Scribd
Upload
69 views

Lecture Notes 13: 2.9 The Covariant Derivative, Lie Bracket, and Rie-Mann Curvature Tensor of R

This document provides lecture notes on curves and surfaces in Rn. It defines concepts like the covariant derivative, Lie bracket, and Riemann curvature tensor of Rn. It then discusses how these concepts extend to surfaces embedded in R3, deriving explicit expressions for the induced covariant derivative on surfaces in local coordinates. Several exercises are provided to illustrate properties of these concepts.

Uploaded by

Sanjeev Shukla
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
69 views

Lecture Notes 13: 2.9 The Covariant Derivative, Lie Bracket, and Rie-Mann Curvature Tensor of R

This document provides lecture notes on curves and surfaces in Rn. It defines concepts like the covariant derivative, Lie bracket, and Riemann curvature tensor of Rn. It then discusses how these concepts extend to surfaces embedded in R3, deriving explicit expressions for the induced covariant derivative on surfaces in local coordinates. Several exercises are provided to illustrate properties of these concepts.

Uploaded by

Sanjeev Shukla
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Math 497C Nov 11, 20041

Curves and Surfaces


Fall 2004, PSU

Lecture Notes 13

2.9 The Covariant Derivative, Lie Bracket, and Rie-


mann Curvature Tensor of Rn
Let A ⊂ Rn , p ∈ A, and W be a tangent vector of A at p, i.e., suppose
there exists a curve γ : (−, ) → A with γ(0) = p and γ  (0) = W . Then
if f : A → R is a function we define the (directional) derivative of f with
respect to W at p as

Wp f := (f ◦ γ) (0) = dfp (W ).

Similarly, if V is a vectorfield along A, i.e., a mapping V : A → Rn , p → Vp ,


we define the covariant derivative of V with respect to W at p as

∇Wp V := (V ◦ γ) (0) = dVp (W ).

Note that if f and V are C 1 , then by definition they may be extended to an


open neighborhood of A. So dfp and dVp , and consequently Wp f and ∇Wp V
are well defined. In particular, they do not depend on the choice of the curve
γ or the extensions of f and V .

Exercise 1. Let Ei be the standard basis of Rn , i.e., E1 := (1, 0, . . . , 0),


E2 := (0, 1, 0, . . . , 0), . . . , En := (0, . . . , 0, 1). Show that for any functions
f : Rn → R and vectorfield V : Rn → Rn

(Ei )p f = Di f (p) and ∇(Ei )p V = Di V (p)

(Hint: Let ui : (−, ) → Rn be given by ui (t) := p + tEi , and observe that


(Ei )p f = (f ◦ ui ) (0), ∇(Ei )p V = (V ◦ ui ) (0)).
1
Last revised: November 29, 2004

1
The operation ∇ is also known as the standard Levi-Civita connection of
R . If W is a tangent vectorfield of A, i.e., a mapping W : A → Rn such
n

that Wp is a tangent vector of A for all p ∈ A, then we set

W f (p) := Wp f and (∇W V )p := ∇Wp V.

Note that W f : A → R is a function and ∇W V is a vectorfield. Further, we


define
(f W )p := f (p)Wp .
Thus f W : A → Rn is a also a vector field.

Exercise 2. Show that it V = (V 1 , . . . , V n ), i.e., V i are the component


functions of V , then

∇W V = (W V 1 , . . . , W V n ).

Exercise 3. Show that if Z is a tangent vectorfield of A and f : A → R is


a function, then

∇W +Z V = ∇W V + ∇Z V, and ∇f W V = f ∇W V.

Further if Z : A → Rn is any vectorfield, then

∇W (V + Z) = ∇W V + ∇W Z, and ∇W (f V ) = (W f )V + f ∇W V.

Exercise 4. Note that if V and W are a pair of vectorfields on A then


V, W  : A → R defined by V, W p := Vp , Wp  is a function on A, and show
that    
ZV, W  = ∇Z V, W + V, ∇Z W .

If V, W : A → Rn are a pair of vector fields, then their Lie bracket is the


vector filed on A defined by

[V, W ]p := ∇Vp W − ∇Wp V.

Exercise 5. Show that if A ⊂ Rn is open, V, W : A → Rn are a pair of


vector fields and f : A → R is a scalar, then

[V, W ]f = V (W f ) − W (V f ).

2
(Hint: First show that V f = V, grad f  and W f = W, grad f  where

grad f := (D1 f, . . . , Dn f ).

Next define
Hess f (V, W ) := V, ∇W grad f ,
and show that Hess f (V, W ) = Hess f (W, V ). In particular, it is enough to
show that Hess f (Ei , Ej ) = Dij f , where {E1 , . . . , En } is the standard basis
for Rn . Then Leibnitz rule yields that

V (W f ) − W (V f )
= V W, grad f  − W V, grad f 
= ∇V W, grad f  + W, ∇V grad f  − ∇W V, grad f  − V, ∇W grad f 
= [V, W ], grad f  + Hess f (W, V ) − Hess f (V, W )
= [V, W ]f,

as desired.)
If V and W are tangent vectorfields on an open set A ⊂ Rn , and Z : A →
n
R is any vectorfield, then

R(V, W )Z := ∇V ∇W Z − ∇W ∇V Z − ∇[V,W ] Z

defines a vectorfield on A. If Y is another vectorfield on A, then we may also


define an associated scalar quantity by
 
R(V, W, Z, Y ) := R(V, W )Z, Y ,

which is known as the Riemann curvature tensor of Rn .


Exercise 6. Show that R ≡ 0.

2.10 The Induced Covariant Derivative on Surfaces;


Gauss’s Formulas revisited
Note that if M ⊂ R3 is a regular embedded surface and V , W : M → R3 are
vectorfields on M . Then ∇W V may no longer be tangent to M . Rather, in
general we have
   ⊥
∇W V = ∇W V + ∇W V ,

3
where (∇W V ) and (∇W V )⊥ respectively denote the tangential and normal
components of ∇W V with resect to M . More explicitly, if for each p ∈ M
we let n(p) be a unit normal vector to Tp M , then
 ⊥      ⊥
∇W V p
:= ∇Wp V, n(p) n(p) and ∇W V := ∇W V − ∇W V .

Let X (M ) denote the space of tangent vectorfield on M . Then We define the


(induced) covariant derivative on M as the mapping ∇ : X (M ) × X (M ) →
X (M ) given by
 
∇W V := ∇W V .

Exercise 7. Show that, with respect to tangent vectorfields on M , ∇ satis-


fies all the properties which were listed for ∇ in Exercises 3 and 4.

Next we derive an explicit expression for ∇ in terms of local coordinates.


Let X : U → M be a proper regular patch centered at a point p ∈ M , i.e.,
X(0, 0) = p, and set
X i := Xi ◦ X −1 .
Then X i are vectorfields on X(U ), and for each q ∈ X(U ), (X i )q forms a
basis for Tq M . Thus on X(U ) we have
 
V = V i X i , and W = W iX i
i i

for some functions V i , W i : X(U ) → R. Consequently, on X(U ),


 
∇W V = ∇( W j X j ) V iX i
j

   
= W j ∇X j V iX i
j i

  
= Wj X j V i + V i ∇X j X i
j i
   
= W j X j V i + W j V i ∇X j X i .
j i

Next note that if we define uj : (−, ) → R2 by uj (t) := tEj , where E1 :=


(1, 0) and E2 := (0, 1). Then X ◦ui : (−, ) → M are curves with X ◦ui (0) =

4
p and (X ◦ ui ) (0) = Xi (0, 0) = X i (p). Thus by the definitions of ∇ and ∇
we have
 
∇(X j )p X i = ∇(X j )p X i
   
= X i ◦ (X ◦ uj ) (0)
 
= (Xi ◦ uj ) (0)

Now note that, by the chain rule,


 
(Xi ◦ uj ) (0) = DXi uj (0) Duj (0) = Xij (0, 0).

Exercise 8. Verify the last equality above.


Thus, by Gauss’s formula,
 
∇(X j )p X i = Xij (0, 0)
 
= Γkij (0, 0)Xk (0, 0) + lij (0, 0)N (0, 0)
k
    
= Γkij X −1 (p) Xk X −1 (p)
k
   
= Γkij X −1 (p) X k p .
k

k
In particular if we set X ij := Xij ◦ X −1 and define Γij : X(U ) → R by
k
Γij := Γkij ◦ X −1 , then we have
   k
∇X j X i = X ij = Γij X k ,
k

which in turn yields


  k

∇W V = j i
W XjV + W V j i
Γij X k .
j i k

Now recall that Γkij depends only on the coefficients of the first fundamental
form gij . Thus it follows that ∇ is intrinsic:

5
Exercise 9. Show that if f : M → M is an isometry, then
 
∇ df (W ) df (V ) = df ∇W V ,

where ∇ denotes the covariant derivative on M (Hint: It is enough to show


 

that ∇ df (X i ) df (X j ), df (X l ) = df ∇X i X j , df (X l )).

Next note that if n : X(U ) → S2 is a local Gauss map then

∇W V, n = −V, ∇W n = −V, dn(W ) = V, S(W ),

where, recall that, S is the shape operator of M . Thus


 ⊥
∇Wp V = V, S(Wp )n(p),

which in turn yields


 
∇W V = ∇W V + V, S(W ) n.

This is Gauss’s formula and implies the expression that we had derived earlier
in local coordinates.

Exercise 10. Verify the last sentence.

You might also like