative jumps, which the reported validation does not fully constrain.

major comments (4)
  1. [Sec. II B, Eq. (2)] The rate k2←1 = 1.095 ns−1 is measured from 1→2 cluster-dissociation first-passage times for clusters of 3–8 vacancies (Sec. II A, SI I) and then assigned to every non-association cooperative jump, including kink formation and cluster-count-preserving moves (Sec. II B). The only direct evidence for this universality is (i) a null trend in six short MD trajectories with admitted scatter and (ii) a statement of nearly identical DFT energy barriers [32,33]. At 1000 K, a 0.1 eV barrier difference changes rates by roughly a factor of three, and the difference grows at lower temperatures. Because the KMC simulations contain clusters of hundreds of vacancies, and because the plateau value and the 9%/14% boundaries are controlled by the slow jump rate, this single-rate assignment is load-bearing. Please provide either direct MD/DFT measurement of kink/preserving-move rates or a sensitivity analy
  2. [Sec. II C, remark (ii)] All rates are sampled at 1000 K, and the text asserts that the three-regime structure depends only weakly on temperature without showing supporting calculations. If the slow and fast processes have different energy barriers, the effective association/dissociation ratio will change with temperature, and the location and sharpness of the 9% and 14% thresholds, as well as the plateau value, could move substantially. This is not merely a presentation issue: the temperature dependence is used to connect the simulation to device operating conditions. A lower-temperature rate calculation or at least a barrier-decomposition analysis is needed to justify the claim.
  3. [Sec. II A, Fig. 2] The first-passage-time fits in Fig. 2 are shown without confidence intervals or goodness-of-fit measures. The total MD dataset is 390 ns distributed over six trajectories, and SI I shows that per-size exponential fits have significant scatter. The reported rate ratio k1←2/k2←1 ≈ 93:1 is the main input controlling clustering and transport, so its uncertainty should be quantified. Please report standard errors or confidence bounds on k2←1 and k1←2 and propagate them to D(c), at least approximately.
  4. [Sec. II C, Fig. 7] The diffusion coefficients are shown without error bars or explicit inter-seed variability, although ten independent simulations per concentration were performed. The claims that D is exactly zero below ~9% and that the onset can be located within numerical accuracy rely on this variability. Please report the seed-to-seed spread of D, and, if possible, give a quantitative criterion for the regime boundaries rather than visual inspection of Fig. 7.
minor comments (5)
  1. [Sec. II C] Typo: “conicides” should be “coincides.”
  2. [Sec. II C, remark (ii)] Typo: “spezialized” should be “specialized.”
  3. [Ref. [62]] The repository title contains “Monte-Marlo”; it should be “Monte-Carlo.”
  4. [SI Section V] The text refers to “Fig. 5 of the main manuscript” when describing diffusion coefficients; the correct reference is Fig. 7 of the main text.
  5. [Fig. S3] The caption reads “10 3μs long KMC simulations”; the formatting is ambiguous and should read “ten 3 μs long KMC simulations” or similar.

Circularity Check

0 steps flagged

No significant circularity: the three-regime D(c) curve is a genuine KMC output from rates measured in MLIP MD, not a fit or a renamed input.

full rationale

The derivation chain is a forward simulation, not a hidden tautology. The KMC rates k2<-1 = 1.095 ns^-1 and k1<-2 = 101.9 ns^-1 (Eqs. 2-3) are fitted to first-passage-time histograms from MLIP MD (Fig. 2), an external atomistic benchmark grounded in DFT; the KMC then generates cluster densities (Fig. 6) and MSDs, from which D(c) is obtained via the Einstein relation (Eq. 5). The 9% onset and 14% plateau are emergent properties of the simulated dynamics rather than prescribed thresholds; the paper explicitly describes them as outputs ('the density of lattice-site clusters reaches a maximum at about 9%', 'the diffusion coefficient begins to increase from zero at ... approximately 9%'). No equation defines the predicted D(c) in terms of the fitted rates by construction, and no fitted parameter is renamed as a prediction. The self-citations (Ref. 36 for the MLIP and for cooperative-association/dissociation motifs) are not load-bearing in a circular sense: the MLIP is openly available, DFT-trained, and externally validated, and the cooperative mechanism is independently supported by DFT [31-33] and TEM [42,43]. The KMC-vs-MD MSD comparison (SI II) is a consistency check, not the target claim. The stated limitations - temperature dependence asserted in Sec. II C remark (ii), 'significant scatter' in size-dependent FPT fits (SI I), and validation only for 4-6 vacancy clusters (SI II) - are robustness/correctness risks, not evidence of circularity, because they do not make the output equal to an input by definition.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

The KMC model is parameterized by exactly two fitted rates; everything else (three regimes, D(c), structure factors) is emergent. The central claim therefore inherits the assumptions above, particularly the universality of the slow rate and the temperature transferability. No new physical entities are postulated; the triangular mobile-region constraint is inherited from prior MD observations [36] and validated against MD MSDs (SI Sect. II).

free parameters (4)
  • k₂←₁ (slow/dissociation rate) = 1.095 ns⁻¹
    Fitted to the exponential FPT histogram of 1→2 cluster dissociation events from 390 ns of MLIP MD at 1000 K (Fig. 2a, Eq. 2). Sets the slow-jump timescale for all non-association cooperative jumps in the KMC.
  • k₁←₂ (fast/association rate) = 101.9 ns⁻¹
    Fitted to the FPT histogram of 2→1 association events (Fig. 2b, Eq. 3). The ~93:1 ratio with k₂←₁ encodes the clustering tendency that drives the reported regimes.
  • MSD regression window = τ ∈ [2.5, 3.5] µs
    Hand-chosen window for linear regression of MSD to obtain D (Eq. 5, Sec. II C); D values depend on this choice, justified only by SI Fig. S5 linearity.
  • Equilibration/sampling times = 4.5 µs equilibration; 0.5 µs cluster-density window
    Hand-chosen simulation protocol; cluster densities in Fig. 6 are sensitive to the sampling interval (2 ns adjacency sampling) and window length.
axioms (6)
  • domain assumption Markovian two-state kinetics: FPTs are exponentially distributed with a single rate (Eq. 1)
    Underlies the extraction of k₂←₁ and k₁←₂ from MD FPT histograms (Sec. II A); multi-exponential or non-Markovian dynamics would invalidate the rate extraction.
  • domain assumption Vacancy migration proceeds exclusively via one-vacancy-assisted cooperative jumps; isolated-vacancy jumps have barriers >2 eV and are suppressed
    KMC includes only cooperative jumps (Sec. II B); inherited from cited DFT/MLIP work [31–36], not re-derived here.
  • domain assumption All non-association cooperative jumps share the dissociation rate k₂←₁
    The slow rate is measured only for 1→2 dissociation but applied to kink formation and other cluster-count-preserving jumps, justified by 'nearly identical energy barriers' from cited DFT [32,33] (Sec. II B, III B).
  • domain assumption Rejection KMC with averaged time increment 1/(N k_max) reproduces continuous-time dynamics
    Time advancement is averaged rather than drawn from the true waiting-time distribution (Sec. III B, Ref. [51]); partially supported by KMC-vs-MD MSD comparison (SI Sect. II).
  • domain assumption Rate universality over cluster size and shape (3–8 vacancies extrapolated to much larger clusters)
    FPTs averaged over 3–8 vacancy clusters; SI Sect. I shows 'significant scatter' in per-size fits and no systematic trend—a null result from six trajectories, extrapolated to clusters of hundreds of vacancies.
  • ad hoc to paper The three-regime structure is weakly temperature-dependent despite all rates being sampled at 1000 K
    Stated as an expectation in Sec. II C remark (ii) with no supporting calculation or activation-energy analysis.

pith-pipeline@v1.3.0-alltime-deepseek · 16096 in / 22990 out tokens · 220046 ms · 2026-08-02T00:37:02.658844+00:00 · methodology

0 comments
read the original abstract

Sulfur vacancy migration has a crucial impact on electronic transport and the functional behavior of MoS$_2$-based devices such as memristors and memtransistors. According to recent atomistic simulations, vacancy migration proceeds via cooperative, vacancy-assisted sulfur jumps, implying strongly correlated defect dynamics. Here, we investigate the collective behavior of sulfur-vacancy clusters in MoS$_2$ using kinetic Monte-Carlo simulations with transition rates derived from machine learning interatomic potential molecular dynamics simulations. We identify three transport regimes: At low concentrations, vacancies are immobile or confined within small clusters, whereas at high concentrations, classical diffusive transport with a constant diffusion coefficient is observed, and vacancies aggregate into anisotropically extended clusters. A well defined intermediate regime is characterized by clusters merging into a connected, fluctuating network with a concentration-dependent diffusion coefficient. This regime is characterized by a broad distribution of cluster sizes. The strong dependence of the vacancy diffusion coefficient on the average defect concentration provides new insights into the origin of memristive behavior observed in MoS$_2$.

Figures

Figures reproduced from arXiv: 2607.14951 by Aaron Fl\"ototto, Benjamin Spetzler, Christian Dre{\ss}ler, Erich Runge, Martin Ziegler.

Figure 1
Figure 1. Figure 1: FIG. 1. Illustration of slow and fast sulfur-vacancy jumps in [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: FIG. 2. Probability density histogram of the first-passage [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: FIG. 3. Time evolution of the number of sulfur-vacancy transitions between lattice site pairs from KMC simulations with [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: FIG. 4. Site-resolved sulfur vacancy occupation function averaged within 0.5 [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: FIG. 5. Vacancy structure factor computed via Eq. 7 from [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: FIG. 6. Density of connected vacancy clusters as a function [PITH_FULL_IMAGE:figures/full_fig_p006_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: FIG. 7. Diffusion coefficient [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

65 extracted references · 3 linked inside Pith

  1. [1]

    K. F. Mak, C. Lee, J. Hone, J. Shan, and T. F. Heinz, Atomically Thin MoS2: A New Direct-Gap Semiconduc- tor, Physical Review Letters105, 136805 (2010)

    2010

  2. [2]

    Radisavljevic, A

    B. Radisavljevic, A. Radenovic, J. Brivio, V. Giacometti, and A. Kis, Single-layer MoS 2 transistors, Nature Nan- otechnology6, 147 (2011)

    2011

  3. [3]

    Q. H. Wang, K. Kalantar-Zadeh, A. Kis, J. N. Cole- man, and M. S. Strano, Electronics and optoelectronics of two-dimensional transition metal dichalcogenides, Na- ture Nanotechnology7, 699 (2012)

    2012

  4. [4]

    Chhowalla, H

    M. Chhowalla, H. S. Shin, G. Eda, L.-J. Li, K. P. Loh, and H. Zhang, The chemistry of two-dimensional lay- ered transition metal dichalcogenide nanosheets, Nature Chemistry5, 263 (2013)

    2013

  5. [5]

    L. Li, Z. Qin, L. Ries, S. Hong, T. Michel, J. Yang, C. Salameh, M. Bechelany, P. Miele, D. Kaplan, M. Chhowalla, and D. Voiry, Role of sulfur vacancies and undercoordinated Mo regions in MoS2 nanosheets toward the evolution of hydrogen, ACS Nano13, 6824 (2019)

    2019

  6. [6]

    J. Hu, L. Yu, J. Deng, Y. Wang, K. Cheng, C. Ma, Q. Zhang, W. Wen, S. Yu, Y. Pan, J. Yang, H. Ma, F. Qi, Y. Wang, Y. Zheng, M. Chen, R. Huang, S. Zhang, Z. Zhao, J. Mao, X. Meng, Q. Ji, G. Hou, X. Han, X. Bao, Y. Wang, and D. Deng, Sulfur vacancy-rich MoS 2 as a catalyst for the hydrogenation of CO 2 to methanol, Na- ture Catalysis4, 242 (2021)

    2021

  7. [7]

    H. Fei, R. Liu, J. Wang, T. Guo, Z. Wu, D. Wang, and F. Liu, Targeted modulation of competitive active sites toward nitrogen fixation via sulfur vacancy engineering over MoS2, Advanced Functional Materials33, 2302501 (2023). 9

    2023

  8. [8]

    X. Jin, T. Lee, A. Soon, and S.-J. Hwang, Vacancy- assisted transformation of MoS 2 nanosheets into de- fective MoS x nanoclusters to regulate sodium-ion elec- trode functionality, Advanced Functional Materials34, 2316446 (2024)

    2024

  9. [9]

    B. A. Chen, A. J. Rowberg, T. A. Pham, T. Ogitsu, P. V. Kamat, and S. Ptasinska, Reactivity of sulfur vacancy- rich MoS 2 to water dissociation, Journal of Physical Chemistry C128, 10379 (2024)

    2024

  10. [10]

    X. Pang, Q. Zhang, X. Liang, W. Qin, H. Xu, and Y. Zhao, Sulfur vacancy synergistically coupling grain cluster for boosting MoS 2 piezoresistive properties, Jour- nal of Applied Physics136, 125106 (2024)

    2024

  11. [11]

    V. K. Sangwan, D. Jariwala, I. S. Kim, K.-S. Chen, T. J. Marks, L. J. Lauhon, and M. C. Hersam, Gate-tunable memristive phenomena mediated by grain boundaries in single-layer MoS 2, Nature Nanotechnology10, 403 (2015)

    2015

  12. [12]

    D. Li, B. Wu, X. Zhu, J. Wang, B. Ryu, W. D. Lu, W. Lu, and X. Liang, MoS 2 memristors exhibiting vari- able switching characteristics toward biorealistic synaptic emulation, ACS Nano12, 9240 (2018)

    2018

  13. [13]

    V. K. Sangwan, H. S. Lee, H. Bergeron, I. Balla, M. E. Beck, K. S. Chen, and M. C. Hersam, Multi-terminal memtransistors from polycrystalline monolayer molybde- num disulfide, Nature554, 500 (2018)

    2018

  14. [14]

    Jadwiszczak, D

    J. Jadwiszczak, D. Keane, P. Maguire, C. P. Cullen, Y. Zhou, H. Song, C. Downing, D. Fox, N. McEvoy, R. Zhu, J. Xu, G. S. Duesberg, Z. M. Liao, J. J. Boland, and H. Zhang, MoS 2 memtransistors fabricated by local- ized helium ion beam irradiation, ACS Nano13, 14262 (2019)

    2019

  15. [15]

    L. Wang, W. Liao, S. L. Wong, Z. G. Yu, S. Li, Y. F. Lim, X. Feng, W. C. Tan, X. Huang, L. Chen, L. Liu, J. Chen, X. Gong, C. Zhu, X. Liu, Y. W. Zhang, D. Chi, and K. W. Ang, Artificial synapses based on multiterminal memtransistors for neuromorphic application, Advanced Functional Materials29, 1901106 (2019)

    2019

  16. [16]

    H. S. Lee, V. K. Sangwan, W. A. Rojas, H. Berg- eron, H. Y. Jeong, J. Yuan, K. Su, and M. C. Her- sam, Dual-gated MoS2 memtransistor crossbar array, Ad- vanced Functional Materials30, 2003683 (2020)

    2020

  17. [17]

    X. Feng, S. Li, S. L. Wong, S. Tong, L. Chen, P. Zhang, L. Wang, X. Fong, D. Chi, and K. W. Ang, Self- selective multi-terminal memtransistor crossbar array for in-memory computing, ACS Nano15, 1764 (2021)

    2021

  18. [18]

    J. Yuan, S. E. Liu, A. Shylendra, W. A. G. Rojas, S. Guo, H. Bergeron, S. Li, H. S. Lee, S. Nasrin, V. K. Sangwan, A. R. Trivedi, and M. C. Hersam, Reconfigurable MoS 2 memtransistors for continuous learning in spiking neural networks, Nano Letters21, 6432 (2021)

    2021

  19. [19]

    S. Fu, J. H. Park, H. Gao, T. Zhang, X. Ji, T. Fu, L. Sun, J. Kong, and J. Yao, Two-terminal MoS 2 memristor and the homogeneous integration with a MoS 2 transistor for neural networks, Nano Letters23, 5869 (2023)

    2023

  20. [20]

    G. M. Marega, H. G. Ji, Z. Wang, G. Pasquale, M. Tri- pathi, A. Radenovic, and A. Kis, A large-scale integrated vector–matrix multiplication processor based on mono- layer molybdenum disulfide memories, Nature Electron- ics6, 991 (2023)

    2023

  21. [21]

    K. Zhu, S. Pazos, F. Aguirre, Y. Shen, Y. Yuan, W. Zheng, O. Alharbi, M. A. Villena, B. Fang, X. Li, A. Milozzi, M. Farronato, M. Mu˜ noz-Rojo, T. Wang, R. Li, H. Fariborzi, J. B. Roldan, G. Benstetter, X. Zhang, H. N. Alshareef, T. Grasser, H. Wu, D. Ielmini, and M. Lanza, Hybrid 2D–CMOS microchips for mem- ristive applications, Nature618, 57 (2023)

    2023

  22. [22]

    S. E. Liu, T. T. Zeng, R. Wu, V. K. Sangwan, and M. C. Hersam, Low-voltage short-channel MoS 2 mem- transistors with high gate-tunability, Journal of Materi- als Research39, 1463 (2024)

    2024

  23. [23]

    S. T. Yang, T. H. Yang, B. W. Liang, H. C. Lo, W. H. Chang, P. Y. Lin, C. Y. Su, and Y. W. Lan, Submicron memtransistors made from monocrystalline molybdenum disulfide, ACS Nano18, 6936 (2024)

    2024

  24. [24]

    T. Zou, S. Heo, G. Byeon, S. Yoo, M. Kim, Y. Reo, S. Kim, A. Liu, and Y. Y. Noh, Two-dimensional tun- neling memtransistor with thin-film heterostructure for low-power logic-in-memory complementary metal-oxide semiconductor, ACS Nano18, 13849 (2024)

    2024

  25. [25]

    H. Li, L. Tao, and J. B. Xu, Intrinsic memristive mecha- nisms in 2D layered materials for high-performance mem- ory, Journal of Applied Physics129, 050902 (2021)

    2021

  26. [26]

    Spetzler, Z

    B. Spetzler, Z. Geng, K. Rossnagel, M. Ziegler, and F. Schwierz, Lateral 2D TMDC memristors - experiment and modeling, inProceedings of 2022 IEEE 16th Inter- national Conference on Solid-State and Integrated Cir- cuit Technology, ICSICT 2022(Institute of Electrical and Electronics Engineers Inc., 2022)

    2022

  27. [27]

    Abdel, M

    D. Abdel, M. Herda, M. Ziegler, C. Chainais-Hillairet, B. Spetzler, and P. Farrell, Numerical analysis and sim- ulation of lateral memristive devices: Schottky, ohmic, and multi-dimensional electrode models, Computers & Mathematics with Applications199, 286 (2025)

    2025

  28. [28]

    Spetzler, D

    B. Spetzler, D. Abdel, F. Schwierz, M. Ziegler, and P. Farrell, The role of vacancy dynamics in two- dimensional memristive devices, Advanced Electronic Materials10, 2300635 (2023)

    2023

  29. [29]

    Spetzler, V

    B. Spetzler, V. K. Sangwan, M. C. Hersam, and M. Ziegler, High-throughput numerical modeling of the tunable synaptic behavior in 2D MoS 2 memristive de- vices, npj 2D Materials and Applications9, 17 (2025)

    2025

  30. [30]

    C. Wu, B. Zhang, J. Chen, B. Zhao, W. Wang, Z. Li, R. Shi, and K. Liu, Vacancy-migration-enabled, two- dimensional gate-free neuromorphic devices for high- temperature applications, Advanced Functional Materi- als , e31998 (2026)

    2026

  31. [31]

    H. P. Komsa, S. Kurasch, O. Lehtinen, U. Kaiser, and A. V. Krasheninnikov, From point to extended defects in two-dimensional MoS 2: Evolution of atomic struc- ture under electron irradiation, Physical Review B - Con- densed Matter and Materials Physics88, 035301 (2013)

    2013

  32. [32]

    C. Gao, X. Yang, M. Jiang, L. Chen, Z. Chen, and C. V. Singh, Defect evolution behaviors from single sulfur point vacancies to line vacancies in monolayer molybde- num disulfide, Physical Chemistry Chemical Physics23, 19525 (2021)

    2021

  33. [33]

    P. Y. Wang, B. A. Chen, Y. C. Lee, and C. C. Chiu, First-principles modeling of the highly dynamical surface structure of a MoS 2 catalyst with S-vacancies, Physical Chemistry Chemical Physics24, 24166 (2022)

    2022

  34. [34]

    Hloˇ zn´ y, J

    A. Hloˇ zn´ y, J. Brndiar, M. Casula, and I. ˇStich, Struc- ture and dynamics of sulfur vacancies in monolayer MoS2 studied by DFT-based machine learning potentials, Jour- nal of Chemical Physics163, 214118 (2025)

    2025

  35. [35]

    Hloˇ zn´ y, J

    A. Hloˇ zn´ y, J. Brndiar, Y. Luo, and I. ˇStich, Upscal- ing DFT-trained machine-learning interatomic poten- tial toward Quantum Monte Carlo accuracy: Sulfur- 10 vacancy migration in monolayer MoS 2 as a testbed (2026), arXiv:2605.22601

    Pith/arXiv arXiv 2026

  36. [36]

    Fl¨ ototto, B

    A. Fl¨ ototto, B. Spetzler, R. von Stackelberg, M. Ziegler, E. Runge, and C. Dreßler, Large-scale cooperative sul- fur vacancy dynamics in two-dimensional MoS2 from ma- chine learning interatomic potentials, Small22, e10679 (2026)

    2026

  37. [37]

    S. M. Gali, A. Pershin, A. Lherbier, J. C. Charlier, and D. Beljonne, Electronic and transport properties in defec- tive MoS2: Impact of sulfur vacancies, Journal of Physi- cal Chemistry C124, 15076 (2020)

    2020

  38. [38]

    Speckmann, K

    C. Speckmann, K. Mustonen, D. Propst, C. Mangler, and J. Kotakoski, Electron irradiation effects on monolayer MoS2 at elevated temperatures, Physical Review B111, 054107 (2025)

    2025

  39. [39]

    P. Y. Wang, C. C. Yeh, M. J. Chiu, and C. C. Chiu, A comparative study on the linear scaling relations for the diffusion of S-vacancies on MoS 2 and WS 2, Physical Chemistry Chemical Physics26, 5070 (2024)

    2024

  40. [40]

    H¨ anseroth, A

    J. H¨ anseroth, A. Fl¨ ototto, M. N. Qaisrani, and C. Dreßler, Fine-tuning unifies foundational machine- learned interatomic potential architectures at ab initio accuracy, The Journal of Physical Chemistry Letters17, 3152 (2026)

    2026

  41. [41]

    H¨ anseroth, A

    J. H¨ anseroth, A. Fl¨ ototto, and C. Dreßler, Universal in- teratomic potentials as configuration-space generators for one-shot and iterative fine-tuning of ab initio-accurate material-specific models (2026), arXiv:2606.23214 [cond- mat.mtrl-sci]

    Pith/arXiv arXiv 2026

  42. [42]

    S. Zhou, S. Wang, H. Li, W. Xu, C. Gong, J. C. Gross- man, and J. H. Warner, Atomic structure and dynam- ics of defects in 2D MoS 2 bilayers, ACS Omega2, 3315 (2017)

    2017

  43. [43]

    Q. Chen, H. Li, S. Zhou, W. Xu, J. Chen, H. Sawada, C. S. Allen, A. I. Kirkland, J. C. Grossman, and J. H. Warner, Ultralong 1D vacancy channels for rapid atomic migration during 2D void formation in monolayer MoS 2, ACS Nano12, 7721 (2018)

    2018

  44. [44]

    Ritort and P

    F. Ritort and P. Sollich, Glassy dynamics of kinetically constrained models, Advances in Physics52, 219 (2003)

    2003

  45. [45]

    Causer, I

    L. Causer, I. Lesanovsky, M. C. Ba˜ nuls, and J. P. Garrahan, Dynamics and large deviation transitions of the XOR-Fredrickson-Andersen kinetically constrained model, Physical Review E102, 052132 (2020)

    2020

  46. [46]

    J¨ ackle and A

    J. J¨ ackle and A. Kronig, A kinetic lattice-gas model for the triangular lattice with strong dynamic correlations. I. Self-diffusion, Journal of Physics: Condensed Matter 6, 7633 (1994)

    1994

  47. [47]

    Kronig and J

    A. Kronig and J. Jackle, A kinetic lattice-gas model for the triangular lattice with strong dynamic correlations. ii. Collective diffusion, Journal of Physics: Condensed Matter6, 7655 (1994)

    1994

  48. [48]

    Humphrey, A

    W. Humphrey, A. Dalke, and K. Schulten, VMD: Visual Molecular Dynamics, Journal of Molecular Graphics14, 33 (1996)

    1996

  49. [49]

    Batatia, D

    I. Batatia, D. P. Kovacs, G. Simm, C. Ortner, and G. Csanyi, MACE: Higher order equivariant message passing neural networks for fast and accurate force fields, inAdvances in Neural Information Processing Systems, Vol. 35 (Curran Associates, Inc., 2022) pp. 11423–11436

    2022

  50. [50]

    Batatia, P

    I. Batatia, P. Benner, Y. Chiang, A. M. Elena, D. P. Kov´ acs, J. Riebesell, X. R. Advincula, M. Asta, M. Avaylon, W. J. Baldwin, F. Berger, N. Bernstein, A. Bhowmik, S. M. Blau, V. C˘ arare, J. P. Darby, S. De, F. D. Pia, V. L. Deringer, R. Elijoˇ sius, Z. El-Machachi, F. Falcioni, E. Fako, A. C. Ferrari, A. Genreith-Schriever, J. George, R. E. A. Goodal...

    Pith/arXiv arXiv 2023

  51. [51]

    Chatterjee and D

    A. Chatterjee and D. G. Vlachos, An overview of spatial microscopic and accelerated kinetic Monte-Carlo meth- ods, Journal of Computer-Aided Materials Design14, 253 (2007)

    2007

  52. [52]

    Fl¨ ototto, E

    A. Fl¨ ototto, E. Runge, and C. Dreßler, Visualization of self-organized defect clustering and concentration- dependent vacancy diffusion in MoS 2,www.doi.org/10. 71758/refodat.88(2026), research data repository (RE- FODAT)

    2026

  53. [53]

    Q. Ma, P. M. Odenthal, J. Mann, D. Le, C. S. Wang, Y. Zhu, T. Chen, D. Sun, K. Yamaguchi, T. Tran, M. Wurch, J. L. McKinley, J. Wyrick, K. Magnone, T. F. Heinz, T. S. Rahman, R. Kawakami, and L. Bartels, Con- trolled argon beam-induced desulfurization of monolayer molybdenum disulfide, Journal of Physics Condensed Matter25, 252201 (2013)

    2013

  54. [54]

    C. C. Cheng, A. Y. Lu, C. C. Tseng, X. Yang, M. N. Hedhili, M. C. Chen, K. H. Wei, and L. J. Li, Activating basal-plane catalytic activity of two-dimensional MoS 2 monolayer with remote hydrogen plasma, Nano Energy 30, 846 (2016)

    2016

  55. [55]

    Bertolazzi, S

    S. Bertolazzi, S. Bonacchi, G. Nan, A. Pershin, D. Beljonne, and P. Samor ` ı, Engineering chemically ac- tive defects in monolayer MoS 2 transistors via ion-beam irradiation and their healing via vapor deposition of alka- nethiols, Advanced Materials29, 1606760 (2017)

    2017

  56. [56]

    C. Tsai, H. Li, S. Park, J. Park, H. S. Han, J. K. Nørskov, X. Zheng, and F. Abild-Pedersen, Electrochemical gen- eration of sulfur vacancies in the basal plane of MoS 2 for hydrogen evolution, Nature Communications8, 15113 (2017)

    2017

  57. [57]

    J. Y. Lee, J. H. Kim, Y. Jung, J. C. Shin, Y. Lee, K. Kim, N. Kim, A. M. van der Zande, J. Son, and G. H. Lee, Evolution of defect formation during atomically precise desulfurization of monolayer MoS2, Communications Ma- terials2, 80 (2021)

    2021

  58. [58]

    P. Man, S. Jiang, K. H. Leung, K. H. Lai, Z. Guang, H. Chen, L. Huang, T. Chen, S. Gao, Y. K. Peng, C. S. Lee, Q. Deng, J. Zhao, and T. H. Ly, Salt-induced high- density vacancy-rich 2D MoS2 for efficient hydrogen evo- lution, Advanced Materials36, 2304808 (2024)

    2024

  59. [59]

    T. D. K¨ uhne, M. Iannuzzi, M. D. Ben, V. V. Ry- bkin, P. Seewald, F. Stein, T. Laino, R. Z. Khal- iullin, O. Sch¨ utt, F. Schiffmann, D. Golze, J. Wil- helm, S. Chulkov, M. H. Bani-Hashemian, V. Weber, U. Borˇ stnik, M. Taillefumier, A. S. Jakobovits, A. Laz- zaro, H. Pabst, T. M¨ uller, R. Schade, M. Guidon, S. Andermatt, N. Holmberg, G. K. Schenter, A....

    2020

  60. [60]

    Fl¨ ototto, B

    A. Fl¨ ototto, B. Spetzler, R. von Stackelberg, M. Ziegler, E. Runge, and C. Dreßler, Machine learning inter- atomic potentials for sulfur vacancy dynamics in single sheet MoS 2,https://refodat.de/receive/refodat_ mods_00000068(2025), research data repository (REFO- DAT)

    2025

  61. [61]

    A. H. Larsen, J. J. Mortensen, J. Blomqvist, I. E. Castelli, R. Christensen, M. Du lak, J. Friis, M. N. Groves, B. Hammer, C. Hargus, E. D. Hermes, P. C. Jennings, P. B. Jensen, J. Kermode, J. R. Kitchin, E. L. Kolsbjerg, J. Kubal, K. Kaasbjerg, S. Lysgaard, J. B. Maronsson, T. Maxson, T. Olsen, L. Pastewka, A. Pe- terson, C. Rostgaard, J. Schiøtz, O. Sch...

    2017

  62. [62]

    Fl¨ ototto, E

    A. Fl¨ ototto, E. Runge, and C. Dreßler, MoS 2 kinetic Monte-Marlo source code repository (2026),https: //github.com/afloetotto/mos2_montecarlo(commit 51448a1)

    2026

  63. [63]

    Fl¨ ototto, Simulation Package for Atomistic Modelling (SPAM) (2026),https://github.com/afloetotto/spam (commit 05185c4)

    A. Fl¨ ototto, Simulation Package for Atomistic Modelling (SPAM) (2026),https://github.com/afloetotto/spam (commit 05185c4). Supplementary Material: Self-organized defect clustering and concentration-dependent vacancy diffusion in MoS 2 Aaron Fl¨ ototto,1, 2,∗ Benjamin Spetzler, 3 Martin Ziegler, 3 Erich Runge, 1, 2 and Christian Dreßler 1, 2 1Technische ...

    2026

  64. [64]

    J¨ ackle and A

    J. J¨ ackle and A. Kronig, A kinetic lattice-gas model for the triangular lattice with strong dynamic correlations. I. Self- diffusion, Journal of Physics: Condensed Matter6, 7633 (1994)

    1994

  65. [65]

    Fl¨ ototto, B

    A. Fl¨ ototto, B. Spetzler, R. von Stackelberg, M. Ziegler, E. Runge, and C. Dreßler, Large-scale cooperative sulfur vacancy dynamics in two-dimensional MoS 2 from machine learning interatomic potentials, Small22, e10679 (2026)

    2026