On Kähler Metrisability of Two-dimensional Complex Projective Structures

3 Kähler metrisability

In this section we will derive necessary conditions for a complex projective structure \([\nabla]\) on a complex surface \((M,J)\) to arise via the Levi-Civita connection of a (pseudo-)Kähler metric. There exists a complex projectively invariant linear first order differential operator acting on \(J\)-hermitian \((2,\! 0)\) tensor fields on \(M\) with weight \(1/3\), i.e sections of the bundle \(S^2_J(TM)\otimes \left(\Lambda^4(T^*M)\right)^{1/3}\). This differential operator has the property that nondegenerate sections in its kernel are in one-to-one correspondence with (pseudo-)Kähler metrics on \(M\) whose Levi-Civita connection is compatible with \([\nabla]\) (see [12, 26, 31]).

3.1 The differential analysis

We will show that in the surface case, the (pseudo-)Kähler metrics on \((M,J,[\nabla])\) whose Levi-Civita connection is compatible with \([\nabla]\) can equivalently be characterised in terms of a differential system on Cartan’s bundle \((\pi : B \to M,\theta)\).

Proposition 3.1

Suppose the (pseudo-)Kähler metric \(g\) is compatible with \([\nabla]\). Then, writing \(\pi^*g=g_{i\bar\jmath}\theta^i_0\circ\overline{\theta^{\jmath}_0}\) and setting \(h_{i\bar\jmath}=g_{i\bar\jmath}\left(g_{1\bar 1}g_{2\bar 2}-|g_{1\bar 2}|^2\right)^{-2/3}\), we have \[\tag{3.1} \mathrm{d}h_{i\bar \jmath}=h_{i\bar\jmath}\left(\theta^0_0+\overline{\theta^0_0}\,\right)+h_{i\bar s}\overline{\theta^s_{j}}+h_{s\bar\jmath}\theta^s_i +h_i\overline{\varepsilon_{sj}\theta^s_0}+\overline{h_{j}}\varepsilon_{si}\theta^s_0\] for some complex-valued functions \(h_i\) on \(B\). Conversely, suppose there exist complex-valued functions \(h_{i\bar \jmath}=\overline{h_{j\bar \imath}}\) and \(h_i\) on \(B\) solving (3.1) and satisfying \(\left(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2\right)\neq 0\), then the symmetric \(2\)-form \[h_{i\bar\jmath}\left(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2\right)^{-2}\theta^i_0\circ \overline{\theta^{\jmath}_0}\] is the \(\pi\)-pullback of a \([\nabla]\)-compatible (pseudo-)Kähler metric on \(M\).

Proof. Let \(g\) be a (pseudo-)Kähler metric on \((M,J)\) and write \(g=g_{i\bar\jmath}\,\mathrm{d}z^i\circ \mathrm{d}\overline{z^\jmath}\) for local holomorphic coordinates \(z=(z^1,z^2) : U \to \mathbb{C}^2\) on \(M\). Denoting by \(\nabla\) the Levi-Civita connection of \(g\), on \(U\) the identity \(\nabla g=0\) is equivalent to \[\frac{\partial g_{k\bar\jmath}}{\partial z^i}=g_{s\bar\jmath}\Gamma^s_{ik}\qquad \text{and}\qquad \frac{\partial g_{k\bar\jmath}}{\partial \overline{z^{\imath}}}=g_{k\bar s}\overline{\Gamma^s_{ij}},\] where \(\Gamma^i_{jk}\) denote the complex Christoffel symbols of \(\nabla\). Abbreviate \(G=\det g_{i\bar\jmath}\), then we obtain \[\frac{\partial G}{\partial z^i}=G\,\Gamma^s_{si}.\] Hence, the partial derivative of \(h_{k\bar\jmath}=g_{k\bar\jmath}\,G^{-2/3}\) with respect to \(z^i\) becomes \[\begin{aligned} \frac{\partial h_{k\bar\jmath}}{\partial z^i}&=g_{l\bar\jmath}\,\Gamma^l_{ik}\,G^{-2/3}-\frac{2}{3}g_{k\bar\jmath}\,\Gamma^s_{si}\,G^{-2/3}=h_{l\bar\jmath}\left(\Gamma^l_{ik}-\frac{2}{3}\Gamma^s_{si}\delta^l_k\right)\\ &=h_{l\bar\jmath}\left(\Gamma^l_{ik}-\frac{1}{3}\Gamma^s_{si}\delta^l_k-\frac{1}{3}\Gamma^s_{sk}\delta^l_i\right)-\frac{1}{3}h_{l\bar\jmath}\left(\Gamma^s_{si}\delta^l_k-\Gamma^s_{sk}\delta^l_i\right). \end{aligned}\] Note that the last two summands in the last equation are antisymmetric in \(i,k\), so that we may write \[-\frac{1}{3}h_{l\bar\jmath}\left(\Gamma^s_{si}\delta^l_k-\Gamma^s_{sk}\delta^l_i\right)=\overline{h_j}\varepsilon_{ik}\] for unique complex-valued functions \(h_i\) on \(U\). We thus get \[\tag{3.2} \frac{\partial h_{k\bar\jmath}}{\partial z^i}=h_{s\bar\jmath}\Pi^s_{ik}+\overline{h_j}\varepsilon_{ik}.\] In entirely analogous fashion we obtain \[\tag{3.3} \frac{\partial h_{k\bar\jmath}}{\partial \overline{z^{\imath}}}=h_{k\bar s}\overline{\Pi^s_{ij}}+h_k\overline{\varepsilon_{ij}}.\] Recall from Theorem 2.2 that the coordinate system \(z : U \to \mathbb{C}^2\) induces a unique section \(\sigma_z : U \to B\) of Cartan’s bundle such that \[\tag{3.4} \left(\sigma_z\right)^*\theta^0_0=0, \qquad \left(\sigma_z\right)^*\theta^i_0=\mathrm{d}z^i, \qquad \left(\sigma_z\right)^*\theta^i_j=\Pi^i_{jk}\mathrm{d}z^k.\] Consequently, using (3.2), (3.3), (3.4) we see that (3.1) is necessary.

Conversely, suppose there exist complex-valued functions \(h_{i\bar\jmath}=\overline{h_{j\bar\imath}}\) and \(h_i\) on \(B\) solving (3.1) for which \[\left(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2\right)\neq 0.\] Setting \(g_{i\bar\jmath}=h_{i\bar\jmath}\left(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2\right)^{-2}\) we get \[\tag{3.5} \mathrm{d}g_{i\bar \jmath}=-g_{i\bar \jmath}\left(\theta^0_0+\bar\theta^0_0\right)+g_{i\bar s}\overline{\theta^s_{j}}+g_{s\bar\jmath}\theta^s_i+g_{i\bar \jmath\bar s}\overline{\theta^s_0}+g_{i\bar \jmath s}\theta^s_0\] with \[g_{i\bar \jmath \bar s}=\frac{(h_{i\bar \jmath}h_{l\bar s}+h_{i\bar s}h_{l\bar \jmath})\varepsilon^{lk}h_k}{(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2)^3}, \quad \text{and} \quad g_{i\bar \jmath k}=\frac{(h_{i\bar \jmath}h_{k\bar s}+h_{k\bar\jmath}h_{i\bar s})\overline{\varepsilon^{su}{h_u}}}{(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2)^3}.\] It follows with (3.5) that there exists a unique \(J\)-Hermitian metric \(g\) on \(M\) such that \(\pi^*g=g_{i\bar\jmath}\,\theta^i_0\circ \overline{\theta^{\jmath}_0}\). Choose local holomorphic coordinates \(z=(z^1,z^2) : U \to \mathbb{C}^2\) on \(M\). By abuse of notation we will write \(g_{i\bar\jmath},g_{i\bar \jmath \bar s},g_{i\bar \jmath k}\) for the pullback of the respective functions on \(B\) by the section \(\sigma_z : U \to B\) associated to \(z\). From (3.5) we obtain \[\frac{\partial g_{i\bar\jmath}}{\partial z^s}=g_{u\bar\jmath}\Pi^u_{is}+g_{i\bar\jmath s}=g_{u\bar\jmath}\left(\Pi^u_{is}+g^{\bar v u}g_{i\bar v s}\right)=g_{u\bar\jmath}\left(\Pi^u_{is}+\delta^u_ib_s+\delta^u_sb_i\right)=g_{u\bar\jmath}\Gamma^u_{is}\] where we write \[b_i=\frac{h_{i\bar s}\overline{\varepsilon^{su}h_u}}{(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2)^{11/3}}\qquad \text{and} \qquad \Gamma^i_{jk}=\Pi^i_{jk}+\delta^i_jb_k+\delta^i_kb_j.\] Likewise we obtain \[\frac{\partial g_{i\bar\jmath}}{\partial \overline{z^{s}}}=g_{i\bar u}\overline{\Gamma^u_{js}}.\] It follows that there exists a complex-linear connection \(\nabla\) on \(U\) defining \([\nabla]\) and whose complex Christoffel symbols are given by \(\Gamma^i_{jk}\). By construction, the connection \(\nabla\) preserves \(g\) and hence must be the Levi-Civita connection of \(g\). Furthermore, \(\nabla\) being complex-linear implies that \(g\) is Kähler. This completes the proof.

3.1.1 First prolongation

Differentiating (3.1) yields \[\tag{3.6} 0=\mathrm{d}^2 h_{i\bar\jmath}=\varepsilon_{si}\overline{\eta_j}\wedge\theta^s_0+\overline{\varepsilon_{sj}}\eta_i\wedge\overline{\theta^s_0}-(h_{s\bar\jmath}W^{s}_{iv\bar u}+h_{i\bar s}\overline{W^{s}_{ju\bar v}})\overline{\theta^u_0}\wedge\theta^v_0\] with \[\eta_k=\mathrm{d}h_k+h_k\left(\overline{\theta^0_0}-\theta^0_0\right)-h_j\theta^j_k+\overline{\varepsilon^{ij}}h_{k\bar\jmath}\overline{\theta^0_i}.\] This implies that we can write \[\tag{3.7} \eta_i=a_{ij}\theta^j_0\] for unique complex-valued functions \(a_{ij}\) on \(B\). Equations (3.6) and (3.7) imply \[\tag{3.8} \varepsilon_{ki}\overline{a_{jl}}-\overline{\varepsilon_{lj}} a_{ik}=\overline{h_{j\bar s}}W^s_{ik\bar l}-h_{i\bar s}\overline{W^s_{jl\bar k}}\] Contracting this last equation with \(\overline{\varepsilon^{jl}}\varepsilon^{ik}\) implies that the function \[h=-\frac{1}{2}\overline{\varepsilon^{ij}a_{ij}}\] is real-valued. We get \[a_{jl}=\varepsilon_{jl}h-\frac{1}{2}\overline{\varepsilon^{iu}}h_{s\bar \imath}W^s_{jl\bar u},\] and thus \[\mathrm{d}h_i=h_i\left(\theta^0_0-\overline{\theta^0_0}\,\right)+h_j\theta^j_i+h_{i\bar s}\overline{\varepsilon^{sl}\theta^0_l}+\left(\varepsilon_{ij}h-\frac{1}{2}\overline{\varepsilon^{uv}} h_{s\bar u}W^s_{ij\bar v}\right)\theta^j_0.\] Plugging the formula for \(a_{ij}\) back into (3.8) yields the integrability conditions \[h_{s\bar\jmath}W^s_{ik\bar l}-h_{i\bar s}\overline{W^s_{jl\bar k}}=\frac{1}{2}\overline{\varepsilon_{lj}\varepsilon^{uv}}h_{s\bar u}W^s_{ik\bar v}-\frac{1}{2}\varepsilon_{ki}\varepsilon^{uv}h_{u\bar s}\overline{W^s_{jl\bar v}}.\] This last equation can be simplified so that we obtain:

Proposition 3.2

A necessary condition for a complex projective surface \((M,J,[\nabla])\) to be Kähler metrisable is that \[\tag{3.9} \overline{h_{j\bar s}}W^s_{ik\bar l}+\overline{h_{l\bar s}}W^s_{ik\bar\jmath}=h_{k\bar s}\overline{W^s_{jl\bar \imath}}+h_{i\bar s}\overline{W^s_{jl\bar k}}\] admits a nondegenerate solution \(\overline{h_{i\bar\jmath}}=h_{j\bar\imath}\).

Remark 3.3

Note that under suitable constant rank assumptions the system (3.9) defines a subbundle of the bundle over \(M\) whose sections are hermitian forms on \((M,J)\). For a generic complex projective structure \([\nabla]\) this subbundle does have rank \(0\).

3.1.2 Second prolongation

We start by computing \[0=\mathrm{d}^2 h_i\wedge\theta^1_0\wedge\theta^2_0=-\left(h_{i\bar\jmath}\overline{\varepsilon^{jk}L_{k}}\right)\theta^1_0\wedge\overline{\theta^1_0}\wedge\theta^2_0\wedge\overline{\theta^2_0}\] which is equivalent to \[\left(\begin{array}{cc} h_{1\bar 1} & h_{1\bar 2} \\ h_{2\bar 1} & h_{2\bar 2}\end{array}\right)\cdot \left(\begin{array}{r} \overline{L_2} \\ -\overline{L_1}\end{array}\right)=0\] which cannot have any solution with \((h_{11}h_{22}-|h_{12}|^2)\neq 0\) unless \(L_1=L_2=0\). This shows:

Theorem 3.4

A necessary condition for a complex projective surface to be Kähler metrisable is that it is Liouville-flat, i.e. its complex projective Liouville curvature vanishes.

Remark 3.5

Note that the vanishing of the Liouville curvature is equivalent to requesting that the curvature of \(\theta\) is of type \((1,\! 1)\) only, which agrees with general results in [9].

Assuming henceforth \(L_1=L_2=0\) we also get \[\tag{3.10} 0=\mathrm{d}^2h_i=\left(\varepsilon_{ij}\eta+\varphi_{ij}\right)\wedge\theta^j_0\] with \[\eta=\mathrm{d}h +2h\mathrm{Re}(\theta^0_0)+2\varepsilon^{ij}\mathrm{Re}(h_i\theta^0_j)-\frac{1}{2}\varepsilon^{kl}h_{k\bar \imath}\overline{\varepsilon^{ij}K_{js\bar l}\theta^s_0}\] and \[\varphi_{ij}=\mathrm{d}r_{ij}+r_{ij}\overline{\theta^0_0}-r_{si}\theta^s_j-r_{sj}\theta^s_i-h_lW^l_{ij\bar s}\overline{\theta^s_0}+\frac{1}{2}\overline{\varepsilon^{uv}}\left(h_{i\bar u}\overline{K_{vs\bar\jmath}}+h_{j\bar u}\overline{K_{vs\bar \imath}}\right)\overline{\theta^s_0}\] where \[r_{ij}=-\frac{1}{2}\overline{\varepsilon^{uv}} h_{s\bar u}W^s_{ij\bar v}.\] It follows with Cartan’s lemma that there are functions \(a_{ijk}=a_{ikj}\) such that \[\varepsilon_{ij}\eta+\varphi_{ij}=a_{ijk}\theta^k_0.\] Since \(\varphi_{ij}\) is symmetric in \(i,j\), this implies \[\eta=\frac{1}{2}\varepsilon^{ji}a_{ijs}\theta^s_0.\] Since \(h\) is real-valued, we must have \[\varepsilon^{ji}a_{ijs}=\overline{\varepsilon^{uv}}\varepsilon^{kl}h_{k\bar u}K_{ls\bar v}.\] Concluding, we get \[\mathrm{d}h=-2h\mathrm{Re}(\theta^0_0)+2\varepsilon^{kl}\mathrm{Re}(h_l\theta^0_k)+\frac{1}{2}\overline{\varepsilon^{ij}}\varepsilon^{kl}\mathrm{Re}(h_{k\bar \imath}K_{ls\bar\jmath}\theta^s_0).\] This completes the prolongation procedure.

Remark 3.6

Note that further integrability conditions can be derived from (3.10), we won’t write these out though.

Using Proposition 3.1 we obtain:

Theorem 3.7

Let \((M,J,[\nabla])\) be a complex projective surface with Cartan geometry \((\pi : B \to M,\theta)\). If \(U\subset B\) is a connected open set on which there exist functions \(h_{i\bar\jmath}=\overline{h_{j\bar\imath}}\), \(h_i\) and \(h\) that satisfy the rank \(9\) linear system \[\tag{3.11} \begin{aligned} \mathrm{d}h_{i\bar\jmath}&=2h_{i\bar\jmath}\mathrm{Re}(\theta^0_0)+h_{i\bar s}\overline{\theta^{s}_j}+h_{s\bar\jmath}\theta^s_i+h_i\overline{\varepsilon_{sj}\theta^{s}_0}+\overline{h_j}\varepsilon_{si}\theta^s_0,\\ \mathrm{d}h_k&=2\mathrm{i}h_k\mathrm{Im}(\theta^0_0)+h_l\theta^l_k+h_{k\bar \imath}\overline{\varepsilon^{ij}\theta^0_j}+\left(\varepsilon_{kl}h-\frac{1}{2}\overline{\varepsilon^{ij}} h_{s\bar \imath} W^s_{kl\bar\jmath}\right)\theta^l_0,\\ \mathrm{d}h&=-2h\mathrm{Re}(\theta^0_0)-2\varepsilon^{lk}\mathrm{Re}(h_l\theta^0_k)+\frac{1}{2}\overline{\varepsilon^{ij}}\varepsilon^{kl}\mathrm{Re}(h_{k\bar \imath}K_{ls\bar\jmath}\theta^s_0), \end{aligned}\] and \((h_{1\bar 1}h_{2 \bar 2}-|h_{1\bar 2}|^2)\neq 0\), then the quadratic form \[g=\frac{h_{i\bar\jmath}\theta^i_0\circ \overline{\theta^{\jmath}_0}}{(h_{1\bar 1}h_{2\bar 2}-|h_{1\bar 2}|^2)^2}\] is the \(\pi\)-pullback to \(U\) of a (pseudo-)Kähler metric on \(\pi(U) \subset M\) that is compatible with \([\nabla]\).

From this we get:

Corollary 3.8

The Kähler metrics defined on some domain \(U\subset \mathbb{CP}^2\) which are compatible with the standard complex projective structure on \(\mathbb{CP}^2\) are in one-to-one correspondence with the hermitian forms on \(\mathbb{C}^3\) whose rank is at least two.

Proof. Suppose the complex projective structure \([\nabla]\) has vanishing complex projective Weyl and Liouville curvature. Then the differential system (3.11) may be written as \[\tag{3.12} \mathrm{d}H+\theta H+H\theta^*=0\] with \[H=H^*=\left(\begin{array}{ccc}h & -\overline{h_2} & \overline{h_1}\\ -h_2 & -h_{22} & h_{21}\\ h_1 & h_{12} & -h_{11}\end{array}\right)\] where \({}^*\) denotes the conjugate transpose matrix. Recall that in the flat case \(\theta=g^{-1}\mathrm{d}g\) for some smooth \(g: B \to \mathrm{PSL}(3,\mathbb{C})\), hence the solutions to (3.12) are \[H=g^{-1}C\left(g^{-1}\right)^*\] where \(C=C^*\) is a constant hermitian matrix of rank at least two. The statement now follows immediately with Theorem 3.7.

Remark 3.9

On can deduce from Corollary 3.8 that a Kähler metric \(g\) giving rise to flat complex projective structures must have constant holomorphic sectional curvature. A result first proved in [37] (in all dimensions).

Remark 3.10

One can also ask for existence of complex projective structures \([\nabla]\) whose degree of mobility is greater than one, i.e. they admit several (non-proportional) compatible Kähler metrics. In [15] (see also [21]) it was shown that the only closed complex projective manifold with degree of mobility greater than two is \(\mathbb{CP}^n\) with the projective structure arising via the Fubini-Study metric.