Nonexistence of Minimizing Harmonic Maps from $2$-Spheres
A FUTAKI
DOI: https://doi.org/10.3792/pjaa.56.291
1980-01-01
Abstract:A harmonic map is, by definition, a critical point of the functional E. A harmonic map is said to be minimizing if it minimizes energy in its connected component of C=(N, M), i.e. in its homotopy class. When dim N= 1, N is a circle S and a harmonic map f" S--M is a closed geodesic. It is well known that every component of C=(S, M) contains a minimizing closed geodesic. In contrast with this, when dim N=2, it is not always true that there exists a minimizing harmonic map in each component of C=(N, M). For instance there exists no minimizing harmonic map of degree +1 from a Riemann surface of genus =>1 to a Riemann sphere whatever metrics are chosen on them ([5]). On the other hand, Sacks and Uhlenbeck [8] established an existence result when N=S. Their result was applied to the proof of Frankel’s conjecture by Siu and Yau [9] and to the study on the topology of 3-manifolds by Meeks and Yau [7]. The following is a result of Sacks and Uhlenbeck refined by Siu and Yau. Let M be a compact 1-connected Riemannian manifold. Let foe C=(S, M). Then there exist minimizing harmonic maps f, ., f e C=(S, M) such that , f,=fo in re(M) and that i=l (1.2) , E(f,)= inf{ E(g,)[p e N, g,= fo in m.(M)}. i=1 i=l i=l