Predicting rotation-vibration levels of isotopically substituted molecules: water as an example

We present the results of prediction of energy levels of various water isotopologues using the fit of mass independent part of potential energy surface (PES) of H$_2$$^{16}$O only. Fit was done in such a way that mass dependent part of PES - adiabatic correction of H$_2$$^{16}$O was used during the fit and also non-adiabatic mass-dependent correction to kinetic energy. When the predicted levels of other isotopologues have been calculated, the same mass independent surface as for H$_2$$^{16}$O was used and both adiabatic and non-adiabatic corrections of the corresponding isotopologues have been employed. As a result the discrepancy between predicted and observed energy levels of H$_2$$^{17}$O and H$_2$$^{18}$O were almost of the same value as for H$_2$$^{16}$O for the levels of H$_2$$^{17}$O and H$_2$$^{18}$O known experimentally. It is natural to suppose, that at least for the levels of H$_2$$^{16}$O included in the fit, the same accuracy of prediction of H$_2$$^{17}$O and H$_2$$^{18}$O levels should be expected. Thus, the procedure transferring experimental knowledge of the major isotopologues to the minor isotopologues without the fitting of the levels of these minor isotopologues , has been developed. For other isotopologues such as D$_2$$^{16}$O and HDO the discrepancy are less perfect, but still within about 0.1 cm$^{-1}$. These results provide us with a very accurate tool for the prediction of energy levels of minor isotopologues