On the existence of the second Dirac operator in Riemannian space