Existing problems in theoretical determination of red-shifted or blue-shifted hydrogen bond

Both Hobza (Int. J. Quantum Chem., 90 (2002) 1071) and Li et al. (J. Am. Chem. Soc., 124 (2002) 9639) have predicted a blue-shifted or improper N-H⋯F hydrogen bond existing in the F2N-H⋯FH complex with the method of Hartree-Fock calculation followed by Møller-Plesset correlation energy correction truncated at second order (MP2/6-31G** and MP2/6-311+G**, respectively). Unexpectedly, our density functional calculation (B3LYP) has predicted a red-shifted or normal N-H⋯F hydrogen bond in this complex whether the counterpoise-corrected gradient optimization and frequency calculation is applied or not. Changing the basis set from 6-311+G(d,p) to 6-311++G(3df,3pd) does not change the red-shifted hydrogen bond prediction. Experimental identification should be provided for evaluating relative performance of these two theoretical methods in predicting correct hydrogen bond stretching vibration frequency shift upon complex formation. X-H bond length change and its vibrational frequency change do not necessarily have monotonic corresponding relationship, i.e. blue-shift does not necessarily mean X-H bond length contraction, and vice versa. Much more attention or consideration should be paid in theoretical hydrogen bond type prediction such as red-shifted or blue-shifted.