MATHEMATICAL PHYSICS

UNIT – 8
Metric Tensor & Chriostofell Symbols

PRESENTED BY: DR. RAJESH MATHPAL

ACADEMIC CONSULTANT
SCHOOL OF SCIENCES
U.O.U. TEENPANI, HALDWANI
UTTRAKHAND
MOB:9758417736,7983713112
 STRUCTURE OF UNIT

 8.1. INTRODUCTION
 8.2. RIEMANNIAN SPACE: METRIC TENSOR
𝒋
 8.3. FUNDAMENTAL TENSORS gjk, gjk AND 𝜹𝒌
 8.4. CHRISTOFELL’S 3-INDEX SYMBOLS
 8.5. GEODESICS
 8.1. INTRODUCTION

 In the mathematical field of differential geometry, one definition of a metric

tensor is a type of function which takes as input a pair of tangent
vectors v and w at a point of a surface (or higher dimensional differentiable
manifold) and produces a real number scalar g(v, w) in a way that generalizes
many of the familiar properties of the dot product of vectors in Euclidean
space. In the same way as a dot product, metric tensors are used to define
the length of and angle between tangent vectors. Through integration, the
metric tensor allows one to define and compute the length of curves on the
manifold.
 A metric tensor is called positive-definite if it assigns a positive value g(v, v) > 0 to every
nonzero vector v. A manifold equipped with a positive-definite metric tensor is known as
a Riemannian manifold. On a Riemannian manifold, the curve connecting two points that
(locally) has the smallest length is called a geodesic, and its length is the distance that a
passenger in the manifold needs to traverse to go from one point to the other. Equipped with
this notion of length, a Riemannian manifold is a metric space, meaning that it has a distance
function d(p, q) whose value at a pair of points p and q is the distance from p to q. Conversely,
the metric tensor itself is the derivative of the distance function (taken in a suitable manner).
Thus the metric tensor gives the infinitesimal distance on the manifold.
 While the notion of a metric tensor was known in some sense to mathematicians such as Carl
Gauss from the early 19th century, it was not until the early 20th century that its properties as
a tensor were understood by, in particular, Gregorio Ricci-Curbastro and Tullio Levi-Civita, who
first codified the notion of a tensor. The metric tensor is an example of a tensor field.
 The components of a metric tensor in a coordinate basis take on the form of
a symmetric matrix whose entries transform covariantly under changes to the
coordinate system. Thus a metric tensor is a covariant symmetric tensor.
From the coordinate-independent point of view, a metric tensor field is
defined to be a nondegenerate symmetric bilinear form on each tangent
space that varies smoothly from point to point.
8.2. RIEMANNIAN SPACE: METRIC TENSOR
An expression which express the distance between two adjacent point is called a metric or
line element. In three dimensional space the line element, i.e., the distance between two
adjacent points (x, y, z) and (x + dx, y + dy, z + dz) in Cartesian coordinates is given by

ds2 = dx2 + dy2 + dz2.

In terms of general curvilinear coordinates, the line element becomes

3 3
ds 2 =  g
j =1 k =1
jk du j duk =g jk du j duk (Using summation convention)

This idea was generalised by Riemann to n-dimensional space.

The distance between two neighbouring points with coordinates x j and xj + dxj is
given by
n n
ds 2 =  g
j =1 k =1
jk dx j dx k =g jk dx j dx k …(8.1)

(Using summation convention)

where the coefficients gjk are the functions of coordinates xj, subject to the restriction g =
determinant of gjk, i.e, 𝑔𝑗𝑘 ≠ 0.
The quadratic differential form gjk dxj dxk is independent of the coordinates system and is called
the Riemannian metric for n dimensional space. The space which is characterised by Riemannian
metric is called Riemannian space. Hence the quantities gjk are the components of a covariant
symmetric tensor of rank two, called the metric tensor or fundamental tensor.
(dx1)2 + (dx2)2 + (dx3)2 + … + (dxn)2 or dxjdxk,
the space is called n-dimensional Euclidean space. It is now obvious that Euclidean spaces are the
particular cases of Riemannian space.
In general theory of relativity (four dimensional space), the line element is given by
Ds2 = gjkdxjdxk (j, k = 1, 2, 3, 4).
In special theory of relativity, the line element is given by
(dx1)2 + (dx2)2 + (dx3)2 + … + (dxn)2 or dxj dxk.
the space is called n-dimensional Euclidean space. It is now obvious that Euclidean spaces
are the particular cases of Riemannian space.
In general theory of relativity (four dimensional space), the line element is given
by
ds2 = gjkdxidxjdxk (j, k = 1, 2, 3, 4).
In special theory of relativity, the line element is given by
ds2 = (dx1)2 + (dx2)2 + (dx3)2 [with x4 = ict, i = √ (-1)]
= dxjdxj (j = 1, 2, 3, 4).
As ds2 = gjk dxj dxk has been defined in general space (i.e., Riemannian space), it
is independent of the coordinate system, i.e., dx2 = gjk dxj dxk is an invariant.
 𝒋
8.3. FUNDAMENTAL TENSORS gjk, ds inAND
(i) Covariant fundamental tensor gjk. The line element or interval
gjk
𝜹𝒌space is given by
Riemannian
ds2 = gjk dxj dxk. …(8.2)
As dxj dxk are contravariant vectors and ds2 is invariant for arbitrary choice of vectors dxj and dxk, it
follows from quotient law that gjk is a covariant tensor, we have
ds2 = gjk dxj dxk in system of variables xj
𝜇 𝑣 𝜇
= 𝑔𝜇𝑣 𝑑𝑥 𝑑𝑥 in system of variables 𝑥
𝜇 𝑣
i.e., = 𝑔𝜇𝑣 𝑑𝑥 𝑑𝑥 = gjk dxj dxk. …(8.3)
Now applying inverse transformation law to dxj and dxj, i.e.,
𝜕𝑥 𝑗 𝜇
𝑑𝑥 𝑗 = 𝜇 𝑑𝑥 etc.
𝜕𝑥
𝜇 𝑣 𝜕𝑥 𝑗 𝜇 𝜕𝑥
𝑘
𝑣
𝑔𝜇𝑣 𝑑𝑥 𝑑𝑥 = 𝑔𝑗𝑘 𝜇 𝑑𝑥 𝑣 𝑑𝑥
𝜕𝑥 𝜕𝑥
𝑗 𝑘
𝜕𝑥 𝜕𝑥 𝜇 𝑣
= 𝑔𝑗𝑘 𝜇 𝑣 𝑑𝑥 𝑑𝑥
𝜕𝑥 𝜕𝑥
𝜕𝑥 𝑗 𝜕𝑥 𝑘 𝜇 𝑣
i.e., 𝑔𝜇𝑣 − 𝑔𝑗𝑘 𝜇 𝑣 𝑑𝑥 𝑑𝑥 = 0 …(8.4)
𝜕𝑥 𝜕𝑥
 𝜇 𝑣
As 𝑑𝑥 and 𝑑𝑥 are arbitrary contravarient vectors, we must have
𝜕𝑥 𝑗 𝜕𝑥 𝑘
𝑔𝜇𝑣 − 𝑐 𝜇 𝑣 =0
𝜕𝑥 𝜕𝑥
𝜕𝑥 𝑗 𝜕𝑥 𝑘
𝑔𝜇𝑣 = 𝜇 𝑣 𝑔𝑗𝑘
𝜕𝑥 𝜕𝑥
Hence 𝑔𝑗𝑘 is a covariant tensor of rank 2.
𝑔𝑗𝑘 may be expressed as
1 1
𝑔𝑗𝑘 = 𝑔𝑗𝑘 + 𝑔𝑘𝑗 + 𝑔𝑗𝑘 − 𝑔𝑘𝑗
2 2
= 𝐴𝑗𝑘 + 𝐵𝑗𝑘
…(8.5)
1
where 𝐴𝑗𝑘 = 𝑔𝑗𝑘 + 𝑔𝑘𝑗 is symmetric tensor
2
1 ቑ …(8.6)
and 𝐵𝑗𝑘 = 2
𝑔𝑗𝑘 − 𝑔𝑘𝑗 is symmetric tensor
∴ ds2 = gjk dxj dxk = (Ajk + Bjk) dxj dxk. …(8.7)
We have
Bjk dxj dxk = Bkj dxk dxj (interchanging dummy indices j and k)
= – Bjk dxj dxk
(cince Bjk is antisymmetric i.e., Bjk = – Bkj)
i.e., 2Bjk dxj dxk = 0.
As dxj and dxk are arbitrary vectors, we have
Bjk = 0
1
i.e., 𝑔𝑗𝑘 + 𝑔𝑘𝑗 = 0
2

i.e., 𝑔𝑗𝑘 + 𝑔𝑘𝑗

i.e., 𝑔𝑗𝑘 is symmetric.
So, we can write 𝑔𝑗𝑘 as
1
𝑔𝜇𝑣 . = 2 𝑔𝜇𝑣 + 𝑔𝑣𝜇

Thus we have proved that the metric tensor gjk is covariant symmetric
tensor of rank 2. This is called covariant fundamental tensor of rank 2.
(ii) Contravariant fundamental tensor gjk.
Let us define gjk as
𝑐𝑜𝑓𝑎𝑐𝑡𝑜𝑟 𝑜𝑓 𝑔𝑗𝑘 𝑖𝑛 𝑔
𝑔 𝑗𝑘 …(8.8)
𝑔
where g is the determinant of gjk, i.e.,
𝑔11 𝑔12 𝑔13 … 𝑔1𝑛
𝑔21 𝑔22 𝑔23 … 𝑔2𝑛
… … … … …
𝑔 = 𝑔𝑗𝑘 =
…
… …
… …
… … …
… …
𝑔𝑛1 𝑔𝑛2 𝑔𝑛3 … 𝑔𝑛𝑛
Since gjk is symmetric, g is symmetric which implies cofactor of gjk in g is symmetric and so
gjk is symmetric.
Let Aj be an arbitrary contravariant vector, then by quotient law,
Ak = gjk Aj …(8.9)
is an arbitrary covariant vector.
Now multiplying eqn. (8.9) by gkl, we get
gkl Ak = gjk gkl Aj. …(8.10)
𝑐𝑜𝑓𝑎𝑐𝑡𝑜𝑟 𝑜𝑓 𝑔𝑘𝑙 𝑖𝑛 𝑔
But gjk gkl = gjk 𝑔

= 𝛿𝑗𝑙 (by theory of determinants).

Therefore equation (8.10) yields
gkl Ak = 𝛿𝑗𝑙 Aj = Al …(8.11)
i.e., the inner product of gkl with an arbitrary covariant vector Ak yields a
contravariant vector. Hence by quotient law gkl is a contravariant tensor of rank 2.
Thus gjk is symmetric contravariant tensor of rank two. This tensor is reciprocal
of gjk and is called conjugate metric tensor or contravariant fundamental tensor of
rank 2.
 𝒋
(iii) Mixed fundamental tensor 𝒈𝒌 or 𝜹𝒍𝒋 . we have
gjk gkl = 𝛿𝑗𝑙 …(8.12)
As gjk and gkl are covariant and contravariant tensors of rank 2 recpectively,
therefore, from quotient law 𝛿𝑗𝑙 is also a tensor of rank 2; it is a mixed tensor,
contravariant in l and covarian in j and is known as mixed fundamental tensor. As
important property of mixed fundamental tensor is that its components have the
same value in all coordinate system, i.e., mixed fundamental tensor is invariant.
𝑗
The three tensors gjk, gjk and 𝛿𝑘 are called the fundamental tensors and are of
basic importance in general theory of relativity.
8.4.CHRISTOFELL’S 3-INDEX SYMBOLS

We now introduce two expressions (not tensors) formed of the fundamental tensors, known as
Christofell’s symbols of first and second kind, namely :
Christofell’s symbol of first kind.
1 𝜕𝑔𝑙𝑗 𝜕𝑔𝑘𝑙 𝜕𝑔𝑗𝑘
𝑗𝑘, 𝑙 = Γl,jk = + − …(8.13)
2 𝜕𝑥 𝑘 𝜕𝑥 𝑗 𝜕𝑥 𝑙
Christofell’s symbol of second kind.
𝑙 1 𝜕𝑔𝑗𝑚 𝜕𝑔 𝜕𝑔𝑗𝑘
= Γ l .jk = g lm + 𝑘𝑚 − 𝑚 …(8.14)
𝑗𝑘 2 𝜕𝑥 𝑘 𝜕𝑥 𝑗 𝜕𝑥

From the symmetry property of gjk it follows that

𝑗𝑘, 𝑙 = 𝑘𝑗, 𝑙 or Γl,jk = Γl.kj …(8.15)
𝑙 𝑙
and = or Γ l .jk = Γ l kj …(8.16)
𝑗𝑘 𝑘𝑗
thereby indicating that Christofell’s symbols Γl.jk and Γ l jk are symmetrical with respect to indices
j and k.
Relations between Christofell’s symbols of first and second kind.
(i) Replacing l by m in eqn. (8.13), we get
1 𝜕𝑔𝑚𝑗 𝜕𝑔𝑘𝑚 𝜕𝑥𝑗𝑘
Γm,jk = 2 𝜕𝑥 𝑘
+ 𝜕𝑥 𝑗
− 𝜕𝑥 𝑚
Multiplying both sides of above equation by glm, we get
1 𝜕𝑔𝑗𝑚 𝜕𝑔𝑘𝑚 𝜕𝑔𝑗𝑘
𝑔𝑙𝑚 Γm,jk = 𝑔𝑙𝑚 + − (since gjm = gmj)
2 𝜕𝑥 𝑘 𝜕𝑥 𝑗 𝜕𝑥 𝑚
l
= Γjk [Using (8.14)]
l
i.e., Γjk = 𝑔𝑙𝑚 Γm,jk …(8.17)
(ii) Interchanging l and m in eqn. (8.14), we get
m 1 𝜕𝑔𝑗𝑙 𝜕𝑔𝑘𝑙 𝜕𝑔𝑗𝑘
Γjk = 2 𝑔𝑙𝑚 𝜕𝑥 𝑘
+ 𝜕𝑥 𝑗
− 𝜕𝑥 𝑙
Multiplying above equation by glm, we get
m 1 𝜕𝑔𝑗𝑙 𝜕𝑔𝑘𝑙 𝜕𝑔𝑗𝑘
𝑔𝑙𝑚 Γjk = 2 𝑔𝑙𝑚 𝑔𝑙𝑚 𝜕𝑥 𝑘
+ 𝜕𝑥 𝑗
− 𝜕𝑥 𝑙
1 𝜕𝑔𝑗𝑙 𝜕𝑔𝑘𝑙 𝜕𝑔𝑗𝑘
= + − (since glm gml = 𝛿𝑙𝑙 = 1)
2 𝜕𝑥 𝑘 𝜕𝑥 𝑗 𝜕𝑥 𝑙
= Γl,jk
 8.5. GEODESICS
In Euclidean three dimensional space the path of shortest distance between two fixed
points is a straight line. Here we shall generalise this fundamental concept to Riemannian
space.

The path of extremum (maximum or minimum) distance between any two points in
Riemannian space is called the geodesic. Thus a geodesic is determined by the condition
that the path between two fixed points A and B given by be extremum, i.e.,
B

 ds B
extremum (or stationary), …(8.18)
A
 ds
A B
  ds = 0
i.e., A …(8.19)

where 𝛿 represents the variation symbol.

In Riemannian space, we have

ds2 = gjk dxj dxk …(8.20)

Keeping the end points A and B fixed, let the path be deformed by giving every
intermediate point an arbitrary infinitesimal displacement 𝛿𝑥 𝑚 , so that expression (8.20)
yields

2 ds 𝛿 (ds) = 𝛿 (gjk) dxj dxk + gjk 𝛿 (dxj) dxk + gjk dxj 𝛿 (dxk)
𝜕 𝑔 𝑗𝑘
= dxj dxk 𝛿𝑥 𝑚 + gjk dxk 𝛿 (dxj) + gjk dxj 𝛿 (dxk).
𝜕𝑥𝑚

Dividing both sides by 2 ds and using the relation

𝜕𝑥𝑗 𝑑
𝛿 = 𝛿𝑥 𝑗 ;
𝑑𝑠 𝑑𝑠

1 𝑑𝑥𝑗 𝑑𝑥𝑘 𝜕 𝑔 𝑗𝑘 𝑑𝑥𝑘 𝑑 𝑑𝑥𝑗 𝑑

We get 𝑑𝑠 = + − 𝛿𝑥 𝑚 + 𝑔𝑗𝑘 𝛿𝑥 𝑗 + 𝑔𝑗𝑘 𝛿𝑥 𝑘 𝑑𝑠.
2 𝑑𝑠 𝑑𝑠 𝜕𝑥𝑚 𝑑𝑠 𝑑𝑠 𝑑𝑠 𝑑𝑠

…..(8.21)

Substituting the value of 𝛿 𝑑𝑠 from (8.21) in (8.19), we get

1  dx j dx k g jk 
B
dx j d dx k d
 
2 A  ds ds x m
 x + g jk
m
(  x ) + g jk
k
(  x j ) ds = 0
ds ds ds ds 
 On changing the dummy indices in the last two terms, we get

 dx j dx k g jk
  dx k  d 

B
dx j
. (  x m ) ds = 0.
1
A   + +
m
x  jm
g g 
 ds ds x
m mk
2   ds ds  ds 


Integrating the second term by parts and remembering that the variation 𝛿 is zero
at the fixed end points A and B,

 dx j dx k g jk
 d  dx k  

B
1 dx j
A  − + g mk   x ds = 0.
m
 g jm
 ds ds x
m
2  ds  ds ds   

As the infinitesimal displacements 𝛿𝑥 𝑚 are arbitrary, therefore for the integral to

be stationary the coefficient of 𝛿𝑥 𝑚 in the integrand must vanish at all points on the path,
i.e.,

1  dx j dx k g jk d  dx j dx k  
 −  g jm + g mk  = 0
2  ds ds x m ds  ds ds  

1 dx j dx k g jk 1 dg jm dx j 1 d 2 xk
− − g = 0.
2 ds ds x m
mk
i.e., 2 ds ds 2 ds 2

1 dg mk dx k 1 d 2 xk
− − g mk = 0.
2 ds ds 2 ds 2 …(8.22)
But we have

dg jm g jm dx k dg mk g mk dx j
= and =
ds x k ds ds x j ds

With these substitutions, equation (8.21) becomes

1 dx j dx k  g jk g jm gmk  1 d 2x j d 2 xk 
 − −  −  jm
g + g  = 0.
2 ds ds  x m x k x j
mk
 2 ds 2 ds 2 
i.e.,

Replacing the dummy indices j and k and l in the second bracketed terms, we get

1 dx j dx k  g jk g jm gmk  1 d 2 xl d 2 xl 
 − −  −  jm 2
g + g  = 0.
2 ds ds  x m x k x j
mk
 2 ds ds 2 

Using symmetry property of glm (i.e., glm = gml) above equation mau be written as

1 dx j dx k  gmj gkm g jk  d 2 xl
 − j − m  lm 2 = 0.
+ g
2 ds ds  x k x x  ds
 Now multiplying throughout by gmp, we get

1 dx j dx k mp  g mj g km g jk  2 l
mp d x
g  + − m 
+ glm g =0
2 ds ds  x k
x j
x  ds 2

1 dx j dx k mp 2 l
p d x
g  m. jk +  l =0
or 2 ds ds ds 2

d 2 x p dx j dx k mp
+ g  m. jk = 0
i.e., ds 2 ds ds …(8.23)

d 2 x p dx j dx k p
+  jk = 0.
or ds 2 ds ds …(8.24)

Equation (8.24) represents the required condition to be satisfied in order that the
integral be stationary. Hence equation (8.24) represents the differential equation of a
geodesic. For p = 1, 2, 3, 4 this equation gives four differential equations which determine
a geodesic.
THANKS