It is true that, on new moon, the projection of the Sun-Earth-Moon angle is zero, but I think that the elongation is defined as the actual angle, not the projected angle.
The elongation will be zero only if the Moon? is on the ecliptic, and then there is a total Solar eclipse.
that's my guess too, I am looking for a (simplified) model to visualize that. But I am stuck to this: there are three bodies,sun,earth and moon; two of them on the ecliptic plane; in case of a new moon the angle separation sun-observer-target of the projection of the moon on the ecliptic plane should be 0 in my thinking; but it is not. At what point I am wrong? Spherical trigonometry? sorry for a not so clear description ...
