2
$\begingroup$

This question goes on where this question ended. Given is an non-empty $(n-1)$-dimensional ($n\ge2$) differentiable submanifold $X\subset\mathbb{R}^n$ such that there exists an open $U\subset\mathbb{R}^n$ with $X\subset U$ and a differentiable function $f:U\to\mathbb{R}$ such that $X=f^{-1}(0)$ and $Df(p)\neq0$ for all $p\in X$. Then $X$ is (as shown in the other question) the topological boundary in $U$ of $A:=\{p\in U\,|\,f(p)<0\}$.

I have to show that $X$ is an orientable manifold. That is, there is an differentiable atlas $\mathcal{A}$ such that $\det(D(h_i\circ h_j^{-1})(h_j(p))>0$ for all $p\in U_i\cap U_j$, for all charts $(U_i,h_,V_i),(U_j,h_j,V_j)\in\mathcal{A}$. Alternatively, there exists a $(n-1)$-form on $X$ which vanishes nowhere.

Now intuitively I feel like this is clear. Let $V\subset X$ be a connected component; I think that it is the case that $X$ is actually the smooth boundary of $A$, since if it wasn't, then $V$ (or some other connected component) would have some sort of kink somewhere, which would contradict it being a submanifold; as $A$ is open in $\mathbb{R}^n$ (since $U$ is open and $A$ is open in $U$), $A$ is orientable (restricting the standard orientation on $\mathbb{R}^n$ to $A$), and an orientation on $A$ induces an orientation on $\partial A=X$. Can this be turned into a formal argument?

Alternatively, how would one approach a proof using the two definitions of orientation given above? As the gradient $Df(p)$ is non-zero everywhere on $X$, and $X$ is the level set of $f$ at $C=0$, the gradient would be an ideal candidate to provide the outward pointing normal vector needed.

Any help is much appreciated. This question has kept me busy for some time now.

$\endgroup$
  • $\begingroup$ It would be helpful if you could explain what $A$ is without referring to some answer to some other question. $\endgroup$ – Thomas May 10 '16 at 19:03
  • $\begingroup$ You are right, I have edited. It wasn't in the answer of some other question, but in the question itself, though. $\endgroup$ – B. Pasternak May 10 '16 at 19:04
1
$\begingroup$

The following $n-1$ form $\omega$ on $X$ is nowhere zero and thus orients $X$:

For $x\in X$ and $v_1,v_2,\cdots,v_{n-1}\in T_x(X)=(\operatorname {grad}f(x) )^\perp\subset T_x(\mathbb R^n)= \mathbb R^n$ define $$\omega_x(v_1,v_2,\cdots,v_{n-1})=\det (v_1,\cdots,v_{n-1},\operatorname {grad}f(x) ).$$

$\endgroup$

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.