Skip to content

Finite Extinction Time

"For any metric \(g_0\) on a closed simply connected three-manifold \(M\), the Ricci flow with surgery starting from \(g_0\) becomes extinct in finite time." — Perelman, Finite extinction time for the solutions to the Ricci flow on certain three-manifolds, arXiv:math/0307245, Theorem 1.1

In Article 04 we proved geometrization via the thin–thick decomposition – Perelman's actual result. For the Poincaré conjecture alone, however, there is a much shorter route: if \(M\) is simply connected, the Ricci flow with surgery already disappears in finite time. This article follows Perelman 0307245 together with the parallel work of Colding & Minicozzi (2005).

1. What does "extinction" mean?

A solution of Ricci flow with surgery is called finitely extinct if there is a \(T < \infty\) with \(M_t = \emptyset\) for all \(t \ge T\) – i.e. from time \(T\) on, every component has been recognized as a spherical space form \(S^3/\Gamma\) and discarded by surgery. In other words: the algorithm of Article 03 terminates instead of accumulating infinitely many surgery times.

This is a topological statement in analytic disguise: if the manifold consists only of spherical components, the flow detects this through a globally growing positive curvature and extinguishes everything.

2. Key idea: area of minimal surfaces under the flow

Perelman's strategy is not analytic (volume or curvature estimates) but topological-geometric: one considers 2-spheres in \(M\) that would function as obstacles to extinction under the flow, and shows that their minimal area decreases monotonically – fast enough to hit zero in finite time.

2.1 The functional \(W_2\) (for \(\pi_2(M) \neq 0\))

Let \(M\) be compact with \(\pi_2(M) \neq 0\). For each non-trivial class \(\alpha \in \pi_2(M)\) define $$ W_2(g, \alpha) := \inf{\,\mathrm{Area}_g(f)\ :\ f : S^2 \to M,\ [f] = \alpha\,}. $$ This is the minimal harmonic 2-sphere area in the class \(\alpha\) – existence by Sacks–Uhlenbeck (1981).

Lemma (Perelman 0307245, §2). Along Ricci flow the function \(t \mapsto W_2(g(t), \alpha)\) satisfies the differential inequality $$ \frac{d}{dt} W_2(g(t), \alpha) \le -4\pi - \tfrac{1}{2} R_{\min}(t)\, W_2(g(t), \alpha), $$ where \(R_{\min}(t) = \min_{x \in M_t} R(x, t)\) is the minimal scalar curvature.

This is the central differential inequality of the proof. It says: even if \(R_{\min} \le 0\), \(W_2\) decreases by at least \(4\pi\) per unit time, because the mean curvature of the minimal surface is bounded by the Gauss–Bonnet constant \(4\pi \chi(S^2) / 2 = 4\pi\). Consequence: \(W_2 \to 0\) in finite time, which is only possible if the class \(\alpha\) vanishes, i.e. the corresponding \(S^2\) in the prime decomposition has been encapsulated by surgery.

2.2 The functional \(W_3\) (for \(\pi_3(M) \neq 0\))

If \(\pi_2(M) = 0\) but \(\pi_1(M) = 0\) and \(M\) is 3-dimensional, the Hurewicz sequence yields \(\pi_3(M) \cong H_3(M; \mathbb{Z}) \cong \mathbb{Z}\). Instead of 2-spheres one then considers families of 2-spheres, parametrized by \(S^1\) (continuous loops in the space of maps \(S^2 \to M\) with constant initial and terminal map). Such families represent classes in \(\pi_1(\Lambda M, *) \cong \pi_3(M)\).

For a non-trivial class \(\beta \in \pi_3(M)\) define: $$ W_3(g, \beta) := \inf_{[\gamma] = \beta}\ \max_{s \in S^1} \mathrm{Area}_g(\gamma(s)). $$

Lemma (Perelman 0307245, §3; Colding–Minicozzi 2005). $$ \frac{d}{dt} W_3(g(t), \beta) \le -4\pi - \tfrac{1}{2} R_{\min}(t)\, W_3(g(t), \beta). $$

The inequality is structurally identical to the one for \(W_2\): it follows from a min-max construction and the Gauss–Bonnet theorem. Again it forces \(W_3 \to 0\) in finite time.

3. Main theorem and corollary

Theorem (Perelman 0307245, Theorem 1.1). Let \(M\) be a closed oriented 3-manifold without aspherical or infinite-\(\pi_1\) factors in its prime decomposition. Then the Ricci flow with surgery from Article 03 becomes extinct in finite time for every initial metric.

Corollary (the case relevant to the proof). If \(M\) is simply connected, the flow becomes extinct in finite time.

Sketch of the corollary: from \(\pi_1(M) = 0\) and Hurewicz one obtains \(\pi_3(M) \neq 0\) (\(\cong \mathbb{Z}\) for a sphere, \(\cong \mathbb{Z}^k\) for a connected sum). The \(W_3\)-inequality forces finite extinction: the flow cannot exist for all times without resolving every non-trivial class in \(\pi_2\) and \(\pi_3\) via surgery. Since the algorithm discards every such resolution as a spherical space form, nothing remains in the end.

4. Two independent proofs

Author Preprint Strategy
Perelman (2003) arXiv:math/0307245, 7 pp. min-max over loops, Gauss–Bonnet, \(W_3\)
Colding & Minicozzi (2005) arXiv:math/0308090, 16 pp. Birkhoff min-max, harmonic replacement

Both use the idea "2-spheres shrink under Ricci flow because of Gauss–Bonnet"; Colding & Minicozzi provide a fully worked-out version that today serves as the standard reference (Annals of Math. 2008).

5. What extinction does not deliver

Finite extinction only detects that the manifold consists of spherical space forms. It yields no structural statement about hyperbolic or Seifert-fibered pieces – for that one still needs the thin–thick decomposition from Article 04. For the full geometrization the extinction argument is not sufficient.

Conjecture needs extinction? needs thin–thick?
Poincaré (simply connected) yes (shortcut) no
Spherical space-form (\(\pi_1\) finite) yes no
Full geometrization no yes

6. Which obstacles fall now

Obstacle (cf. Article 01) Resolution
O5: read topology off the limit, especially for simply connected \(M\) \(\pi_3 \neq 0 \Rightarrow\) \(W_3\)-monotonicity \(\Rightarrow\) finite extinction
O3' (variant): rule out infinitely many surgery times \(W_3 \to 0\) in finite time terminates the algorithm

Cross-references

Sources

  • Perelman, G. (2003). Finite extinction time for the solutions to the Ricci flow on certain three-manifolds. arXiv:math/0307245.
  • Colding, T. H. & Minicozzi, W. P. (2005). Estimates for the extinction time for the Ricci flow on certain 3-manifolds and a question of Perelman. J. Amer. Math. Soc. 18, 561–569. arXiv:math/0308090.
  • Colding, T. H. & Minicozzi, W. P. (2008). Width and finite extinction time of Ricci flow. Geom. Topol. 12, 2537–2586.
  • Sacks, J. & Uhlenbeck, K. (1981). The existence of minimal immersions of 2-spheres. Ann. of Math. 113, 1–24.
  • Morgan, J. & Tian, G. (2007). Ricci Flow and the Poincaré Conjecture. AMS, Chapter 18 – worked-out extinction proof.
  • Kleiner, B. & Lott, J. (2008). Notes on Perelman's papers. Geom. Topol. 12, §§94–96.