Ma · Mathematics

When Is a Proper Injective Map a Homeomorphism onto Its Image?

Published By the 9to5Science editorial team

At a glance

  1. 1. Check Hausdorffness

    First countability alone is insufficient; assume the target space is also Hausdorff.

  2. 2. Use convergent sequences

    In a first-countable space, continuity can be tested using sequences.

  3. 3. Build a compact set

    A convergent sequence together with its limit is compact.

  4. 4. Restrict the map

    Properness makes the relevant preimage compact, so the restricted bijection is a homeomorphism.

Let f:X→Yf:X\to Y be continuous, injective, and proper, where proper means that the inverse image of every compact subset of YY is compact. The natural question is whether ff is a homeomorphism from XX onto its image f(X)f(X), equipped with the subspace topology.

The answer is yes if YY is Hausdorff and first-countable. The Hausdorff assumption is essential: first countability alone does not suffice.

The corrected theorem

Assume that:

  • f:X→Yf:X\to Y is continuous;
  • ff is injective;
  • ff is proper;
  • YY is Hausdorff and first-countable.

Then

f:X⟶f(X)f:X\longrightarrow f(X)

is a homeomorphism, where f(X)f(X) has the subspace topology inherited from YY.

Because ff is already continuous and injective, only continuity of the inverse

f−1:f(X)→Xf^{-1}:f(X)\to X

needs to be proved.

Why first countability helps

A map from a first-countable space is continuous if it preserves convergent sequences. More precisely, since f(X)f(X) is a subspace of the first-countable space YY, it is also first-countable. Therefore it is enough to show:

Whenever yn→yy_n\to y in f(X)f(X), the corresponding sequence f−1(yn)f^{-1}(y_n) converges to f−1(y)f^{-1}(y) in XX.

Set

x=f−1(y),xn=f−1(yn).x=f^{-1}(y),\qquad x_n=f^{-1}(y_n).

The proof will use properness and compactness.

The compact convergent-sequence argument

Define

K={y}∪{yn:n∈N}.K=\{y\}\cup\{y_n:n\in\mathbb N\}.

The set KK is compact. Indeed, if an open cover contains a set covering yy, convergence yn→yy_n\to y implies that this one open set covers all but finitely many yny_n. The remaining finitely many points can be covered by finitely many members of the cover.

Since ff is proper,

C=f−1(K)C=f^{-1}(K)

is compact in XX.

The restriction

f∣C:C→Kf|_C:C\to K

is continuous and bijective. It is bijective because ff is injective and because every point of KK belongs to f(X)f(X). Since CC is compact and KK is Hausdorff, the standard compact-to-Hausdorff theorem says that f∣Cf|_C is a homeomorphism.

Consequently, its inverse is continuous, and therefore

xn=(f∣C)−1(yn)⟶(f∣C)−1(y)=x.x_n=(f|_C)^{-1}(y_n)\longrightarrow (f|_C)^{-1}(y)=x.

Thus f−1f^{-1} preserves convergent sequences. Since f(X)f(X) is first-countable, f−1f^{-1} is continuous. Hence ff is a homeomorphism onto its image.

A fully worked example

Consider

f:R→R2,f(t)=(t,t2).f:\mathbb R\to\mathbb R^2, \qquad f(t)=(t,t^2).

This map is continuous and injective. It is also proper. To see this, let K⊆R2K\subseteq\mathbb R^2 be compact. Its first-coordinate projection is compact and therefore bounded, so there is an M>0M>0 such that every (x,y)∈K(x,y)\in K satisfies ∣x∣≤M|x|\le M. If f(t)∈Kf(t)\in K, then ∣t∣≤M|t|\le M, and hence

f−1(K)⊆[−M,M].f^{-1}(K)\subseteq[-M,M].

The set f−1(K)f^{-1}(K) is closed in the compact interval [−M,M][-M,M], because KK is closed and ff is continuous. Therefore f−1(K)f^{-1}(K) is compact.

The theorem now shows that

t⟼(t,t2)t\longmapsto(t,t^2)

is a homeomorphism from R\mathbb R onto the parabola

{(x,x2):x∈R}\{(x,x^2):x\in\mathbb R\}

with its subspace topology.

Why Hausdorffness cannot be omitted

First countability by itself is not enough. Let

X={a,b}X=\{a,b\}

with the discrete topology, and let

Y={0,1}Y=\{0,1\}

with the indiscrete topology {∅,Y}\{\varnothing,Y\}. Define a bijection by

f(a)=0,f(b)=1.f(a)=0,\qquad f(b)=1.

The space YY is finite, so it is compact and first-countable. The map ff is continuous because the only open subsets of YY are ∅\varnothing and YY. It is proper: every subset of the finite space YY is compact, and its inverse image is a finite subset of the discrete space XX, hence compact.

However, ff is not a homeomorphism. Its inverse is not continuous, because

(f−1)−1({a})={0},(f^{-1})^{-1}(\{a\})=\{0\},

and {0}\{0\} is not open in the indiscrete topology on YY.

Thus the original claim is false for arbitrary first-countable target spaces, but it becomes true after adding the Hausdorff condition.

Common mistakes

Assuming compact sets are enough without Hausdorffness

A continuous bijection from a compact space to a non-Hausdorff space need not be a homeomorphism. The compact-to-Hausdorff theorem specifically requires the target to be Hausdorff.

Applying properness to arbitrary subsets

Properness gives compactness of f−1(K)f^{-1}(K) only when KK is compact. In the proof, the relevant compact set is the convergent sequence together with its limit:

K={y}∪{yn:n∈N}.K=\{y\}\cup\{y_n:n\in\mathbb N\}.

Forgetting the subspace topology

The conclusion is that XX is homeomorphic to f(X)f(X) with the topology inherited from YY. It does not claim that f(X)f(X) is open or closed in YY, unless additional assumptions imply that.

Confusing first countability with sequentiality

First countability is used to conclude that sequential continuity implies continuity. Without a condition of this kind, proving that the inverse preserves sequences may not be enough.

References

  1. Topology, James R. Munkres
  2. General Topology, John L. Kelley

Spotted an error? Send a correction. Read how we write and check articles in our editorial policy.