FULL PAPER Comparison of density functionals for the study of the high spin low spin gap in Fe(III) spin crossover complexes Jitnapa Sirirak 1 | Darunee Sertphon 2 | Wasinee Phonsri 2† | Phimphaka Harding 2 | David J. Harding 2 1 Department of Chemistry, Faculty of Science, Silpakorn University, Muang, Nakhon Pathom 73000, Thailand 2 Functional Materials and Nanotechnology Center of Excellence, Walailak University, Nakhon Si Thammarat, Thasala 80160, Thailand Correspondence David J. Harding, Functional Materials and Nanotechnology Center of Excellence, Walailak University, Nakhon Si Thammarat, Thasala 80161, Thailand. Email: hdavid@g-mail.wu.ac.th Funding information We thank Walailak University for a Postdoctoral Fellowship to JS, the Thailand Research Fund (RSA5580028 and RSA5880048) and the National Science and Technology Development Agency (grant no.: P-10-11181) for funding this research. Abstract A detailed investigation of the accuracy of different quantum mechanical methods for the study of iron(III) spin crossover complexes is presented. The energy spin state gap between the high and low spin states; DE(HS-LS) of nine iron(III) quinolylsalicylaldiminate complexes were calculated with nine different DFT functionals, then compared. DFT functionals: B3LYP, B3LYP-D3, B3LYP*, BH&HLYP, BP86, OLYP, OPBE, M06L, and TPSSh were tested with six basis sets: 3-21G*, dgdzvp, 6-31G**, cc-pVDZ, Def2TZVP, and cc-pVTZ. The cations from the X-ray crystal structures of [Fe (qsal-OMe) 2 ]ClMeCNH 2 O, [Fe(qsal-OMe) 2 ]Cl2MeOH0.5H 2 O, [Fe(qsal-OMe) 2 ]BF 4 MeOH, [Fe (qsal-OMe) 2 ]NCSCH 2 Cl 2 , [Fe(qsal-F) 2 ]NCS, [Fe(qsal-Cl) 2 ]NCSMeOH, [Fe(qsal-Br) 2 ]NCSMeOH, [Fe(qsal-I) 2 ]OTfMeOH, and [Fe(qsal) 2 ]NCSCH 2 Cl 2 were used as starting structures. The results show that B3LYP, B3LYP-D3, OLYP, and OPBE with a 6-31G**, Def2TZVP, and cc-pVTZ basis set give reasonable results of DE(HS-LS) compared with the experimental data. The enthalpy of [Fe (qsal-I) 2 ] 1 calculated with an OLYP functional and cc-pVTZ basis set (1.48 kcal/mol) most closely matches the experimental data (1.34 kcal/mol). B3LYP* yields an enthalpy of 5.92 kcal/mol sug- gesting it may be unsuitable for these Fe(III) complexes, mirroring recent results by Kepp (Inorg. Chem., 2016, 55, 2717–2727). KEYWORDS basis sets, DFT calculations, functionals, iron(III) complexes, spin crossover 1 | INTRODUCTION Spin crossover is an elegant example of bistability with switching occurring between a high spin (HS) and a low spin (LS) state. [1–4] A par- ticular advantage of these materials is that a variety of external stimuli including temperature, pressure or light may be used to enable switching. [5–7] Moreover, their relative simplicity and ready processabil- ity mean that they are attractive targets in the fields of molecular elec- tronics and high density data storage. [8,9] To be spin crossover active requires that the HS and LS states are close in energy, something that depends on the ligand, the metal and its oxidation state. [10] Accurately computing such spin state energies to enable better design or insight into systems where subtle ligand changes radically alter magnetic behavior is, however, computationally challenging. [11–13] Part of the problem lies in the fact that the spin state gap is often within the error associated with the calculation itself. While computational systems exist which can reduce this error they are typically computationally expensive and often require expert knowledge to use correctly. [14,15] Moreover, semiempirical methods such as CAPST2 have been found to have strong dependence on the value chosen for the ionization potential—electron affinity shift. [16,17] DFT calculations have become a popular alternative to these time consuming HF methods. While successful to some degree the choice of functional used has been shown to have a significant impact on the energy gap. Thus, pure functionals such BLYP and BP86 invariably favour LS states while hybrid functionals such as B3LYP tend to overstabilize HS states. [18] Another factor shown to be critical for most common func- tionals is that of the basis sets. Swart has demonstrated that large basis sets are typically required to accurately reproduce experimental data. [19] In relation to spin crossover compounds, the favouring of the HS state by B3LYP has been compensated for by adjusting the amount of † Current address: School of Chemistry, Monash University, Clayton, Melbourne, Victoria, 3800, Australia Int J Quantum Chem. 2017;e25362. https://doi.org/10.1002/qua.25362 http://q-chem.org V C 2017 Wiley Periodicals, Inc. | 1 of 00 Received: 13 August 2016 | Revised: 10 January 2017 | Accepted: 2 February 2017 DOI 10.1002/qua.25362