sign in sign up
<<
Home , Mathematical analysis

arXiv:1805.02216v2 [math.DG] 30 Aug 2018 eety hr a ena nesv td ftesupersymm the of study intensive fun an wi action been spaces has the target there in Recently, to [2]. two-form in maps additional studied Dirac-harmonic an were to account led [7]. into is also Taking one and spinors the 11] additional in [3, contains quartic action is criti full that the Its term of model. part capture nonlinear this only of maps points Dirac-harmonic fields. critical However, spinor The with . map the Dir the along and defined Riemannian is two between map maps a for Dirac-harmonic action for action The introduced. h uesmercnonlinear supersymmetric The hsmteaia td a ntae n[2,weetenot the where [12], in initiated was version. mathematical study geome this mathematical the in This from interest growing tools t a well-established language is using the under studied invariance in be the formulated also neglecting usually when is supersymm varihand, and it by domain dictated literature the is on functional diffeomorphisms action under its invariance of form precise The . rva,maigta h a atmp oapitadtevect the and point a name to maps direction, supers part the opposite map of the the points that critical in meaning the trivial, aiming of are solutions fo that we [17]. result ensuring article in existence established this an be In method could heat-flow boundary with the surface Dira Employing for achieved. [1]. the be in of could solutions uncoupled result theorem existence index several maps, harmonic o h ipetmteaia eso ftesupersymmetri rem the the of as version such mathematical simplest u properties invariant the special For is share s action points is the critical model case its sigma this supersymmetric in since the two-dimensional sup of the functional of action on The [15]. mathematical survey the the the of on see superpartner overview the recent is a which For [16] gravitino a to e od n phrases. and words Key 2010 Date AIHN EUTFRTESPRYMTI NONLINEAR SUPERSYMMETRIC THE FOR RESULT VANISHING A uut3,2018. 31, August : ahmtc ujc Classification. Subject Mathematics igrta w n htacraneeg ssffiinl small sufficiently is t energy certain that a assuming that inequality, and two Sobolev than type o bigger manifolds Euclidean Riemannian an non-compact admit complete on model sigma Abstract. epoeavnsigrsl o rtclpit ftesupers the of points critical for result vanishing a prove We IM OE NHGE DIMENSIONS HIGHER IN MODEL SIGMA uesmercnniersgamdl opeeRiemannia complete model; sigma nonlinear supersymmetric 1. σ mdli nipratmdli oenqatmfil theory. field quantum modern in model important an is -model nrdcinadresults and Introduction OKRBRANDING VOLKER 82;5C7 35J61. 53C27; 58E20; 1 ia-amncmp ihcraueterm curvature with maps Dirac-harmonic mercnonlinear ymmetric vlo sltdsingularities. isolated of oval drcnomltasomtosand transformations conformal nder em:Cnieigas curvature a also Considering terms: ossso h elsuidDirichlet well-studied the of consists -amncmp ol eobtained be could maps c-harmonic rccluu fvrainadthere and variation of tric . u ymtisa o xml the example for as symmetries ous nonlinear c edmnino h oanis domain the of dimension he fspremty nteother the On supergeometry. of aigueo h Atiyah-Singer the of use Making a onso h supersymmetric the of points cal lntrlycul h harmonic the couple naturally al oiieRcicrauethat curvature Ricci positive f o of ion h domain. the eili h oanmnfl is domain the if pecial ti nonlinear etric cato o a for action ac asomtos hsmdlcan model this ransformations, rymti nonlinear ersymmetric yw att ieacriterion a give to want we ly h oanbigacompact a being domain the r toa h eutn equations resulting the ctional htrinaeaaye n[4]. in analyzed are torsion th rsio aihsidentically. vanishes spinor or tytasomtos nthe In transformations. etry ia-amncmaps Dirac-harmonic mercnonlinear ymmetric σ mdl hti Dirac- is that -model, aiod Liouville manifold; n etrspinor vector σ σ mdlms be must -model mdlcoupled -model σ -model see , that was 2 VOLKER BRANDING

Let us now describe the mathematical structure that we are using in more detail. In the following we will assume that (M, g) is a complete non-compact Riemannian spin manifold of dimension n = dim M ≥ 3 and (N, h) a second complete Riemannian manifold of bounded . Whenever we will make use of indices we use Latin letters for indices related to M and Greek letters for indices related to N. The spin assumption guarantees the existence of the (complex) spinor bundle ΣM and sections in this bundle are called spinors. The spinor bundle is a vector bundle over the manifold M that is equipped with a connection ∇ΣM and a hermitian scalar product. On the spinor bundle there exists the of Clifford multiplying a spinor with a tangent vector, which is skew-symmetric

hX · ψ,ξiΣM = −hψ, X · ξiΣM for all X ∈ T M and ψ,ξ ∈ Γ(ΣM). Moreover, the Clifford relations X · Y · ψ + Y · X · ψ = −2g(X,Y )ψ hold for all X,Y ∈ T M and ψ ∈ Γ(ΣM). Let φ: M → N be a map and let φ∗TN be the pull-back of the tangent bundle from N. We consider the twisted bundle ΣM ⊗ φ∗TN, on this bundle we obtain a connection induced from ΣM and φ∗TN, which will be denoted by ∇˜ . Sections in ΣM ⊗ φ∗TN are called vector spinors. On ΣM ⊗ φ∗TN we have a scalar product induced from ΣM and φ∗TN, we will denote its real part by h·, ·i. The twisted Dirac operator acting on vector spinors is defined by n ˜ D/ := ei · ∇ei , Xi=1 where ei, i = 1,..., dim M is an orthonormal basis of T M. Note that the operator D/ is elliptic. Moreover, we assume that the connection on φ∗TN is metric, thus D/ is also self-adjoint with respect to the L2- if M is compact. In this article we want to focus on the system of partial differential equations that arises as critical points of the full supersymmetric nonlinear sigma model without referring to its concrete structure, see [5] for a previous analysis. The critical points satisfy a coupled system of the following form τ(φ)=A(φ)(dφ, dφ)+ B(φ)(dφ, ψ, ψ)+ C(φ)(ψ,ψ,ψ,ψ), (1.1) Dψ/ =E(φ)(dφ)ψ + F (φ)(ψ, ψ)ψ. (1.2) ∗ Here, τ(φ) := trg∇dφ ∈ Γ(φ TN) denotes the tension field of the map φ and the other terms represent the analytical structure of the right hand side. We will always assume that the endomorphisms A,B,C,E and F are bounded. The structure of this system is motivated by the scaling behavior of φ and ψ. More precisely, it is the most general system one can build when ones gives weight 0 to the map part and weight 1/2 to the vector spinors. The system (1.1), (1.2) includes the cases of Dirac-harmonic maps, which are solutions of n 1 τ(φ)= RN (ψ, e · ψ)dφ(e ), 2 i i Xi=1 Dψ/ = 0 and also Dirac-harmonic maps with curvature term, which are solutions of n 1 1 τ(φ)= RN (ψ, e · ψ)dφ(e ) − h(∇RN )♯(ψ, ψ)ψ, ψi, 2 i i 12 Xi=1 1 Dψ/ = RN (ψ, ψ)ψ. (1.3) 3 Here, RN denotes the curvature on N. Note that if we consider magnetic Dirac-harmonic maps [2] then A 6= 0 and in the case of Dirac-harmonic maps with torsion [4] we have E 6= 0. A VANISHING RESULT FOR THE SUPERSYMMETRIC NONLINEAR SIGMA MODEL 3

Recently, an existence result for a system of the form (1.1), (1.2) could be obtained in the case that the domain manifold is a Lorentzian spacetime that expands sufficiently fast [10]. In this article we will give a vanishing result for finite energy solutions of the system (1.1), (1.2) on higher-dimensional complete Riemannian manifolds. In the proof of our main result we will apply an Euclidean type of the following form −2 2n/(n−2) n 2 ( |u| dV ) n ≤ C2 |∇u| dV (1.4) ZM ZM 1,2 for all u ∈ W (M) with compact , where C2 is a positive constant that depends on the geometry of M. Such an inequality holds in Rn and is well-known as Gagliardo-Nirenberg inequality. However, if one considers a non-compact complete Riemannian manifold of infinite one has to make additional assumptions to ensure that an equality of the form (1.4) holds. One way of formulating these assumptions is to introduce the notion of a positive Green’s function on complete non-compact Riemannian manifolds. Let Ω be a bounded open of M such that x ∈ Ω and G be the solution of

∆g = δx in Ω, (G =0 on ∂Ω. Ω Ω We Gx (y) = G(y) when y ∈ Ω and Gx (y) = 0 otherwise. In addition, we set Gx(y) = Ω supΩ s.t. x∈Ω Gx (y),y ∈ M. Then we know that [14, Theorem 3.28], (1) either Gx(y)=+∞, ∀y ∈ M, (2) or Gx(y) < ∞, ∀y ∈ M \ {x}, where Gx is the positive Green’s function of pole x. In the first case the manifold M is called parabolic, in the second case it is called non-parabolic. At this point we may state [14, Theorem 3.29]. Theorem 1.1. Let (M, g) be a complete Riemannian manifold of dimension n that has infinite volume. Then the following two statements are equivalent (1) The Euclidean type Sobolev inequality (1.4) holds. (2) The manifold (M, g) is non-parabolic and there exists K > 0 such that for any x ∈ M and any t> 0 −n/(n−2) Volg {y ∈ M s.t. Gx(y) >t} ≤ Kt , where Gx is the positive minimal Green’s function of pole x. Another way of ensuring that (1.4) holds is the following: If (M, g) is a complete, non-compact Riemannian manifold of dimension n with nonnegative Ricci curvature, and if for some point x ∈ M Volg(BR(x)) lim n > 0 R→∞ ωnR n holds, then (1.4) holds true, see [19]. Here, ωn denotes the volume of the unit ball in R . For further geometric conditions ensuring that (1.4) holds we refer to [14, Section 3.7]. In this article we will give the following result: Theorem 1.2. Let (M, g) be a complete and non-compact Riemannian manifold of dimension dim M = n > 2 with positive Ricci curvature that admits an Euclidean type Sobolev inequality of the form (1.4). Moreover, let N be a Riemannian manifold of bounded geometry. Assume that (φ, ψ) is a smooth solution of the system (1.1), (1.2). If

(|dφ|n + |ψ|2n) dV < ε (1.5) ZM with ε> 0 sufficiently small, then φ must be trivial and ψ vanishes identically. Note that due to a classical result of Yau [21] a complete non-compact Riemannian manifold with nonnegative Ricci curvature has infinite volume. 4 VOLKER BRANDING

Remark 1.3. In one is interested in supersymmetric sigma models with a two- dimensional domain since one is looking for a “generalized” surface of minimal in a given target . However, supersymmetric sigma models also exist for domain manifolds with dimension bigger than two, see for example [13, Chapter 3]. Although we are neglecting su- persymmetry in our analysis our main result Theorem 1.2 may serve as a guideline to derive vanishing results for the supersymmetric models from physics.

Remark 1.4. The small number ε in condition (1.5) in Theorem 1.2 can also be made explicit, it has to satisfy the inequality

n 2 2 2n − 2 2 1+ n 2 3 2 n N 2 −1 ε ≤ max{ n|F | ∞ + (3|C| ∞ + |B| ∞ )+ (|R | ∞ + |E| ∞ ) , n2C L 1+ n/2 L 2 L 2+ n L L  s  2n − 4 1+ n 3 −2+ n 1+ n 3  N ∞ 2 ∞ 2 ∞ 2 ∞ 2 ∞ 2 |R |L + (3|A|L + |B|L )+ (3|C|L + |B|L ) n Cs 2 2 2+ n 2 2  n  2 n N 2 −1 + (|R | ∞ + n|E| ∞ )] }. 1+ n/2 L L  Moreover, we want to mention that demanding the smallness of the Ln-norms of dφ and |ψ|2 is a natural condition since these norms scale in the correct way.

Remark 1.5. If one considers the system (1.1), (1.2) for the domain being a closed surface several vanishing results similar to Theorem 1.2 have been obtained. More precisely, suppose that (φ, ψ) is a smooth solution of the system (1.1), (1.2) with small energy, that is

(|dφ|2 + |ψ|4) dV < ε. ZM (1) If (φ, ψ) is a Dirac-harmonic map (A = C = E = F = 0) then φ is constant and ψ is a standard harmonic spinor [12, Proposition 4.2]. (2) If (φ, ψ) is a Dirac-harmonic map with curvature term (A = E = 0) and there do not exist harmonic spinors on M then φ is constant and ψ vanishes identically, see [3, Lemma 4.9] and [9, Corollary 3.7].

This article is organized as follows: In section 2 we present the proof of the main result and in the last section we make a short comment on Dirac-harmonic maps with curvature term that are also critical with respect to the domain metric.

2. Proof of theorem 1.2 Our method of proof is inspired from a global pinching lemma established in [22] and a recent result on biharmonic maps between complete Riemannian manifolds [8]. Throughout the proof we will employ the usual summation convention, that is we will sum over repeated indices. We will make use of a cutoff function 0 ≤ η ≤ 1 on M that satisfies C η(x)=1 for x ∈ B (x ), η(x)=0 for x ∈ B (x ), |∇η|≤ for x ∈ M, R 0 2R 0 R where BR(x0) denotes the geodesic ball around x0 with radius R. Moreover, we employ the Weitzenb¨ock formula for the twisted Dirac-operator D/, that is 1 1 D/ 2ψ = −∆˜ ψ + SM ψ + e · e · RN (dφ(e ), dφ(e ))ψ. (2.1) 4 2 i j i j Here, ∆˜ denotes the connection Laplacian on ΣM ⊗ φ∗TN, SM denotes the scalar curvature on M and RN is the curvature tensor on N. This formula can be deduced from the general Weitzenb¨ock formula for twisted Dirac operators, see [18, Theorem II.8.17]. A VANISHING RESULT FOR THE SUPERSYMMETRIC NONLINEAR SIGMA MODEL 5

Lemma 2.1. Suppose that ψ ∈ Γ(ΣM ⊗ φ∗TN) is a smooth solution of (1.2) and fix q > 2. Then the following inequality holds

C q 1 2 ˜ 2 q−2 2 ˜ 2 q−4 2 |ψ| dV ≥( − δ1 − δ2) η |∇ψ| |ψ| dV + (q − 2) η |h∇ψ, ψi| |ψ| dV R M 2 M M Z Z Z (2.2) 1 + η2SM |ψ|q dV − C η2|ψ|q|dφ|2 dV − C η2|ψ|q+4 dV, 4 1 2 ZM ZM ZM where δ1, δ2 are two positive numbers that may be chosen arbitrarily and

2 2 N n|E|L∞ n|F |L∞ C1 := n|R |L∞ + , C2 := . 4δ1 4δ2 Proof. Combining (1.2) and (2.1) we obtain

∆˜ ψ = − E(φ)(dφ)Dψ/ − F (φ)(ψ, ψ)Dψ/ −∇(E(φ)(dφ)) · ψ −∇(F (φ)(ψ, ψ)) · ψ 1 1 + SM ψ + e · e · RN (dφ(e ), dφ(e ))ψ. 4 2 i j i j

Testing this equation with η2|ψ|q−2ψ and integrating over M we find

1 η2h∆˜ ψ, ψi|ψ|q−2 dV = η2SM |ψ|q dV 4 ZM ZM 1 + η2|ψ|q−2he · e · RN (dφ(e ), dφ(e ))ψ, ψi dV 2 i j i j ZM − η2|ψ|q−2hE(φ)(dφ)Dψ,ψ/ i dV ZM − η2|ψ|q−2hF (φ)(ψ, ψ)Dψ,ψ/ i dV. ZM Note that the terms involving of E(φ) and F (φ) vanish due to the skew-symmetry of the Clifford multiplication. Moreover, using and the properties of the cutoff function η, we find

η2h∆˜ ψ, ψi|ψ|q−2 dV = − η2|∇˜ ψ|2|ψ|q−2 dV − (q − 2) η2|h∇˜ ψ, ψi|2|ψ|q−4 dV ZM ZM ZM − 2 h∇˜ ψ, ψi|ψ|q−2η∇η dV ZM C 1 ≤ |ψ|q dV − η2|∇˜ ψ|2|ψ|q−2 dV R2 2 ZM ZM − (q − 2) η2|h∇˜ ψ, ψi|2|ψ|q−4 dV. ZM Combining both equations we get

C 1 |ψ|q dV ≥ η2|∇˜ ψ|2|ψ|q−2 dV + (q − 2) η2|h∇˜ ψ, ψi|2|ψ|q−4 dV R2 2 ZM ZM ZM 1 1 + η2SM |ψ|q dV + η2|ψ|q−2he · e · RN (dφ(e ), dφ(e ))ψ, ψi dV 4 2 i j i j ZM ZM − η2|ψ|q−2hE(φ)(dφ)Dψ,ψ/ i dV − η2|ψ|q−2hF (φ)(ψ, ψ)Dψ,ψ/ i dV. ZM ZM 6 VOLKER BRANDING

The result follows by using the inequalities

2 2 −n|E|L∞ 2 2 −hE(φ)(dφ)Dψ,ψ/ i≥−δ1|∇˜ ψ| − |dφ| |ψ| , 4δ1 2 2 n|F |L∞ 6 −hF (φ)(ψ, ψ)Dψ,ψ/ i≥−δ2|∇˜ ψ| − |ψ| , 4δ2 N N 2 2 hei · ej · R (dφ(ei), dφ(ej ))ψ, ψi≥−n|R |L∞ |dφ| |ψ| , where δ1, δ2 > 0 may be chosen arbitrarily. 

As a next step we derive a similar inequality for the differential of the map φ.

Lemma 2.2. Suppose that φ: M → N is a smooth solution of (1.1) and fix r > 2. Then the following inequality holds

C 1 |dφ|r dV ≥ η2hdφ(RicM (e )), dφ(e )i|dφ|r−2 dV + η2|∇dφ|2|dφ|r−2 dV (2.3) R2 i i 2 ZM ZM ZM r − 2 + η2|h∇dφ, dφi|2|dφ|r−4 dV 2 ZM 2 r+2 2 8 r−2 − C3 η |dφ| dV − C4 η |ψ| |dφ| dV, ZM ZM where

N 1+ r 2 3 2 1+ r 2 3 2 C :=|R | ∞ + (3|A| ∞ + |B| ∞ ), C := (3|C| ∞ + |B| ∞ ). 3 L 2 L 2 L 4 2 L 2 L Proof. Recall that for a map between two Riemannian manifolds the following Bochner formula holds ([20, Proposition 1.34])

M N (∆dφ)(ei)= dφ(Ric (ei)) + R (dφ(ej ), dφ(ei))dφ(ej )+ ∇ei τ(φ).

2 r−2 Testing this equation with η |dφ| dφ(ei) and integrating over M we find

2 r−2 2 M r−2 η h∆dφ, dφi|dφ| dV = η hdφ(Ric (ei)), dφ(ei)i|dφ| dV ZM ZM 2 N r−2 + η hR (dφ(ei), dφ(ej ))dφ(ei), dφ(ej )i|dφ| dV ZM + η2h∇τ(φ), dφi|dφ|r−2 dV. ZM Using integration by parts we may rewrite

η2h∇τ(φ), dφi|dφ|r−2 dV = − 2 η∇ηhτ(φ), dφi|dφ|r−2 dV − η2|τ(φ)|2|dφ|r−2 dV ZM ZM ZM − (r − 2) η2hτ(φ), dφi|dφ|r−4hdφ, ∇dφi dV ZM and also

η2h∆dφ, dφi|dφ|r−2 dV = − 2 η∇ηh∇dφ, dφi|dφ|r−2 dV − η2|∇dφ|2|dφ|r−2 dV ZM ZM ZM − (r − 2) η2|h∇dφ, dφi|2|dφ|r−4 dV. ZM A VANISHING RESULT FOR THE SUPERSYMMETRIC NONLINEAR SIGMA MODEL 7

This allows us to deduce the following inequality

2 M r−2 2 2 r−2 η hdφ(Ric (ei)), dφ(ei)i|dφ| dV + η |∇dφ| |dφ| dV ZM ZM + (r − 2) η2|h∇dφ, dφi|2|dφ|r−4 dV ZM =2 η∇ηhτ(φ), dφi|dφ|r−2 dV − 2 η∇ηh∇dφ, dφi|dφ|r−2 dV ZM ZM 2 N r−2 2 2 r−2 − η hR (dφ(ei), dφ(ej ))dφ(ei), dφ(ej )i|dφ| dV + η |τ(φ)| |dφ| dV ZM ZM + (r − 2) η2hτ(φ), dφi|dφ|r−4hdφ, ∇dφi dV ZM C 1 1+ r ≤ |dφ|r dV + η2|∇dφ|2|dφ|r−2 dV + η2|τ(φ)|2|dφ|r−2 dV R2 2 2 ZM ZM ZM r − 2 − + |RN | ∞ η2|dφ|r+2 dV + η2|h∇dφ, dφi|2|dφ|r 4 dV, L 2 ZM ZM which yields

1 η2hdφ(RicM (e )), dφ(e )i|dφ|r−2 dV + η2|∇dφ|2|dφ|r−2 dV i i 2 ZM ZM r − 2 + η2|h∇dφ, dφi|2|dφ|r−4 dV 2 ZM C r 1+ r 2 2 r−2 N 2 r+2 ≤ |dφ| dV + η |τ(φ)| |dφ| dV + |R | ∞ η |dφ| dV. R2 2 L ZM ZM ZM Moreover, making use of (1.1) we may estimate

2 2 3 2 4 2 3 2 8 |τ(φ)| ≤ (3|A| ∞ + |B| ∞ )|dφ| + (3|C| ∞ + |B| ∞ )|ψ| . L 2 L L 2 L The result follows by combining both equations. 

We may combine Lemmas 2.1 and 2.2 to obtain the following

Lemma 2.3. Suppose that the pair (φ, ψ) is a smooth solution of (1.1), (1.2). Then the fol- lowing inequality holds

C 1 (|ψ|4p + |dφ|2p) dV ≥ η2hdφ(RicM (e )), dφ(e )i|dφ|2p−2 dV + η2SM |ψ|4p dV R2 i i 4 ZM ZM ZM 2p − 1 p − 1 + η2 d|ψ|2p 2 dV + η2 d|dφ|p 2 dV 2p2 p2 ZM ZM 1 + ( − δ − δ ) η2| ∇˜ ψ|2|ψ|4p−2 dV 2 1 2 ZM 1 + η2|∇dφ|2|dφ|2p−2 dV 2 ZM 2C pC − (C + 4 + 1 ) η2|ψ|4p+4 dV 2 1+ p 1+ p ZM −1+ p C − (C + C + 1 ) η2|dφ|2p+2 dV, (2.4) 3 4 1+ p 1+ p ZM where p> 1. 8 VOLKER BRANDING

Proof. First, we add up (2.2) with q = 4p and (2.3) with r = 2p to get the following inequality C 1 (|ψ|4p + |dφ|2p) dV ≥ η2hdφ(RicM (e )), dφ(e )i|dφ|2p−2 dV + η2SM |ψ|4p dV R2 i i 4 ZM ZM ZM + (4p − 2) η2|h∇˜ ψ, ψi|2|ψ|4p−4 dV ZM 1 + ( − δ − δ ) η2|∇˜ ψ|2|ψ|4p−2 dV 2 1 2 ZM + (p − 1) η2|h∇dφ, dφi|2|dφ|2p−4 dV ZM 1 + η2|∇dφ|2|dφ|2p−2 dV 2 ZM 2 4p 2 2 4p+4 − C1 η |ψ| |dφ| dV − C2 η |ψ| dV ZM ZM 2 2p+2 2 8 2p−2 − C3 η |dφ| dV − C4 η |ψ| |dφ| dV. ZM ZM In order to estimate the terms on the right hand side containing both ψ and dφ we use the general Young inequality to obtain 2 −1+ p |ψ|8|dφ|2p−2 ≤ |ψ|4p+4 + |dφ|2p+2, 1+ p 1+ p 1 p |dφ|2|ψ|4p ≤ |dφ|2p+2 + |ψ|4p+4. 1+ p 1+ p In addition, we have 1 1 |h∇˜ ψ, ψi|2|ψ|4p−4 = d|ψ|2 2|ψ|4p−4 = d|ψ|2p 2, 4 4p2 1 1 |h∇dφ, dφi|2|dφ|2p−4 = d|dφ|2 2|dφ|2p−4 = d|dφ|p 2. 4 p2  Combining all the estimates then yields the claim. Making use of (1.4) we can now give the following estimate: Lemma 2.4. Let (M, g) be a complete non-compact Riemannian manifold of infinite volume that admits an Euclidean type Sobolev inequality. Assume that f is a positive function on M. For dim M = n> 2 and p> 1 arbitrary the following inequality holds 2 1 η2f 2p+2 dV ≤ C f n dV n f 2p dV + η2|df p|2 dV , (2.5) s R2 ZM ZM ZM ZM where the positive constant Cs depends on n,p and the geometry of M.  Proof. By H¨older’s inequality we find

1 2r r−1 η2f 2p+2 dV ≤ (ηf p)2r dV r f r−1 dV r . ZM ZM ZM n   Now, we choose r = n−2 and apply (1.4) to obtain

2n n−2 p −2 n p 2 (ηf ) n dV ≤ C2 |d(ηf )| dV. ZM ZM Hence, we find  2 η2f 2p+2 dV ≤ C f n dV n |d(ηf p)|2 dV. ZM ZM ZM Using the properties of the cutoff functionη we find  C |d(ηf p)|2 dV ≤ f 2p dV + 2 η2|df p|2 dV, R2 ZM ZM ZM A VANISHING RESULT FOR THE SUPERSYMMETRIC NONLINEAR SIGMA MODEL 9 which completes the proof.  At this point we are ready to give the proof of the main result.

1 Proof of Theorem 1.2. We make use of the inequality (2.4), where we choose δ1 = δ2 = 4 . Moreover, we apply (2.5) to f = |dφ|, |ψ|2 taking into account the smallness condition (1.5). Choosing p = n/2 we obtain the following inequality C (|ψ|2n + |dφ|n) dV R2 ZM 1 ≥ η2hdφ(RicM (e )), dφ(e )i|dφ|n−2 dV + η2SM |ψ|2n dV i i 4 ZM ZM 2 2n − 2 2C4 nC1 2 n 2 + − ε n C (C + + ) η d|ψ| dV n2 s 2 1+ n/2 2+ n   ZM 2 2n − 4 −2+ n C1 2 n 2 + − ε n C (C + C + ) η d |dφ| 2 dV n2 s 3 4 2+ n 1+ n/2   ZM 2 2C4 nC1 1 2n − ε n C (C + + ) |ψ| dV s 2 1+ n/2 2+ n R2 ZM 2 −2+ n C1 1 n − ε n C (C + C + ) |dφ| dV. s 3 4 2+ n 1+ n/2 R2 ZM Choosing ε small enough and taking the R →∞ we obtain 1 0 ≥ hdφ(RicM (e )), dφ(e )i|dφ|n−2 dV + SM |ψ|2n dV. i i 4 ZM ZM By assumption M has positive Ricci curvature and thus RicM becomes a positive definite non- degenerate bilinear form on T M. Moreover, the positivity of the Ricci curvature also implies that the scalar curvature is positive, which completes the proof. 

3. A remark on Dirac-harmonic maps with curvature term in higher dimensions In this section we want to make a comment on Dirac-harmonic maps with curvature term from domain manifolds with dim M ≥ 3. The action functional for Dirac-harmonic maps with curvature term is given by 1 1 S(φ, ψ)= (|dφ|2 + hψ, Dψ/ i− hRN (ψ, ψ)ψ, ψi) dV. (3.1) 2 6 ZM Here, the indices are contracted as N α γ β δ hR (ψ, ψ)ψ, ψi = Rαβγδhψ , ψ ihψ , ψ i, α,β,γ,δX which ensures that the functional is real-valued. The critical points of the action functional (3.1) are given by n 1 1 τ(φ)= RN (ψ, e · ψ)dφ(e ) − h(∇RN )♯(ψ, ψ)ψ, ψi, (3.2) 2 i i 12 Xi=1 1 Dψ/ = RN (ψ, ψ)ψ, (3.3) 3 where ♯: φ∗T ∗N → φ∗TN represents the musical . For a derivation of the critical points see [11, Section II] and [3, Proposition 2.1]. Solutions (φ, ψ) of the system (3.2), (3.3) are called Dirac-harmonic maps with curvature term. It is obvious that Dirac-harmonic maps with curvature term have the analytical structure of the system (1.1), (1.2). In physics one often considers the metric on the domain of the action functional (3.1) as a dynamic field of the theory. For this reason one also varies the action functional with respect 10 VOLKER BRANDING to the domain metric. As the corresponding Euler-Lagrange equation one gets the vanishing of the energy-momentum tensor 1 1 0= S =2hdφ(e ), dφ(e )i− g |dφ|2 + hψ, e · ∇˜ ψ + e · ∇˜ ψi− g hRN (ψ, ψ)ψ, ψi. ij i j ij 2 i ej j ei 6 ij (3.4) For a derivation of the energy-momentum tensor see [6, Lemma 5.3]. In the mathematics literature one calls an action functional that is critical with respect to the domain metric stationary. A vanishing result for stationary Dirac-harmonic maps with curvature term with finite energy from complete Riemannian manifolds has been obtained in [6, Theorem 5.4]. For a smooth Dirac-harmonic map with curvature term that is also critical with respect to the domain metric, that is a solution of the system (3.2), (3.3), (3.4), we obtain the following vanishing result Theorem 3.1. Let the triple (φ, ψ, g) be a smooth Dirac-harmonic map with curvature term that is critical with respect to the domain metric. Suppose that there do not exist harmonic spinors on M, that is solutions of ∂/Ψ = 0, where ∂/ is the standard Dirac operator and Ψ ∈ Γ(ΣM). If dim M ≥ 3 and the target has positive sectional curvature then φ is constant and ψ vanishes identically. Proof. Taking the trace of (3.4) and using (3.3) we find n 0 =(2 − n)|dφ|2 + hψ, Dψ/ i− hRN (ψ, ψ)ψ, ψi 6 1 =(2 − n)(|dφ|2 + hRN (ψ, ψ)ψ, ψi). 6 At this point we need to distinguish two cases. (1) First, suppose that φ: M → N is a constant map. Consider v ∈ φ∗T N, Ψ ∈ Γ(ΣM) and set ψ := Ψ ⊗ v. It is easy to check that this pair (φ, ψ) satisfies hRN (ψ, ψ)ψ, ψi = hRN (v, v)v, vi|Ψ|4 = 0 due to the skew symmetry of the Riemann curvature tensor regardless of any curvature assumptions on the target. Since dφ = 0 the equations (3.2), (3.3) reduce to ∂/Ψ = 0. But as we are assuming that M does not admit harmonic spinors we can deduce that ψ = 0. (2) Now, we are considering a pair (φ, ψ) that is not of the form from above, in this case the term hRN (ψ, ψ)ψ, ψi will be different from zero. A careful inspection reveals that for n> 2 and N having positive sectional curvature we have 1 |dφ|2 + hRN (ψ, ψ)ψ, ψi ≥ 0, 6 see [11, Proof of Theorem 1.2] for more details. This allows us to conclude that under the assumptions of the theorem the map φ is constant and the vector spinor ψ is trivial. Note that we actually do not require that the map φ satisfies (3.2) here. 

Remark 3.2. It is clear that such a result does not hold if n = 2 since the action functional (3.1) is conformally invariant in this case. The above result would also hold if we consider an additional two-form in the action functional as in [2] since this does not give a contribution to the energy-momentum tensor.

Acknowledgements: The author gratefully acknowledges the support of the Austrian Fund (FWF) through the project P30749-N35 “Geometric variational problems from string theory”. A VANISHING RESULT FOR THE SUPERSYMMETRIC NONLINEAR SIGMA MODEL 11

References [1] Bernd Ammann and Nicolas Ginoux. Dirac-harmonic maps from index theory. Calc. Var. Partial Differential Equations, 47(3-4):739–762, 2013. [2] Volker Branding. Magnetic Dirac-harmonic maps. Anal. Math. Phys., 5(1):23–37, 2015. [3] Volker Branding. Some aspects of Dirac-harmonic maps with curvature term. Differential Geom. Appl., 40:1–13, 2015. [4] Volker Branding. Dirac-harmonic maps with torsion. Commun. Contemp. Math., 18(4):1550064, 19, 2016. [5] Volker Branding. Energy estimates for the supersymmetric nonlinear sigma model and applications. Potential Anal., 45(4):737–754, 2016. [6] Volker Branding. Nonlinear Dirac equations, monotonicity formulas and Liouville . arXiv:1605.03453, 2016. [7] Volker Branding. On conservation laws for the supersymmetric sigma model. Results Math., 72(4):2181–2201, 2017. [8] Volker Branding. A Liouville-type theorem for biharmonic maps between complete Riemannian manifolds with small energies. Arch. Math. (Basel), 111(3):329–336, 2018. [9] Volker Branding. A note on twisted Dirac operators on closed surfaces. Differential Geom. Appl., 60:54–65, 2018. [10] Volker Branding and Klaus Kr¨oncke. Global existence of Dirac- maps with curvature term on expanding spacetimes. Calc. Var. Partial Differential Equations, 57(5):57:119, 2018. [11] Q. Chen, J. Jost, and G. Wang. Liouville theorems for Dirac-harmonic maps. J. Math. Phys., 48(11):113517, 13, 2007. [12] Qun Chen, J¨urgen Jost, Jiayu Li, and Guofang Wang. Dirac-harmonic maps. Math. Z., 254(2):409–432, 2006. [13] Pierre Deligne and Daniel S. Freed. Supersolutions. In Quantum fields and strings: a course for mathemati- cians, Vol. 1, 2 (Princeton, NJ, 1996/1997), pages 227–355. Amer. Math. Soc., Providence, RI, 1999. [14] Emmanuel Hebey. Sobolev spaces on Riemannian manifolds, volume 1635 of Lecture Notes in Mathematics. Springer-Verlag, Berlin, 1996. [15] J¨urgen Jost, Enno Kessler, J¨urgen Tolksdorf, Ruijun Wu, and Miaomiao Zhu. From harmonic maps to the nonlinear supersymmetric sigma model of quantum field theory. at the interface of theoretical physics, riemannian geometry and nonlinear analysis. https://arxiv.org/abs/1710.01519, 2017. [16] J¨urgen Jost, Enno Kessler, J¨urgen Tolksdorf, Ruijun Wu, and Miaomiao Zhu. Regularity of Solutions of the Nonlinear Sigma Model with Gravitino. Comm. Math. Phys., 358(1):171–197, 2018. [17] J¨urgen Jost, Lei Liu, and Miaomiao Zhu. A global weak solution of the Dirac-harmonic map flow. Ann. Inst. H. Poincar´eAnal. Non Lin´eaire, 34(7):1851–1882, 2017. [18] H. Blaine Lawson, Jr. and Marie-Louise Michelsohn. Spin geometry, volume 38 of Princeton Mathematical . Princeton University Press, Princeton, NJ, 1989. [19] Zhong Min Shen. Some rigidity phenomena for Einstein metrics. Proc. Amer. Math. Soc., 108(4):981–987, 1990. [20] Yuanlong Xin. Geometry of harmonic maps. Progress in Nonlinear Differential Equations and their Appli- cations, 23. Birkh¨auser Boston Inc., Boston, MA, 1996. [21] Shing Tung Yau. Some function-theoretic properties of complete Riemannian manifold and their applications to geometry. Indiana Univ. Math. J., 25(7):659–670, 1976. [22] Zhen-rong Zhou and Qun Chen. Global pinching lemmas and their applications to geometry of submanifolds, harmonic maps and Yang-Mills fields. Adv. Math. (China), 32(3):319–326, 2003.

University of Vienna, Faculty of Mathematics, Oskar-Morgenstern-Platz 1, 1090 Vienna, Austria, E-mail address: [email protected]