Let $f: (M, g) \to (M, g)$ be a conformal diffeomorphism of the riemannian manifold $(M, g)$, with

$$ g(f(p))(Df(p) \cdot v_1, Df(p) \cdot v_2) = \mu^2(p) g(p)(v_1, v_2), \quad \forall p \in M, \, \forall v_1, v_2 \in T_p M, $$

for a certain function $\mu \in C^{\infty}(M)$. Is it possible to conformally change the metric of $M$ so as to $f$ become an isometry?

Explicitly, does there exist a metric $\tilde{g} = \alpha g$ in $M$ such that

$$\tilde{g}(f(p))(Df(p) \cdot v_1, Df(p) \cdot v_2) = \tilde{g}(p)(v_1, v_2), \quad \forall p \in M, \, \forall v_1, v_2 \in T_p M \, \text{ ?}$$

Plugging $\tilde{g} = \alpha g$ in the above equation, we obtain that $\alpha$ must satisfy

$$ \alpha(p) = \mu^2(p) \alpha(f(p)), \quad \forall p \in M. $$

Can we continue?


Theorem. Let $C(M)$ be the conformal group of a Riemannian manifold $M$ with $dim(M)=n \ge 2$. If $M$ is not conformally equivalent to $S^n$ or $E^n$, then $C(M)$ is inessential, i.e. can be reduced to a group of isometries by a conformal change of metric.

This theorem has a long and complicated history, you can find its proof and the historic discussion in

J. Ferrand, The action of conformal transformations on a Riemannian manifold. Math. Ann. 304 (1996), no. 2, 277–291.

In view of this theorem, whenever $f: (M,g)\to (M,g)$ is a conformal automorphism without a fixed point, then there exists a positive function $\alpha$ on $M$ such that $f: (M,\alpha g)\to (M,\alpha g)$ is an isometry. I will prove it in the case when $(M,g)$ is conformal to the sphere and leave you the case of $E^n$ as it is similar.

Every conformal automorphism $f$ of the standard sphere which does not have a fixed point in $S^n$ has to have a fixed point $p$ in the unit ball $B^{n+1}$. (I am using the Poincare extension of conformal transformations of $S^n$ to the hyperbolic $n+1$-space in its unit ball model.) After conjugating $f$ via an automorphism $q$ of $S^n$ (sending $p$ to the center of the ball), we obtain $h=q f q^{-1}$ fixing the origin in $B^{n+1}$, which implies that $f\in O(n+1)$ and, thus, preserves the standard spherical metric $g_0$ on $S^n$. Now, use the fact that $g_0$ is conformal to $g$ and $q^*(g_0)$ is conformal to $g$ as well. qed


This not always possible if $f$ has a fixed point. Consider $R^n$ endowed with the euclidean metric. Let $f$ defined by $f(x)=2x$. The map if is conformal. Suppose that there is a function $g$ which turns the euclidean metric $\langle \cdot,\cdot\rangle$ to a metric invariant by $f$. For every $u,v$ in the tangent space of $0$, you would have $g(0)\langle 2u,2v\rangle =g(0)\langle u,v\rangle$. This is clearly impossible.

  • 1
    $\begingroup$ I don't get where you took the $n$ from. But yes, if $f$ has a fixed point, it is not always possible. What if $f$ has no fixed points? $\endgroup$ Mar 3 '16 at 17:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.