Another way to fill in the “missing link” in Fig. 5.48 is to look for a magnetostatic analog to Eq. 2.21. The obvious candidate would be
A ( r )=\int_{ O }^{ r }( B \times d l ) .
V( r ) \equiv-\int_{ O }^{ r } E \cdot d l (2.21)
(a) Test this formula for the simplest possible case—uniform B (use the origin as your reference point). Is the result consistent with Prob. 5.25? You could cure this problem by throwing in a factor of \frac{1}{2} , but the flaw in this equation runs deeper
(b) Show that \int( B \times d l ) is not independent of path, by calculating \oint( B \times d l ) around the rectangular loop shown in Fig. 5.63.
As far as I knoW ,^{28} the best one can do along these lines is the pair of equations
\text { (i) } V( r )=- r \cdot \int_{0}^{1} E (\lambda r ) d \lambda ,
\text { (ii) } A ( r )=- r \times \int_{0}^{1} \lambda B (\lambda r ) d \lambda .
[Equation (i) amounts to selecting a radial path for the integral in Eq. 2.21; equation (ii) constitutes a more “symmetrical” solution to Prob. 5.31.]
(c) Use (ii) to find the vector potential for uniform B.
(d) Use (ii) to find the vector potential of an infinite straight wire carrying a steady current I. Does (ii) automatically satisfy ∇ · A = 0? [Answer: \left(\mu_{0} I / 2 \pi s\right)(z \hat{ s }-s \hat{ z }) ]