Since U(x1,y1)= 0 outside of Σ (see figure III-1), we can rewrite the integral defined in Eq. II-2A by integrating from to :
General approximations:
distance z >> the largest linear dimension of Σ.
distance z >> the largest linear dimension of the region of observation.
With those two hypothesis, we obtain:
The term r01 in the exponent cannot be replaced by z as it is done in the denominator because the error which would result from this approximation would be multiplied by and would lead to errors on the phase far larger than radians.