Preserving positive intermediate curvature

Consider a compact manifold $N$ (with or without boundary) of dimension $n$. Positive $m$-intermediate curvature interpolates between positive Ricci curvature ($m = 1$) and positive scalar curvature ($m = n-1$), and it is obstructed on partial tori $N^n = M^{n-m} \times \mathbb{T}^m$. Given Riemannian metrics $g, \bar{g}$ on $(N, \partial N)$ with positive $m$-intermediate curvature and $m$-positive difference $h_g - h_{\bar{g}}$ of second fundamental forms we show that there exists a smooth family of Riemannian metrics with positive $m$-intermediate curvature interpolating between $g$ and $\bar{g}$. Moreover, we apply this result to prove a non-existence result for partial torical bands with positive $m$-intermediate curvature and strictly $m$-convex boundaries.


Introduction
The existence of Riemannian metrics with positive curvature implies obstructions on the topology of closed manifolds: Manifolds with topology N n = M n−1 × S 1 do not admit metrics of positive Ricci curvature by the Theorem of Bonnet-Myers, while manifolds with topology N n = T n do not admit metrics of positive scalar curvature by the resolution of the Geroch conjecture due to R. Schoen and S.-T.Yau [12] (for 3 ≤ n ≤ 7 by using minimal hypersurfaces) and M. Gromov and H.-B. Lawson [7] (by using spinors and the Atyah-Singer index theorem).
The above results yield obstructions for positive curvature on the the partial tori N n = M n−m × T m for the limit cases m = 1 and m = n − 1.Recently, S. Brendle, S. Hirsch and the second author introduced the notion of positive intermediate curvature (see Section 2 for a precise definition), which interpolates between positive Ricci curvature (m = 1) and positive scalar curvature (m = n − 1).They obtained the following obstruction result on partial tori: Theorem 1.1 (Generalized Geroch conjecture, Theorem 1.5 in [1]).Assume n ≤ 7 and 1 ≤ m ≤ n − 1.Let N n be a closed and orientable manifold of dimension n, and suppose that there exists a closed and orientable manifold M n−m and a map F : N n → M n−m × T m with non-zero degree.Then the manifold N does not admit a metric with positive m-intermediate curvature.
The associated rigidity question was studied by J. Chu, K-K.Kwong and M.-C. Lee [5] in ambient dimension at most five.S. Chen [3] extended the obstruction result to manifolds with arbitrary ends.Moreover, K. Xu [15] showed sharpness of the result in [1] by constructing counterexamples for dimensions n > 7 and 3 ≤ m ≤ n − 3. The above result was recently used by M. L. Labbi [8] to compute the Riemann invariant of products of spheres and tori.
In this work, we study the interaction of the internal geometry and the boundary geometry for metrics of positive intermediate curvature.The corresponding question for positive scalar curvature and mean curvature on the boundary dates back to work by M. Gromov and H.-B. Lawson [6].A similar interaction appears in the proof of the positive mass theorem by R. Schoen and S.-T.Yau [13] -planes with positive mean curvature act as barriers for minimal hypersurfaces in the bulk region with positive scalar curvature.Related is work by Y. Shi and L.-F.Tam [14], where they proved an estimate for the integral of the mean curvature over the boundary in manifolds with nonnegative scalar curvature.P. Miao [9] proved a positive mass theorem on manifolds with corners by performing a suitable intrinsic bending construction.A pertubation argument, which made the boundary totally geodesic while keeping the scalar curvature non-decreasing, allowed S. Brendle, F.C. Marques and A. Neves [2] to construct counterexamples to the Min-Oo conjecture.
Recently, the first author proved a general result [4, Main Theorem 1] on the interaction of internal geometry and boundary geometry by suitably gluing Riemannian metrics.The result applies to a wide range of positive curvature conditions, for example to metrics with positive curvature operator, PIC 2, PIC 1 (with convex boundary), positive isotropic curvature (with two-convex boundary) and positive scalar curvature (with mean-convex boundary).For a different approach to the gluing problem, see also the thesis by A. Schlichting, [11].
The first result of our work extends the gluing result of the first author to m-intermediate curvature with the natural condition of m-convexity on the boundary: Suppose that N n is a compact smooth manifold with smooth boundary ∂N of dimension dim N = n.Let g, g be Riemannian metrics on N , such that g = g on the boundary ∂N .
Then there exists λ 0 > 0, a family of smooth Riemannian metrics {ĝ λ } λ>λ 0 , and a neighborhood U of the boundary ∂N , such that the metric ĝλ agrees with the metric g outside of U , and the metric ĝλ agrees with the metric g in a neighbourhood of ∂N .Additionally, we have ĝλ → g as λ → ∞ in C α for any α ∈ (0, 1).[6]).Suppose (N, g) is an orientable compact smooth Riemannian manifold with smooth boundary ∂N .Assume the metric g has positive scalar curvature and is strictly mean convex (i.e.H ∂N > 0) with respect to the outward unit normal.Then the double of N carries a metric of positive scalar curvature.This lemma (in conjunction with the nonexistence of positive scalar curvature metrics on the torus) then implies the following obstruction to positive scalar curvature on torical bands: Theorem 1.4 (Boundaries of a torical band, M. Gromov and H.-B. Lawson, [6]).Consider the smooth manifold with boundary N = T n−1 × [−1, 1] and let g be a Riemannian metric on N with positive scalar curvature.Then the boundary ∂N cannot be strictly mean convex.D. Räde extended the above result to a scalar curvature and mean curvature comparison result on more general bands [10].
We extend the above result on scalar curvature, mean curvature and torical bands to partially torical bands by proving a generalization of the doubling lemma of M. Gromov and H.-B. Lawson.The work is structured as follows: In Section 2 we introduce our notation and recall the definition of intermediate curvature.In Section 3 we prove an algebraic lemma connecting the cone of positive m-intermediate curvature and m-convexity.In Section 4 and 5 we prove the gluing result and in Section 6 we perform the doubling constructions.

Preliminaries
Let (V, •, • ) be a n-dimensional real inner product space.The space of algebraic curvature tensors on V denoted by C B (V ) is given by multilinear maps R : for all v 1 , v 2 , v 3 , v 4 ∈ V , and satisfying the first Bianchi identity, i.e.
We denote the space of symmetric bilinear maps T : Following Definition 1.1 in work of the S. Brendle, S. Hirsch and the second author [1] we define the cone C m (V ) of non-negative m-intermediate curvature in the space of algebraic curvature tensors by T (e p , e q , e p , e q ) ≥ 0 for all orthonormal bases For a Riemannian manifold (N n , g) with boundary ∂N we consider its Levi-Civita connection D and its Riemann curvature tensor Rm N given by the formula Let ν be the inward pointing unit normal vector field on the boundary ∂N .The scalar-valued second fundamental form h g : T (∂N ) ⊗ T (∂N ) → C ∞ (∂N ) of the boundary ∂N with respect to the Riemannian metric g is defined by With this convention the scalar-valued second fundamental form is positive on the standard sphere S n ⊂ R n+1 with respect to the inward pointing unit normal vector ν = −x.
We say that the boundary ∂N is strictly m-convex (where For m = 1 we recover the notion of strict convexity, and for m = n − 1 we recover the notion of strict mean convexity.Let (V, •, • ) be a n-dimensional inner product space and S : V × V → R be a symmetric bilinear form.Let W ⊂ V be a (n − 1)-dimensional subspace, and let ν ∈ W ⊥ be a unit vector.Let S| W be the restriction of Proof.
Suppose that the bilinear form S| W is m-positive.We extend the bilinear form S| W to a bilinear T on V by setting T (v, w) := S| W (v , w ) for v, w ∈ V .Here v denotes the orthogonal projection from V to W .
The assumption on the m-positivity of the bilinear form S| W on W implies that the bilinear form T is (m + 1)-positive on V .
We first want to show that T being Let {e 1 , . . ., e n } be an orthonormal basis of V with respect to the inner product •, • .We denote the components of the vector ν with respect to this orthonormal basis by a p , i.e. a p = ν, e p .
We have T ∧ (ν b ⊗ ν b ) (e p , e q , e p , e q ) = a 2 p T (e q , e q ) + a 2 q T (e p , e p ) − 2a p a q T (e p , e q ) for 1 ≤ p, q ≤ n by definition of the Kulkarni-Nomizu product.We observe the identity We evaluate the first term in the above sum: We evaluate the second term in the above sum: Now suppose that a p = 0 for some m + 1 ≤ p ≤ n.We define the unit vector a q e q .
With this definition we deduce from equation (1) the identity T (e q , e q ) + T (w, w) The first term involving the trace tr V (T ) is positive as above.The term in the bracket is positive, since the bilinear form T is (m + 1)-positive, and w ⊥ span{e 1 , . . ., e m } by construction.Hence the sum is positive and we deduce On the other hand, by the construction of T , the restriction of S − T to the subspace W vanishes.Therefore, we may write where ω is a suitable 1-form.Note that The other implication in the equivalence follows by taking the orthonormal basis {f 1 , . . ., f n−1 , ν} of the vector space V , where {f 1 , . . ., f n−1 } is an orthonormal basis of the subspace W .

Preserving Curvature Conditions
In this section, we assume that g and g are Riemannian metrics on N such that g − g = 0 along ∂N .We describe our choice of perturbation as in work of S. Brendle, F.C. Marques and A. Neves [2].We fix a neighborhood U of the boundary ∂N and a smooth boundary defining function ρ : N → [0, ∞) by taking it to be the distance function from the boundary ∂N with respect to the metric g.Then we have |Dρ| g = 1.Since g − g = 0 along the boundary ∂N , we can find a symmetric (0,2)-tensor S such that g = g + ρS in a neighborhood of ∂N and S = 0 outside U .The scalar-valued second fundamental forms and the boundary defining function satisfy for all X, Y ∈ Γ(T (∂N )).This implies that the identity holds on the boundary ∂N for all X, Y ∈ Γ(T (∂N )).

Moreover, we choose a smooth cut-off function
For λ > 0 sufficiently large we define a smooth metric ĝλ on the manifold N by the formula In the sequel, we will show that ĝλ preserves positive m-intermediate curvature of g and g for sufficiently large λ > 0. Note that we have the identity ĝλ = g in the region {ρ ≤ e −2λ 2 } and ĝλ = g outside the neighbourhood U .Moreover, from the construction it follows that ĝλ → g as λ → ∞ in C α for any α ∈ (0, 1).

Proof.
We fix a point x ∈ N such that ρ(x) ≥ e −λ 2 .Let {e 1 , . . ., e n } be a geodesic normal frame around the point x with respect to the metric ĝλ .Let ϕ be a two-form.We write ϕ = i,j ϕ ij e i ∧ e j for coefficients ϕ ij , which are anti-symmetric in i and j.In the following the Einstein summation convention will be adopted freely.Since ϕ is in particular a (2,0)-tensor, ϕ induces by the fundamental principle of tensor calculus a linear map [ϕ] : (T x N ) * → T x N via the action [ϕ]w := ϕ ij w(e i )e j .Equation ( 5) in work of the first author [4] yields the estimate where C > 0 is a positive constant independent of λ.
To pick up the sectional curvatures, we choose for 1 ≤ p, q ≤ n the two-form ϕ pq by specifying the components (ϕ pq ) ij = δ i p δ j q − δ j p δ i q .This implies the identities [ϕ pq ]dρ = ∇ p ρ e q − ∇ q ρ e p and 4 Rm(e p , e q , e p , e q ) = Rm(ϕ pq , ϕ pq ).
Recall that S is supported in U , so Dρ and D 2 ρ are uniformly bounded with respect to the metric ĝλ .Thus the above estimate implies after summation the inequality (4) m p=1 n q=p+1 (Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q )) S ∇ p ρ e q − ∇ q ρ e p , ∇ p ρ e q − ∇ q ρ e p − Cχ ′ (λρ) − Cλ −1 in the region {ρ ≥ e −λ 2 }.Here the constant C is independent of λ, but it does depend on S, g, ρ, N .
By our assumption on the scalar-valued second fundamental forms and Proposition 3.1 we deduce (with respect to the metric g) at each point on ∂N .This allows us to fix a small number a > 0 such that S ∧ (dρ ⊗ dρ) − 4a|∇ρ| 2 δ ∧ δ ∈ C m (with respect to the metric g) in a small neighborhood of ∂N where ρ ≥ e −λ 2 .From the construction of the metric ĝλ we deduce (with respect to the metric ĝλ ) in a small neighborhood of ∂N where ρ ≥ e −λ 2 .Moreover, we observe that S ∇ p ρ e q − ∇ q ρ e p , ∇ p ρ e q − ∇ q ρ e p = S ∧ (dρ ⊗ dρ)(e p , e q , e p , e q ).(Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q )) We split the region into two sub-regions as follows.Let us fix a real number s 0 ∈ [0, 1) such that Cχ ′ (s 0 ) < ǫ 2 .By the construction of the cut-off function χ, we have inf 0≤s≤s 0 (−χ ′′ (s)) > 0. This implies in the region {e (Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q )) Thus, we obtain (Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q )) (Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q )) (Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q ))   ≥ −ǫ if λ > 0 is sufficiently large.Putting the above together, we conclude that (Rm ĝλ (e p , e q , e p , e q ) − Rm(e p , e q , e p , e q ))   ≥ −ǫ if λ > 0 is sufficiently large.This completes the proof of Proposition 4.1.
In the region {ρ < e −λ 2 }, we have ĝλ = g + hλ , where hλ is defined by Let {e 1 , . . ., e n } be a geodesic normal frame around x with respect to the metric ĝλ .Equation (12) in work of the first author [4] implies for any two-form ϕ = ϕ ij e i ∧ e j and a positive constant L > 0 independent of λ.Choosing for 1 ≤ p, q ≤ n the two-form ϕ pq by (ϕ pq ) ij = δ i p δ j q − δ j p δ i q in equation ( 7) and summing over p, q, we obtain m p=1 n q=p+1 (Rm ĝλ (e p , e q , e p , e q ) − Rm g(e p , e q , e p , e q )) S ∇ p ρ e q − ∇ q ρ e p , ∇ p ρ e q − ∇ q ρ e p − Lλ −1 .
Proceeding similarly as in the proof of Proposition 4.1, we have m p=1 n q=p+1 S ∇ p ρ e q − ∇ q ρ e p , ∇ p ρ e q − ∇ q ρ e p ≥ 2a|∇ρ| 2 g .
From this, the assertion follows.
Combining Proposition 4.1 and Proposition 4.2, we can summarize the results in this section: Rm ĝλ (E p , E q , E p , E q ) > 0.
We divide the proof into two cases: In the first case x ∈ N is in the inner gluing region {ρ ≥ e −λ 2 }; and in the second case x ∈ N is in the outer gluing region {ρ < e −λ 2 }.
For the first case, we have g = ĝλ − h λ where h λ (x) = λ −1 χ(λρ)S(x).Fix a point x ∈ N .We evolve the orthonormal basis {E 1 , . . ., E n } in the tangent space T x N by the linear ordinary differential equation Then we see that the basis {E i (s)} remains orthonormal with respect to the metric g s = ĝλ −sh λ .We define e i := E i (1).Then the basis {e 1 , . . ., e n } is orthonormal with respect to the metric g = ĝλ −h λ .It follows that m p=1 n q=p+1 Rm g (e p , e q , e p , e q ) > 0.
In view of Corollary 4.3, it suffices to show that m p=1 n q=p+1 Rm g (E p , E q , E p , E q ) > 0.

By writing E
where C is a positive constant depending only on N, g, g and χ.This implies that Rm g (E p , E q , E p , E q ) − Rm g (e p , e q , e p , e q ) ≤ 1 Hence we have Rm g (e p , e q , e p , e q ) > 0, if λ > 0 is sufficiently large.
Combining the two cases together, Corollary 4.3 implies that m p=1 n q=p+1 Rm ĝλ (E p , E q , E p , E q ) > 0 for sufficiently large λ > 0. Since the point x ∈ N and the orthonormal basis {E 1 , ..., E n } ⊂ T x N are arbitrary, we conclude that Rm ĝλ ∈ Int(C m ) on N for sufficiently large λ > 0. This finishes the proof of the theorem.

Proof of Theorem 1.5 on boundaries of partially torical bands
In this section we prove Theorem 1.5 by using the strategy outlined by M. Gromov and H.-B. Lawson.We prove a generalization of their doubling lemma (as stated in Lemma 1.3).
Once we have established the doubling lemma Theorem 1.5 follows directly by the generalization of the Geroch conjecture, Theorem 1.1.We mimic the doubling process by M. Gromov and H.-B. Lawson, see Theorem 5.7 in [6].We will closely follow their notations and constructions.For completeness we describe their construction in detail.
We first shrink the manifold N a little bit while preserving its boundary condition as follow: let N 1 = N \C where C is a thin collar of the boundary ∂N and N 1 is chosen so that ∂N 1 is still strictly m-mean convex.We then consider the Riemannian product N × I with Riemannian metric g N + dt ⊗ dt and define D(N ) = {p ∈ N × I | dist(p, N 1 ) = ǫ}, where 0 < ǫ ≪ 1.Note that D(N ) is homeomorphic to the double of N .Now we fix a point x ∈ ∂N 1 , and let σ be the geodesic segment in N 1 emanating orthogonally from ∂N 1 at x. Then the product σ × I will be totally geodesic in the product N × I.
Let µ 1 , ...µ n−1 be the principal curvatures of ∂N 1 at x.By the construction of M. Gromov and H.-B. Lawson the principle curvatures of D(N ) will be of the following form for a point x θ corresponding to an angle θ (see Figure 8 in [6]): As in the Figure 8 in [6], we have a natural polar coordinates describing these x θ , let us denote pr N (x θ ) the projection of x θ to the corresponding point on ∂N 1 .Since the bilinear forms h ∂N 1 at pr N (x θ ) and h D(N ) at x θ are diagonalized simultaneously in this construction, we have the following relation: .
We apply the Gauss equation to D(N ) as a submanifold of N × I at the point x θ : We have for any orthonormal basis {e i } n i=1 at the point x θ the relation Rm D(N ) (e i , e j , e i , e j ) = Rm N ×I (e i , e j , e i , e j ) + h D(N ) (e i , e i )h D(N ) (e j , e j ) − h D(N ) (e i , e j ) 2 .(11)

( 1 )
the Riemannian manifolds (N, g) and (N, g) have positive m-intermediate curvature, (2) the difference h g − h g is strictly m-convex (i.e.strictly m-positive), then the Riemannian manifold (N, ĝλ ) has positive m-intermediate curvature for all λ > λ 0 .The main ingredient in the proof is Proposition 3.1 relating the cone of positive m-intermediate curvature to the Kulkarni-Nomizu product of m-convex symmetric two-tensors.In the second part of the paper we consider Riemannian manifolds with m-positive intermediate curvature and strictly m-convex boundary: Let us first recall the doubling lemma by M. Gromov and H.-B. Lawson for manifolds with positive scalar curvature and strictly mean convex boundaries.Lemma 1.3 (Doubling of positive scalar curvature metrics, M. Gromov and H.-B. Lawson

Theorem 1 . 5 (
Boundaries of a partially torical band).Let n ≤ 7 and 1 ≤ m ≤ n − 1. Suppose M n−m is a closed orientable manifold.Consider the smooth manifold with boundary N = M n−m × T m−1 × [−1, 1] and a Riemannian metric g with positive m-intermediate curvature on N .Then the boundary ∂N cannot be strictly m-convex.

3 .
Connecting m-convexity and m-intermediate curvature In this section we prove a lemma in linear algebra, which allows us to connect the cone of positive m-intermediate curvature and strict m-convexity.Proposition 3.1 (m-intermediate curvature cone and m-convexity).

TTTT
∧ (ν b ⊗ ν b ) (e p , e q , e p , e q ) (e q , e q ) − 2T ∧ (ν b ⊗ ν b ) (e p , e q , e p , e q ) = tr V (T )If a p = 0 for all m + 1 ≤ p ≤ n, then we deduce the estimate ∧ (ν b ⊗ ν b ) (e p , e q , e p , e q ) = tr V (T ) since the bilinear form T is (m + 1)-positive by construction and hence n-positive.Hence we deduce T ∧ (ν b ⊗ ν b ) ∈ Int(C m ) in this case.

Proposition 4 . 2 (
Curvature estimates in outer gluing region).Suppose that h g − h g is m-positive on the boundary ∂N .Let ǫ > 0 be an arbitrary positive real number.If λ = λ(ǫ, β) > 0 is sufficiently large, then

Corollary 4. 3 . 5 .
Suppose that h g − h g is m-positive on the boundary ∂N .Let ǫ > 0 be an arbitrary positive real number.If λ = λ(ǫ, χ, β) > 0 is sufficiently large, then we have the pointwise inequality m p=1 n q=p+1 Rm ĝλ (x)(e p , e q , e p , e q ) Rm g (x)(e p , e q , e p , e q ), x)(e p , e q , e p , e q ) any ĝλ -orthonormal frame {e 1 , . . ., e n } and any x ∈ N .Proof of Theorem 1.2 on preserving positive intermediate curvature Suppose that h g − h g is m-positive on the boundary ∂N .Suppose also that (N, g) and (N, g) have positive m-intermediate curvature.Fix a point x ∈ N and let {E 1 , . . ., E N } be an orthonormal basis of the tangent space T x N with respect to the Riemannian metric ĝλ .We want to show that

Lemma 6 . 1 (
Doubling for positive m−intermediate curvature).Suppose (N n , g) is an orientable compact smooth Riemannian manifold with smooth boundary ∂N , such that the metric g has positive m-intermediate curvature, and the boundary ∂N is strictly mconvex (possibly finitely many connected components).Then the double of N carries a metric of positive m-intermediate curvature.