3
$\begingroup$

Let $(M^3,g)$ be a complete and orientable Riemannian $3$-manifold and let $\Sigma^2 \subset M$ be a compact orientable totally geodesic surface embedded in $M$. For $f \in C^{2,\alpha}(\Sigma)$ with small $C^0$ norm, let

$$\Sigma_f = \{ \operatorname{exp}_p(f(p)N(p)) : p \in \Sigma \}$$

be the normal graph of $f$ over $\Sigma$, where, of course, $N$ is a unit normal field for $\Sigma$. What equation does $f$ obey if $\Sigma_f$ is a minimal surface in $(M,g)$?

My thoughts: try to write the first variation formula for the area of $\Sigma_f$. For this, define

$$\varphi_t(p) = \operatorname{exp}_p((f(p) + t\eta(p))N(p)), \quad p \in \Sigma$$

where $\eta \in C^{\infty}(\Sigma)$ is any smooth function and $t$ is small. To compute the area of $\Sigma_f(t) := \varphi_t(\Sigma)$, we need to evaluate the following integral:

$$\operatorname{Area}(\Sigma_f(t)) = \int_\Sigma \sqrt{E_tG_t - F_t^2} dA$$

where $E_t, F_t$ and $G_t$ are the coefficients of the first fundamental form of $\varphi_t$. For instance, if $\{e_1, e_2\}$ is a local orthonormal frame for $\Sigma$, then

$$E_t = g\left( \mathrm{d} \varphi_t(e_1), \mathrm{d} \varphi_t(e_1) \right).$$

However, this involves the derivatives of the exponential map with respect to both base point and vector argument. How do I proceed? Do you have a faster argument?

$\endgroup$
1
  • 2
    $\begingroup$ See Appendix A (p. 66) here arxiv.org/abs/1803.02716 (this computation probably exists in many places in the literature, but it was important for us to carry out the full details so you might find it useful). I guess the only difference between your idea and what we do is that we use Fermi coordinates over $\Sigma$. Note that this computation is bound to be somewhat complicated if you really want to know all of the terms, there's not really any getting around this. Note that in most situations you can get away without the exact equation just using the second variation formula etc $\endgroup$ Commented Jul 12, 2021 at 17:44

0

You must log in to answer this question.