TY - JOUR

T1 - Identification of the best DFT functionals for a reliable prediction of lignin vibrational properties

AU - Barsberg, Soren

PY - 2015/2/21

Y1 - 2015/2/21

N2 - Lignin is the most abundant aromatic plant polymer on earth. Useful information on its structure and interactions is gained by vibrational spectroscopy and relies on the quality of band assignments. B3LYP predictions were recently shown to support band assignments. Further progress calls for a comprehensive study of the quality of available theoretical methods in relation to the task of predicting lignin vibrational properties. The present study examined more than 50 functionals for prediction of IR vibrations of an appropriate lignin model. Based on a basis set incompleteness study, the pc-2 basis set was used. B98, X3LYP and B97-1 were the overall best-performing functionals, and “fingerprint” band positions were predicted by single-factor scaling of harmonic frequencies to an average error of ±3 cm−1 by optimized scaling factors of 1.017, 1.021 and 1.016, respectively. Their performance using instead explicit anharmonic correction was slightly worse giving an error of ca. ±5 cm−1. The X3LYP and B97-1 functionals offer also good description of hydrogen bonding. Error compensation from, e.g., insufficient treatment of solvation is likely to affect these results, and thus at this stage no single functional stands out. These results provide a needed basis for further theoretical developments in relation to vibrational assignments of Infrared and Raman spectra of lignin.

AB - Lignin is the most abundant aromatic plant polymer on earth. Useful information on its structure and interactions is gained by vibrational spectroscopy and relies on the quality of band assignments. B3LYP predictions were recently shown to support band assignments. Further progress calls for a comprehensive study of the quality of available theoretical methods in relation to the task of predicting lignin vibrational properties. The present study examined more than 50 functionals for prediction of IR vibrations of an appropriate lignin model. Based on a basis set incompleteness study, the pc-2 basis set was used. B98, X3LYP and B97-1 were the overall best-performing functionals, and “fingerprint” band positions were predicted by single-factor scaling of harmonic frequencies to an average error of ±3 cm−1 by optimized scaling factors of 1.017, 1.021 and 1.016, respectively. Their performance using instead explicit anharmonic correction was slightly worse giving an error of ca. ±5 cm−1. The X3LYP and B97-1 functionals offer also good description of hydrogen bonding. Error compensation from, e.g., insufficient treatment of solvation is likely to affect these results, and thus at this stage no single functional stands out. These results provide a needed basis for further theoretical developments in relation to vibrational assignments of Infrared and Raman spectra of lignin.

KW - Lignin

KW - Infrared

KW - Raman

KW - Density functional theory

KW - Assignments

U2 - 10.1007/s00214-015-1638-2

DO - 10.1007/s00214-015-1638-2

M3 - Journal article

VL - 134

JO - Theoretical Chemistry Accounts

JF - Theoretical Chemistry Accounts

SN - 1432-881X

M1 - 33

ER -