# Flux surface

A given smooth surface S with normal n is a flux surface of a smooth vector field B when

$\vec B \cdot \vec n = 0$

everywhere on S. In other words, the magnetic field does not cross the surface S anywhere, i.e., the magnetic flux traversing S is zero. It is then possible to define a scalar flux function (f) such that its value is constant on the surface S, and

$\vec B \cdot \vec \nabla f = 0$

In three dimensions, the only closed flux surface corresponding to a non-vanishing vector field is a topological toroid.  This fact lies at the basis of the design of magnetic confinement devices.

Assuming the flux surfaces have this toroidal topology, the function f defines a set of nested surfaces, so it makes sense to use this function to label the flux surfaces, i.e., f may be used as a "radial" coordinate. Each toroidal surface f encloses a volume V(f). The surface corresponding to an infinitesimal volume V is essentially a line that corresponds to the toroidal axis (called magnetic axis when B is a magnetic field).

The flux F through an arbitrary surface S is given by

$F = \int_S{\vec B \cdot \vec n dS}$

When B is a magnetic field with toroidal nested flux surfaces, two magnetic fluxes can be defined from two corresponding surfaces.  The poloidal flux is defined by

$\psi = \int_{S_p}{\vec B \cdot \vec n dS}$

where Sp is a ring-shaped ribbon stretched between the magnetic axis and the flux surface f. (Complementarily, Sp can be taken to be a surface spanning the central hole of the torus.) Likewise, the toroidal flux is defined by

$\phi = \int_{S_t}{\vec B \cdot \vec n dS}$

where St is a poloidal section of the flux surface. It is natural to use ψ or φ to label the flux surfaces instead of the unphysical label f.