Solid state 15N NMR and theoretical studies of primary and secondary geometric H/D isotope effects on low-barrier NHN-hydrogen bonds

Using a combination of high resolution and dipolar solid state N-15 NMR we have determined H/D isotope effects on the nitrogen-hydron (L = H, D) distances and N-15 chemical shielding tensors of strongly hydrogen bonded bisisocyanide salts of the type [(CO)(5)Cr-C=N ... L ... N=C-Cr(CO)(5)]X--(+), where X+ = AsPh4+ (2) and X+ = NPr4+ (3). These compounds have been modeled theoretically by the linear system [C=N ... L ... N=C]Li--(+) (1). The crystal field acting on the anion was generated by a variety of fixed C ... Li distances. For the calculation of dynamical corrections of geometries and NMR chemical shifts, an iterative procedure based on the crude adiabatic approximation was employed, consisting of (i) ab initio calculation of the energy hypersurface at the MP2/6-31+G(d,p) level, (ii) solution of the Schrodinger equation for the anharmonic collinear hydron motion, and (iii) NMR chemical shift calculations using the IGLO-method. The two hydrogen bond distances r(1) = N ... L and r(2) = L ... N are found to change in a correlated way when H is replaced by D, as a function of X+, i.e., of the electric field at the hydrogen bond site. The correlation r(1) = f(r(2)) established here experimentally and theoretically for very strong NHN-hydrogen bonds shows a good agreement with a correlation established previously (Steiner, Th. J. Chem. Soc., Chem. Commun. 1995, 1331) based on the neutron diffraction structures of a number of weakly hydrogen bonded solids. A plot of the sum q(2) = r(1) + r(2)-corresponding in a linear hydrogen bond to the heavy atom separation-as a function of the proton dislocation from the hydrogen bond center q(1) = 1/2(r(1)-r(2)) exhibits a minimum value at about 2.54 Angstrom for the symmetric low-barrier hydrogen bond at q(1) = 0. This situation is realized experimentally for 2. When q(1) not equal 0 anharmonic single well hydrogen bonds are obtained, typical for 3. The geometric H/D isotope effects can be split into a primary effect referring to the hydron position q(1) = 1/2(r(1)-r(2)) and a secondary effect referring to the heavy atom position q(2). Secondary effects have been reported previously by Ubbelohde. Both isotope effects are shown to be related in a simple empirical way to the hydrogen bond geometries and to the isotopic fractionation factors. Finally, it is shown that the chemical shielding of the nuclei in the hydrogen bridge is a qualitative probe for the primary and secondary geometric isotope effects.