Uniqueness of tangent cones for semicalibrated integral 2-cycles (Q971951)

The main purpose of this paper is to introduce the concept of semicalibration \(\omega\) on a closed \(m\)-dimensional Riemannian manifold \((M,g)\) and to show a regularity property for tangent cones for semicalibrated integral cycles. Such semicalibration is a \(k\)-form \(\omega\) of comass one, defined as in case of a calibration, but with no need to be closed. The notion of smooth calibrated submanifold is extended to the non-smooth case by defining a semicalibrated submanifold as an integral rectifiable \(k\)-dimensional current \(C\), a linear functional acting on the space of smooth \(k\)-forms with compact support, such that on each open set \(U\) of \(M\), \(\langle C \llcorner U,\omega \rangle =M(C \llcorner U)\), where the mass \(M(C)\) extends the notion of volume of a submanifold. Considering dilatations of a semicalibrated current \(C\) under the maps \((x-x_0)/r\), with \(r\to 0\), by an almost monotonicity formula derived by the authors, the mass of \((x-x_0)/r)_*C \llcorner B_1(x_0)\) (that equals \(r^kM(C\llcorner B_r(x_0))\)) is uniformly bounded, and applying a compacteness theorem of Federer and Fleming, they conclude that there exists a sequence \(r_m\to 0\) for which the sequence of currents \((x-x_0)/{r_m})_*C \llcorner B_1(x_0)\) converges weakly to a semicalibrated \(k\)-current \(C_{\infty, x_0}\) (the tangent cone to \(C\) at \(x_0\)). The proof consists in locally expressing \(\omega(x)\) as a perturbation of a constant form \(\omega_0\), by defining it as a sum \(\omega(x)=\omega_0(x)+\omega_1(x)\), and to show that for small radii the perturbation terms are not significant. In Theorem 1 the authors prove that this blow-up limit is unique (for each \(x_0\)) when \(k=2\) and \(\omega\) is \(C^2\) but not necessarily closed. The uniqueness, in case \(k=2\), is obtained by recalling the structure of a tangent cone calibrated by a constant 2-dimensional calibration \(\omega_0\), that is the Kähler form associated to an orthogonal complex structure of \(\mathbb{R}^{2n}\). As an application of this theorem on obtains the uniqueness of the tangent cone for special Legendrian integral cycles in \(\mathbb{S}^5\). The authors generalize a estimate due to \textit{B. White} [Duke Math. J. 50, 143--160 (1983; Zbl 0538.49030)] for the case \(C\) is an integral area minimizing 2-cycle, but using a completely different approach. They prove in Theorem 2 that there exist constants \(r_0>0\), \(C_1>0\) and \(\gamma\in (0,1]\) such that for any integer-rectifiable 2-cycle, \(\omega\)-semicalibrated \(C\) and any \(0<r<r_0\), one has \(M(C\llcorner B_r(x_0))/(\pi r^2) -\Theta(\|C\|, x_0)\leq C_1 r^{\gamma}\), where \(\Theta(\|C\|,x_0)\) is the density of \(C\) at \(x_0\). They use a first-order elliptic problem (that comes from the definition of semicalibrated submanifold), introducing a suitable new technique, and not a second-order one as in the case of area-minimizing current, and they believe they can extend this approach to higher dimensions. ``Furthermore, we show that semicalibrated 2-cycles are essentially semicalibrated by a constant 2-form and hence almost holomorphic cycles in some \(\mathbb{C}^m\)'', and analyzing the restriction of such cycles to small balls around \(x_0\), that can be pushed forward onto \(\mathbb{CP}^{m-1}\), ``show that their support cannot fill all of \(\mathbb{CP}^{m-1}\)'', giving a (well defined) current of small mass ``... we are able to estimate the mass of the dilated current viewed as a function of the dilating radius \(r\) in terms of its derivative. From a suitable choice of radii, this is in turn estimated by the mass of the dilated current albeit with a gap proportional to \(r\) around \(r_0\). Filling the hole and applying a standard iteration argument, we obtain the theorem''. A perturbation argument proves the theorem for \(\omega\) non-closed.