2024-02-25

490: For Diffeomorphism from C Manifold with Boundary onto Neighborhood of Point Image on C Manifold with Boundary, Differential at Point Is 'Vectors Spaces - Linear Morphisms' Isomorphism

<The previous article in this series | The table of contents of this series | The next article in this series>

A description/proof of that for diffeomorphism from C manifold with boundary onto neighborhood of point image on C manifold with boundary, differential at point is 'vectors spaces - linear morphisms' isomorphism

Topics


About: C manifold

The table of contents of this article


Starting Context



Target Context


  • The reader will have a description and a proof of the proposition that for any diffeomorphism from any C manifold with boundary onto any neighborhood of any point image on any C manifold with boundary, the differential of the diffeomorphism or any its codomain extension at the point is a 'vectors spaces - linear morphisms' isomorphism.

Orientation


There is a list of definitions discussed so far in this site.

There is a list of propositions discussed so far in this site.


Main Body


1: Note


For any diffeomorphism between any C manifolds with boundary, the corresponding proposition is well-known; this proposition is on the case that the codomain is not any whole C manifold with boundary.

The domain has to be the whole C manifold with boundary, because the definition of differential requires that gf is a C function over the whole C manifold with boundary, unless the definition of differential is extended somehow.


2: Description


For any C manifolds with (possibly empty) boundary, M1,M2, any point, pM1, and any diffeomorphism, f:M1Nf(p), where Nf(p)M2 is any neighborhood of f(p), the differential of f at p, df|p:TpM1Tf(p)M2 is a 'vectors spaces - linear morphisms' isomorphism.

In fact, for any subset, SM2, such that Nf(p)S (especially, S=M2), and the extension of f on the codomain, f:M1S, such that f=f, df|p:TpM1Tf(p)M2 is a 'vectors spaces - linear morphisms' isomorphism.


3: Proof


Let us prove that df|p is injective.

Let vvTpM1 be any tangent vector.

df|pvdf|pv? Is there a C function, g:M2R, such that df|pv(g)df|pv(g)?

df|pv(g)=v(gf); df|pv(g)=v(gf).

There is a C function, h:M1R, such that v(h)v(h).

As M2 is locally compact Hausdorff, by the proposition that any C manifold with boundary is locally compact, there is an open neighborhood, Uf(p)M2, of f(p) such that Uf(p)Nf(p), by the proposition that for any locally compact Hausdorff topological space, in any neighborhood around any point, there is an open neighborhood of the point whose (the open neighborhood's) closure is compact and contained in the former neighborhood.

f1|Uf(p):Uf(p)M1 is C over Uf(p), by the proposition that for any map between any arbitrary subsets of any C manifolds with boundary Ck at any point, where k includes , the restriction on any domain that contains the point is Ck at the point. hf1|Uf(p):Uf(p)R is C over Uf(p), by the proposition that for any maps between any arbitrary subsets of any C manifolds with boundary Ck at corresponding points, where k includes , the composition is Ck at the point.

There is a C extension, g:M2R, such that g|Uf(p)=hf1|Uf(p), by the proposition that for any C function from any closed subset of any C manifold with boundary and any open neighborhood of the closed subset, there is a C extension over the manifold with boundary that is supported in the open neighborhood.

As f is continuous at p, there is an open neighborhood, UpM1, of p such that f(Up)Uf(p). gf|Up=hf1|Uf(p)f|Up=h|Up. As any tangent vector is a local operator, its action depends only on any open neighborhood of the point, and so, v(gf)=v(h)v(h)=v(gf).

So, yes, there is a g such that df|pv(g)=v(gf)v(gf)=df|pv(g), and yes, df|pvdf|pv, and df|p is injective.

Let us prove that df|p is surjective.

As any tangent vector on Tf(p)M2 is realized by a C curve, γ:JUf(p), where J=(ϵ,ϵ)[0,ϵ), or (ϵ,0], by the proposition that any tangent vector at any point on any C manifold with boundary is realized by a C curve, let us specify a C curve, γ:JM1, such that γ=fγ.

γ:=f1γ will do, because fγ=ff1γ=γ. γ is indeed C, by the proposition that for any maps between any arbitrary subsets of any C manifolds with boundary Ck at corresponding points, where k includes , the composition is Ck at the point.

When p is a boundary point, whatever the type of J is, γ realizes a tangent vector on TpM1.

When p is not any boundary point and J=(ϵ,ϵ), γ realizes a tangent vector on TpM1.

When p is not any boundary point and J=[0,ϵ) or (ϵ,0], there are a positive ϵϵ and a chart, (UpM1,ϕp), and J:=[0,ϵ) or (ϵ,0] and ϕpγ|J:Jϕp(Up) is C, and so, there is a C extension, γ:(ϵ,ϵ)Rd, of ϕpγ|J, where ϵ can be and will be chosen such that ϵϵ, but as γ is continuous at 0, ϵ can be chosen such that γ((ϵ,ϵ))ϕp(Up)Rd because ϕp(Up) is open on Rd, and let us define γ:(ϵ,ϵ)M1=ϕp1γ, which realizes a tangent vector on TpM1, still satisfying γ|[0,ϵ) or (ϵ,0]=fγ|[0,ϵ) or (ϵ,0], because fγ|[0,ϵ) or (ϵ,0]=fϕp1γ|[0,ϵ) or (ϵ,0]=fϕp1ϕpγ|[0,ϵ) or (ϵ,0]=fγ|[0,ϵ) or (ϵ,0]=γ|[0,ϵ) or (ϵ,0]. In fact, γ could be taken to be from (ϵ,ϵ) in the first place.

df|p indeed maps the tangent vector realized by γ (or γ, but we will call both γ hereafter) to the tangent vector realized by γ, because df|pdγ/dt|0(g)=dγ/dt|0(gf)=d(gfγ)/dt|0=d(fγ)/dt|0(g)=dγ/dt|0(g).

So, df|p is surjective.

df|p is a 'vectors spaces - linear morphisms' isomorphism, by the proposition that any bijective linear morphism is a 'vectors spaces - linear morphisms' isomorphism.

For df|p, in fact, df|p=df|p, because df|pv(g)=v(gf)=v(gf)=df|pv(g).

So, df|p is a 'vectors spaces - linear morphisms' isomorphism.


References


<The previous article in this series | The table of contents of this series | The next article in this series>