You are on page 1of 6

PREREQUISITE GEOMETRY FOR GENERAL RELATIVITY

MATTHEW RAYMOND

1. Directional Derivatives
Definition 1.1. Let M be a smooth manifold, and γ : R → M , be a smooth curve.
Given γ(λ0 ) = p, define the directional derivative operator of γ at p ∈ M as the
linear map,
Xγ,p : C ∞ (M ) → R : f 7→ Xγ,p f ≡ (f ◦ γ)0 (λ0 ).
Remark 1.2. For intuitive reasons, we may refer to Xγ,p as the tangent vector (or
velocity) to γ at p.
Definition 1.3. Let the tangent space of M at a point p ∈ M be the set
Tp M ≡ {Xγ,p | γ is a smooth curve through p}.
This may be equipped with an R-vector space structure by defining
⊕ : Tp M × Tp M → Hom(C ∞ , R) : (Xγ,p ⊕ Xδ,p )(f ) ≡ Xγ,p (f ) + Xδ,p (f ),
: R × Tp M → Hom(C ∞ , R) : (α Xγ,p )(f ) ≡ α · Xγ,p (f ).
Theorem 1.4. Let (M, OM , AM ) be a C ∞ manifold, and Tp M the tangent space
to M at p ∈ M . Then ∃τ : Xτ,p = Xγ,p ⊕ Xδ,p and ∃σ : Xσ,p = α Xγ,p .
Proof. Let σ : M → R : λ 7→ γ(λ) ≡ γ(αλ + λ0 ), and µα : R → R : r 7→ µα (r) ≡
αr + λ0 . It follows that σ(λ) = (γ ◦ µα )(λ). Then, if γ(λ0 ) = p
Xσ,p f = (f ◦ γ ◦ µα )0 (0) = µ0α (0) · (f ◦ γ)0 (λ0 ) = α · Xγ,p f.
Let (U, x) ∈ AM , and τ : R → M : λ 7→ τ (λ) ≡ x−1 ((x ◦ γ)(λ + λ0 ) + (x ◦ δ)(λ +
λ1 ) − (x ◦ γ)(λ0 )), assuming γ(λ0 ) = γ(λ1 ) = p. It follows that τ (0) = p. Then
Xτ,p f = (f ◦ x−1 ◦ x ◦ τ )0 (0)
0
= ∂i (f ◦ x−1 )(x(p)) · (x ◦ τ )i (0)
0 0
= ∂i (f ◦ x−1 )(x(p))(x ◦ γ)i (λ0 ) + ∂j (f ◦ x−1 )(x(p))(x ◦ δ)j (λ1 )
= (f ◦ γ)0 (λ0 ) + (f ◦ δ)0 (λ1 )
= Xγ,p f + Xδ,p f.
Since this construction is not ill defined, it completes the proof. 

2. Tangent Spaces
Definition 2.1. Let (U, x) ∈ AM and Xγ,p : C ∞ (M ) → R. We may express Xγ,p
under its components under (U, x). Assuming that γ(λ0 ) = p,
0
Xγ,p f = ∂i (f ◦ x−1 )(x(p))(x ◦ γ)i (λ0 ) ≡ γ̇(x)
i
(λ0 ) (∂f )/(∂xi ) p .


Theorem 2.2. B ≡ {(∂/∂xi )p } ⊆ Tp U , where i = 1, . . . d = dim(Tp M ) is a basis


for Tp U .

Date: November 12, 2019.


1
2 MATTHEW RAYMOND

Proof. In order for B to be a hamal basis for Tp U , we must have linear indepen-
dence. Indeed, for j = 1, . . . d,
 

0 = λi xj = λi ∂i (xj ◦ x−1 )(x(p)) = λi δ ji = λj .
∂xi p
The last equality follows as xj ◦ x−1 : Rd → R. This completes the proof. 
Definition 2.3. Let Tp∗ M be the dual space to Tp M , called the cotangent space
of M at p, defined by

Tp∗ M ≡ { d p f : Tp M −
→ R | f ∈ C ∞ (M ) ∧ p ∈ M }.
Where d p f ∈ Tp∗ M , called the gradient operator of f at p acts on X ∈ Tp M by
virtue of
X 7→ d p f X ≡ Xf.
Theorem 2.4. Let (U, x), (V, y) ∈ AM , with {U, V } =
S S
6 ∅. Then, X ∈ Tp {U, V }
transits by virtue of  k
k ∂x j
X(y) = X(x) .
∂y k p

Proof. Proof follows from the application of d p xj to the chart representation of X.


   
∂ ∂
k
X(y) x j
= X i
(x) d p x j i
= X(x) δ ji = X(x)
j
.
∂y k p ∂xi p
A similar, but inverse transformation rule may be derived for elements of Tp∗ M . 
Remark 2.5. The d p xj , with j = 1, . . . , d is a dual basis for Tp∗ M . Proof is trivial
and left to the reader.

3. Bundles
Definition 3.1. A bundle is a tuple (E, M, π), where E is a smooth manifold,
’the total space’, M is a smooth manifold, ’the base space’ and π : E → M is a
surjective smooth map, ’the projection’.
Definition 3.2. Let (E, M, π) be a bundle. We can ’place’ fibres over p ∈ M ,
π −1 (p) ≡ preimπ ({p}).
Definition 3.3. Let (E, M, π) be a bundle. We may define a section σ over M as
σ : M → E : σ ◦ π = idM .
Definition 3.4. Let (T M, M, π) be a tuple, where the set T M is defined by
[
TM ≡ {Tp M },
p∈M

and π : T M → M : X 7→ p : X ∈ Tp M . Then we have the immediate corollary:


π −1 (p) = Tp M . Let the topology OT M on T M be the ’coarsest’ topology as to
make π continuous,
OT M ≡ {preimπ (U ) | U ∈ OM }.
Let d = dim M . Then construct AT M from AM by, where T U ⊆ T M , and ξx :
T U → ξx (T U ) ⊆ R2d .
AT M ≡ {(T U, ξx ) | (U, x) ∈ AM }.
PREREQUISITE GEOMETRY FOR GENERAL RELATIVITY 3

Theorem 3.5. Let (T M, M, π) be as described as above. Then T M is a smooth


manifold with dim T M = 2d.
Proof. It is clear from definition that π is surjective. It is also clear that ∅ ∈ OM ,
and preimπ (∅) = ∅, so ∅ ∈ OT M . Similarly, preimπ (M ) = T M , so T M ∈ OT M .
Let U, V ∈ OM , and preim(U ), preim(V ) ∈ OT M . Then,
\ \ 
{preim(U ), preim(V )} = preim {U, V } .
Since {U, V } ∈ OM , preim ( {U, V }) ∈ OT M . A similar method may be used
T T
to confirm OT M is closed under arbitrary unions. Construct ξx : T M ⊇ T U →
ξx (T U ) ⊆ R2d , with p = π(x)
X 7→ (π ◦ x1 )(X), . . . , (π ◦ xd )(X), d p x1 X, . . . , d p xd X .


Then construct the map ξx−1 : R2d ⊇ ξx−1 (T U ) → T U ⊆ T M as


 
1 d 1 d i ∂
(α , . . . , α , β , . . . , β ) 7→ β ,
∂xi x−1 (α1 ,...αd )
with p = x−1 (α1 , . . . αd ). Consider the the chart transition map ξy ◦ ξx−1 ,
  !
−1 1 d 1 d i ∂
(ξy ◦ ξx )(α , . . . , α , β , . . . , β ) = ξy β
∂xi p
For brevity, now we only consider the n-th and m-th elements,
  !   !   ! !
∂ ∂ ∂
ξy β i = . . . , (y n ◦ π) β i , . . . , β i d p ym ,...
∂xi p ∂xi p ∂xi p
 m ! !
n −1 1 d i ∂y
= . . . , (y ◦ x )(α , . . . , α ), . . . , β ,...
∂xi p
These maps are smooth by definition. This completes the proof. 
Corollary 3.6. It immediately follows that the tuple (T M, M, π) is a bundle, called
the tangent bundle.
Theorem 3.7. The tuple (Tp∗ M, M, π) is a bundle, with,
[
TM ≡ Tp∗ M .
p∈M

Proof. Proof follows the same outline as above, and is left to the reader. 

4. Fields
Definition 4.1. Recall the tuple (C ∞ (M ), +, ·) is an infinite-dimensional vector
space when equipped with · : R × C ∞ (M ) → C ∞ (M ). Redefine · : C ∞ (M ) ×
C ∞ (M ) → C ∞ (M ). Then (C ∞ (M ), +, ·) is a noncommutative ring.
Definition 4.2. Let Γ(T M ) be the set of sections over T M . Equip Γ(T M ) with
⊕ : Γ(T M ) × Γ(T M ) → Γ(T M ) : (χ, χ̃) 7→ (χ ⊕ χ̃)(f ) ≡ χf + χ̃f,
: C ∞ (M ) × Γ(T M ) → Γ(T M ) : (g, χ) 7→ (g χ)(f ) ≡ (χf )g.
Then (Γ(T M ), ⊕, ) is a C ∞ (M )-module (vector space over a ring).
4 MATTHEW RAYMOND

Definition 4.3. A vector field χ is a section of T M , χ ∈ Γ(T M ). A covector field


ω is a section of T ∗ M . A (p, q)-tensor field is a section of T pq M .

5. Connections

Definition 5.1. Let M be a smooth manifold. Then a connection ∇ on M is a


map which satisfies the following;
i ∇X f = Xf ∀f ∈ C ∞ (M );
ii ∇X (T + S) = ∇X T + ∇X S;
iii ∇X (T ⊗ S) = T ⊗ ∇X S + ∇X T ⊗ S;
iv ∇f X+Y T = f · ∇X T + ∇Y T ∀f ∈ C ∞ (M ).
A quintuple (M, O, A , ∇) is called a smooth manifold with a connection ∇.
Definition 5.2. Let (U, x) ∈ A . Then the dim(M )3 connection coefficient func-
tions are the smooth maps Γijk : M ⊇ U → R,
 

p 7→ Γijk (p) ≡ dxi ∇ ∂k (p).
∂x ∂xj

Theorem 5.3. Let X, Y ∈ Γ(T M ), then ∇X Y ∈ Γ(T M ) has the chart represen-
tation,
   
∂ m ∂ q ∂
∇X Y = X i · Y · + X i
· Y m
· Γ mi .
∂xi ∂xm ∂xq
Proof. Proof follows directly from the definition of ∇.

∇X Y = ∇X i ∂ Ym
∂xi ∂xm

= Xi · ∇ ∂i Y m m
∂x ∂x  
i

m
 ∂ i m ∂
= X · ∇ ∂i Y · + X · Y · ∇ ∂i m
∂x ∂xm ∂x ∂x
   
i ∂ m ∂ i m q ∂
=X · Y · + X · Y · Γ mi q .
∂xi ∂xm ∂x


Remark 5.4. On a change of index, the above result becomes,


 
a i ∂ a
(∇X Y ) = X · Y + X i · Y m · Γami .
∂xi
Theorem 5.5. Let T ijk ∈ T 12 M , and (U, x) ∈ A . Then, by the Leibniz rule,

∇X T ijk = X(T ijk ) + Γism T sjk X m − Γsjm T isk X m − Γskm T ijs X m .

Proof. Recall in (U, x), T ijk ≡ T ( dxs , ∂x∂ q , ∂x∂ p ). Application of the Leibniz com-
pletes the proof. 

Theorem 5.6. Given (U, x) ∈ A , we may write ∇ ∂m dxi = Σijm dxj . These Σs
∂x
are not independent from the connection coefficient functions.
PREREQUISITE GEOMETRY FOR GENERAL RELATIVITY 5

Proof. We best know dxi by applying it to a vector field. Proof then follows from
the Leibniz rule,
    
i ∂ i
 ∂
i ∂
∇ ∂m dx = ∇ ∂m dx + dx ∇ ∂m j
∂x ∂xj ∂x ∂xj ∂x ∂x
 ∂  

q ∂
= ∇ ∂m dxi + Γ jm dx i
∂x ∂xj ∂xq
  ∂
= ∇ ∂m dxi + Γqjm δ iq .
∂x ∂xj
On rearrangement, we find the Γs differ from the Σs by a sign,
 
i j

i

j ∂
−Γ jm dx = ∇ ∂m dx dx = ∇ ∂m dxi .
∂x ∂xj ∂x


Theorem 5.7. Let ω ∈ Γ(T ∗ M ) and X ∈ Γ(T M ). Then, by a similar calculation,
 

(∇X ω)b = X i · ωb − ωs · X
m
· Γsbm .
∂xi
Proof. Proof follows in a similar way to above. 
Theorem 5.8. Let (U, x), (V, y) ∈ A , with {U, V } =
S
6 ∅. Then the coordinate
coefficient functions transform by,
∂y i ∂ 2 xq ∂y i ∂xs ∂xp q
Γi(y)jk = + Γ .
∂xq ∂y k ∂y j ∂xq ∂y j ∂y k (x)sp
Proof. Proof follows from the definition of ∇.
 

Γi(y)jk = dy i ∇ ∂k j
∂y ∂y

i
∂xs ∂
 
∂y q
= dx ∇ ∂xp ∂
∂xq ∂y k ∂x ∂y ∂x
p j s

∂y i ∂xp ∂xs ∂xs


    
q ∂ ∂
= dx ∇ ∂ + ∇ ∂
∂xq ∂y k ∂xp ∂y j ∂xs ∂y j ∂xp ∂xs

∂y i ∂xp ∂ ∂xs ∂ ∂xp ∂xs ∂ q


 
q
= dx + k j Γ
∂xq ∂y k ∂xp ∂y j ∂xs ∂y ∂y ∂xq (x)sq
∂y i ∂ ∂xs q ∂y i ∂xp ∂xs q q
= δ s + δ Γ
∂xq ∂y k ∂y j ∂xq ∂y k ∂y j q (x)sq
∂y i ∂ 2 xq ∂y i ∂xp ∂xs q
= + Γ .
∂xq ∂y k y j ∂xq ∂y k ∂y j (x)sp

Definition 5.9. Where T ijk is the torsion tensor; a connection ∇ of (M, O, A , ∇)
is called torsion free if,
Γi[jk] ≡ T ijk = 0.
Theorem 5.10. Let (M, O, A , ∇) be a smooth manifold with a connection ∇, and
p ∈ M . Then one can construct a chart (U, x̃) with p ∈ U such that
Γ̃i(jk) (p) = 0.
6 MATTHEW RAYMOND

at the point p.
Proof. Let (V, x) ∈ A with p ∈ V . In general, the above does not hold. Then
consider (U, x̃) ∈ A to which one transits by virtue of
1
(x̃ ◦ x−1 )i (α1 , . . . , αd ) ≡ αi − Γi(mk) (p) · αm αk .
2
The commutativity of R-multiplication automatically symmetrises the lower in-
dices. Then,
 i
∂ x̃ 1
j
= δ ij − Γi(mk) (p) · (δ mj αk + αm δ kj )
∂x p 2
1
= δ ij − (Γi(jk) (p) · αk + Γi(mj) (p) · αm )
2
1
= δ j − (Γi(jk) (p) · αk + Γi(jm) (p) · αm )
i
2
= δ ij − Γi(jk) (p) · αk .
In a similar manner,
∂ 2 x̃i
 
= −Γi(jk) (p) · δ kn = −Γi(jn) (p).
∂xj ∂xn p

Without loss of generality, fix x(p) = (0, . . . , 0). Then by the Γ transformation rule,
Γ̃ijk (p) = Γijk (p) − Γi(jn) (p) · δ nk = Γijk (p) − Γi(jk) (p) = T ijk .


Trinity Grammar School

You might also like