This problem works through three of the structural results behind the vector potential
PHYSICS · Multi-part · Fermi University Physics · Electricity, Magnetism & Electromagnetism · Magnetostatics · Maxwell's equations and conservation laws
The Boundary-Value Problem, Ch The vector potential and the magnetostatic boundary-value problem
Problem
This problem works through three of the structural results behind the vector potential, using only ∇⋅B=0, the definition of the curl and divergence, and the free-space Green function of Poisson's equation.
((a)) Existence. A steady field B(r) satisfies ∇⋅B=0 everywhere. Consider the trial potential, with x0=y0=z0=0 taken as the lower limits of every integral, Ax(x,y,z)Ay(x,y,z)Az=∫0zBy(x,y,z′)dz′−∫0yBz(x,y′,0)dy′,=−∫0zBx(x,y,z′)dz′,=0. Show, using ∇⋅B=0 only where it is needed, that ∇×A=B at every point, so that a vector potential of B exists.
((b)) Reaching the Coulomb gauge. In a region where the true field is the uniform B=B0z^ (B0 a constant, unit T), someone proposes A=Cxx^+B0xy^, with C a constant (unit T). (i) Verify that ∇×A=B0z^, so this is a legitimate vector potential of the field. (ii) Show that ∇⋅A=C=0 in general, so this A is not in the Coulomb gauge. (iii) Find a gauge function λ(x,y,z) solving ∇2λ=−∇⋅A that removes the divergence, and give the resulting A′=A+∇λ. Confirm ∇×A′=B0z^ still.
((c)) Recovering Biot-Savart. A steady current density J(r′), confined to a bounded region and satisfying ∇⋅J=0, produces A(r)=4πμ0∫∣r−r′∣J(r′)dτ′. Take the curl of this integral with respect to the field point r (the integrand's r′-dependence is untouched by ∇) and show that the result is the Biot-Savart law, B(r)=4πμ0∫J(r′)×∣r−r′∣3r−r′dτ′. State where the identity ∇×[fJ(r′)]=(∇f)×J(r′) is used and why the term f(∇×J) that the product rule would otherwise add is absent.