Corrigendum: Optimal control of semilinear elliptic equations with pointwise constraints on the gradient of the state

Contrary to the claim in our paper [1], equation (6.11) cannot be used to define the normal derivative 8.A.(P) for elements p ~ Vt(f~) because the term (A'p, z)a depends on the representation of A*p as an element of Co~(f~) ' for elements z s W2'~'(f~); see Lemma 2. Therefore Theorem 9 is not properly formulated and then the solution of the adjoint state equation (6.7) must be intended in a variational sense, as for the classical Neumann problem with a datum belonging to (HI(f~)) '. More precisely