Correction to “Force Fields for Molecular Modeling of Sarin and its Simulants: DMMP and DIMP”

While all the force field parameters for diisopropyl methylphosphonate (DIMP) reported in the main text of our work are correct, we made mistakes in the input files for molecular dynamics (MD) simulations with the DIMP molecule. These mistakes resulted in incorrect values of liquid densities and self-diffusion coefficient, reported in Table 14 and Figure S1. More specifically, we updated the LAMMPS input file: the description of the partial charges (for TraPPE with HF charges), the force constants for bonded interactions, and the parameters describing the dihedral angles (for both TraPPE with HF and with the combined charges). The charges in the TraPPE HF input files must correspond to the HF charges from Table S1. The parameter @dihedral:HOPO2 from the TraPPE force field (defined in tra.lt line 56) is assigned for the H–O–P–O₂ dihedral interaction. The dihedral angle @dihedral:HOPO2 has had the correct parameters applied, but the single-bonded oxygen atom was mislabeled: $atom:O_2 instead of $atom:O2. We also corrected the force constants for harmonic bonds of 501.91 kcal/mol, as the initial input files had the wrong value of 717.02 kcal/mol. The dihedral angles had the phase shifts applied with the wrong signs, which have also been corrected. These parameters are now correctly applied to the charges, bonded interactions, and to the dihedral angles. We reran the MD simulations with the corrected parameters and computed the liquid densities of DIMP at 298 and 303 K. We have updated Figure S1 to reflect these new results. All of the trends remained the same, but the resulting values of density changed slightly, showing better agreement between our MD predictions and experimental data from the literature. We also recomputed the self-diffusivity values of liquid DIMP at 293 K. The mistakes in the force field files may have affected MD predictions of other properties of DIMP, such as surface tensions and viscosities, reported in refs (1) and (2). Importantly, the parameters reported in the main text are correct, and none of the Monte Carlo simulation results are affected. All previous conclusions remain unchanged. Additionally, Table 8 in the main text containing parameters for the sarin molecule had the P atom incorrectly labeled as CH₃. The corrected table is given below.