--- title: 'Entire Grauert Tube: Complex Structure & Rigid Geometry' url: https://www.emergentmind.com/topics/entire-grauert-tube type: topic --- # Entire Grauert Tube: Complex Structure & Rigid Geometry Searching arXiv for recent papers on entire Grauert tubes and closely related rigidity/topology results. {"query":"all: \"entire Grauert tube\"","max_results":10,"sort_by":"relevance"} {"query":"id:2101.04368 OR id:2501.15682 OR id:1705.03359 OR id:2508.03352 OR id:2605.05582 OR id:2410.05685 OR id:1803.03579","max_results":10,"sort_by":"relevance"} An **entire Grauert tube** is the global form of the adapted complexification of a real-analytic Riemannian manifold. For a real-analytic Riemannian manifold \((M,g)\), one first constructs, on a sufficiently small disk bundle in \(TM\), a unique complex structure adapted to the foliation by complexified geodesics; the tube is called **entire** when this adapted complex structure extends to all of \(TM\). In that case the tangent bundle becomes a Stein complex manifold with a canonical strictly plurisubharmonic exhaustion satisfying a homogeneous complex Monge–Ampère equation, and the existence of this global complex structure has strong consequences for curvature, geodesic dynamics, topology, rigidity, and algebraization [2101.04368] [2605.05582]. ## 1. Adapted complexification on the tangent bundle Let \((M^n,g)\) be a real-analytic Riemannian manifold and let \(\pi:TM\to M\) be the tangent-bundle projection. The tangent bundle \(T(TM)\) has canonical vertical and horizontal subbundles. At \(v\in TM\), \[ V(v)=\ker(d\pi_v), \qquad H(v)=\ker K_v, \] where \(K\) is the connection map defined by the Levi-Civita connection, and \[ T_v(TM)=V(v)\oplus H(v). \] There are natural identifications \[ d\pi_v:H(v)\simeq T_{\pi(v)}M, \qquad K_v:V(v)\simeq T_{\pi(v)}M. \] The basic geometric input is the foliation of \(TM\setminus M\) by complexified geodesics. If \(y:\mathbb R\to M\) is a unit-speed geodesic with \(y'(0)=v\), then \[ w+i\,t\longmapsto t\,y'(w)\in T_{y(w)}M \] is a holomorphic immersion of a strip in \(\mathbb C\) into \(TM\). Requiring these leaves to be genuine complex submanifolds yields the adapted complex structure [2101.04368]. For \(R>0\), \[ T^R M=\{\,v\in TM\mid g(v,v)0\) [2101.04368]. Uniqueness permits the definition of the maximal radius function \[ \rho(v)=\sup\{\;r>0\mid v\in T^rM \text{ and the adapted complex structure extends to a neighborhood of }v\}. \] The manifold has an **entire Grauert tube** precisely when \[ \rho(v)=+\infty\quad\text{for every }v\in TM, \] equivalently, when the adapted complex structure extends to all of \(TM\) [2101.04368]. A complementary intrinsic formulation uses the fiberwise norm \(u(v)=|v|_g\). In a Grauert tube \(T^rM\), \(u\) is strictly plurisubharmonic and satisfies \[ (i\partial\bar\partial u)^n=0\quad\text{on }T^rM\setminus M. \] Writing points as \(z=\exp_x(i\,v)\), one sets \(\rho(z)=|v|_g\), and \[ \omega=i\partial\bar\partial(\rho^2) \] is a Kähler form extending the Liouville symplectic form [2501.15682]. ## 2. Stein, Kähler, and Monge–Ampère structure An entire Grauert tube is not merely a complexification in a local or formal sense. It makes \((TM,J)\) into a Stein complex manifold of complex dimension \(n\), with zero-section \(M\) a totally real maximal-dimensional submanifold and with strictly plurisubharmonic exhaustion \(\rho=u^2\) [2605.05582]. The defining PDE is the homogeneous complex Monge–Ampère equation \[ (i\,\partial\bar\partial\,u)^n=0 \quad\text{on }TM\setminus M, \] or, in the radius-\(r\) notation, \[ (i\partial\bar\partial u)^n=0 \quad\text{on }T^rM\setminus M. \] This Monge–Ampère foliation is the complex-geometric expression of the complexified geodesic flow [2501.15682] [2508.03352]. The “entire” condition can also be stated globally: there exists a global Kähler metric \[ \omega=i\partial\bar\partial\phi \] for some exhaustion \(\phi\) agreeing with \(\rho^2\) near the zero section, and \(\phi\) solves the HCMA outside \(M\) [2501.15682]. In that setting \(TM\) is Stein, \(\bar\partial\)-cohomology vanishes, and geodesic-flow leaves compactify holomorphically [2501.15682]. This complex-analytic package is unusually rigid. Song shows that for an entire Grauert tube the Riemannian metric extends meromorphically to a nondegenerate symmetric \((2,0)\)-tensor \[ \mathbf g\in H^0(TM,K_{TM}^{-2}), \] called the complexified metric, and its pole locus \(S\subset TM\) becomes the natural obstruction in affineness questions [2605.05582]. This places entire Grauert tubes at the interface of Stein geometry, Kähler geometry, and algebraic geometry. A common misconception is to identify “Grauert tube” with “entire Grauert tube.” In fact, many papers on Grauert tubes work at finite radius. For compact negatively curved real-analytic manifolds, one studies a maximal radius \(\tau_0>0\) on which the Grauert tube \(M_\tau\) is strictly pseudoconvex, and this finite-radius complexification is sufficient for the holomorphic continuation of Laplace eigenfunctions and the study of complex nodal sets [1803.03579]. Entireness is therefore a much stronger global condition than the existence of a local or finite-radius tube. ## 3. Examples, non-examples, and metric dependence Known examples of entire Grauert tubes arise from compact Lie groups with special metrics. Any compact Lie group \(G\) with a bi-invariant metric has entire tube, any product \(G\times\mathbb R^k\) does as well, and quotients of such spaces by free isometric actions inherit entire tubes [2101.04368]. These examples already show that entireness is compatible with substantial symmetry and with homogeneous and cohomogeneity-one settings. The standard round sphere provides the basic non-example. For the round metric on \(S^n\) of sectional curvature \(+1\), the adapted complex structure exists only on \[ T^{\pi/2}S^n=\{\,v\in TS^n:|v|<\tfrac{\pi}{2}\}, \] because Jacobi-field matrices develop poles at finite imaginary time. Hence \(S^n\) is not entire [2101.04368]. This is important conceptually: even positive sectional curvature does not imply entireness. The strongest metric-level classification presently available in a nontrivial compact Lie-group case is for left-invariant metrics on \(\mathrm{SU}(2)\). Writing the metric as \[ A=\operatorname{diag}(\lambda_1,\lambda_2,\lambda_3), \qquad 0<\lambda_1\le\lambda_2\le\lambda_3, \] the geodesic flow is the free rigid-body motion with Euler equations \[ \begin{cases} \dot v_1=(\lambda_2-\lambda_3)\,\dfrac{v_2v_3}{\lambda_2\lambda_3},\\[6pt] \dot v_2=(\lambda_3-\lambda_1)\,\dfrac{v_3v_1}{\lambda_3\lambda_1},\\[6pt] \dot v_3=(\lambda_1-\lambda_2)\,\dfrac{v_1v_2}{\lambda_1\lambda_2}. \end{cases} \] This system is completely integrable via the quadratic first integrals \[ E(v)=\tfrac12\sum_{i=1}^3\lambda_i v_i^2, \qquad M(v)=\tfrac12\sum_{i=1}^3\lambda_i^2 v_i^2. \] If \(\lambda_1,\lambda_2,\lambda_3\) are all distinct, then the intersection of the corresponding quadrics in \(\mathbb P^3\) is a smooth genus-one curve, and Picard’s theorem gives an obstruction to entireness. The resulting classification is that the tube is entire if and only if, up to relabelling, \[ \lambda_1=\lambda_2=1,\quad \lambda_3=\Lambda,\quad 0<\Lambda\le 1. \] If \(\Lambda>1\), the maximal radius is finite; if all three eigenvalues are distinct, the tube is never entire [1705.03359]. This classification makes clear that entireness depends delicately on the metric, not merely on the underlying manifold. On \(S^3\simeq \mathrm{SU}(2)\), some left-invariant metrics are entire and others are not [1705.03359]. ## 4. Curvature, geodesic counting, and rational ellipticity A decisive structural fact is that entireness forces nonnegative sectional curvature: Lempert–Szőke observed that if \(T^RM\equiv TM\) is entire, then \((M,g)\) must have nonnegative sectional curvature [2101.04368]. Chen proves that under compactness and simple connectivity this strong complex-geometric hypothesis implies equally strong rational-homotopy consequences. If \((M^n,g)\) is compact, simply connected, real-analytic, and has entire Grauert tube, then \(M\) is **rationally elliptic**: all but finitely many of its homotopy groups are finite; equivalently, the Betti numbers of the based loop space \(\Omega M\) have polynomial growth [2101.04368]. As a corollary, any compact manifold with entire Grauert tube satisfies \[ \chi(M)\ge 0, \] after passing to the universal cover and using Cheeger–Gromoll when \(\pi_1(M)\) is infinite [2101.04368]. The proof follows the Morse-theoretic/Gromov–Bott strategy. For \[ D_T(x)=\{\,v\in T_xM:g(v,v)\le T^2\}, \qquad n_T(x,y)=\#\{\,v\in D_T(x)\mid \exp_x(v)=y\}, \] one studies the number of geodesics of length \(\le T\) from \(x\) to \(y\). Jacobi fields \(J_j(t)\) along \(\gamma(t)=\exp_x(t\,v)\), with \(J_j(0)=0\) and \(J_j'(0)=e_j\), satisfy the integral identity \[ \int_M n_T(x,y)\,dy = \int_{S_xM}\det\!\bigl(g(J_i(v),J_j(v))_{i,j=1}^{n-1}\bigr)\,d\omega(v). \] Under entireness, the real-analytic functions \(g(J_i(t),J_j(t))\) analytically continue to meromorphic matrix-valued functions \(f_{ij}(z)\) on the upper half-plane. A Fatou-type theorem yields \[ \frac{d}{dz}\bigl[-f(z)^{-1}\bigr] = A+\int_\mathbb R\frac{d\mu(t)}{(z-t)^2}, \qquad z\in\mathbb C^+, \] with \(A\) positive semidefinite and \(\mu\) a positive matrix-valued measure supported at the poles. From this one obtains polynomial growth of the Jacobi-field determinant, hence polynomial growth of \(\int_M n_T(x,y)\,dy\). Combining that with the Berger–Bott–Grove–Halperin inequality \[ \sum_{j=0}^{k-1}\dim H_j(\Omega M;\mathbb Q) \le C\,\mathrm{Vol}(M)\,\int_M n_{Ck}(x,y)\,dy \] implies polynomial growth of loop-space Betti numbers and therefore rational ellipticity [2101.04368]. The significance is twofold. First, the global complexification of geodesics controls the oscillation of Jacobi fields. Second, that analytic control translates into a topological restriction that matches the Bott–Grove–Halperin expectation in this stronger complex-geometric regime [2101.04368]. ## 5. Zero entropy and manifold-level classification consequences The geodesic-counting bounds have a dynamical counterpart. If \((M,g)\) is a closed real-analytic Riemannian manifold with entire Grauert tube, then the topological entropy of the geodesic flow vanishes: \[ h_{\mathrm{top}}(g)=0. \] The proof combines Mañé’s formula \[ h_{\mathrm{top}}(g) = \lim_{T\to\infty}\frac1T\log\int_{M\times M}n_T(x,y)\,dx\,dy \] with the polynomial bound on the integrated counting function obtained from the analytic continuation of Jacobi fields [2410.05685]. An equivalent approach uses the Jacobian-growth formula for the geodesic flow and the same polynomial bound on vertical Jacobi fields [2410.05685]. Zero entropy yields several restrictions. If \((M,g)\) admits an entire Grauert tube, then \(\pi_1(M)\) has subexponential, in fact virtually polynomial, growth [2410.05685]. For simply connected analytic manifolds with entire Grauert tube, \(\dim H_j(\Omega M;\mathbb Q)\) grows only polynomially in \(j\) [2410.05685], consistent with rational ellipticity. In dimension \(3\), a closed analytic \(3\)-manifold admits an entire Grauert tube if and only if it is modelled on one of the Thurston geometries \[ \mathbb S^3,\qquad \mathbb S^2\times\mathbb E,\qquad \mathbb E^3. \] Equivalently, these are precisely the \(3\)-manifolds admitting good complexifications [2410.05685]. In dimension \(4\), under the entire-tube hypothesis, several classification statements are obtained for simply connected manifolds, elliptic surfaces, and manifolds with infinite fundamental group of polynomial growth [2410.05685]. In dimension \(5\), a closed analytic simply connected \(5\)-manifold admits an entire Grauert tube if and only if it is diffeomorphic to one of \[ S^5,\quad S^3\times S^2,\quad \text{the nontrivial }S^3\text{-bundle over }S^2,\quad \mathrm{SU}(3)/\mathrm{SO}(3), \] again matching the known nonnegatively curved examples [2410.05685]. These results also clarify a possible misunderstanding: rational ellipticity is necessary in the simply connected setting, but it is not sufficient. The Brieskorn \(5\)-manifold \[ M_2=\{(z_1,\dots,z_4)\in S^7\subset\mathbb C^4: z_1^2+z_2^3+z_3^3+z_4^3=0\} \] is simply connected and rationally elliptic, yet its loop-space homology over \(\mathbb F_2\) grows exponentially, so it does not admit any metric with entire Grauert tube [2410.05685]. ## 6. Zoll manifolds, compactification at infinity, and rigidity For Zoll manifolds, the entire condition becomes rigid enough to determine the metric. A Zoll manifold is one for which all geodesics are closed, of common minimal period \(\ell\), and each is embedded on \((0,\ell)\). By Bott–Samelson, its cohomology ring matches exactly one of the five CROSS models \[ S^n,\quad \mathbb RP^n,\quad \mathbb CP^n,\quad \mathbb HP^m,\quad \mathbb OP^2. \] When the type is \(\mathbb CP^n\), Li–Song prove that a compact real-analytic Zoll manifold of type \(\mathbb CP^n\) with entire Grauert tube is isometric to \((\mathbb CP^n,c\,g_{FS})\) for some \(c>0\) [2501.15682]. The proof begins by compactifying each leaf of the Riemann foliation. Periodicity of geodesics shows that every leaf \(L\simeq\mathbb C\) compactifies by adding two points to \[ L\cup\{0,\infty\}\simeq \mathbb CP^1, \] and the added divisor \(D\) yields a compact complex manifold \[ X=TM\cup D. \] In the \(\mathbb CP^n\) case, \(D\) is ample in \(X\), and cohomological calculations together with the adjunction formula and an index count show \[ c_1(X)=(n+1)[D]. \] Hence \(X\) is Fano of dimension \(2n\), Picard number \(2\), and index \(n+1\). Wiśniewski’s classification then identifies \(X\) with \(\mathbb CP^n\times\mathbb CP^n\), with \(D\) a divisor of bi-degree \((1,1)\). Comparing the given exhaustion with the model log-cosh potential along compactified leaves forces agreement of the Kähler metrics and therefore the zero-section metric is exactly Fubini–Study up to scaling [2501.15682]. A parallel argument now exists for the quaternionic projective plane. If a Zoll manifold of type \(\mathbb HP^2\) admits an entire Grauert tube, then it is isometric, up to rescaling, to the canonical \(\mathbb HP^2\) [2508.03352]. In that case the projective algebraization \(X=TM\sqcup D\) is an \(8\)-dimensional Fano manifold with \[ H^2(X)\cong\mathbb Z,\qquad -K_X=\mathcal O_X(6D),\qquad \rho(X)=1,\qquad r=6. \] By Mukai’s classification of coindex-\(3\) Fano manifolds and comparison of Betti numbers, \(X\) must be \(\mathrm{Gr}(2,6)\), and comparison of the canonical exhaustion on \(\mathrm{Gr}(2,6)\setminus D_{\mathrm{std}}\cong T\mathbb HP^2\) with the given exhaustion again identifies the zero-section metric [2508.03352]. These rigidity theorems show that entireness, combined with Zoll periodicity and CROSS-type cohomology, is strong enough to recover the canonical metric rather than merely the underlying manifold. ## 7. Algebraization and Burns’ conjecture Burns asked in 1982 whether an entire Grauert tube is an affine algebraic variety. Song gives a partial answer: after removing a suitable codimension-one analytic subset, an entire Grauert tube becomes affine [2605.05582]. The central object is the pole locus \(S\) of the meromorphic complexified metric \[ \mathbf g\in H^0(TM,K_{TM}^{-2}), \] called the **tube singularity**. For every point \(p\in TM\setminus S\), there exists a holomorphic function \[ b_p\in\mathcal O(TM),\qquad b_p(p)=1,\qquad b_p|_S=0, \] such that \[ TM\setminus b_p^{-1}(0) \] is biholomorphic to a smooth affine variety on which \(1/b_p\) is a regular function [2605.05582]. Thus \(TM\setminus S\) is covered by affine Zariski opens, and the codimension-one pole set is the only identified obstruction to global affineness. The proof uses a revised version of Demailly’s affineness criterion for Stein manifolds. If \(X\) is Stein with strictly plurisubharmonic exhaustion \(\phi\), finite Monge–Ampère volume \[ \int_X(dd^c\phi)^n<\infty, \] and an auxiliary function \(\psi\in L^1_{\mathrm{loc}}(X,\mathbb R)\) satisfying \[ Ric\bigl(dd^c(e^\phi)\bigr)+\tfrac12\,dd^c\psi\ge 0, \qquad \psi\le A\phi+B, \] then the singularity set \[ S=\{\,x\in X\mid e^{-2\psi}\notin L^1_{\mathrm{loc}}(x)\} \] controls affineness. If \(S=\varnothing\), then \(X\) itself is biholomorphic to an affine variety [2605.05582]. For entire Grauert tubes, Song takes \[ \phi=\log(1+\cosh u), \qquad u=\sqrt{2E}, \] and defines \[ \psi:=-\log\|\det(\mathbf g)\|_{dd^c(e^\phi)}. \] This produces the required curvature inequality and growth bound, so the revised criterion applies directly [2605.05582]. This has immediate consequences for known rigid cases. When the tube singularity vanishes, the entire Grauert tube is globally affine. Song notes that in Zoll-metric cases, where Burns–Leung prove \(S=\varnothing\), Burns’ conjecture follows [2605.05582]. More generally, the result shows that entire Grauert tubes are already very close to affine algebraic varieties, even when the global conjecture remains open. A broader implication is that entire Grauert tubes occupy a rare position among geometric complexifications: they begin from real-analytic Riemannian data, are governed by the homogeneous complex Monge–Ampère equation, constrain curvature and entropy, often force rational ellipticity or rigid model geometry, and in favorable cases admit projective or affine algebraization [2101.04368] [2501.15682] [2605.05582]. Source: https://www.emergentmind.com/topics/entire-grauert-tube