7.23.3 Tensor Component Coordinate Boundary

The Tensor Component Coordinate Boundary sets limits on how tensor components are expressed within a coordinate system.

Tensor Component Coordinate Boundary is the edge of the domain, within a chosen coordinate system, over which the coordinate functions themselves remain well-defined and invertible, and hence the edge beyond which tensor components expressed in those coordinates cease to correspond meaningfully to the underlying tensor. It is a boundary attributable specifically to the coordinate system, as distinct from the tensor, the vector space, or the manifold on which the tensor lives.

Coordinate Systems Have Limited Domains

Coordinates as Local Maps

A coordinate system assigns to each point of a region of space a tuple of real numbers, and this assignment is required to be a smooth, invertible correspondence — a diffeomorphism — only over some domain, typically a proper subset of the full space. The coordinate boundary is the edge of that domain: the set of points at which the coordinate map either stops being defined, stops being invertible, or stops being smooth.

Global Coordinates Are the Exception, Not the Rule

Many spaces of interest, including spheres, tori, and other curved or topologically nontrivial manifolds, admit no single coordinate system covering the entire space without some form of boundary or degeneracy; this is a topological fact about the space, not a deficiency of any particular choice of coordinates, so multiple overlapping coordinate charts are the norm rather than the exception.

How Tensor Components Behave at the Coordinate Boundary

Divergence of Component Values

As a point approaches a coordinate boundary, the numerical values of a tensor's components in that coordinate system can grow without bound, even though the tensor itself remains a perfectly regular, finite object at that point. This divergence stems from the transformation law for components, which involves partial derivatives of the coordinate functions, and these derivatives can become singular exactly where the coordinate map itself degenerates.

v^{i'} = \sum_{i} \frac{\partial x^{i'}}{\partial x^{i}} v^{i}

A Concrete Example: Spherical Coordinates at the Poles

In spherical coordinates on a sphere, the coordinate map degenerates at the north and south poles, where the azimuthal angle becomes undefined; a vector field's components expressed in this coordinate system can appear singular exactly at the poles, while the same vector field, expressed in a different chart centered on the pole, has perfectly finite components there. The poles are the coordinate boundary of the spherical chart, not a genuine feature of any tensor field defined on the sphere.

Diagram of a Coordinate Boundary

Distinguishing a Coordinate Boundary from a Genuine Singularity of the Tensor

The Invariant-Scalar Test

Because divergence of components near a coordinate boundary can be purely an artifact of the coordinate choice, the reliable way to determine whether the tensor itself is singular at that point is to evaluate a coordinate-independent scalar constructed from the tensor, such as a full contraction with the metric. If this scalar stays finite as the boundary is approached, the coordinate system was simply inadequate there, not the tensor.

Historical Example from General Relativity

The Schwarzschild coordinate description of a black hole exhibits exactly this phenomenon at its event horizon: certain tensor components computed in Schwarzschild coordinates diverge there, which for a long period was mistaken for a physical singularity, until coordinate systems that remain regular across the horizon (such as Eddington–Finkelstein coordinates) showed the divergence to be a coordinate boundary effect, with the associated curvature invariants remaining finite at the horizon.

Practical Handling

Restricting Formulas to Their Valid Domain

Any formula for a tensor's components derived in a specific coordinate system carries an implicit domain of validity equal to that coordinate system's own domain; applying the formula outside this domain, including at or beyond its coordinate boundary, is not a valid operation, regardless of whether the formula appears to produce a numerical output there.

Atlas Construction to Eliminate a Boundary's Effect

The standard resolution is to supply an atlas: a collection of coordinate charts whose domains collectively cover the entire space, with each chart's coordinate boundary lying strictly inside the domain of at least one other chart in the collection, together with transition functions relating the tensor's components across overlapping charts, so that the tensor is described without leaving any point uncovered by some coordinate system in which its components remain regular.