Bochner technique and curvature vanishing theorems

Intuition [Beginner]

On a curved surface, a vector field that tries to be "constant" (parallel) faces an obstacle: the curvature bends it. The Bochner technique measures exactly how much energy such a field must carry, and shows that if the curvature is positive enough, the field cannot exist at all.

Think of a rubber band stretched over a sphere. If the sphere curves outward strongly enough, the rubber band snaps back. The Bochner technique quantifies this snapping: it integrates a curvature-weighted energy over the whole manifold and shows the total must be zero, forcing the field to vanish everywhere.

This idea produces vanishing theorems. A vanishing theorem says "if the curvature satisfies condition X, then the manifold cannot carry objects of type Y." The most famous example: positive Ricci curvature kills harmonic 1-forms on compact manifolds.

Why does this concept exist? The Bochner technique converts a topological question ("does a harmonic form exist?") into an analytic one ("can a function satisfy both the harmonic equation and an energy inequality?"). This conversion is the engine behind some of the deepest results in Riemannian geometry.

Visual [Beginner]

A curved surface with several arrows (vector field) drawn tangent to it. The arrows try to point in a consistent direction, but the curvature of the surface twists them. A shaded region represents the Ricci curvature contribution, with positive shading indicating regions where the curvature pushes back against the vector field. The vector arrows fade to zero in the high-curvature region.

Positive Ricci curvature forces harmonic vector fields to vanish by an integrated energy argument.

Worked example [Beginner]

The circle $S^1$. The circle has zero curvature. A harmonic 1-form on $S^1$ is just $d\theta$, the angular form. It does not vanish.

Step 1. The Ricci curvature of $S^1$ is zero (the circle is flat). The Bochner technique gives no constraint when curvature is zero.

Step 2. Now consider $S^2$, the round sphere. The Ricci curvature is positive everywhere (it equals the metric up to a constant). The Bochner technique says: there are no harmonic 1-forms on $S^2$.

Step 3. Check this against the topology. The first Betti number $b_1(S^2)=0$, so indeed there are no non-product harmonic 1-forms. The Bochner technique gives a geometric proof of a topological fact.

What this tells us: the Bochner technique replaces topological arguments with curvature-based analytic ones. When curvature is positive, harmonic objects vanish, constraining the topology.

Check your understanding [Beginner]

Formal definition [Intermediate+]

Let $(M,g)$ be a compact Riemannian manifold of dimension $n$ with Levi-Civita connection $\nabla$. For a smooth 1-form $\omega$, the rough Laplacian (also called the connection Laplacian or Bochner Laplacian) is

$${\Delta }^{\nabla }\omega =-{\mathrm{tr}}_g({\nabla }^2\omega )=-g^{ij}{\nabla }_i{\nabla }_j\omega .$$

The Hodge Laplacian is ${\Delta }_H=d\delta +\delta d$ where $\delta =d^{\ast }$ is the codifferential.

Definition (Bochner-Weitzenbock formula). For a 1-form $\omega$ on a Riemannian manifold:

$${\Delta }_H\omega ={\Delta }^{\nabla }\omega +\mathrm{Ric}(\omega )$$

where $\mathrm{Ric}(\omega )$ denotes the Ricci curvature acting on $\omega$: in components, $(\mathrm{Ric}(\omega ){)}_k=R_{kj}{g}^{ji}{\omega }_i$.

This identity says that the difference between the Hodge Laplacian and the rough Laplacian is measured by the Ricci tensor. For $k$-forms with $k\ge 2$, the curvature term becomes more complicated (involving the full Riemann tensor), but the structural pattern is the same: the Hodge and rough Laplacians differ by a zeroth-order curvature term.

Definition (Bochner identity for the energy). For a 1-form $\omega$, let $f=\frac{1}{2}|\omega {|}^2$. Then:

$$\Delta f=|\nabla \omega {|}^2-⟨{\Delta }^{\nabla }\omega ,\omega ⟩$$

Combining with the Weitzenbock formula and assuming $\omega$ is harmonic (${\Delta }_H\omega =0$):

$$\Delta f=|\nabla \omega {|}^2+\mathrm{Ric}(\omega ,\omega )$$

Counterexamples to common slips

• The Bochner technique only works for 1-forms. False. The Weitzenbock formula generalises to $k$-forms for any $k$, and also to spinor fields (giving the Lichnerowicz formula). The curvature term is more complex for $k\ge 2$ but the strategy is the same.
• Positive sectional curvature is needed for vanishing. False. Positive Ricci curvature suffices for 1-forms. For $k$-forms with $k\ge 2$, one needs a suitable curvature operator to be positive, which is weaker than positive sectional curvature.
• The Bochner technique requires compactness. The integration step uses compactness (to discard the boundary term from integrating $\Delta f$). On non-compact manifolds, additional decay conditions at infinity can substitute, but the cleanest results are on closed manifolds.

Key theorem with proof [Intermediate+]

Theorem (Bochner vanishing). Let $(M^n,g)$ be a compact Riemannian manifold with $\mathrm{Ric}\ge \lambda g$ for some $\lambda >0$. Then every harmonic 1-form $\omega$ on $M$ vanishes identically: $H_{dR}^1(M)=0$.

Proof. Let $\omega$ be harmonic: ${\Delta }_H\omega =0$. The Bochner identity gives

$$\Delta f=|\nabla \omega {|}^2+\mathrm{Ric}(\omega ,\omega )$$

where $f=\frac{1}{2}|\omega {|}^2$. Since $\mathrm{Ric}\ge \lambda g$ with $\lambda >0$:

$$\mathrm{Ric}(\omega ,\omega )\ge \lambda |\omega {|}^2=2\lambda f$$

Therefore:

$$\Delta f\ge |\nabla \omega {|}^2+2\lambda f\ge 0$$

Integrating over $M$ (which has no boundary):

$$0={\int }_M\Delta fdV\ge {\int }_M|\nabla \omega {|}^{2}dV+2\lambda {\int }_{M}fdV\ge 0$$

The left side equals zero by the divergence theorem (${\int }_M\Delta f=0$ on a closed manifold). Since both integrands are non-negative and their total integral is zero, each must vanish: $|\nabla \omega |=0$ and $f=0$. Therefore $|\omega |=0$ everywhere, so $\omega \equiv 0$. $\square$

Bridge. This theorem builds toward the Lichnerowicz eigenvalue bound 03.02.15 in the Master section, and the foundational reason the argument works is that the Weitzenbock formula splits the Hodge Laplacian into a topological part (the rough Laplacian) and a geometric part (the Ricci tensor). This is exactly the mechanism by which curvature constrains topology: positive Ricci curvature kills first cohomology, and the bridge is that the Bochner-Weitzenbock identity translates the harmonic equation into an energy inequality whose only solution is zero. The result pairs with the Myers theorem 03.02.08 which uses positive Ricci curvature to constrain the fundamental group, giving two independent routes from Ricci positivity to topological restrictions.

Exercises [Intermediate+]

Advanced results [Master]

Theorem 1 (Bochner vanishing for $k$-forms). *Let $(M^n,g)$ be a compact Riemannian manifold. If the curvature operator $\mathcal{R}:{\Lambda }^2{T}^{\ast }M\to {\Lambda }^2{T}^{\ast }M$ is positive-definite, then $H_{dR}^k(M)=0$ for all $1\le k\le n-1$.*

This is the strongest Bochner-type vanishing for $k$-forms: positive curvature operator kills all intermediate cohomology.

Theorem 2 (Lichnerowicz formula for spinors). On a compact spin Riemannian manifold $(M^n,g)$ with scalar curvature $S$, the Dirac operator $D$ satisfies $D^2 = \nabla^ \nabla + \tfrac{1}{4}S$.If$S > 0$everywhere,then$M$ admits no harmonic spinors.*

Theorem 3 (Gromov-B Lawson). A compact simply-connected spin manifold with metrics of positive scalar curvature is exactly the class of spin manifolds whose $\hat{A}$-genus vanishes in dimensions congruent to 1 or 2 mod 8. The obstruction comes from the Lichnerowicz formula combined with the Atiyah-Singer index theorem applied to the Dirac operator.

Theorem 4 (Meyer's theorem). If $(M^n,g)$ is compact with positive curvature operator (stronger than positive sectional curvature), then $M$ is a homology sphere: $H^k(M,\mathbb{R})=0$ for $1\le k\le n-1$.

Theorem 5 (Gallot-Meyer). The $k$-th Betti number vanishes if the $(k)$-th curvature operator (acting on ${\Lambda }^k$) is pointwise positive-definite. This refines Meyer's theorem by targeting specific Betti numbers.

Theorem 6 (Kodaira vanishing via Bochner). On a compact Kahler manifold with positive line bundle $L$, the Kodaira vanishing theorem $H^q(M,K_M\otimes L)=0$ for $q>0$ can be proved by the Bochner technique applied to harmonic $(0,q)$-forms with values in $L$.

Synthesis. The Bochner technique is the foundational reason that curvature constrains topology on compact Riemannian manifolds; the central insight is that the Weitzenbock formula decomposes the Hodge Laplacian into a purely analytic part (the rough Laplacian) and a purely geometric part (the curvature operator). Putting these together with integration over the closed manifold, the analytic part contributes a non-negative term and the geometric part carries the sign information that forces vanishing. This pattern recurs throughout Riemannian geometry, appearing in the Lichnerowicz formula for spinors (where positive scalar curvature kills harmonic spinors), the Gallot-Meyer refinement (where the curvature operator on ${\Lambda }^k$ targets specific Betti numbers), and the bridge to index theory via the Dirac operator where the Atiyah-Singer index theorem turns the Bochner-Lichnerowicz vanishing into an obstruction to positive scalar curvature metrics. The generalisation from 1-forms to arbitrary tensor fields and spinor fields is what makes the technique so powerful: any section of a bundle with a connection admits a Weitzenbock-type identity, and the sign of the curvature term determines whether harmonic sections can exist.

Full proof set [Master]

Proposition (Weitzenbock formula for 1-forms). On a Riemannian manifold $(M,g)$ with Levi-Civita connection $\nabla$, the Hodge Laplacian and rough Laplacian on 1-forms are related by ${\Delta }_H\omega ={\Delta }^{\nabla }\omega +\mathrm{Ric}(\omega )$.

Proof. In local coordinates, the Hodge Laplacian on a 1-form $\omega ={\omega }_{i}d{x}^i$ is

$$({\Delta }_H\omega {)}_k=-g^{ij}{\nabla }_i{\nabla }_j{\omega }_k+R_{kj}{g}^{ji}{\omega }_i$$

The first term is $-({\Delta }^{\nabla }\omega {)}_k$ by definition. The second term is $(\mathrm{Ric}(\omega ){)}_k=R_{kj}{g}^{ji}{\omega }_i$. The Ricci tensor arises from commuting covariant derivatives: ${\nabla }_i{\nabla }_j{\omega }_k-{\nabla }_j{\nabla }_i{\omega }_k=-R_{ijk}{}^l{\omega }_l$. Contracting $i$ with $j$ via $g^{ij}$ and using the first Bianchi identity produces the Ricci contraction. $\square$

Proposition (Parallel forms on flat manifolds). If $(M,g)$ is a compact flat Riemannian manifold, then every harmonic form is parallel.

Proof. When $\mathrm{Ric}=0$, the Bochner identity for a harmonic 1-form $\omega$ reduces to $\Delta f=|\nabla \omega {|}^2$. Integrating: $0=\int |\nabla \omega {|}^2$, so $\nabla \omega =0$. The same argument applies to $k$-forms using the Weitzenbock formula with zero curvature: the curvature term vanishes identically, leaving only $|\nabla \omega {|}^2$ to integrate. $\square$

Connections [Master]

• Hermitian manifold and the Kahler form 03.02.11. The Bochner technique on Kahler manifolds uses the Kahler condition $\nabla J=0$ from 03.02.11 to obtain refined vanishing theorems for $(p,q)$-forms. The Kodaira vanishing theorem, proved via Bochner, is a direct application where the Kahler metric's Ricci form controls the cohomology of holomorphic line bundles.

• Killing fields and infinitesimal isometries 03.02.07. The Bochner technique applies directly to Killing vector fields, which satisfy a second-order equation compatible with the Weitzenbock framework. Bochner's original 1946 paper studied vector fields, and the vanishing of Killing fields under positive Ricci curvature is a direct consequence of the same energy identity.

• Myers-Steenrod and isometry groups 03.02.08. The Myers theorem gives a different constraint from positive Ricci curvature: the fundamental group is finite. Combined with the Bochner result that $b_1=0$, the two theorems give a comprehensive picture of how positive Ricci curvature restricts both the abelian and non-abelian topology of a compact manifold.

Historical & philosophical context [Master]

Salomon Bochner introduced the technique in 1946 ^{[Bochner 1946]}, in a paper titled "Vector fields and Ricci curvature" published in the Annals of Mathematics. Bochner's original result was for harmonic 1-forms: he showed that positive Ricci curvature forces the first Betti number to vanish. The key insight was to compute the Laplacian of the energy density $|\omega {|}^2$ of a harmonic form and integrate.

Andre Lichnerowicz extended the technique to spinor fields in 1963, obtaining the Lichnerowicz formula $D^2={\nabla }^{\ast }\nabla +\frac{1}{4}S$ for the Dirac operator. This extension, combined with the Atiyah-Singer index theorem, produced the first obstruction to positive scalar curvature metrics on spin manifolds: the $\hat{A}$-genus must vanish.

The generalisation to $k$-forms via the curvature operator is due to Meyer (1971), and the definitive treatment of the technique in book form appears in Besse's "Einstein Manifolds" (1987) ^{[Besse 1987]}. The Bochner technique remains one of the most widely used tools in Riemannian geometry, with applications ranging from vanishing theorems in algebraic geometry (Kodaira vanishing) to the study of Einstein metrics and Ricci flow.

Bibliography [Master]

@article{bochner1946,
  author = {Bochner, Salomon},
  title = {Vector fields and {R}icci curvature},
  journal = {Ann. of Math.},
  volume = {47},
  pages = {192--201},
  year = {1946}
}

@book{lichnerowicz1958,
  author = {Lichnerowicz, Andr{\'e}},
  title = {G{\'e}om{\'e}trie des groupes de transformations},
  publisher = {Dunod},
  address = {Paris},
  year = {1958}
}

@book{besse1987,
  author = {Besse, Arthur L.},
  title = {Einstein Manifolds},
  publisher = {Springer},
  year = {1987}
}

@book{petersen2016,
  author = {Petersen, Peter},
  title = {Riemannian Geometry},
  edition = {3},
  publisher = {Springer},
  year = {2016}
}