A reactive neural network framework for water-loaded acidic zeolites
Open accessThe DOI matches its Crossref record; 109 of the 112 references it cites were found in a registry or an index, and 1 are not checked yet; 27 of the 64 sources its claims cite were checked for support.
Abstract
Under operating conditions, the dynamics of water and ions confined within protonic aluminosilicate zeolite micropores are responsible for many of their properties, including hydrothermal stability, acidity and catalytic activity. However, due to high computational cost, operando studies of acidic zeolites are currently rare and limited to specific cases and simplified models. In this work, we have developed a reactive neural network potential (NNP) attempting to cover the entire class of acidic zeolites, including the full range of experimentally relevant water concentrations and Si/Al ratios. This NNP has the potential to dramatically improve sampling, retaining the (meta)GGA DFT level accuracy, with the capacity for discovery of new chemistry, such as collective defect formation mechanisms at the zeolite surface. Furthermore, we exemplify how the NNP can be used as a basis for further extensions/improvements which include data-efficient adoption of higher-level (hybrid) references via Δ-learning and the acceleration of rare event sampling via automatic construction of collective variables. These developments represent a significant step towards accurate simulations of realistic catalysts under operando conditions.
Verified 2026-11-22 · 1 reference not checked yet
Overview
CoChat Summary
## TL;DR
The paper develops a reactive neural network potential (NNP) for water-loaded Brønsted acidic zeolites (BAS) that aims to cover broad chemical and configurational space, including varied water loadings and Si/Al ratios. It reports that the model retains roughly meta-GGA DFT-level accuracy across diverse tests and can also support Δ-learning and machine-learned collective variables for rare-event sampling.
## Key findings
- The final SCAN+D3(BJ) training database contained 248,439 structures; an ensemble of six SchNet NNPs trained on it achieved average test RMSEs of 5.3 meV atom−1 for energies and 186 meV Å−1 for forces.
- Across 2700 test configurations, the NNP gave an average reaction-energy error of about 6 kJ mol−1 for FAU reaction-path tests, and the authors say this is within DFT accuracy.
- For the in-domain Δ-learning example (proton jump in CHA), 150 training structures were sufficient to reach test-set RMSEs of 1.3 meV atom−1 and 69 meV Å−1, with the ΔNNP deviating by less than 3 kJ/mol from the ωB97X-D3(BJ) reference.
- In out-of-domain tests, performance remained good but worsened modestly: proton jump in FAU showed about 5 kJ mol−1 relative-energy error, and Al–O(H) bond dissociation in FAU reached up to 8 kJ mol−1 error.
- The model captured multiple reactive behaviors not in the training set, including a collective hydrolysis mechanism at an external MFI surface and a high-temperature GIS liquid-state run with thousands of bond-breaking events.
## Methods
- The authors built a broad training set from 500 short (10 ps) AIMD trajectories plus lattice deformations, bulk-water simulations, water clusters, ice/alumina polymorphs, and prior CHA trajectories; the resulting reference dataset used SCAN+D3(BJ) single-point labels.
- Generalization was tested with MD and NEB on unseen systems, including FAU, MFI, GIS, bulk-water-solvated fragments (Si(OH)4 and Al(OH)3), and high-temperature simulations at 1600 K and 3000 K.
- They trained an ensemble of six SchNet-based NNPs and also built two extensions: Δ-learning to ωB97X-D3(BJ) and machine-learned collective variables via a variational autoencoder on NNP representations.
## Limitations
- The paper states that finite-size effects hinder the interpretation of confined-water dynamics in FAU because a small primitive cell was used primarily for benchmarking.
- The authors note that the very high-temperature conditions used in some tests (1600 K and 3000 K) are not applicable to standard applications of water-zeolite systems; this is an evident limitation of the stress tests, not the target use case.
- For the biased-dynamics free-energy calculations, the authors explicitly say their restraints may bias the reported free-energy profiles and that the goal was to demonstrate capability rather than maximal accuracy.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures



Verification
Claims the paper makes
17 Valid · 2 Not valid · 15 Not checked · 66 Own results and interpretations
Some cited sources support this claim and some don’t — read each source’s verdict below.
In the paper
“As a result, the proper mechanistic understanding of BAS-water-zeolite interactions is still lacking, limited to either static calculations at ultra-high vacuum conditions13,14 or exploratory (ab initio) dynamical simulations of narrow scope15–18.”
§ A reactive neural network framework for water-loaded acidic zeolites > Introduction
What the sources say
- Not checked — no full text of the sourcecites [13] Silaghi M-C Regioselectivity of Al–O bond hydrolysis during zeolites dealumination unified by brønsted–evans–polanyi relationship ACS Catal. 2015 5 11 15 10.1021/cs501474u source
- Not checked — no full text of the sourcecites [14] Silaghi M-C Chizallet C Sauer J Raybaud P Dealumination mechanisms of zeolites and extra-framework aluminum confinement J. Catal. 2016 339 242 255 10.1016/j.jcat.2016.04.021 source
- Partly supportedcites [15] Heard CJ Fast room temperature lability of aluminosilicate zeolites Nat. Commun. 2019 10 1 7 10.1038/s41467-019-12752-y 31619677 PMC6795794 sourceExperimental — accuracy not yet measured
The Results (Computational studies) describe biased ab initio molecular dynamics of the initial hydrolysis of water-loaded chabazite, so this paper is an example of a dynamical study with a specific scope. It does not state that mechanistic understanding of Brønsted acid site–water–zeolite interactions is generally lacking or that prior work is limited to the two approaches named in the citing sentence. (LLM, conf=0.84)
- Partly supportedcites [16] Nielsen M Collective action of water molecules in zeolite dealumination Catal. Sci. Technol. 2019 9 3721 3725 10.1039/C9CY00624A sourceExperimental — accuracy not yet measured
The Introduction of Nielsen et al. (2019) supports the 'static calculations' half of the claim: 'so far all DFT studies of zeolite dealumination have employed a static approach, in which structures are located as minima in the potential energy landscape. Free energies were subsequently derived from nuclear vibrations in the harmonic approximation, which has obvious and severe limitations, in particular at the high temperatures employed in dealumination', and 'the static DFT approach in practice does not allow adequate sampling of all relevant states when taking the collective action of multiple water molecules into account'. However, our copy never describes those static studies as being at 'ultra-high vacuum conditions' (the limitation it names is the harmonic free-energy treatment and the gas-phase reference state for adsorbed water), and it does not characterise prior work as 'exploratory (ab initio) dynamical simulations of narrow scope' — it states the opposite, that all previous dealumination DFT studies were static, with its own DFT-MD/umbrella-sampling work presented as the first in situ dynamical investigation at realistic steaming conditions (Results and discussion; Conclusions). The claim's framing is also broader (BAS–water–zeolite interactions in general) than the dealumination-specific statements in the source. (LLM, conf=0.72)
- Disputedcites [17] Grifoni E Confinement effects and acid strength in zeolites Nat. Commun. 2021 12 2630 10.1038/s41467-021-22936-0 33976197 PMC8113345 sourceExperimental — accuracy not yet measured
The Introduction (page unavailable) describes full ab initio metadynamics across four zeolite frameworks and multiple water loadings, yielding thermodynamic and mechanistic information at atomistic detail—not a dynamical study of narrow scope. The Results and discussion report general findings about BAS solvation, contrary to the claim that mechanistic understanding remains limited to such studies. (LLM, conf=0.85)
Checked by a reasoning model — may be wrong.
- Not checked — no full text of the sourcecites [18] Jin M The role of water loading and germanium content in germanosilicate hydrolysis J. Phys. Chem. C 2021 125 23744 23757 10.1021/acs.jpcc.1c06873 source
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“A standard tool for accelerating the reactive sampling in zeolite-water systems, and thus reaching experimentally relevant timescale or realistic levels of model complexity is the class of reactive analytical potentials, for example, ReaxFF19. However, due to their fixed functional form, these potentials have limited transferability to systems with different chemical composition20.”
§ A reactive neural network framework for water-loaded acidic zeolites > Introduction
What the sources say
- Partly supportedcites [19] Senftle TP The ReaxFF reactive force-field: development, applications and future directions npj Comput. Mater. 2016 2 1 14 10.1038/npjcompumats.2015.11 sourceExperimental — accuracy not yet measured
The section on ReaxFF parameterisations and development branches (page unavailable in our copy) says that parameter sets cannot be combined arbitrarily and that transferring them between branches requires extensive refitting. It does not attribute this limitation to a fixed functional form; it discusses the branches within a common functional form and points instead to ReaxFF’s lack of atom typing as a complication. (LLM, conf=0.86)
- Supportedcites [20] Vandermause J Xie Y Lim JS Owen CJ Kozinsky B Active learning of reactive Bayesian force fields applied to heterogeneous catalysis dynamics of H/Pt Nat. Commun. 2022 13 5183 10.1038/s41467-022-32294-0 36055982 PMC9440250 source
The Introduction says ReaxFF models require expert fine-tuning for each system and can deviate substantially from ab initio results because of their limited parametric functional form. That supports the citing sentence’s point that the potentials do not readily transfer between chemically different systems. (LLM, conf=0.86)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The sources with a verdict support this claim; the rest weren’t checked yet, have no full text we could read, or their verdicts aren’t rated yet.
In the paper
“This relationship governs many features of BAS zeolites including (i) proton solvation, and thus acidity7,8”
§ A reactive neural network framework for water-loaded acidic zeolites > Introduction
What the sources say
- Not checked — no full text of the sourcecites [7] Pfriem N Role of the ionic environment in enhancing the activity of reacting molecules in zeolite pores Science 2021 372 952 957 10.1126/science.abh3418 33958482 source
- Supportedcites [8] Bates JS Bukowski BC Greeley J Gounder R Structure and solvation of confined water and water–ethanol clusters within microporous Brønsted acids and their effects on ethanol dehydration catalysis Chem. Sci. 2020 11 7102 7122 10.1039/D0SC02589E 33250979 PMC7690318 source
The Introduction (page not available in our copy) says water stabilizes solvated protons and that solvent polarity changes protonation equilibrium constants in zeolites. It also reports that solvent properties can change the strength of Brønsted acid sites, supporting the link between water interactions, proton solvation, and acidity. (LLM, conf=0.88)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“(iii) water mobility and clustering in zeolite pores10”
§ A reactive neural network framework for water-loaded acidic zeolites > Introduction
What the sources say
- Partly supportedcites [10] Fasano M Interplay between hydrophilicity and surface barriers on water transport in zeolite membranes Nat. Commun. 2016 7 12762 10.1038/ncomms12762 27694935 PMC5477497 sourceExperimental — accuracy not yet measured
The Results section (‘Intracrystalline diffusion in zeolites’; page not specified) reports that water–zeolite interactions affect water mobility in MFI pores at low hydration, while water–water interactions predominate at higher hydration. It examines water–water hydrogen bonds but does not establish that Brønsted acid sites govern water clustering. (LLM, conf=0.86)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“Two paradigms dominate the MLP field currently: (i) training of universal MLP that cover large parts of the chemical space with dozens of elements, but with limited coverage of the configuration space, e.g., not considering all relevant chemical reactions with the associated transition states, e.g., OC22, CHGNet, and others25–29”
§ A reactive neural network framework for water-loaded acidic zeolites > Introduction
What the sources say
- Partly supportedcites [25] Takamoto S Towards universal neural network potential for material discovery applicable to arbitrary combination of 45 elements Nat. Commun. 2022 13 2991 10.1038/s41467-022-30687-9 35637178 PMC9151783 sourceExperimental — accuracy not yet measured
The Abstract and Methods describe a potential spanning 45 elements and a training dataset with no transition-state structures, supporting the claim about broad chemical coverage and a gap in training coverage. But the Results report that the model predicts transition-state energies and evaluates chemical-reaction pathways, so the paper does not support an unqualified claim that it does not consider reactions or their transition states. The indexed copy provides section labels but no page numbers. (LLM, conf=0.85)
- Partly supportedcites [26] Tran R The open catalyst 2022 (OC22) dataset and challenges for oxide electrocatalysts ACS Catal. 2023 13 3066 3084 10.1021/acscatal.2c05426 sourceExperimental — accuracy not yet measured
Our copy supports the chemical-space half of the characterization: the OC22 slabs are built from 51 elements (Chemical Diversity of Open Catalyst 2022, Fig. 11) and the work targets 'a general purpose potential, rather than separate models for each specific property' (Discussion), i.e. it is a legitimate instance of the universal-MLP paradigm. However, the source nowhere says its configuration-space coverage is limited by omitting relevant reactions or transition states; on the contrary it states the dataset 'spans the configurational complexity for oxide surfaces... varying surface terminations, adsorbate+slab configurations and coverage, and non-stoichiometric substitutions and vacancies' (Introduction) and only concedes that 'OC22 structures may not always be the most stable or pertinent for a particular reaction pathway of interest' (The OC22 Dataset). Searches for 'transition state', reaction-barrier coverage returned no such statement, so that qualifier of the claim is not attributable to this source. (LLM, conf=0.72)
- Partly supportedcites [27] Deng B CHGNet as a pretrained universal neural network potential for charge-informed atomistic modelling Nat. Mach. Intell. 2023 5 1031 1041 10.1038/s42256-023-00716-3 sourceExperimental — accuracy not yet measured
Our copy supports the first half of the attributed characterisation: CHGNet is presented as a pretrained universal MLIP trained on the Materials Project Trajectory (MPtrj) dataset, and the 'Materials Project Trajectory Dataset' section states the data cover '~146,000 inorganic materials composed of 89 elements', 'over 100,000 occurrences for 60 different elements' and 'comprehensive coverage of all chemistries, excluding only the noble gases and actinoids'. However, no passage states the second half — limited coverage of configuration space or the omission of chemical reactions and their transition states; the paper's own limitations discussion concerns magmom/charge inference and charge neutrality ('Discussion'), not reaction-pathway sampling (the closest content is the 'Data parsing' filtering of structures 'higher than 1 eV per atom ... relative to the relaxed structure', which is never framed as a transition-state coverage limitation). So only the chemical-space part of what this citation is offered for is in the source. (LLM, conf=0.72)
- Partly supportedcites [28] Zhang, D. et al. DPA-2: Towards a universal large atomic model for molecular and material simulation. Preprint at 10.48550/arXiv.2312.15492 (2023). sourceExperimental — accuracy not yet measured
The Introduction of DPA-2 supports the general shape of the claim: it describes 'universal'/'fundamental' large atomic models built on extensive DFT-labeled data covering dozens of elements (M3GNet across 89 chemical species, PFP on 45 elements, ALIGNN on 89 elements, CHGNet/MACE-MP-0 on 1.58M MPtrj inorganic crystal frames) and states such single-task models are only directly applicable 'in which the explored configurations are effectively covered by the training data', with DFT-labeled data 'expensive and sparse in the configurational and chemical spaces' and MLPs struggling to generalize beyond their training data. However, the specific exemplification attributed here — that these universal models do not cover relevant chemical reactions and their associated transition states — is not stated in our copy; the limitations it does name are inconsistent DFT labeling settings (e.g. MPtrj vs ANI-1x, restricting molecular generalizability) and GGA-vs-meta-GGA accuracy, and its own multi-task corpus actually includes reactive adsorbate/catalyst data (OC2M AIMD trajectories) and a Transition-1x dataset rather than a claim that reactions/transition states are absent from universal MLP training sets. (LLM, conf=0.72)
- Partly supportedcites [29] Batatia, I. et al. A foundation model for atomistic materials chemistry. Preprint at 10.48550/arXiv.2401.00096 (2024). sourceExperimental — accuracy not yet measured
Our copy supports the chemical-space half of the characterisation: Section 4 'Related work: general purpose MLIPs' lists the general-purpose models and their elemental reach (MEGNet/CHGNet/M3GNet on MP-based data covering most of the 89 elements, DPA-1 and DPA-2 with 56 and 73 elements, PFP with 45 then 72 elements, ALIGNN-FF with 89), and states they were 'created primarily for the purpose of "materials discovery", i.e. predicting thermodynamic stability of hypothetical inorganic crystals', trained on 'snapshots of DFT relaxations of MP structures' (Appendix Fig. S66, p. 141, confirms MPtrj is MP relaxation trajectories). What the source does not state is the specific configuration-space charge as worded — that these universal potentials do not include the relevant chemical reactions and their transition states; the nearest statements are about its own model, namely that Section 2.2 Catalysis shows it generalising 'to out-of-domain catalysis tasks from bulk training configurations' with barriers that are quantitatively wrong without fine-tuning, and that the field 'has yet to be a comprehensive demonstration that a single ML potential can describe solid, liquid, and gaseous systems ... well beyond the distribution of the underlying training set'. So the transition-state clause is inferred from the training-data description rather than asserted. (LLM, conf=0.70)
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
49 more claims · 4 with their cited sources checked
Create a free account to see all 100 claims and to verify the 19 references still open.
- Probably not a substantial claim. 14
- It may state two separate claims as one. 38
Citations
Cites
Generalized Gradient Approximation Made Simple
Perdew, J. P., Burke, K., & Ernzerhof, M. (1996) · Physical Review Letters · doi:10.1103/physrevlett.77.3865 · Cited by 215,824 · cited 2×
“All AIMD simulations and structure optimizations were performed at the computationally less demanding PBE+D3(BJ)[29] level employing the dispersion correction of Grimme et al.”
MethodEfficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set
Kresse, G., & Furthmüller, J. (1996) · Physical Review B · doi:10.1103/physrevb.54.11169 · Cited by 123,652
Projector augmented-wave method
Blöchl, P. E. (1994) · Physical Review B · doi:10.1103/physrevb.50.17953 · Cited by 92,732
From ultrasoft pseudopotentials to the projector augmented-wave method
Kresse, G., & Joubert, D. (1999) · Physical Review B · doi:10.1103/physrevb.59.1758 · Cited by 85,307 · cited 1×
“4) along with standard versions of the Projector Augmented-Wave (PAW) potentials.[79, 80] Calculations at constant volume were performed with an energy cutoff of 400 eV.”
BackgroundKresse, G., & Furthmüller, J. (1996) · Computational Materials Science · doi:10.1016/0927-0256(96)00008-0 · Cited by 75,638
Grimme, S., Antony, J., Ehrlich, S., & Krieg, H. (2010) · The Journal of Chemical Physics · doi:10.1063/1.3382344 · Cited by 59,686
Ab initio molecular dynamics for liquid metals
Kresse, G., & Hafner, J. (1993) · Physical Review B · doi:10.1103/physrevb.47.558 · Cited by 45,727
Effect of the damping function in dispersion corrected density functional theory
Grimme, S., Ehrlich, S., & Goerigk, L. (2011) · Journal of Computational Chemistry · doi:10.1002/jcc.21759 · Cited by 28,221 · cited 2×
“The bulk of the data-base is derived from 500 short (10 ps) ab initio molecular dynamics (AIMD) trajectories (at PBE +D3(BJ) level) 35 – 37 using a set of BAS zeolite models.”
MethodCanonical dynamics: Equilibrium phase-space distributions
Hoover, W. G. (1985) · Physical Review A · doi:10.1103/physreva.31.1695 · Cited by 23,465 · cited 2×
“The AIMD equilibration used the canonical (NVT) ensemble along with a 1 fs time step, the Nosé–Hoover thermostat[70, 71] and with hydrogen being replaced by tritium.”
MethodKresse, G., & Hafner, J. (1994) · Physical Review B · doi:10.1103/physrevb.49.14251 · Cited by 23,019
A unified formulation of the constant temperature molecular dynamics methods
Nosé, S. (1984) · The Journal of Chemical Physics · doi:10.1063/1.447334 · Cited by 19,152 · cited 2×
“The AIMD equilibration used the canonical (NVT) ensemble along with a 1 fs time step, the Nosé–Hoover thermostat[70, 71] and with hydrogen being replaced by tritium.”
MethodThe atomic simulation environment—a Python library for working with atoms
Hjorth Larsen, A., Jørgen Mortensen, J., Blomqvist, J., Castelli, I. E., Christensen, R., Dułak, M., Friis, J., Groves, M. N., Hammer, B., Hargus, C., et al. (2017) · Journal of Physics: Condensed Matter · doi:10.1088/1361-648x/aa680e · Cited by 4,464 · cited 1×
“0) and the atomic simulation environment (ASE).[83]”
Strongly Constrained and Appropriately Normed Semilocal Density Functional
Sun, J., Ruzsinszky, A., & Perdew, J. P. (2015) · Physical Review Letters · doi:10.1103/physrevlett.115.036402 · Cited by 3,596 · cited 2×
“These structures were used for SCAN+D3(BJ)[35] single-point (SP) calculations to create the bulk of the training database.”
MethodPLUMED 2: New feathers for an old bird
Tribello, G. A., Bonomi, M., Branduardi, D., Camilloni, C., & Bussi, G. (2014) · Computer Physics Communications · doi:10.1016/j.cpc.2013.09.018 · Cited by 3,479 · cited 1×
“For the biased dynamics, we used well-tempered metadynamics from the PLUMED package 75,99 .”
MethodA Theory of Water and Ionic Solution, with Particular Reference to Hydrogen and Hydroxyl Ions
Bernal, J. D., & Fowler, R. H. (1933) · The Journal of Chemical Physics · doi:10.1063/1.1749327 · Cited by 3,255 · cited 1×
“Finally, the unit cells of two ice polymorphs (Ice II 89 and Ice I h 90 ) were deformed in the same way for sampling of low-energy structures of crystalline water.”
MethodWell-Tempered Metadynamics: A Smoothly Converging and Tunable Free-Energy Method
Barducci, A., Bussi, G., & Parrinello, M. (2008) · Physical Review Letters · doi:10.1103/physrevlett.100.020603 · Cited by 3,184
On representing chemical environments
Bartók, A. P., Kondor, R., & Csányi, G. (2013) · Physical Review B · doi:10.1103/physrevb.87.184115 · Cited by 2,572
The ReaxFF reactive force-field: development, applications and future directions
Senftle, T. P., Hong, S., Islam, M. M., Kylasa, S. B., Zheng, Y., Shin, Y. K., Junkermeier, C., Engel-Herbert, R., Janik, M. J., Aktulga, H. M., et al. (2016) · npj Computational Materials · doi:10.1038/npjcompumats.2015.11 · Cited by 2,363 · cited 7×
“A standard tool for accelerating the reactive sampling in zeolite-water systems, and thus reaching experimentally relevant timescale or realistic levels of model complexity is the class of reactive analytical potentials, for example, ReaxFF 19 .”
MethodInfluentialMardirossian, N., & Head-Gordon, M. (2017) · Molecular Physics · doi:10.1080/00268976.2017.1333644 · Cited by 2,307 · cited 2×
“…we chose the range-separated hybrid DFT functional ω B97X complemented with the empirical dispersion correction D3(BJ): a functional that shows considerably better performance for water cluster binding energies and reaction barriers 74,77,78 than our baseline reference SCAN+D3(BJ) functional.”
BackgroundMethodE(3)-equivariant graph neural networks for data-efficient and accurate interatomic potentials
Batzner, S., Musaelian, A., Sun, L., Geiger, M., Mailoa, J. P., Kornbluth, M., Molinari, N., Smidt, T. E., & Kozinsky, B. (2022) · Nature Communications · doi:10.1038/s41467-022-29939-5 · Cited by 2,285
The General Utility Lattice Program ( GULP )
Gale, J. D., & Rohl, A. L. (2003) · Molecular Simulation · doi:10.1080/0892702031000104887 · Cited by 2,265 · cited 1×
“Single-point calculations using ReaxFF[41] were performed with GULP.[81] NNP training and NNP level simulations used the Python packages SchNetPack[82] (version 1.”
MethodSchNet – A deep learning architecture for molecules and materials
Schütt, K. T., Sauceda, H. E., Kindermans, P.-J., Tkatchenko, A., & Müller, K.-R. (2018) · The Journal of Chemical Physics · doi:10.1063/1.5019779 · Cited by 2,217
Unke, O. T., Chmiela, S., Sauceda, H. E., Gastegger, M., Poltavsky, I., Schütt, K. T., Tkatchenko, A., & Müller, K.-R. (2021) · Chemical Reviews · doi:10.1021/acs.chemrev.0c01111 · Cited by 1,647 · cited 2×
“An emerging alternative to analytical force fi elds is represented by machine learning potentials (MLPs), which interpolate the potential energy surface (PES) at the level of an ab initio training set 22 – 24 .”
BackgroundThe hydrothermal synthesis of zeolites: Precursors, intermediates and reaction mechanism
Cundy, C. S., & Cox, P. A. (2005) · Microporous and Mesoporous Materials · doi:10.1016/j.micromeso.2005.02.016 · Cited by 1,590 · cited 1×
“…and thus acidity 7,8 , (ii) hydrolytic bond dissociation and defect formation, which controls catalyst durability and activity 9 , (iii) water mobility and clustering in zeolite pores 10 , and (iv) the synthesis of zeolites from precursor gels containing silica fragments, water and cations 11 .”
BackgroundFluid catalytic cracking: recent developments on the grand old lady of zeolite catalysis
Vogt, E. T. C., & Weckhuysen, B. M. (2015) · Chemical Society Reviews · doi:10.1039/c5cs00376h · Cited by 1,075 · cited 2×
“In this work, we developed a neural network potential (NNP) for the entire class of protonic aluminosilicate (H-AS) zeolites, which are one of the cornerstones of existing petrochemical processes, [3] as well as one of the main candidates for emerging applications of zeolites in sustainable chemistry.”
BackgroundMardirossian, N., & Head-Gordon, M. (2014) · Physical Chemistry Chemical Physics · doi:10.1039/c3cp54374a · Cited by 986 · cited 1×
“…we chose the range-separated hybrid DFT functional ω B97X complemented with the empirical dispersion correction D3(BJ): a functional that shows considerably better performance for water cluster binding energies and reaction barriers 74,77,78 than our baseline reference SCAN+D3(BJ) functional.”
MethodCHGNet as a pretrained universal neural network potential for charge-informed atomistic modelling
Deng, B., Zhong, P., Jun, K., Riebesell, J., Han, K., Bartel, C. J., & Ceder, G. (2023) · Nature Machine Intelligence · doi:10.1038/s42256-023-00716-3 · Cited by 944 · cited 2×
“Also, commonly,theuniversalMLPsaretrainedondatasetscomprisingonly close-to-equilibrium structures 25,27 – 29 , hence caution is needed when applying them to activated events.”
BackgroundBig Data Meets Quantum Chemistry Approximations: The Δ-Machine Learning Approach
Ramakrishnan, R., Dral, P. O., Rupp, M., & von Lilienfeld, O. A. (2015) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.5b00099 · Cited by 911 · cited 4×
“…that the learned representations of the NNP baseline models can be used for data-ef fi cient learning of higher (hybrid) DFT level corrections ( Δ -learning) 33 for speci fi c use cases, in addition to developing machine-learned collective variables for the acceleration of rare event sampling 34 .”
MethodInfluentialApplications of Zeolites in Sustainable Chemistry
Li, Y., Li, L., & Yu, J. (2017) · Chem · doi:10.1016/j.chempr.2017.10.009 · Cited by 880 · cited 2×
“In this work, we developed a neural network potential (NNP) for the entire class of protonic aluminosilicate (H-AS) zeolites, which are one of the cornerstones of existing petrochemical processes, [3] as well as one of the main candidates for emerging applications of zeolites in sustainable chemistry.[2] Our NNP provi…”
BackgroundCombining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems
Keith, J. A., Vassilev-Galindo, V., Cheng, B., Chmiela, S., Gastegger, M., Müller, K.-R., & Tkatchenko, A. (2021) · Chemical Reviews · doi:10.1021/acs.chemrev.1c00107 · Cited by 849
Structures and transformation mechanisms of the η, γ and θ transition aluminas
Zhou, R.-S., & Snyder, R. L. (1991) · Acta Crystallographica Section B Structural Science · doi:10.1107/s0108768191002719 · Cited by 811 · cited 1×
“…same set of 210 lattice deformations for six dense BAS zeolite polymorphs, namely, four alumina polymorphs α -Al 2 O 3 (corundum) 82 , θ -Al 2 O 3 83 , γ -AlO(OH) (Boehmite) 84 , and α -Al(OH) 3 (Gibbsite) 85 , as well as two aluminosilicate polymorphs, Si 3 Al 2 O 12 H 3 (H 3 O-Natrolite) 86…”
BackgroundPotential and challenges of zeolite chemistry in the catalytic conversion of biomass
Ennaert, T., Van Aelst, J., Dijkmans, J., De Clercq, R., Schutyser, W., Dusselier, M., Verboekend, D., & Sels, B. F. (2016) · Chemical Society Reviews · doi:10.1039/c5cs00859j · Cited by 806 · cited 2×
“Recently, great experimentaland theoretical efforts have been madeto go beyond the traditional applications of zeolites, for example in converting sustainable bio-feedstocks into chemicals 4 – 6 .”
BackgroundMethodThe farthest point strategy for progressive image sampling
Eldar, Y., Lindenbaum, M., Porat, M., & Zeevi, Y. Y. (1997) · IEEE Transactions on Image Processing · doi:10.1109/83.623193 · Cited by 693
Jinnouchi, R., Lahnsteiner, J., Karsai, F., Kresse, G., & Bokdam, M. (2019) · Physical Review Letters · doi:10.1103/physrevlett.122.225701 · Cited by 635 · cited 2×
“…, and (ii) active-learning procedures to develop system-speci fi c MLPs to accelerate simulations for a speci fi c model or thermodynamic state point 30 – 32 , which capture the details of the PES, including transition states, but have little or no transferability to systems with different chemical…”
BackgroundMethodA foundation model for atomistic materials chemistry
Batatia, I., Benner, P., Chiang, Y., Elena, A. M., Kovács, D. P., Riebesell, J., Advincula, X. R., Asta, M., Avaylon, M., Baldwin, W. J., et al. (2024) · arXiv · doi:10.48550/arxiv.2401.00096 · Cited by 604 · cited 2×
“Also, commonly,theuniversalMLPsaretrainedondatasetscomprisingonly close-to-equilibrium structures 25,27 – 29 , hence caution is needed when applying them to activated events.”
BackgroundSorption thermal storage for solar energy
Yu, N., Wang, R. Z., & Wang, L. W. (2013) · Progress in Energy and Combustion Science · doi:10.1016/j.pecs.2013.05.004 · Cited by 589 · cited 1×
“The dynamic behaviour of zeolite-confined water containing solvated protons is of high interest due to (potential) application of zeolites in water purification [10, 43], heat storage[44] or reaction optimization under humid conditions (such as biomass conversion).”
BackgroundA foundation model for atomistic materials chemistry
Batatia, I., Benner, P., Chiang, Y., Elena, A. M., Kovács, D. P., Riebesell, J., Advincula, X. R., Asta, M., Avaylon, M., Baldwin, W. J., et al. (2025) · The Journal of Chemical Physics · doi:10.1063/5.0297006 · Cited by 566
Advances in theory and their application within the field of zeolite chemistry
Van Speybroeck, V., Hemelsoet, K., Joos, L., Waroquier, M., Bell, R. G., & Catlow, C. R. A. (2015) · Chemical Society Reviews · doi:10.1039/c5cs00029g · Cited by 516 · cited 1×
“Such systems are relevant for modeling potential precursors of zeolite synthesis or products of zeolite degradation in hot liquid water (de-silication and de-alumination) 5,9,63,64 .”
BackgroundNew trends in tailoring active sites in zeolite-based catalysts
Shamzhy, M., Opanasenko, M., Concepción, P., & Martínez, A. (2019) · Chemical Society Reviews · doi:10.1039/c8cs00887f · Cited by 508 · cited 1×
“[4, 5, 6] A further critical consideration for both existing and emerging applications is the interaction between H-AS and water.”
BackgroundHautier, G., Ong, S. P., Jain, A., Moore, C. J., & Ceder, G. (2012) · Physical Review B · doi:10.1103/physrevb.85.155208 · Cited by 506 · cited 2×
“To put these NNP errors in context, note that standard GGA-level DFT functionals show, for 27 formation reactions that involve silica, an RMSE of about 28 meV atom − 1 with respect to experiment 57 .”
MethodOn-the-fly active learning of interpretable Bayesian force fields for atomistic rare events
Vandermause, J., Torrisi, S. B., Batzner, S., Xie, Y., Sun, L., Kolpak, A. M., & Kozinsky, B. (2020) · npj Computational Materials · doi:10.1038/s41524-020-0283-z · Cited by 459 · cited 1×
“…structures whichismuch larger compared to datasets used typically to train MLPs for speci fi c systems/reactions (e.g., those with fi xed chemical composition) 30,31 yet much smaller than databases (e.g., OC22) 26 that aim to span the whole periodic table containing millions of DFT data points.”
MethodTakamoto, S., Shinagawa, C., Motoki, D., Nakago, K., Li, W., Kurata, I., Watanabe, T., Yayama, Y., Iriguchi, H., Asano, Y., et al. (2022) · Nature Communications · doi:10.1038/s41467-022-30687-9 · Cited by 446 · cited 2×
“…con fi guration space, e.g., not considering all relevant chemical reactions with the associated transition states, e.g., OC22, CHGNet, and others 25 – 29 , and (ii) active-learning procedures to develop system-speci fi c MLPs to accelerate simulations for a speci fi c model or thermodynamic state…”
BackgroundSchNetPack: A Deep Learning Toolbox For Atomistic Systems
Schütt, K. T., Kessel, P., Gastegger, M., Nicoli, K. A., Tkatchenko, A., & Müller, K.-R. (2019) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.8b00908 · Cited by 433
Bryantsev, V. S., Diallo, M. S., van Duin, A. C. T., & Goddard, W. A. (2009) · Journal of Chemical Theory and Computation · doi:10.1021/ct800549f · Cited by 403 · cited 1×
“org) and four isomers of (H2O)20.[67] All clusters were first optimized (constant volume conditions) with a unit cell ensuring a distance between equivalent periodic images of at least 1 nm.”
MethodNanosized microporous crystals: emerging applications
Mintova, S., Jaber, M., & Valtchev, V. (2015) · Chemical Society Reviews · doi:10.1039/c5cs00210a · Cited by 385 · cited 1×
“This makes zeolites a versatile material class with applications ranging from thermal energy storage to gas separation and water purification, but predominantly in heterogeneous catalysis.[1, 2] The presence of aluminium, and in particular the necessary charge compensation add another layer of complexity to the struct…”
BackgroundBenchmark Structures and Binding Energies of Small Water Clusters with Anharmonicity Corrections
Temelso, B., Archer, K. A., & Shields, G. C. (2011) · The Journal of Physical Chemistry A · doi:10.1021/jp2069489 · Cited by 334 · cited 2×
“Inaddition,weusedsinglewaterandwaterclusters in vacuo taken from the BEGDB database 50 .”
MethodImbalzano, G., Anelli, A., Giofré, D., Klees, S., Behler, J., & Ceriotti, M. (2018) · The Journal of Chemical Physics · doi:10.1063/1.5024611 · Cited by 330 · cited 1×
“The seven hypothetical zeolite frameworks were selected from a siliceous zeolite database 38 by Farthest Point Sampling (FPS) 39,40 using the smooth overlap of atomicpositions (SOAP) 41 kernel as metric to obtain con fi gurations with structurally distinct atomic environments (see “ Methods ”…”
MethodThe Open Catalyst 2022 (OC22) Dataset and Challenges for Oxide Electrocatalysts
Tran, R., Lan, J., Shuaibi, M., Wood, B. M., Goyal, S., Das, A., Heras-Domingo, J., Kolluru, A., Rizvi, A., Shoghi, N., et al. (2023) · ACS Catalysis · doi:10.1021/acscatal.2c05426 · Cited by 329 · cited 3×
“Theresultingdatabase contains248,439 structures whichismuch larger compared to datasets used typically to train MLPs for speci fi c systems/reactions (e.g., those with fi xed chemical composition) 30,31 yet much smaller than databases (e.g., OC22) 26 that aim to span the whole periodic table containing millions of DFT d…”
MethodNajibi, A., & Goerigk, L. (2018) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.8b00842 · Cited by 320
DPA-2: a large atomic model as a multi-task learner
Zhang, D., Liu, X., Zhang, X., Zhang, C., Cai, C., Bi, H., Du, Y., Qin, X., Peng, A., Huang, J., et al. (2024) · npj Computational Materials · doi:10.1038/s41524-024-01493-2 · Cited by 303
Accurate structure analysis with synchrotron radiation. The electron density in Al2O3 and Cu2O
Kirfel, A., & Eichhorn, K. (1990) · Acta Crystallographica Section A Foundations of Crystallography · doi:10.1107/s0108767389012596 · Cited by 286 · cited 1×
“Apart from microporous structures and two-dimensional H-AS, we also added the same set of 210 lattice deformations for six dense H-AS polymorphs, namely, four alumina polymorphs α-Al2O3 (corundum),[60]”
BackgroundThe central role of density functional theory in the AI age
Huang, B., von Rudorff, G. F., & von Lilienfeld, O. A. (2023) · Science · doi:10.1126/science.abn3445 · Cited by 237 · cited 1×
“First, we considered a common application domain/target of the Δ -learning approach, in which one trains and deploys the correction ( Δ NNP) model on the same reaction/process (the in-domain case) 79 , i.e., the proton jump in CHA.”
MethodDealumination mechanisms of zeolites and extra-framework aluminum confinement
Silaghi, M.-C., Chizallet, C., Sauer, J., & Raybaud, P. (2016) · Journal of Catalysis · doi:10.1016/j.jcat.2016.04.021 · Cited by 236 · cited 3×
“O(H) bonds 14,66,73 (see Supplementary Fig.”
BackgroundMethodMechanism of the Acid-Catalyzed Si−O Bond Cleavage in Siloxanes and Siloxanols. A Theoretical Study
Cypryk, M., & Apeloig, Y. (2002) · Organometallics · doi:10.1021/om011055s · Cited by 219 · cited 1×
“2b),which is a standard precursor for Si-O bond hydrolysis reported in multiple publications before 15,66 – 68 .”
BackgroundThermochemistry of Pure-Silica Zeolites
Piccione, P. M., Laberty, C., Yang, S., Camblor, M. A., Navrotsky, A., & Davis, M. E. (2000) · The Journal of Physical Chemistry B · doi:10.1021/jp002148a · Cited by 214 · cited 1×
“…silica structures (three polymorphs: α -quartz, α -cristobalite, tridymite; four zeolites: AFI, FER, IFR, MTW) at the NNP level (including the unit cell) for which experimental data (structures and enthalpies) 96 – 98 and DFT data (PBE+D3(BJ), SCAN+D3(BJ)) 38 are available (see Supplementary Fig.”
MethodFactors that Determine Zeolite Stability in Hot Liquid Water
Zhang, L., Chen, K., Chen, B., White, J. L., & Resasco, D. E. (2015) · Journal of the American Chemical Society · doi:10.1021/jacs.5b07398 · Cited by 208 · cited 1×
“These fi ndings arealsoin line with previous experimental studies on the hydrolysis of MFI zeolites in hot liquid waterthatsuggestthatzeolitedegradationpredominantly startsatthe externalsurface, in which water autoprotolysisand silanol groups play a crucialrole 69,70 .”
ResultConfinement effects and acid strength in zeolites
Grifoni, E., Piccini, G., Lercher, J. A., Glezakou, V.-A., Rousseau, R., & Parrinello, M. (2021) · Nature Communications · doi:10.1038/s41467-021-22936-0 · Cited by 203 · cited 2×
“Importantly, both the degree of deprotonation and the distance of the proton from the BAS as a function of water loading are in very good agreement with the biased AIMD simulations of Grifoni et al. for protonic FAU zeolite 17 .”
ResultMachine-learned interatomic potentials by active learning: amorphous and liquid hafnium dioxide
Sivaraman, G., Krishnamoorthy, A. N., Baur, M., Holm, C., Stan, M., Csányi, G., Benmore, C., & Vázquez-Mayagoitia, Á. (2020) · npj Computational Materials · doi:10.1038/s41524-020-00367-7 · Cited by 194
Role of the ionic environment in enhancing the activity of reacting molecules in zeolite pores
Pfriem, N., Hintermeier, P. H., Eckstein, S., Kim, S., Liu, Q., Shi, H., Milakovic, L., Liu, Y., Haller, G. L., Baráth, E., et al. (2021) · Science · doi:10.1126/science.abh3418 · Cited by 164 · cited 2×
“Thedynamicbehaviorofzeolite-con fi ned water-containing solvated protons is of high interest due to the (potential) applications of zeolites in water puri fi cation 10,59 , heat storage 60 or reaction optimization under humid conditions (such as biomass conversion) 7 .”
BackgroundIce. II. A proton-ordered form of ice
Kamb, B. (1964) · Acta Crystallographica · doi:10.1107/s0365110x64003553 · Cited by 162 · cited 1×
“Finally, the unit cells of two ice polymorphs (Ice II 89 and Ice I h 90 ) were deformed in the same way for sampling of low-energy structures of crystalline water.”
MethodSilicon Liquid Structure and Crystal Nucleation from Ab Initio Deep Metadynamics
Bonati, L., & Parrinello, M. (2018) · Physical Review Letters · doi:10.1103/physrevlett.121.265701 · Cited by 159 · cited 2×
“Next, we conducted NNP performance tests for modeling of reaction pathways in FAU (primitive unit cell, Si/Al=47) using NNP level climbing image NEB calculations.”
MethodMachine-learned interatomic potentials by active learning: amorphous and liquid hafnium dioxide
Sivaraman, G., Krishnamoorthy, A. N., Baur, M., Holm, C., Stan, M., Cśanyi, G., Benmore, C. J., & Vázquez‐Mayagoitia, Á. (2019) · Cambridge University Engineering Department Publications Database · doi:10.17863/cam.55408 · Cited by 158
Zeolite (In)Stability under Aqueous or Steaming Conditions
Heard, C. J., Grajciar, L., Uhlík, F., Shamzhy, M., Opanasenko, M., Čejka, J., & Nachtigall, P. (2020) · Advanced Materials · doi:10.1002/adma.202003264 · Cited by 157 · cited 2×
“Such systems are relevant for modeling potential precursors of zeolite synthesis or products of zeolite degradation in hot liquid water (de-silication and de-alumination) 5,9,63,64 .”
BackgroundSelf-citationDPA-2: a large atomic model as a multi-task learner
Zhang, D., Liu, X., Zhang, X., Zhang, C., Cai, C., Bi, H., Du, Y., Qin, X., Peng, A., Huang, J., et al. (2023) · arXiv · doi:10.48550/arxiv.2312.15492 · Cited by 155
First-principles study of the OH-stretching modes of gibbsite
Balan, E., Lazzeri, M., Morin, G., & Mauri, F. (2006) · American Mineralogist · doi:10.2138/am.2006.1922 · Cited by 152 · cited 1×
“…for six dense BAS zeolite polymorphs, namely, four alumina polymorphs α -Al 2 O 3 (corundum) 82 , θ -Al 2 O 3 83 , γ -AlO(OH) (Boehmite) 84 , and α -Al(OH) 3 (Gibbsite) 85 , as well as two aluminosilicate polymorphs, Si 3 Al 2 O 12 H 3 (H 3 O-Natrolite) 86 and Al 2 Si 2 O 5 (OH) 4 (Dickite) 87 .”
BackgroundSierka, M., & Sauer, J. (2001) · The Journal of Physical Chemistry B · doi:10.1021/jp004081x · Cited by 149
Ryder, J. A., Chakraborty, A. K., & Bell, A. T. (2000) · The Journal of Physical Chemistry B · doi:10.1021/jp9943427 · Cited by 149
Vandermause, J., Xie, Y., Lim, J. S., Owen, C. J., & Kozinsky, B. (2022) · Nature Communications · doi:10.1038/s41467-022-32294-0 · Cited by 147 · cited 1×
“However, due to their fi xed functional form, these potentials have limited transferability to systems with different chemical composition 20 .”
BackgroundInteraction of water with zeolites: a review
Resasco, D. E., Crossley, S. P., Wang, B., & White, J. L. (2021) · Catalysis Reviews · doi:10.1080/01614940.2021.1948301 · Cited by 140 · cited 1×
“Such systems are relevant for modeling potential precursors of zeolite synthesis or products of zeolite degradation in hot liquid water (de-silication and de-alumination) 5,9,63,64 .”
BackgroundToward High-Throughput Zeolite Membranes
Tsapatsis, M. (2011) · Science · doi:10.1126/science.1205957 · Cited by 131 · cited 1×
“Thedynamicbehaviorofzeolite-con fi ned water-containing solvated protons is of high interest due to the (potential) applications of zeolites in water puri fi cation 10,59 , heat storage 60 or reaction optimization under humid conditions (such as biomass conversion) 7 .”
BackgroundInterplay between hydrophilicity and surface barriers on water transport in zeolite membranes
Fasano, M., Humplik, T., Bevilacqua, A., Tsapatsis, M., Chiavazzo, E., Wang, E. N., & Asinari, P. (2016) · Nature Communications · doi:10.1038/ncomms12762 · Cited by 123 · cited 2×
“…and thus acidity 7,8 , (ii) hydrolytic bond dissociation and defect formation, which controls catalyst durability and activity 9 , (iii) water mobility and clustering in zeolite pores 10 , and (iv) the synthesis of zeolites from precursor gels containing silica fragments, water and cations 11 .”
BackgroundBaerlocher, C., Meier, W. M., & OLSON, D. H. L. (2001) · Elsevier eBooks · doi:10.1016/b978-044450701-3/50351-7 · Cited by 122
Fast room temperature lability of aluminosilicate zeolites
Heard, C. J., Grajciar, L., Rice, C. M., Pugh, S. M., Nachtigall, P., Ashbrook, S. E., & Morris, R. E. (2019) · Nature Communications · doi:10.1038/s41467-019-12752-y · Cited by 121 · cited 8×
“In our previous work 15 , a fairly similar Si-O bond-breaking mechanisminvolvingprotontransferwasfoundtobecharacterizedby a free energy ofactivation amounting to approx.”
BackgroundMethodInfluentialSelf-citationZhang, H., Xiao, Q., Guo, X., Li, N., Kumar, P., Rangnekar, N. D., Jeon, M. Y., Al-Thabaiti, S. A., Narasimharao, K., Basahel, S. N., et al. (2016) · Angewandte Chemie International Edition · doi:10.1002/anie.201601135 · Cited by 120 · cited 2×
“This zeolite model resembles 2D-MFI nanosh-eets which have been successfully prepared by exfoliation of a multi-lamellar MFI zeolite 65 .”
MethodBates, J. S., Bukowski, B. C., Greeley, J., & Gounder, R. (2020) · Chemical Science · doi:10.1039/d0sc02589e · Cited by 117 · cited 1×
“This relationship governs many features of H-AS including i) proton solvation, and thus acidity,[7, 8] ii) hydrolytic bond dissociation and defect formation, which controls catalyst durability and activity,[9] iii) water mobility and clustering in zeolite pores[10] and iv) the synthesis of zeolites from precursor gels…”
BackgroundSilaghi, M.-C., Chizallet, C., Petracovschi, E., Kerber, T., Sauer, J., & Raybaud, P. (2015) · ACS Catalysis · doi:10.1021/cs501474u · Cited by 108
Joshi, K. L., Psofogiannakis, G., van Duin, A. C. T., & Raman, S. (2014) · Phys. Chem. Chem. Phys. · doi:10.1039/c4cp02612h · Cited by 97 · cited 5×
“Table 1 and Figure 1c also show the performance of the reactive analytical force field ReaxFF specialized for water-loaded H-AS systems.[41] Table 1.”
BackgroundMethodResultInfluentialA machine-learned interatomic potential for silica and its relation to empirical models
Erhard, L. C., Rohrer, J., Albe, K., & Deringer, V. L. (2022) · npj Computational Materials · doi:10.1038/s41524-022-00768-w · Cited by 95
Cooperative Role of Water Molecules during the Initial Stage of Water-Induced Zeolite Dealumination
Stanciakova, K., Ensing, B., Göltl, F., Bulo, R. E., & Weckhuysen, B. M. (2019) · ACS Catalysis · doi:10.1021/acscatal.9b00307 · Cited by 91 · cited 2×
“O(H) bonds 14,66,73 (see Supplementary Fig.”
MethodDeuteration of Crystalline Hydroxides. Hydrogen Bonds of gamma-AlOO(H,D) and gamma-FeOO(H,D).
Christensen, A. N., Lehmann, M. S., Convert, P., et al. (1982) · Acta chemica Scandinavica/Acta chemica Scandinavica. B, Organic chemistry and biochemistry/Acta chemica Scandinavica. A, Physical and inorganic chemistry/Acta chemica Scandinavica. Series B. Organic chemistry and biochemistry/Acta chemica Scandinavica. Series A, Physical and inorganic chemistry · doi:10.3891/acta.chem.scand.36a-0303 · Cited by 86 · cited 1×
“…for six dense BAS zeolite polymorphs, namely, four alumina polymorphs α -Al 2 O 3 (corundum) 82 , θ -Al 2 O 3 83 , γ -AlO(OH) (Boehmite) 84 , and α -Al(OH) 3 (Gibbsite) 85 , as well as two aluminosilicate polymorphs, Si 3 Al 2 O 12 H 3 (H 3 O-Natrolite) 86 and Al 2 Si 2 O 5 (OH) 4…”
BackgroundReaxFF Parameter Optimization with Monte-Carlo and Evolutionary Algorithms: Guidelines and Insights
Shchygol, G., Yakovlev, A., Trnka, T., van Duin, A. C. T., & Verstraelen, T. (2019) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.9b00769 · Cited by 85
Collective action of water molecules in zeolite dealumination
Nielsen, M., Hafreager, A., Brogaard, R. Y., De Wispelaere, K., Falsig, H., Beato, P., Van Speybroeck, V., & Svelle, S. (2019) · Catalysis Science & Technology · doi:10.1039/c9cy00624a · Cited by 76 · cited 2×
“[15, 16, 17, 18] These investigations demonstrate the importance of capturing dynamics under operating conditions, being able to discover unexpected reaction mechanisms and defective species that hitherto eluded structural identification, but are not sufficiently economical for a global exploration of structural and r…”
BackgroundBocus, M., Goeminne, R., Lamaire, A., Cools-Ceuppens, M., Verstraelen, T., & Van Speybroeck, V. (2023) · Nature Communications · doi:10.1038/s41467-023-36666-y · Cited by 67 · cited 6×
“…one may train a correction model on top of the baseline model, using a computationally more demanding but moreaccurate leveloftheoryfor a small subset of data points that typically covers only the speci fi c process/reaction of interest (e.g., a proton jump in a speci fi c zeolite framework) 48 .”
MethodResultInfluentialTreps, L., Gomez, A., de Bruin, T., & Chizallet, C. (2020) · ACS Catalysis · doi:10.1021/acscatal.9b05103 · Cited by 63
Pressure-controlled polytypism in hydrous layered materials
Dera, P., Prewitt, C. T., Japel, S., et al. (2003) · American Mineralogist · doi:10.2138/am-2003-1006 · Cited by 62 · cited 1×
“…for six dense BAS zeolite polymorphs, namely, four alumina polymorphs α -Al 2 O 3 (corundum) 82 , θ -Al 2 O 3 83 , γ -AlO(OH) (Boehmite) 84 , and α -Al(OH) 3 (Gibbsite) 85 , as well as two aluminosilicate polymorphs, Si 3 Al 2 O 12 H 3 (H 3 O-Natrolite) 86 and Al 2 Si 2 O 5 (OH) 4 (Dickite) 87 .”
BackgroundDefect‐Mediated Ordering of Condensed Water Structures in Microporous Zeolites
Bukowski, B. C., Bates, J. S., Gounder, R., & Greeley, J. (2019) · Angewandte Chemie International Edition · doi:10.1002/anie.201908151 · Cited by 59 · cited 1×
“…this interaction is not adequately viewed as a simple bulk-liquid interface, but rather a collection of complex binding, clustering, exchange, and reaction steps between variously sized water clusters and an inhomogeneous surface that is complicated by topology-dependent con fi nement effects 12 .”
BackgroundAccurate large-scale simulations of siliceous zeolites by neural network potentials
Erlebach, A., Nachtigall, P., & Grajciar, L. (2022) · npj Computational Materials · doi:10.1038/s41524-022-00865-w · Cited by 58
Self-citationMachine learning potential era of zeolite simulation
Ma, S., & Liu, Z.-P. (2022) · Chemical Science · doi:10.1039/d2sc01225a · Cited by 57 · cited 1×
“An emerging alternative to analytical force fi elds is represented by machine learning potentials (MLPs), which interpolate the potential energy surface (PES) at the level of an ab initio training set 22 – 24 .”
BackgroundKubicki, J. D., Xiao, Y., & Lasaga, A. C. (1993) · Geochimica et Cosmochimica Acta · doi:10.1016/0016-7037(93)90338-w · Cited by 55
The effect of water on the validity of Löwenstein's rule
Heard, C. J., Grajciar, L., & Nachtigall, P. (2019) · Chemical Science · doi:10.1039/c9sc00725c · Cited by 48 · cited 4×
“Wealsosubsampled(usingSOAP-FPS)AIMDtrajectoriesofzeolite CHA taken from our previous works on non-Löwenstein pairs (Al-O-Al) with various water loadings (0, 1, 15 water molecules) 45 and biased AIMD runs of Si-O(H) and Al-O(H) bond cleavage mechanisms 15 .”
BackgroundMethodResultInfluentialSelf-citationCombining phonon accuracy with high transferability in Gaussian approximation potential models
George, J., Hautier, G., Bartók, A. P., Csányi, G., & Deringer, V. L. (2020) · The Journal of Chemical Physics · doi:10.1063/5.0013826 · Cited by 44 · cited 16×
“These errors are similar to other reactive (rotationally invariant) MLPs 42,44,52 and about thesameasfor ourpreviously developedsilica NNPs 38 .”
BackgroundMethodInfluentialZSM-5 decrystallization and dealumination in hot liquid water
Maag, A. R., Tompsett, G. A., Tam, J., Ang, C. A., Azimi, G., Carl, A. D., Huang, X., Smith, L. J., Grimm, R. L., Bond, J. Q., et al. (2019) · Physical Chemistry Chemical Physics · doi:10.1039/c9cp01490j · Cited by 40 · cited 1×
“These fi ndings arealsoin line with previous experimental studies on the hydrolysis of MFI zeolites in hot liquid waterthatsuggestthatzeolitedegradationpredominantly startsatthe externalsurface, in which water autoprotolysisand silanol groups play a crucialrole 69,70 .”
ResultCombining phonon accuracy with high transferability in Gaussian approximation potential models.
George, J., Hautier, G., Bartók, A. P., et al. (2020) · Cambridge University Engineering Department Publications Database · doi:10.17863/cam.62481 · Cited by 37
Predicting Structural Properties of Pure Silica Zeolites Using Deep Neural Network Potentials
Sours, T. G., & Kulkarni, A. R. (2023) · The Journal of Physical Chemistry C · doi:10.1021/acs.jpcc.2c08429 · Cited by 36 · cited 1×
“These errors are similar to other reactive (rotationally invariant) MLPs 42,44,52 and about thesameasfor ourpreviously developedsilica NNPs 38 .”
BackgroundUnderstanding Water–Zeolite Interactions: On the Accuracy of Density Functionals
Stanciakova, K., Louwen, J. N., Weckhuysen, B. M., Bulo, R. E., & Göltl, F. (2021) · The Journal of Physical Chemistry C · doi:10.1021/acs.jpcc.1c04270 · Cited by 35 · cited 1×
“…an average over the MD trajectory) - this strong stabilization is in qualitative agreement with the static single water adsorption energies in CHA zeolite reported in the literature ranging from − 50 to almost − 100 kJmol − 1 depending on the type of exchange-correlation DFT functional adopted 61 .”
ResultMechanism of Zeolite Hydrolysis under Basic Conditions
Jin, M., Liu, M., Nachtigall, P., Grajciar, L., & Heard, C. J. (2021) · Chemistry of Materials · doi:10.1021/acs.chemmater.1c02799 · Cited by 33 · cited 3×
“O(H) bonds 14,66,73 (see Supplementary Fig.”
BackgroundMethodSelf-citationMolecular Dynamics with Chemical Accuracy─Alkane Adsorption in Acidic Zeolites
Berger, F., Rybicki, M., & Sauer, J. (2023) · ACS Catalysis · doi:10.1021/acscatal.2c05493 · Cited by 30 · cited 1×
“…pathways indicates both that the general NNP baseline model is rather robust across the BAS zeolite PES (with water) and that the correction surface is rather simple (low-dimensional) as, e.g., shown recently for the DFT-to-MP2 correction surface of alkane adsorption in protonic zeolites 80 .”
BackgroundThe Brønsted acidity of three- and two-dimensional zeolites
Thang, H. V., Vaculík, J., Přech, J., Kubů, M., Čejka, J., Nachtigall, P., Bulánek, R., & Grajciar, L. (2019) · Microporous and Mesoporous Materials · doi:10.1016/j.micromeso.2019.03.033 · Cited by 28 · cited 1×
“The resulting surface model is similar to previously used modelsystems 94,95 of synthesized MFI nanosheets 65 .”
MethodSelf-citationDefect‐Mediated Ordering of Condensed Water Structures in Microporous Zeolites
Bukowski, B. C., Bates, J. S., Gounder, R., & Greeley, J. (2019) · Angewandte Chemie · doi:10.1002/ange.201908151 · Cited by 24
Klemm, H. W., Prieto, M. J., Xiong, F., Hassine, G. B., Heyde, M., MENZEL, D., Sierka, M., Schmidt, T., & Freund, H. (2020) · Angewandte Chemie International Edition · doi:10.1002/anie.202002514 · Cited by 22 · cited 1×
“This bilayer resembles silicatene, a two-dimensional double-six-ring layer 81 , but contains four, fi ve-, six, and ten-rings (see Supplementary Table 1).”
BackgroundThe Role of Water Loading and Germanium Content in Germanosilicate Hydrolysis
Jin, M., Veselý, O., Heard, C. J., Kubů, M., Nachtigall, P., Čejka, J., & Grajciar, L. (2021) · The Journal of Physical Chemistry C · doi:10.1021/acs.jpcc.1c06873 · Cited by 21
Self-citationConstructing Collective Variables Using Invariant Learned Representations
Šípka, M., Erlebach, A., & Grajciar, L. (2023) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.2c00729 · Cited by 21
Self-citationZheng, M., & Bukowski, B. C. (2024) · The Journal of Physical Chemistry C · doi:10.1021/acs.jpcc.4c01087 · Cited by 18 · cited 3×
“…(ii) active-learning procedures to develop system-speci fi c MLPs to accelerate simulations for a speci fi c model or thermodynamic state point 30 – 32 , which capture the details of the PES, including transition states, but have little or no transferability to systems with different chemical…”
BackgroundMethodInfluentialStuckenschmidt, E., Joswig, W., & BAUR, W. H. (1996) · European Journal of Mineralogy · doi:10.1127/ejm/8/1/0085 · Cited by 12
Borgmans, S., Rogge, S. M. J., Vanduyfhuys, L., & Van Speybroeck, V. (2023) · Journal of Chemical Theory and Computation · doi:10.1021/acs.jctc.3c01028 · Cited by 8 · cited 1×
“12 or a related recent work 76 ).”
BackgroundSaha, I., Erlebach, A., Nachtigall, P., Heard, C. J., & Grajciar, L. (2023) · ChemRxiv · doi:10.26434/chemrxiv-2022-d1sj9-v3 · Cited by 6
Self-citationEnhanced sampling of robust molecular datasets with uncertainty-based collective variablesPreprint
Tan, A. R., Dietschreit, J. C. B., & Gómez‐Bombarelli, R. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2402.03753 · Cited by 4 · cited 1×
“However, the inclusion of the biased AIMD runs in the dataset is important and computationally ef fi cient when generating training data for system-speci fi c/reaction-speci fi c MLPs 47 – 49 .”
MethodLearning a reactive potential for silica-water through uncertainty attributionPreprint
Roy, S., Dürholt, J. P., Asche, T. S., et al. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2307.01705 · Cited by 2 · cited 1×
“Such systems are relevant for modeling potential precursors of zeolite synthesis or products of zeolite degradation in hot liquid water (de-silication and de-alumination) 5,9,63,64 .”
BackgroundMACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields
Batatia, I., Kovacs, D. P., Simm, G., Ortner, C., & Csanyi, G. (2022) · Advances in Neural Information Processing Systems 35 · doi:10.52202/068431-0830
Research works printed but not linked: 3
- Schütt, K., Unke, O. & Gastegger, M. Equivariant message passing for the prediction of tensorial properties and molecular spectra. In Proceedings of the 38th International Conference on Machine Learning 9377–9388 (PMLR, 2021).
- van der Maaten L Hinton G Visualizing Data using t-SNE J. Mach. Learn Res. 2008 9 2579 2605
- Kingma, D. P. & Ba, J. Adam: a method for stochastic optimization. In 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings (eds Bengio, Y. & LeCun, Y.) (ICLR, 2015).
Cited by
Application of machine learning interatomic potentials in heterogeneous catalysis
Olajide, G., Baral, K., Ezendu, S., Soyemi, A., & Szilvási, T. (2025) · Journal of Catalysis · doi:10.1016/j.jcat.2025.116202 · Cited by 34
Bocus, M., Vandenhaute, S., & Speybroeck, V. V. (2024) · Angewandte Chemie International Edition · doi:10.1002/anie.202413637 · Cited by 28
Wang, X., Wang, Q., Wang, C., Chu, Y., Hu, M., Deng, F., Yu, J., & Xu, J. (2025) · Journal of the American Chemical Society · doi:10.1021/jacs.5c01756 · Cited by 24
Learning a reactive potential for silica-water through uncertainty attribution
Roy, S., Dürholt, J. P., Asche, T. S., Zipoli, F., & Gómez‐Bombarelli, R. (2024) · Nature Communications · doi:10.1038/s41467-024-50407-9 · Cited by 28
“Several NNIP have been constructed for pure water 18 – 20 , amorphous and liquid silica 21,22 and water-fi lled aluminosilicate zeolites 23 .”
Zeolite–Water Chemistry: Characterization Methods to Unveil Zeolite Structure
Almeida, J. N., Song, L., Askarli, S., Chung, S., & Ruiz-Martínez, J. (2025) · Chemistry - Methods · doi:10.1002/cmtd.202400076 · Cited by 29
Metadata
- Authors
- Erlebach, A., Šípka, M., Saha, I., Nachtigall, P., Heard, C. J., & Grajciar, L.
- Affiliations
- Charles University (CZ)
- Corresponding author
- Erlebach, Andreas
- Grajciar, Lukáš
- ORCID
- Venue
- Nature Communications
- Published
- 2024-05-17
- Volume
- 15
- Issue
- 1
- Pages
- 4215-4215
- ISSN
- 2041-1723 (ISSN-L)
- PMC
- PMC11101627
- OpenAlex
- W4396999021
- Cited by
- 47
- References
- 112
- Citations by year
- 2026: 13 · 2025: 20 · 2024: 13
- FWCI
- 4.47
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- cc-by
- creativecommons.org/licenses/by/4.0 (tdm)
- creativecommons.org/licenses/by/4.0 (vor)
- Locations
- Nature Communications (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by-nc-sa)
- PubMed (publishedVersion)
- arXiv (Cornell University) (submittedVersion, cc-by-nc-sa)
- RePEc: Research Papers in Economics (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Univerzita Karlova v Praze (PRIMUS/20/SCI/004, UNCE/SCI/014)
- Ministerstvo Školství, Mládeže a Tělovýchovy (ID:90254, CZ.02.1.01/0.0/0.0/15 003/0000417)
- Grantová Agentura České Republiky (23-07616S)
- Topics
- Physical Sciences › Materials Science › Materials Chemistry › Machine Learning in Materials Science
- Physical Sciences › Chemistry › Inorganic Chemistry › Zeolite Catalysis and Synthesis
- Physical Sciences › Physics and Astronomy › Atomic and Molecular Physics, and Optics › Advanced Chemical Physics Studies
- Keywords
- acid zeolites, zeolite acidity, hydrothermal stability, Si/Al ratio, delta learning, rare event sampling, collective variables, machine learning interatomic potentials
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol and qwen3.8-omni-flash).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.