LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM

1. Generalized Inverse Function Theorem

We will start with
Theorem 1.1 (Generalized Inverse Function Theorem, compact version).
Let f : M → N be a smooth map that is one-to-one on a compact submanifold X of
M . Moreover, suppose dfx : Tx M → Tf (x) N is a linear diffeomorphism for each x ∈ X.
Then f maps a neighborhood U of X in M diffeomorphically onto a neighborhood V of
f (X) in N .

Note that if we take X = {x}, i.e. a one-point set, then is the inverse function
theorem we discussed in Lecture 2 and Lecture 6. According to that theorem, one can
easily construct neighborhoods U of X and V of f (X) so that f is a local diffeomor-
phism from U to V . To get a global diffeomorphism from a local diffeomorphism, we
will need the following useful lemma:
Lemma 1.2. Suppose f : M → N is a local diffeomorphism near every x ∈ M . If f
is invertible, then f is a global diffeomorphism.

Proof. We need to show f −1 is smooth. Fix any y = f (x). The smoothness of f −1 at

y depends only on the behaviour of f −1 near y. Since f is a diffeomorphism from a
neighborhood of x onto a neighborhood of y, f −1 is smooth at y. 

Proof of Generalized IFT, compact version. According to Lemma 1.2, it is enough to

prove that f is one-to-one in a neighborhood of X. We may embed M into RK , and
consider the “ε-neighborhood” of X in M :
X ε = {x ∈ M | d(x, X) < ε},
where d(·, ·) is the Euclidean distance. Note that X ε is a bounded set in RK since X
is compact. And we have X = k>0 X 1/k , again since X is compact.
T

If f is not one-to-one on each X 1/k , then one can find ak 6= bk ∈ X 1/k such that
f (ak ) = f (bk ). Since all ak ’s lie in a bounded closed set in RK , which is compact,
one can find a subsequence such that aki → a∞ ∈ X. Similarly there is a subsequence
bkij → b∞ ∈ X. Since f (a∞ ) = f (b∞ ), one must have a∞ = b∞ , since f is one-to-one on
X. So in any neighborhood of a∞ , f is not one-to-one. But dfa∞ is linear isomorphism
implies that f is a local diffeomorphism near a∞ , which is a contradiction. 

By staring at the proof, one can see that in the proof, we only need to require that
X is a compact subset in M . So we in fact proved

1
2 LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM

Theorem 1.3 (Generalized Inverse Function Theorem, compact subset version).

Let f : M → N be a smooth map that is one-to-one on a compact subset X of M .
Moreover, suppose dfx : Tx M → Tf (x) N is a linear diffeomorphism for each x ∈ X.
Then f maps a neighborhood U of X in M diffeomorphically onto a neighborhood V of
f (X) in N .
It is natural to ask: what if we remove the “compactness” assumption? In general,
the theorem is not true, even if we assume that X is a smooth submanifold. Here is a
simple example:
Example. Consider the map
f : R2 → T2 , (t, s) 7→ (eit , eis ).
It is easy to see that f is a local diffeomorphism near any (t, s) ∈ R2 . However, if we
take X be the “irrational-slope line”
√
X = {(t, 2t) | t ∈ R}
in R2 (which is a perfectly nice smooth submanifold), then there is no hope to find
neighborhoods U of X in R2 and V of f (X) in T2 so that f maps U diffeomorphically
onto V , since f (X) is dense in T2 .
Note that if X is a compact submanifold of X, then f (X) is in fact a compact
smooth submanifold of N (since f |X : X → N is an embedding), and f is a diffeo-
morphism from X onto f (X). In the example above, the bad thing is that the image
f (X) of X is not a smooth submanifold of T2 (but only an immersed submanifold),
and thus f is not a diffeomorphism from X onto f (X) (with subspace topology from
the target).
It turns out that if we require f (X) to be a smooth submanifold of N , and require
f : X → f (X) to be a diffeomorphism, then we still have the generalized inverse
function theorem near X:
Theorem 1.4 (Generalized Inverse Function Theorem, non-compact version).
Let f : M → N be a smooth map that is one-to-one on a smooth submanifold X of M .
Moreover, suppose f maps X diffeomorphically onto f (X) (which is assumed to be a
smooth submanifold in N ), and suppose dfx : Tx M → Tf (x) N is a linear diffeomorphism
for each x ∈ X. Then f maps a neighborhood U of X in M diffeomorphically onto a
neighborhood V of f (X) in N .
Proof. The idea is standard: Any non-compact manifold can be written as a union
of many “compact stripes”. By using S∞any positive smooth −1 exhaustion function g on
f (X), we may decompose f (X) = k=1 Kk , where Kk = g ([k, k + 1]). Since f |X :
X → f (X) is a diffeomorphism, each Jk := f |−1 X (Kk ) is compact in X and give such a
decomposition for X. In particular, by Theorem 1.3, there is an open neighborhood U ek
of Jk−1 ∪Jk ∪Jk+1 in M , such that f is a diffeomorphism on Uk . Since f (X) is a smooth
e
K
submanifold in N , by embeddingS N into R and using the induced distance function,
one may prove dk := dist(Kk , j>k+1 Kj ) > 0 (Here we need to use the compactness
of Kk and use the fact that f (X) is a smooth submanifold).
 LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM 3

Now we choose a decreasing sequence of positive numbers ε1 > ε2 > · · · > 0 such
that εk < dk /2 for each k. Note that this implies Klεl ∩ Kkεk = ∅ if |l − k| > 1. Define
Uk = Uek−1 ∩ U
ek ∩ Uek+1 ∩ f −1 (K εk ).
k
Then Uk is an open neighborhood of Jk , Vk = f (Uk ) is an open neighborhood of
Kk , and f maps Uk diffeomorphically onto Vk . Moreover, the definition of Uk implies
Uk−1 ∪ Uk ∪ Uk+1 ⊂ Uek .
S S
We define U = k≥1 Uk and V = k≥1 Vk . Then U is an open neighborhood
of X in M , V is an open neighborhood of f (X) in N , and f : U → V is a local
diffeomorphism everywhere. It remains to show f is injective. Suppose x, y ∈ U
and f (x) = f (y). Then there exists k such that f (x) = f (y) ∈ Vk ⊂ Kkεk . It follows
x, y ∈ Uk−1 ∪Uk ∪Uk+1 ⊂ Uek . Since f is a diffeomorphism on U
ek , we conclude x = y. 

2. Tubular Neighborhood Theorem

Now let X ⊂ M be a smooth submanifold. It is natural to ask: What does M
looks like “near” X? The famous tubular neighborhood theorem claims that X always
admits a “tubular” neighborhood inside M . Moreover, the tubular neighborhood looks
like a neighborhood of X inside its “normal bundle”. This gives some kind of “canonical
form” of a neighborhood of any smooth submanifold.
We first prove the following “Euclidean version” of the tubular neighborhood the-
orem, which gives us a nice neighborhood of any submanifold in RK :
Theorem 2.1 (ε-Neighborhood Theorem). Let ι : X → RK be a smooth submanifold.
Then there exists a smooth positive function ε : X → R+ , such that if we let X ε be the
ε-neighborhood of X,
X ε := {y ∈ RK | |y − x| < ε(x) for some x ∈ X},
then
(1) each y ∈ X ε possesses a unique closest point πε (y) in X;
(2) the map π : X ε → X is a submersion.
(Note: If X is compact, then ε can be taken to be a constant.)
The ε-neighborhood described in the above theorem looks like

We will prove the theorem by constructing a diffeomorphism between a neighbor-

hood of X inside its normal bundle (which can be think of as a nice “non-intersecting”
4 LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM

way to put the green lines above together to form a smooth manifold) and a neigh-
borhood of X inside RK . So let’s first recall the conception of the normal bundle
of a manifold embedded in RK , which is defined in PSet 3 Part 1 Problem 4. Let
ι : X ,→ RK be a smooth submanifold of dimension m. For each x ∈ X, we can
identify Tx X with an m-dimensional vector subspace in RK via
Tx X ' dιx (Tx X) ⊂ Tx RK ' RK .
Let Nx X be the orthogonal complement of Tx X in RK ,
Nx X := {v ∈ Tx RK ' RK | v ⊥ Tx X},
which is a (K − m)-dimensional vector subspace. Define
N X = {(x, v) ∈ RK × RK | x ∈ X, v ∈ Nx X} ⊂ T RK .
In PSet 3 Part 1 Problem 4 you are supposed to show that N X is a K-dimensional
smooth submanifold in T RK , called the normal bundle of X. Moreover, the canonical
projection map
π : N X → X, (x, v) 7→ x
is a submersion.
Remark. The conception of normal bundle N X is extrinsic: it depends on the ambient
space RK and also depends on the way of embedding ι : M → RK .

Proof of the ε-Neighborhood Theorem. Define a map

h : N X → RK , (x, v) 7→ x + v.
Then at any point (x, 0) ∈ N X, dh is non-singular, because dh(x,0) maps T(x,0) (X ×
{0}) ⊂ T(x,0) N X bijectively onto Tx X ⊂ Tx RK , and maps the tangent space of
T(x,0) ({x} × Nx X) bijectively onto Nx X ⊂ Tx RK .
Also by definition, h maps X × {0} ⊂ N X diffeomorphically onto X ⊂ RK .
According to the generalized Inverse Function Theorem above, h maps a neighborhood
U of X × {0} in N X diffeomorphically onto a neighborhood V of X in RK .
To show that there exists ε ∈ C ∞ (X) such that X ε ⊂ V , for each x ∈ X we
let ε̃(x) = sup{r ≤ 1 | Br (x) ⊂ V }. Then ε̃ is a positive continuous function on X.
(Can you prove the continuity of ε̃?) By using Whitney approximation theorem for
functions (See Lecture 4), we can find positive ε ∈ C ∞ (X) such that |ε − ε̃/2| < ε̃/4.
In particular we have ε(x) < ε̃(x). It follows that for such ε, X ε ⊂ V .
Now we define
πε : X ε → X, y 7→ πε (y) = π ◦ h−1 (y).
It is a submersion since π is a submersion and h−1 is a diffeomorphism on V . It remains
to prove that πε (y) is the unique closest point to y in X. In fact, let z ∈ X be the
closest point in X to y. Then the sphere centered at y with radius |y − z| is tangent to
X at z. It follows that the vector y − z is perpendicular to X at z, i.e. y − z ∈ Nz X.
So we have y = z + (y − z) = h(z, y − z), i.e. πε (y) = z. This completes the proof. 
 LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM 5

In general, if X ⊂ M is a smooth submanifold. Then one can still define the normal
bundle N (X, M ) (with respect to the ambient space M ) as the set of points of the form
(x, v), where for any x ∈ X, v is any vector in the quotient space
Nx (X, M ) = Tx M/Tx X.
One can show that N (X, M ) is a smooth manifold whose dimension equals dim M . .
To get a more extrinsic (and geometric) description of N (X, M ), we may embed M
into RK . Then we will get an inclusion Tx X ⊂ Tx M ⊂ Tx RK , and the quotient space
Tx M/Tx X can be identified as the space of vectors in Tx M which are perpendicular to
Tx X. It follows
N (X, M ) ' {(x, v) | x ∈ X, v ∈ Tx M and v ⊥ Tx X}.
Note that by using this identification, we have
T(x,0) N (X, M ) ' Tx X ⊕ Tx⊥ X,
where Tx⊥ X is the orthogonal complement of Tx X inside Tx M .
Now we prove
Theorem 2.2 (Tubular Neighborhood Theorem). Let X ⊂ M be a smooth submani-
fold. Then there exists a diffeomorphism from an open neighborhood of X in N (X, M )
onto an open neighborhood of X in M .
Proof. Embed M into RK . Let πε : M ε → M be as in the ε-neighborhood theorem
(for the embedding ι : M ,→ RK ). Again consider the map
h : N (X, M ) → RK , h(x, v) → x + v.
Then W := h−1 (M ε ) is an open neighborhood of X in N (X, M ). Now consider the
composition
hε = πε ◦ h : W −→ M.
Then hε is smooth, and is the identity map on X ⊂ N (X, M ). Moreover, according to
the decomposition of T(x,0) N (X, M ) above, (dhε )(x,0) maps T(x,0) N (X, M ) bijectively
onto Tx M . So the theorem follows from the generalized IFT. 

3. Smooth Approximations of Continuous Maps

Recall that the Whitney approximation theorem for functions in Lecture 4 tells us
that one can approximate any continuous function by a smooth function in arbitrary
accuracy. Now let M, N be smooth manifolds and let g : M → N be a continuous map.
A natural question to ask is: Can we “approximate” g by a smooth map f : M → N ?
Of course there is one issue in proposing the above question: what do we mean by
“f approximate g”? If N is RK , one can require |f − g| < ε which is exactly what
we did in Lecture 4. But in the general case, f (x) and g(x) are points in a smooth
manifold N , and it makes no sense to talk about f (x) − g(x). (Of course one can
embed N into RK . But then |f − g| will depend on the way of embedding, and thus
is not a good way of measuring whether f and g are “close”, since f and g themselves
has noting to do with the way of embedding N into RK .)
6 LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM

In topology, we do have one way to talk about the relation between two maps,
which is very useful in many contexts. That is the conception of homotopy. Roughly
speaking, we say two maps f0 and f1 are homotopic if one can continuously deform f0
to f1 . More precisely,
Definition 3.1. Let X, Y be topological spaces. We say f0 , f1 ∈ C 0 (X, Y ) are homo-
topic if there exists F ∈ C 0 (X × [0, 1], Y ) so that
F (·, 0) = f0 (·) and F (·, 1) = f1 (·).

Now we prove the Whitney Approximation Theorem for maps, which claims that
any continuous map between smooth manifolds can be continuously deformed to a
smooth map:
Theorem 3.2 (Whitney Approximation Theorem for Continuous Maps).
Given any continuous mapping g ∈ C 0 (M, N ), one can find a smooth mapping f ∈
C ∞ (M, N ) which is homotopic to g. Moreover, if g is smooth on a closed subset A ⊂ M ,
then one can choose f so that f = g on A.

Proof. We embed N into RK , and consider N ε , the ε-neighborhood of N in RK . Since

ε is a positive smooth function on N , the composition ε = ε ◦ g is a positive continuous
function on M . Think of g as a continuous function from M to RN . According to the
Whitney approximation theorem in Lecture 4, there is a smooth function f˜ : M → RK
which is ε-close to g, i.e.
|f˜(x) − g(x)| < ε(g(x)), ∀x ∈ M.
So f˜(x) ∈ B(g(x), ε(g(x))). It follows that
(1 − t)g(x) + tf˜(x) ∈ B(g(x), ε(g(x))) ⊂ N ε , ∀ 0 ≤ t ≤ 1.
Now define F : M × [0, 1] → N by
F (x, t) = πε ((1 − t)g(x) + tf˜(x)).
Then F is a homotopy that connects the continuous map g to the smooth map
f = F (·, 1) = πε ◦ f˜ : M → N.
Finally note that if g is smooth on a closed subset A, then the smooth function
f˜ can be chosen so that f˜ = g on A. It follows that f = g = F (·, t) on A. (In other
words, the homotopy connecting g to f can be chosen to be relative to A.) 

Since in this course, we are mainly interested in smooth objects (smooth manifolds,
smooth submanifolds, smooth functions, smooth maps, smooth vector fields, smooth
vector bundle, smooth differential form etc), we are interested in homotopies connecting
two smooth maps via “smooth path”, i.e.
Definition 3.3. We say f0 , f1 ∈ C ∞ (M, N ) are smoothly homotopic if there exists
F ∈ C ∞ (M × [0, 1], N ) so that
F (·, 0) = f0 and F (·, 1) = f1 .
 LECTURE 10: TUBULAR NEIGHBORHOOD THEOREM 7

Of course if f0 and f1 are smoothly homotopic, then they are homotopic. Con-
versely, we have
Theorem 3.4 (Homotopy ≡ Smooth homotopy).
Suppose f0 , f1 ∈ C ∞ (M, N ) are homotopic, then they are smoothly homotopic.
Proof. Let F : M × [0, 1] → N be a homotopy connecting f0 and f1 . Continuously
extend F to a mapping Fe : M × R → N by defining
Fe(x, t) = F (x, 0) if t ≤ 0, and Fe(x, t) = F (x, 1) if t ≥ 1.
Then Fe is a continuous map from M × R to N , and is smooth on closed subsets
M × {0} and M × {1}. (Recall that by definition in Lecture 4, F̃ is smooth on M × {0}
means there is a smooth function G̃ defined on M × (−ε, ε) so that G̃(x, 0) = F̃ (x, 0).
We don’t require F̃ to be smooth in a neighborhood of M × {0}.) So by Whitney
Approximation Theorem for maps above, there exists a smooth map F : M × R → N
(that is homotopic to Fe, which we don’t need here), such that F = Fe on M × {0} and
M × {1}, i.e.
F (·, 0) = f0 and F (·, 1) = f1 .
It follows that F is the desired smooth homotopy connecting f0 and f1 . 
Note that homotopy is an equivalence relation on the space of continuous maps from
M to N . The equivalence classes are called homotopy classes of maps. Theorem 3.4
implies that smooth homotopy is an equivalence relation on the space of smooth maps
from M to N . Moreover, combining Theorem 3.2 and Theorem 3.4, we immediately
see that homotopy classes of continuous maps coincides with the smooth homotopy
classes of smooth maps.