Pith. sign in

REVIEW 3 major objections 5 minor 74 references

Interpretable, Physics-Informed Learning Reveals Sulfur Adsorption and Poisoning Mechanisms in 13-Atom Icosahedra Nanoclusters

T0 review · 3 major / 5 minor · reviewed 2026-08-03 · deepseek-v4-flash

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

Pith's one-line read The paper's central claim: across 30 icosahedral 13-atom transition-metal clusters, sulfur adsorption is dominated by the electronic interaction term, and the isoelectronic Ti13, Zr13, Hf13 group offers a sulfur-tolerant balance between str

desk verdict Solid 30-metal S-adsorption survey with a real internal inconsistency in the SO2 validation table; worth refereeing after fixes. read the letter →

arxiv 2601.13845 v1 pith:2FYHT4QT submitted 2026-01-20 physics.atm-clus cond-mat.mtrl-sci

Raiane Ferreira Monteiro , João Marcos T. Palheta , Tulio Gnoatto Grison , Octávio Rodrigues Filho , Renato Luis Tame Parreira , Diego Guedes-Sobrinho , Celso R. C. Rêgo , Alexandre C. Dias
show 2 more authors
Krys Elly de Araújo Batista Maurício J. Piotrowski
This is my paper
classification physics.atm-cluscond-mat.mtrl-sci PACS 31.15.E68.43.-h82.65.+r
keywords sulfurpoisoningtransition-metalnanoclustersicosahedralclustersdensityfunctionaltheoryinterpretablemachinelearningd-bandcenteradsorptionenergydecompositionSO2dissociation
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

This paper tries to establish a transferable, data-driven picture of how atomic sulfur poisons 13-atom transition-metal nanoclusters and which metals withstand it best. The authors fix all 30 metals (3d–5d series) to the icosahedral cluster motif and combine dispersion-corrected density functional theory with interpretable machine learning. They find that adsorption energy is dominated by the metal–sulfur interaction term, while distortion of the cluster cage is usually moderate; a few metals, notably Cr, Mo, W, and Re, show large negative distortion energies, meaning their icosahedral scaffold can reorganize into more stable motifs on adsorption. Clustering and regression analyses single out the isoelectronic trio Ti13, Zr13, and Hf13 as a middle regime: sulfur binds strongly enough to activate S-containing molecules yet not so strongly that the cluster falls apart. Explicit SO2 calculations on these three clusters show dissociative, strong adsorption with the icosahedral cage largely preserved, offered as validation and as design guidance for sulfur-tolerant subnanometer catalysts.

What carries the argument

Central engine is the fixed icosahedral (ICO) 13-atom motif—one central atom plus 12 surface atoms—used for all 30 metals, and the adsorption-energy decomposition Eads = ΔEint + ΔEdis(TM13), which splits the electronic interaction between frozen fragments from the energy penalty or gain of distorting the cluster. Around this decomposition, the paper organizes DFT-derived descriptors (binding energy, bond length and effective coordination, d-band center, HOMO–LUMO gap, charge transfer, vibrational frequencies and zero-point energy) into a descriptor space that feeds k-means clustering and leave-one-feature-out regression to rank which descriptors generalize best.

What would settle it

Perform a global structure search (e.g., basin hopping or genetic algorithm) for each of the 30 TM13 clusters without constraining to the icosahedral motif, then recompute atomic sulfur adsorption on the lowest-energy isomer. If for any of the metals flagged with negative distortion energies the non-ICO isomer gives significantly different adsorption energies or a different preferred site, the claim that ICO-based trends represent the operative sulfur-poisoning behavior would be overturned.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central claim is that on the fixed icosahedral (ICO) 13-atom motif, sulfur adsorption on all 30 transition metals is governed primarily by the electronic metal–sulfur interaction, with moderate geometric distortion of the metal cage for most metals, and that an interpretable machine-learning analysis of the DFT-derived descriptors identifies the isoelectronic Ti13, Zr13, Hf13 group as a chemically resilient intermediate-binding regime. Explicit SO2 calculations on those three clusters find strong dissociative adsorption (S–O bonds stretched past 4 Å) while the ICO cage remains essentially intact, which the authors take as direct validation of the trend-guided se

Load-bearing premise

Every conclusion is computed on clusters that were started in, and mostly keep, the icosahedral shape; if for some metals—especially Cr, Mo, W, and Re—the true lowest-energy 13-atom cluster has a non-icosahedral structure, the adsorption energies, site preferences, and poisoning picture could be different.

Editorial extensions

If this is right

  • If correct, the dominance of the interaction term means sulfur-tolerance trends can be predicted mainly from electronic descriptors such as the d-band center, not just from geometric rigidity.
  • The Ti13, Zr13, Hf13 trio becomes a concrete family of candidate sulfur-tolerant subnanometer catalysts; their dissociative SO2 adsorption suggests they may activate sulfur–oxygen bonds rather than merely accumulate sulfur.
  • The descriptor ranking from leave-one-feature-out analysis gives a short menu of pristine-cluster properties that generalize to predict sulfur adsorption on new clusters, enabling cheaper screening beyond the 30 metals.
  • For Cr, Mo, W, and Re, negative distortion energies imply sulfur adsorption can drive the cluster toward a different motif, so their poisoning behavior may depend strongly on isomer choice or environmental conditions.
  • The flattened binding-energy curve after sulfur adsorption suggests that adsorbed sulfur tends to equalize metal cluster stabilities, so poisoning risk may be more uniform across metals than bulk trends imply.

Reading between the lines

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

  • Editorial inference: because the analysis fixes the ICO motif, the predicted poisoning mechanisms and the Ti/Zr/Hf designation hold only if ICO is a representative shape for each metal; a global isomer search for the anomalous Cr, Mo, W, Re cases would test whether the descriptor rankings survive.
  • Editorial inference: the same descriptor pipeline could be applied to other poisoning species (e.g., H2S or phosphorus) or to bimetallic 13-atom clusters to see whether the interaction/distortion decomposition and the Ti/Zr/Hf resilience persist.
  • Editorial inference: the dissociative SO2 adsorption at low coverage suggests these clusters are worth testing in catalytic settings such as hydrodesulfurization or SO2 electroreduction, provided the dissociated fragments can be removed rather than remaining permanently bound.
  • Editorial inference: the clustering groups could double as an experimental roadmap—if size-selected Ti13, Zr13, and Hf13 clusters can be synthesized, direct measurements of sulfur binding and SO2 activation would provide a strong test of the computational trends.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The paper presents a combined DFT and machine-learning study of atomic sulfur adsorption on 13-atom icosahedral transition-metal nanoclusters (TM13) for 30 metals across the 3d–5d series. Using PBE-D3 with vibrational analysis, the authors compute adsorption energies and decompose them into interaction and distortion contributions, together with structural, electronic, and vibrational descriptors. Unsupervised k-means clustering and supervised regression with LOFO feature ranking are used to identify Ti13, Zr13, and Hf13 as a chemically resilient triad, and explicit SO2 adsorption calculations on these three clusters are presented as validation. The central claim is that these isoelectronic metals combine strong S binding with limited structural perturbation, making them promising sulfur-tolerant subnanometer catalysts.

Significance. If the results hold, the paper provides a broad, open dataset of S adsorption descriptors for 30 transition-metal icosahedral nanoclusters and a transparent workflow connecting DFT descriptors to poisoning trends. The computational protocol is standard and well documented, with convergence tests, vibrational confirmation of minima, and a public repository for inputs, outputs, and ML workflows. The energy decomposition is exact bookkeeping, and the LOFO analysis across three regressors is a reasonable interpretability strategy. However, the SO2 validation table contains internal inconsistencies that directly affect the Ti/Zr/Hf recommendation, and the fixed-ICO assumption is not verified for the selected metals. These issues must be resolved before the central claims can be accepted.

major comments (3)
  1. [Table 1; Eq. (4)] The Hf13 rows in Table 1 violate the decomposition Eq. (4). For dissociative Hf13, ΔE_int + 13·ΔE_TM13_dis + ΔE_SO2_dis = -25.100 + 13(0.342) + 5.879 = -14.775 eV (or -18.879 eV if ΔE_TM13_dis is interpreted as total), which does not reproduce the reported E_ads = -12.681 eV. For molecular Hf13, the same sum gives +3.185 eV (per-atom) or -0.272 eV (total), versus E_ads = -6.826 eV. The Zr13 dissociative row only satisfies Eq. (4) if ΔE_TM13_dis = 0.360 eV is treated as a total energy, contradicting the column header '(eV/atom)'. Because Table 1 is the direct DFT validation of the Ti/Zr/Hf 'resilient platform' claim and of the 'strong, predominantly dissociative' SO2 conclusion, these inconsistencies are load-bearing. The authors must correct the units and recompute/report the decomposition for all six rows, and re-examine the Hf13 SO2 results in particular.
  2. [§2 Atomic Configurations; §3.2; Conclusions] The paper fixes all 30 TM13 clusters to the ICO motif, stating that it is the lowest-energy configuration for several TM13 NCs (refs [33,59]), but it does not verify that ICO is the global or even a representative minimum for each element. The manuscript's own §3.2 reports that Cr, Mo, W, and Re can relax to motifs more stable than the pristine ICO geometry, so for these metals the reported S adsorption energies on the relaxed ICO may not be the operative ones. More importantly for the central recommendation, no ground-state check is reported for Ti, Zr, and Hf. If their lowest-energy 13-atom isomers are not ICO, the statement that the ICO framework 'remains largely preserved' upon SO2 adsorption is not evidence of catalytic resilience. I request a concrete test: compare S adsorption on low-energy non-ICO motifs (e.g., decahedral, fcc/hcp, amorphous) for Ti13, Zr13, and Hf13, or perform
  3. [§2 Machine Learning Models; §3.3] The k-means clustering uses k=10 chosen 'based on the idea' that small chemically similar groups would form, without an elbow/silhouette analysis or stability check. The Ti/Zr/Hf selection is attributed to these cluster assignments (Figure 4 and §3.3), and the paper is titled around interpretable, physics-informed learning. Because the ML grouping is a central part of the selection rationale, the arbitrary k is a gap. The authors should report the model-selection criterion and a stability analysis showing that Ti/Zr/Hf remain in the same cluster for a range of k (e.g., 3–12) and across initializations. If the grouping is not robust, the claim that the ML provides 'a solid, data-driven reason' for the triad selection is overstated.
minor comments (5)
  1. [Throughout] There are several typos: 'Nanoclustes' in the keywords, 'physic-informed' in §2, and inconsistent spacing in 'V ASP'. A careful proofread is needed.
  2. [Table 1] The ΔE_SO2_dis column is labeled 'eV/atom' but the distortion energy of an adsorbate is defined per molecule in Eq. (7). Please correct the units. The vibrational frequencies at the bottom of the table should be labeled with the mode number or assignment, and the units (cm^-1) made explicit.
  3. [§2 Computational Details] The text says SOC was 'included and tested' but also says all findings originated from PBE-D3 optimizations. Please clarify explicitly whether the production energies include SOC or only the tested protocols in the SI.
  4. [Figure 6; §2 ML Models] The group cross-validation procedure is described only vaguely. Please specify the group definition (e.g., by d-series or by cluster), the split ratio, and report R²/MAE values for each model. The bar chart in Figure 6 would benefit from error bars or a scatter plot with parity line.
  5. [LOFO Analysis; Figure 5] The LOFO panels report ΔR², ΔMAE, and ΔRMSE, but the heatmap color scales are not defined in the caption, and negative deltas (redundant features) need interpretation. Also, the list of descriptors used in the models should be stated explicitly in the main text, not only in the Methods.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: energy decomposition is exact bookkeeping, SO2 validation is independent DFT; Table 1 Hf inconsistency is a correctness issue, not circularity.

full rationale

The derivation chain is not circular. Eq. 3 defines Eads, and Eq. 4 is an algebraic identity (Eads = ΔEint + 13ΔE_TM13_dis + ΔE_SO2_dis) whose terms are separately computed single-point energies; the statement that ΔEint dominates is an interpretation of the computed numbers, not a fitted constraint. The ML models are trained on the 30 DFT points but evaluated with group cross-validation and LOFO, so the abstract's phrase 'predict adsorption for new samples' is an overstatement of what group-CV generalization shows, but it is not a case of in-sample fit being relabeled as prediction. The Ti13/Zr13/Hf13 selection comes from k-means clustering of DFT-derived descriptors, and the SO2 adsorption results are new explicit DFT calculations (AIMD plus relaxation), so the validation is genuinely external to the ML fit. The fixed-ICO assumption is an explicit input, justified by prior literature (including self-citations [33,59]) as the lowest-energy configuration for several TM13 clusters; the paper does not claim to derive the motif, scopes the study to composition on ICO, and the conclusions are phrased within that motif. Some self-citation is present but is not load-bearing in a circular sense. Separate from circularity, Table 1's Hf13 rows do not satisfy the paper's own Eq. 4 (e.g., dissociative Hf13: Eads=-12.681 eV vs ΔEint+13ΔE_TM13_dis+ΔE_SO2_dis=-14.775 eV if ΔE_TM13 is per atom, or -18.879 eV if total; molecular Hf13: Eads=-6.826 eV vs +3.185 eV by either convention), so the SO2 validation data for Hf contain an internal inconsistency. That is a correctness/support defect, not a circular step, and does not raise the circularity score.

Assumptions & free parameters 1 free parameters · 6 assumptions · 0 invented entities

The central claims rest on a standard DFT protocol, two interpretive heuristics (d-band model, Sabatier principle), and one structural assumption (fixed ICO motif) that is load-bearing and only partially justified. The only hand-set numerical parameter in the analysis chain is k=10 for k-means. No new physical entities are postulated. The dominant epistemic risk is the motif assumption plus PBE accuracy, not any invented mechanism.

free parameters (1)
  • k (number of k-means clusters) = 10
    Set by hand in §2 (Machine Learning Models): 'We set the number of clusters to k = 10, based on the idea that small groups of chemically similar elements would form compact regions across the 30 transition metals.' With N=30, k=10 yields ~3 metals per cluster, which structurally favors triad groupings like Ti/Zr/Hf; the clustering output then feeds the selection later 'validated' by SO2 DFT.
assumptions (6)
  • domain assumption The icosahedral motif is the operative geometry for all 30 TM13 clusters
    §2 (Atomic Configurations): clusters are 'initially set to the ICO, which is the lowest energy configuration for several TM13 NCs' (refs [33,59]). Ground-state status is not verified per element; for Cr, Mo, W, Re the authors themselves report negative ΔE_dis^TM suggesting post-adsorption motifs more stable than pristine ICO (§3.2).
  • domain assumption PBE+D3 energetics are accurate enough for relative adsorption trends across 30 metals
    §2: the authors quote PBE accuracy of 0.28 eV on an isomerization dataset and assert 'better accuracy' for relative energies within related NC families. No experimental adsorption energies on these clusters are offered as benchmarks.
  • domain assumption The d-band model (εd as reactivity descriptor) applies to these finite clusters
    §3.2 invokes Hammer–Nørskov d-band reasoning (ref 66) and asserts that modest εd shifts upon adsorption mean 'the fundamental premises of the d-band model remain applicable'.
  • domain assumption The Sabatier principle is transferable to subnanometer clusters
    §3.3/§3.3.1: intermediate binding strength is equated with optimal activation/poisoning balance per Sabatier (refs 74,75). No catalytic cycle or turnover is computed.
  • standard math Eads decomposition identities (Eqs. 3–7) are exact
    Telescopes to Eads by construction; standard bookkeeping, but fails numerically in Table 1 for the Hf13 rows.
  • domain assumption Parabolic binding-energy trends explained by the cluster-orbital model
    §3.1 interprets parabolic Eb vs atomic number via ref 73; interpretive, not independently tested here.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Interpretable, Physics-Informed Learning Reveals Sulfur Adsorption and Poisoning Mechanisms in 13-Atom Icosahedra Nanoclusters." pith.science (2026). https://pith.science/paper/2FYHT4QT

@misc{pith2026260113845,
  author       = {Pith},
  title        = {Pith review of: Interpretable, Physics-Informed Learning Reveals Sulfur Adsorption and Poisoning Mechanisms in 13-Atom Icosahedra Nanoclusters},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/2FYHT4QT}},
  note         = {Machine review of arXiv:2601.13845}
}
abstract

Transition-metal nanoclusters exhibit structural and electronic properties that depend on their size, often making them superior to bulk materials for heterogeneous catalysis. However, their performance can be limited by sulfur poisoning. Here, we use dispersion-corrected density functional theory (DFT) and physics-informed machine learning to map how atomic sulfur adsorbs and causes poisoning on 13-atom icosahedral clusters from 30 different transition metals (3$d$ to 5$d$). We measure which sites sulfur prefers to adsorb to, the thermodynamics and energy breakdown, changes in structure, such as bond lengths and coordination, and electronic properties, such as $\varepsilon_d$, the HOMO-LUMO gap, and charge transfer. Vibrational analysis reveals true energy minima and provides ZPE-based descriptors that reflect the lattice stiffening upon sulfur adsorption. For most metals, the metal-sulfur interaction mainly determines adsorption energy. At the same time, distortion penalties are usually moderate but can be significant for a few metals, suggesting these are more likely to restructure when sulfur is adsorbed. Using unsupervised \textit{k}-means clustering, we identify periodic trends and group metals based on their adsorption responses. Supervised regression models with leave-one-feature-out analysis identify the descriptors that best predict adsorption for new samples. Our results highlight the isoelectronic triad \ce{Ti}, \ce{Zr}, and \ce{Hf} as a balanced group that combines strong sulfur binding with minimal structural change. Additional DFT calculations for \ce{SO2} adsorption reveal strong binding and a clear tendency toward dissociation on these clusters, linking electronic states, lattice response, and poisoning strength. These findings offer data-driven guidelines for designing sulfur-tolerant nanocatalysts at the subnanometer scale.

Figures

Figures reproduced from arXiv: 2601.13845 by the authors.

Figure 1
Figure 1. Schematic workflow illustrating the combined first-principles and machine-learning strategy adopted [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Vibrational and energetic trends of TM13 and S/TM13 across the 3d–5d series. (a) Vibrational frequencies for pristine TM13 (blue) and S/TM13 (black) systems. (b) Per-atom E TM13 b (blue) and E S/TM13 b (black), compared with bulk cohesive energies (E TM coh , green). 3.2 Sulfur Adsorption on Nanoclusters The S adsorption on TM13 NCs was systematically investigated by considering the three basic adsorption sites on I… view at source ↗
Figure 3
Figure 3. Energetic, structural, and electronic descriptors for [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figures from the paper (4 more)
Figure 4
Figure 4. Figure 4: (a) Two-dimensional PCA projections of standardized DFT-derived descriptors for pristine [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]
Figure 5
Figure 5. Figure 5: Leave-one-feature-out (LOFO) analysis measures how removing each pristine nanocluster descriptor [PITH_FULL_IMAGE:figures/full_fig_p012_5.png]
Figure 6
Figure 6. Figure 6: Comparison between DFT adsorption energies with machine learning predictions for atomic sulfur [PITH_FULL_IMAGE:figures/full_fig_p013_6.png]
Figure 7
Figure 7. Figure 7: The lowest-energy adsorption configurations and electronic structure of [PITH_FULL_IMAGE:figures/full_fig_p014_7.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

74 extracted references · 1 linked inside Pith

  1. [1]

    A. P . Alivisatos, Science, 1996, 271, 933–937

    1996

  2. [2]

    Schmid, Nanoparticles: From Theory to Application , Wiley-VCH, 2nd edn., 2011

    G. Schmid, Nanoparticles: From Theory to Application , Wiley-VCH, 2nd edn., 2011. [3] S. Maity, S. Kolay, S. Chakraborty, A. Devi, Rashi and A. Patra, Chemical Society Reviews , 2025, 54, 1585–1844

    2011

  3. [4]

    Jena and Q

    P . Jena and Q. Sun, Chemical Reviews, 2008, 118, 5755–5870

    2008

  4. [5]

    A. W. Castleman and P . Jena, Proceedings of the National Academy of Sciences , 2006, 103, 10554–10559

    2006

  5. [6]

    A. T. Bell, Science, 2003, 299, 1688–1691

    2003

  6. [7]

    Aslam, V

    U. Aslam, V . G. Rao, S. Chavez and S. Linic, Nature Catalysis, 2018, 1, 656–665

    2018

  7. [8]

    J. Li, X. Li, H.-J. Zhai and L.-S. Wang, Science, 2003, 299, 864–867

    2003

  8. [9]

    Cao and Y

    G. Cao and Y . Wang,Nanostructures and Nanomaterials: Synthesis, Properties, and Applications , World Scientific, 2nd edn., 2011

    2011

Show all 74 references
  1. [10]

    Y oon, H

    B. Y oon, H. Häkkinen, U. Landman, A. S. Wörz, J.-M. Antonietti, S. Abbet, K. Judai and U. Heiz, Science, 2005, 307, 403–407

    2005

  2. [11]

    Sun and P

    G. Sun and P . Sautet, Accounts of Chemical Research, 2021, 54, 3841–3849

    2021

  3. [12]

    C. R. C. Rêgo, P . Tereshchuk, L. N. Oliveira and J. L. F. Da Silva, Physical Review B, 2017, 95, year

    2017

  4. [13]

    C. T. Campbell, Accounts of Chemical Research, 2013, 46, 1712–1719

    2013

  5. [14]

    C. H. Bartholomew, Applied Catalysis A: General , 2001, 212, 17–60. 16

    2001

  6. [15]

    Mesilov, S

    V . Mesilov, S. Dahlin, S. L. Bergman, P . S. Hammershøi, S. Xi, L. J. Pettersson and S. L. Bernasek, Catalysis Science & Technology, 2021, 11, 5619–5632

    2021

  7. [16]

    D.-J. Liu, P . M. Spurgeon, J. Lee, T. L. Windus, P . A. Thiel and J. W. Evans, Physical Chemistry Chemical Physics, 2019, 21, 26483–26491

    2019

  8. [17]

    Fielicke, A

    A. Fielicke, A. Kirilyuk, C. Ratsch, J. Behler, M. Sche ffler, G. von Helden and G. Meijer, Physical Review Letters, 2004, 93, 023401

    2004

  9. [18]

    Heiz and U

    U. Heiz and U. Landman, Nanocatalysis, 2007, vol. 1

    2007

  10. [19]

    J. K. Nørskov, F. Abild-Pedersen, F. Studt and T. Bligaard, Proceedings of the National Academy of Sciences, 2011, 108, 937–943

    2011

  11. [20]

    Häkkinen, Nature Chemistry, 2012, 4, 443–455

    H. Häkkinen, Nature Chemistry, 2012, 4, 443–455

    2012

  12. [21]

    Ferrando, J

    R. Ferrando, J. Jellinek and R. L. Johnston, Chemical Reviews, 2008, 108, 845–910

    2008

  13. [22]

    C. S. T. Peraça, G. R. Nagurniak, R. P . Orenha, R. L. T. Parreira and M. J. Piotrowski, Dalton Transactions, 2020, 49, 492

    2020

  14. [23]

    M. J. Piotrowski, J. M. T. Palheta and R. Fournier, Phys. Chem. Chem. Phys. , 2024, 26, 13172–13181

    2024

  15. [24]

    Z. Say, M. Kaya, C. Kadero ¨glu, C. Langhammer, Y . Koçak, K. E. Ercan, A. T. Sika-Nartey, A. Jalal, A. A. Turk, C. Langhammer, M. J. V arjovi, E. Durgun and E. Ozensoy,Journal of the American Chemical Society , 2022, 144, 8848–8860

    2022

  16. [25]

    P . Yin, X. Luo, Y . Ma, S.-Q. Chu, S. Chen, X. Zheng, J. Lu, X.-J. Wu and H.-W. Liang, Nature Communications, 2021, 12, 3135

    2021

  17. [26]

    V . L. Deringer, M. A. Caro and G. Csányi, Advanced Materials, 2019, 31, 1902765

    2019

  18. [27]

    Y . Zuo, C. Chen, X. Li, Z. Deng, Y . Chen, J. Behler, G. Csányi, A. V . Shapeev, A. P . Thompson, M. A. Wood and S. P . Ong,Journal of Physical Chemistry A , 2020, 124, 731–745

    2020

  19. [28]

    Chmiela, H

    S. Chmiela, H. E. Sauceda, K.-R. Müller and A. Tkatchenko, Nature Communications, 2018, 9, 3887

    2018

  20. [29]

    Xie and J

    T. Xie and J. C. Grossman, Physical Review Letters, 2018, 120, 145301

    2018

  21. [30]

    Baletto and R

    F. Baletto and R. Ferrando, Reviews of Modern Physics , 2005, 77, 371–423

    2005

  22. [31]

    R. L. Johnston, Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 1998, 356, 211–230

    1998

  23. [32]

    Imaoka, H

    T. Imaoka, H. Kitazawa, W.-J. Chun, S. Omura, K. Albrecht and K. Y amamoto, Journal of the American Chemical Society, 2013, 135, 13089–13095

    2013

  24. [33]

    M. J. Piotrowski, P . Piquini and J. L. F. D. Silva, Phys. Rev. B, 2010, 81, 155446

    2010

  25. [34]

    Hohenberg and W

    P . Hohenberg and W. Kohn, Phys. Rev., 1964, 136, B864–B871

    1964

  26. [35]

    Kohn and L

    W. Kohn and L. J. Sham, Phys. Rev., 1965, 140, A1133–A1138

    1965

  27. [36]

    Kresse and J

    G. Kresse and J. Hafner, Phys. Rev. B, 1993, 48, 13115–13126

    1993

  28. [37]

    Kresse and J

    G. Kresse and J. Furthmüller, Phys. Rev. B, 1996, 54, 11169–11186

    1996

  29. [38]

    P . E. Blöchl, Phys. Rev. B, 1994, 50, 17953–17979

    1994

  30. [39]

    Kresse and D

    G. Kresse and D. Joubert, Phys. Rev. B, 1999, 59, 1758–1775

    1999

  31. [40]

    J. P . Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett., 1996, 77, 3865–3868. 17

    1996

  32. [41]

    J. P . Perdew, K. Burke and M. Ernzerhof, Physical Review Letters, 1997, 78, 1396–1396

    1997

  33. [42]

    Grimme, J

    S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104

    2010

  34. [43]

    Grimme, A

    S. Grimme, A. Hansen, J. G. Brandenburg and C. Bannwarth, Chem. Rev., 2016, 116, 5105–5154

    2016

  35. [44]

    C. R. C. Rêgo, L. N. Oliveira, P . Tereshchuk and J. L. F. Da Silva, Journal of Physics: Condensed Matter , 2015, 27, 415502

    2015

  36. [45]

    D. D. Koelling and B. N. Harmon, J. Phys. C: Solid State Phys. , 1977, 10, 3107

    1977

  37. [46]

    Takeda, Z

    T. Takeda, Z. Phys. B: Condens. Matter Quanta , 1978, 32, 43–48

    1978

  38. [47]

    Lejaeghere, G

    K. Lejaeghere, G. Bihlmayer, T. Björkman, P . Blaha, S. Blügel, V . Blum, D. Caliste, I. E. Castelli, S. J. Clark, A. D. Corso, S. de Gironcoli, T. Deutsch, J. K. Dewhurst, I. D. Marco, C. Draxl, M. Dulak, O. Eriksson, J. A. Flores-Livas, K. F. Garrity, L. Genovese, P . Gianno...

    2016

  39. [48]

    Mardirossian and M

    N. Mardirossian and M. Head-Gordon, Mol. Phys., 2017, 115, 2315–2372

    2017

  40. [49]

    Nosé, Mol

    S. Nosé, Mol. Phys., 1984, 52, 255–268

    1984

  41. [50]

    W. G. Hoover, Phys. Rev. A, 1985, 31, 1695–1697

    1985

  42. [51]

    C. R. Caldeira Rego, J. Schaarschmidt, T. Schlöder, M. Penaloza-Amion, S. Bag, T. Neumann, T. Strunk and W. Wenzel, Frontiers in Materials, 2022, 9, year

    2022

  43. [52]

    Schaarschmidt, J

    J. Schaarschmidt, J. Y uan, T. Strunk, I. Kondov, S. P . Huber, G. Pizzi, L. Kahle, F. T. Bölle, I. E. Castelli, T. V egge, F. Hanke, T. Hickel, J. Neugebauer, C. R. C. Rêgo and W. Wenzel, Advanced Energy Materials, 2021, 12, year

    2021

  44. [53]

    C. M. d. O. Bastos, A. C. Dias, A. C. Ferreira de Brito, I. D. Barcelos, A. L. da Rosa, D. N. Silveira, M. J. Piotrowski, W. Wenzel, C. R. C. Rêgo and D. Guedes-Sobrinho, Journal of Materials Chemistry A , 2025

    2025

  45. [54]

    C. R. C. Rêgo, W. Wenzel, M. J. Piotrowski, A. C. Dias, C. Maciel de Oliveira Bastos, L. O. de Araujo and D. Guedes-Sobrinho, 2025

    2025

  46. [55]

    L. O. de Araujo, C. R. Caldeira Rego, W. Wenzel, M. J. Piotrowski, A. C. Dias and D. Guedes-Sobrinho, npj Computational Materials, 2024, 10, year

    2024

  47. [56]

    Soleymanibrojeni, C

    M. Soleymanibrojeni, C. R. Caldeira Rego, M. Esmaeilpour and W. Wenzel, Journal of Materials Chemistry A, 2024, 12, 2249–2266

    2024

  48. [57]

    Bekemeier, C

    S. Bekemeier, C. R. Caldeira Rêgo, H. L. Mai, U. Saikia, O. Waseda, M. Apel, F. Arendt, A. Aschemann, B. Bayerlein, R. Courant, G. Dziwis, F. Fuchs, U. Giese, K. Junghanns, M. Kamal, L. Koschmieder, S. Leineweber, M. Luger, M. Lukas, J. Maas, J. Mertens, B. Mieller, L. Overmey...

    2025

  49. [58]

    J. F. Dalmedico, D. N. Silveira, L. O. de Araujo, W. Wenzel, C. R. C. Rêgo, A. C. Dias, D. Guedes-Sobrinho and M. J. Piotrowski, ChemPhysChem, 2024, 25, year

    2024

  50. [59]

    A. S. Chaves, M. J. Piotrowski and J. L. F. D. Silva, Phys. Chem. Chem. Phys. , 2017, 19, 15484–15502. 18

    2017

  51. [60]

    A. F. Y onezawa, G. R. Nagurniak, R. P . Orenha, E. H. da Silva, R. L. T. Parreira and M. J. Piotrowski, J. Phys. Chem. A , 2021, 125, 4805–4818

    2021

  52. [61]

    K. A. P . Sousa, F. M. Morawski, C. E. M. de Campos, R. L. T. Parreira, M. J. Piotrowski, G. R. Nagurniak and C. L. Jost, Electrochim. Acta, 2022, 408, 139945

    2022

  53. [62]

    J. P . C. S. Felix, K. E. A. Batista, W. O. Morais, G. R. Nagurniak, R. P . Orenha, C. R. C. Rêgo, D. Guedes- Sobrinho, R. L. T. Parreira, M. M. Ferrer and M. J. Piotrowski, J. Comput. Chem., 2023, 44, 1040–1051

    2023

  54. [63]

    Hoppe, Angew

    R. Hoppe, Angew. Chem. Int. Ed., 1970, 9, 25–34

    1970

  55. [64]

    Hoppe, Z

    R. Hoppe, Z. Kristallogr ., 1979, 150, 23–52

    1979

  56. [65]

    Momma and F

    K. Momma and F. Izumi, J. Appl. Cryst., 2008, 41, 653–658

    2008

  57. [66]

    Hammer and J

    B. Hammer and J. K. Nørskov, Advances in Catalysis, Academic Press Inc, San Diego, 2000

    2000

  58. [67]

    R. F. W. Bader, Atoms in Molecules: A Quantum Theory , Clarendon Press, 1994

    1994

  59. [68]

    W. Tang, E. Sanville, and G. Henkelman, J. Phys.: Condens. Matter , 2009, 21, 084204

    2009

  60. [69]

    Pedregosa, G

    F. Pedregosa, G. V aroquaux, A. Gramfort, V . Michel, B. Thirion, O. Grisel, M. Blondel, P . Prettenhofer, R. Weiss, V . Dubourg, J. V anderplas, A. Passos, D. Cournapeau, M. Brucher, M. Perrot and E. Duchesnay, Journal of Machine Learning Research , 2011, 12, 2825–2830

    2011

  61. [70]

    Zou and T

    H. Zou and T. Hastie, Journal of the Royal Statistical Society: Series B (Statistical Methodology) , 2005, 67, 301–320

    2005

  62. [71]

    H. Nori, S. Jenkins, P . Koch and R. Caruana, InterpretML: A Unified Framework for Machine Learning Interpretability, 2019, https://arxiv.org/abs/1909.09223

    2019 arXiv

  63. [72]

    Y . Lou, R. Caruana, J. Gehrke and G. Hooker, Proceedings of the 19th ACM SIGKDD international conference on Knowledge discovery and data mining, 2013, p. 623–631

    2013

  64. [73]

    D. G. Pettifor, Bonding and Structure of Molecules and Solids , Oxford University Press, Oxford, 1995

    1995

  65. [74]

    Sabatier,La Catalyse en Chimie Organique , Encyclopédie de Science Chimique Appliquée, Librairie Polytechnique, Paris et Liége, 1913

    P . Sabatier,La Catalyse en Chimie Organique , Encyclopédie de Science Chimique Appliquée, Librairie Polytechnique, Paris et Liége, 1913

    1913

  66. [75]

    A. J. Medford, A. V ojvodic, J. S. Hummelshøj, J. V oss, F. Abild-Pedersen, F. Studt, T. Bligaard, A. Nilsson and J. K. Nørskov, J. Catal., 2015, 328, 36–42. 19 Supporting Information: Interpretable, Physics-Informed Learning Reveals Sulfur Adsorption and Poisoning Mechanisms ...

    2015

Pith tools

Reviewed August 3, 2026 · model on record in the stance chip above.