On the accuracy of pseudopotentials designed for QMC calculations on molecules of the first row atoms