research36 min studyLesson 839 in Mathematics

Compactness and lower semicontinuity for currents

State the support and boundary bounds behind integral compactness and measure lost oscillation in a flat limit.

Before you begin

On this page

The question for today

What you’ll understand

State a global integral-current compactness theorem with all three uniform controls.
Distinguish weak current convergence, flat convergence, and convergence of mass.
Compute an explicit filling bound for oscillatory polygonal graphs.

Weak geometric limits can lose oscillatory area or cancel nearby oppositely oriented sheets. To extract a limit that still belongs to the integral-current class, we need more than a bound on area. Boundary mass and support control enter the compactness theorem, while lower semicontinuity determines which variational inequalities survive. A sequence of tiny triangular teeth will converge flatly to a segment even though its lengths stay strictly larger than the limiting length.

Integral compactness controls mass, boundary, and location

In Euclidean space, consider integral k-currents whose supports lie in one fixed compact set and whose masses plus boundary masses are uniformly bounded. Federer–Fleming compactness gives a subsequence converging to an integral current; in this global compact-support setting one can take flat convergence and hence weak current convergence. The theorem preserves integrality, a feature not supplied by a generic weak compactness statement for arbitrary distributions.

Each assumption has a role. Mass controls the amount of current, boundary mass controls small-scale fragmentation at its edges, and a common compact region prevents material from escaping to infinity. Local versions replace global bounds with bounds on every relatively compact subdomain and give the corresponding local conclusion. They should not be quoted as a global flat-compactness theorem for a family of translated objects whose supports run away.

\operatorname{spt}T_j\subset K\text{ compact},\quad \sup_j\{\mathbf M(T_j)+\mathbf M(\partial T_j)\}<\infty\quad\Longrightarrow\quad T_{j_\ell}\to T\in\mathbf I_k\text{ flatly}.

This global statement requires integral currents and a common compact support, not just arbitrary finite-mass currents.↔ Scroll wide equations horizontally.

A filling can make flat distance small without small mass difference

The flat norm decomposes a current S as a residual R plus the boundary of a filling Q and minimizes the sum of their masses. A thin strip can have a long boundary but tiny filling area, so its boundary is small in this topology. Weak convergence tests every compactly supported smooth form, while flat convergence additionally provides this quantitative decomposition. Under the stated bounds and setting, these topologies connect through the compactness theorem.

Boundary is continuous under weak current convergence because testing the boundary is testing the original current on an exterior derivative. Mass is lower semicontinuous: every bounded-comass test gives a lower bound on approximating masses, and taking the supremum gives the limit inequality. There is no general reverse inequality. A minimizing sequence therefore can lose excess mass in its limit, which is useful for existence proofs but prevents silently replacing lower semicontinuity by equality.

\mathcal F(S)=\inf_{S=R+\partial Q}\{\mathbf M(R)+\mathbf M(Q)\},\qquad \mathbf M(T)\le\liminf_j\mathbf M(T_j).

The decomposition has a k-dimensional residual and a (k+1)-dimensional filling; finite mass is required for the displayed competitors.↔ Scroll wide equations horizontally.

A sawtooth separates filling size from arclength

Build a graph over the unit interval using j identical triangular teeth, each rising and falling with slopes plus and minus one. Its height tends to zero, but its speed has magnitude square root of two on every edge. The oriented graph and the straight base segment have the same endpoint boundary. Their difference bounds the union of thin triangles between them, producing a flat-norm estimate from the filling area.

This example contains neither unbounded support nor unbounded boundary mass, so the compactness theorem applies. It shows that uniform graph convergence and flat convergence need not preserve length. By contrast, translating a unit segment farther and farther away gives zero as a local weak limit, because each fixed compact test eventually misses it. That escape is precisely what the common-support hypothesis excludes from the global statement.

T_j-T_0=\partial Q_j\quad\Longrightarrow\quad\mathcal F(T_j-T_0)\le\mathbf M(Q_j).

A filling estimate proves flat convergence even when the masses of the currents do not converge to the limiting mass.↔ Scroll wide equations horizontally.

Work it through

Tiny teeth keep excess length but have vanishing filling area

Over [0,1], use j triangular teeth of width 1/j, height 1/(2j), and slopes ±1, oriented left to right. Let T_j be the graph current and T_0 the base segment. Compute current masses, boundary masses, a common support set, and a flat-distance bound.

Each tooth has two edges of length $\sqrt2/(2j)$ , so the total graph mass is $\sqrt2$ . All intermediate endpoints cancel; the boundary is the endpoint difference of the unit segment, with mass two. The base segment has mass one.
All graphs lie in the fixed compact rectangle $[0,1]\times[0,1/2]$ . Each filled tooth has area $(1/2)(1/j)(1/(2j))=1/(4j^2)$ . The j triangles give a filling of mass $1/(4j)$ , with orientation chosen to bound $T_j-T_0$ .
The flat distance is at most this filling mass and tends to zero. Lower semicontinuity reads $1\le\sqrt2$ , strictly. The length difference does not disappear, even though the fillings and graph heights tend to zero.

\mathbf M(T_j)=\sqrt2,\quad\mathbf M(\partial T_j)=2,\quad\mathcal F(T_j-T_0)\le\frac1{4j}\longrightarrow0.

The example satisfies support and boundary bounds while demonstrating strict mass loss in the limit.↔ Scroll wide equations horizontally.

The result

T_j converges flatly to the unit segment with bound 1/(4j), while its masses remain √2 and the limiting mass is one.

The living notebook

Explore the model

Interactive lab

Interactive lab · vectors

Compare a zigzag’s length with its displacement

Pause and Reset, then set angle to 45 degrees and magnitude to 1. Record components (1/√2,1/√2). For a fresh comparison Reset and set angle to −45 degrees and magnitude to 1. Adding the two recorded vectors gives displacement (√2,0), while the traversed length is two, a factor √2 larger than straight displacement. This is one tooth’s geometry, not a display of current convergence or a filling. Edits retain the clock; Reset and both settings are required for each comparison.

Vectorx / y projectionsTrace marker

x component: 1.393
y component: 0.975
Magnitude: 1.7

Angle35°Magnitude1.7

t = 0 s

Playback traces a fixed vector and its projections. Angle and magnitude set the geometry; the moving marker does not change them. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forSmall geometric excursions can retain excess arclength; a flat-convergence claim needs an explicit filling estimate.

Pause & reason

Check your understanding

Interactive checks need JavaScript. You can still read the explanation.

Read the checkpoint explanation

The stated theorem uses all three controls and integrality. Local weak limits can exist under different hypotheses, but that is not the global flat-compactness conclusion being claimed.

Your turn

Try it yourself

Translate the oriented unit segment to run from (j,0) to (j+1,0). Give its mass and boundary mass. Explain its local weak limit and why the global common-support theorem does not apply.

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

A fixed compactly supported test form eventually misses the translated segment. That observation does not supply one compact set containing the entire sequence.

Compare with the solution

Each current has mass one and boundary mass two. Every fixed compactly supported form eventually misses it, so the local weak limit is zero, but there is no common compact support. To check the flat norm, choose a compactly supported form equal to $dx$ along the segment with both $\|\omega\|_{\mathrm{comass}}\le1$ and $\|d\omega\|_{\mathrm{comass}}\le1$ , using a broad transverse cutoff. Its value on the segment is one, which bounds every flat decomposition's total mass below by one. Taking the segment itself as residual gives the reverse bound. Thus each flat norm is one, and there is no global flat convergence to zero.

Keep the idea

Take it with you

Global integral compactness needs uniform mass, boundary mass, and support control.
Flat distance can be small because of a small filling rather than a small mass difference.
Mass is lower semicontinuous, and oscillatory excess can disappear in the limit.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.