Citations and licenses
Atomli citation
Section titled “Atomli citation”Paper in preparation. Please cite the software and the version you ran. Source code will be open after publication of the paper.
Atomli, version 0.1.4. Saltandfish LLC. https://atom.li
@software{atomli, title = {Atomli: an ASE-shaped Python API for atomistic simulation}, author = {{Saltandfish LLC}}, year = {2026}, url = {https://atom.li}, note = {Version 0.1.4}}Use atomli.__version__ to find the installed version. Cite the methods used in the calculation as well.
License
Section titled “License”Atomli is licensed under the
PolyForm Noncommercial License 1.0.0.
The full text ships as LICENSE in the repository and in the wheel.
PolyForm Noncommercial is not an open source license. It permits use for any noncommercial purpose and nothing else. That restriction fails the Open Source Definition, so Atomli is not open source and is not described as open source anywhere. Nothing on this site should be read as an open source grant.
What the license does permit is broad. Personal study, hobby projects, research, experiment and testing for the benefit of public knowledge are all permitted purposes. So is use by any charitable organization, educational institution, public research organization, public safety or health organization, environmental protection organization, or government institution, regardless of how that institution is funded. Changes, new works and redistribution are all permitted for those purposes. Read the license itself rather than this summary.
Commercial use needs a separate license. Ask.
Version 0.1.1 license
Section titled “Version 0.1.1 license”Version 0.1.1 was published to PyPI under the Apache License 2.0. That grant cannot be withdrawn. Anyone who received 0.1.1 under Apache-2.0 keeps every right Apache-2.0 gave them, permanently, including commercial use. Relicensing does not reach backwards.
PolyForm Noncommercial governs 0.1.2 and every later release. The two facts sit side by side and both are true.
| Version | Terms |
|---|---|
| 0.1.1 | Apache License 2.0, irrevocably |
| 0.1.2 and later | PolyForm Noncommercial License 1.0.0 |
Reference engines
Section titled “Reference engines”One row per engine on the benchmark page. Each citation is the one that project asks for, taken from its own repository or documentation.
| Engine | Cite |
|---|---|
| PySCF | Sun et al., J. Chem. Phys. 2020, 153, 024109 |
| GPU4PySCF | Li et al., arXiv:2407.09700; Wu et al., arXiv:2404.09452 |
| tblite (GFN2-xTB) | Bannwarth, Ehlert, Grimme, J. Chem. Theory Comput. 2019, 15, 1652 |
| g-xTB | Froitzheim, Müller, Hansen, Grimme, ChemRxiv 2025, doi:10.26434/chemrxiv-2025-bjxvt |
| Skala | Luise et al., arXiv:2506.14665 |
| Nequix | Koker, Kotak, Smidt, arXiv:2508.16067 |
| ASE | Larsen et al., J. Phys.: Condens. Matter 2017, 29, 273002 |
The PySCF README states which paper must be cited.
@article{pyscf2020, title = {Recent developments in the {PySCF} program package}, author = {Sun, Qiming and Zhang, Xing and Banerjee, Samragni and Bao, Peng and Barbry, Marc and Blunt, Nick S. and Bogdanov, Nikolay A. and Booth, George H. and Chen, Jia and Cui, Zhi-Hao and Eriksen, Janus J. and Gao, Yang and Guo, Sheng and Hermann, Jan and Hermes, Matthew R. and Koh, Kevin and Koval, Peter and Lehtola, Susi and Li, Zhendong and Liu, Junzi and Mardirossian, Narbe and McClain, James D. and Motta, Mario and Mussard, Bastien and Pham, Hung Q. and Pulkin, Artem and Purwanto, Wirawan and Robinson, Paul J. and Ronca, Enrico and Sayfutyarova, Elvira R. and Scheurer, Maximilian and Schurkus, Henry F. and Smith, James E. T. and Sun, Chong and Sun, Shi-Ning and Upadhyay, Shiv and Wagner, Lucas K. and Wang, Xiao and White, Alec and Whitfield, James Daniel and Williamson, Mark J. and Wouters, Sebastian and Yang, Jun and Yu, Jason M. and Zhu, Tianyu and Berkelbach, Timothy C. and Sharma, Sandeep and Sokolov, Alexander Yu. and Chan, Garnet Kin-Lic}, journal = {The Journal of Chemical Physics}, volume = {153}, number = {2}, pages = {024109}, year = {2020}, doi = {10.1063/5.0006074}}PySCF does not implement density functionals itself. It asks that the functional library be cited as well. The benchmark runs use Libxc:
@article{libxc2018, title = {Recent developments in libxc: A comprehensive library of functionals for density functional theory}, author = {Lehtola, Susi and Steigemann, Conrad and Oliveira, Micael J. T. and Marques, Miguel A. L.}, journal = {SoftwareX}, volume = {7}, pages = {1}, year = {2018}, doi = {10.1016/j.softx.2017.11.002}}GPU4PySCF
Section titled “GPU4PySCF”The GPU4PySCF README lists two references and asks for both. Cite the PySCF paper above as well.
@misc{li2024introducing, title = {Introducing GPU-acceleration into the Python-based Simulations of Chemistry Framework}, author = {Li, Rui and Sun, Qiming and Zhang, Xing and Chan, Garnet Kin-Lic}, year = {2024}, eprint = {2407.09700}, archivePrefix = {arXiv}, primaryClass = {physics.comp-ph}, url = {https://arxiv.org/abs/2407.09700}}
@misc{wu2024enhancing, title = {Enhancing GPU-acceleration in the Python-based Simulations of Chemistry Framework}, author = {Wu, Xiaojie and Sun, Qiming and Pu, Zhichen and Zheng, Tianze and Ma, Wenzhi and Yan, Wen and Yu, Xia and Wu, Zhengxiao and Huo, Mian and Li, Xiang and Ren, Weiluo and Gong, Sheng and Zhang, Yumin and Gao, Weihao}, year = {2024}, eprint = {2404.09452}, archivePrefix = {arXiv}, primaryClass = {physics.comp-ph}, url = {https://arxiv.org/abs/2404.09452}}tblite and GFN2-xTB
Section titled “tblite and GFN2-xTB”tblite ships no citation file. Its documentation attributes each parametrization
to a paper, and the GFN2-xTB parameter record inside tblite names its own
reference as DOI: 10.1021/acs.jctc.8b01176. That is the paper to cite for a
GFN2-xTB number, from either engine.
@article{bannwarth2019, title = {{GFN2-xTB}---An Accurate and Broadly Parametrized Self-Consistent Tight-Binding Quantum Chemical Method with Multipole Electrostatics and Density-Dependent Dispersion Contributions}, author = {Bannwarth, Christoph and Ehlert, Sebastian and Grimme, Stefan}, journal = {Journal of Chemical Theory and Computation}, volume = {15}, number = {3}, pages = {1652--1671}, year = {2019}, doi = {10.1021/acs.jctc.8b01176}}tblite originated in the xtb program package. Cite the xTB framework review
when the framework itself is the subject:
@article{bannwarth2020, title = {Extended tight-binding quantum chemistry methods}, author = {Bannwarth, Christoph and Caldeweyher, Eike and Ehlert, Sebastian and Hansen, Andreas and Pracht, Philipp and Seibert, Jakob and Spicher, Sebastian and Grimme, Stefan}, journal = {WIREs Computational Molecular Science}, volume = {11}, pages = {e01493}, year = {2020}, doi = {10.1002/wcms.1493}}The g-xTB distribution points to its ChemRxiv preprint.
@article{gxtb2025, title = {g-xTB: A General-Purpose Extended Tight-Binding Electronic Structure Method For the Elements H to Lr (Z=1--103)}, author = {Froitzheim, Thomas and M{\"u}ller, Marcel and Hansen, Andreas and Grimme, Stefan}, journal = {ChemRxiv}, year = {2025}, doi = {10.26434/chemrxiv-2025-bjxvt}}The reference g-xTB run is driven by xtb 6.7.1. Cite bannwarth2020 above for
that driver.
Skala is the neural exchange-correlation functional used by the QC calculator
at xc="SKALA-1.1". The Skala repository ships a CITATION.cff naming this
preprint as its preferred citation.
@misc{skala2025, title = {Accurate and scalable exchange-correlation with deep learning}, author = {Luise, Giulia and Huang, Chin-Wei and Vogels, Thijs and Kooi, Derk P. and Ehlert, Sebastian and Lanius, Stephanie and Giesbertz, Klaas J. H. and Karton, Amir and Gunceler, Deniz and Battaglia, Stefano and Simm, Gregor N. C. and Szab{\'o}, P. Bern{\'a}t and Stanley, Megan and Bruinsma, Wessel P. and Huang, Lin and Wei, Xinran and Garrido Torres, Jos{\'e} and Katbashev, Abylay and Chavez Zavaleta, Rodrigo and M{\'a}t{\'e}, B{\'a}lint and Kaba, S{\'e}kou-Oumar and Sordillo, Roberto and Chen, Yingrong and Williams-Young, David B. and Bishop, Christopher M. and Hermann, Jan and van den Berg, Rianne and Gori-Giorgi, Paola}, year = {2025}, eprint = {2506.14665}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2506.14665}}The author list above follows the current arXiv version. The CITATION.cff
snapshot in older Skala checkouts lists one author fewer.
Nequix
Section titled “Nequix”The Nequix project names the foundation-model paper as the reference for
nequix-mp-1.
@misc{nequix2025, title = {Training a Foundation Model for Materials on a Budget}, author = {Koker, Teddy and Kotak, Mit and Smidt, Tess}, year = {2025}, eprint = {2508.16067}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2508.16067}}The -pft variants in the model catalog, and the
nequix-omat-1 and nequix-oam-1 weights, are attributed by the project to the
phonon fine-tuning paper instead.
@misc{pft2026, title = {PFT: Phonon Fine-tuning for Machine Learned Interatomic Potentials}, author = {Koker, Teddy and Gangan, Abhijeet and Kotak, Mit and Marian, Jaime and Smidt, Tess}, year = {2026}, eprint = {2601.07742}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2601.07742}}Atomli mirrors the ASE API, and the benchmark harness drives every reference engine through its ASE calculator. The ASE FAQ names the paper to cite.
@article{ase2017, title = {The atomic simulation environment---a Python library for working with atoms}, author = {Larsen, Ask Hjorth and Mortensen, Jens J{\o}rgen and Blomqvist, Jakob and Castelli, Ivano E and Christensen, Rune and Du{\l}ak, Marcin and Friis, Jesper and Groves, Michael N and Hammer, Bj{\o}rk and Hargus, Cory and Hermes, Eric D and Jennings, Paul C and Jensen, Peter Bjerre and Kermode, James and Kitchin, John R and Kolsbjerg, Esben Leonhard and Kubal, Joseph and Kaasbjerg, Kristen and Lysgaard, Steen and Maronsson, J{\'o}n Bergmann and Maxson, Tristan and Olsen, Thomas and Pastewka, Lars and Peterson, Andrew and Rostgaard, Carsten and Schi{\o}tz, Jakob and Sch{\"u}tt, Ole and Strange, Mikkel and Thygesen, Kristian S and Vegge, Tejs and Vilhelmsen, Lasse and Walter, Michael and Zeng, Zhenhua and Jacobsen, Karsten W}, journal = {Journal of Physics: Condensed Matter}, volume = {29}, number = {27}, pages = {273002}, year = {2017}, doi = {10.1088/1361-648X/aa680e}}