Scribd
Upload
25 views

An Short and Simple Intro To Index Notation

The document discusses the Levi-Civita symbol, a useful tool for converting cross products and curls to tensor notation. It defines the symbol and describes its key properties, including that it has 27 components but only 6 are non-zero. It also shows how the symbol can be used to express cross products, curls, determinants and other operations in tensor notation.

Uploaded by

Yasir Lone
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
25 views

An Short and Simple Intro To Index Notation

The document discusses the Levi-Civita symbol, a useful tool for converting cross products and curls to tensor notation. It defines the symbol and describes its key properties, including that it has 27 components but only 6 are non-zero. It also shows how the symbol can be used to express cross products, curls, determinants and other operations in tensor notation.

Uploaded by

Yasir Lone
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Physics 209

Fall 2002
Notes 3
The Levi-Civita Symbol

The Levi-Civita symbol is useful for converting cross products and curls into the lan-
guage of tensor analysis, and for many other purposes. The following is a summary of its
most useful properties in three-dimensional Euclidean space.
The Levi-Civita symbol is defined by

 1, if (ijk) is an even permutation of (123);
ijk = −1, if (ijk) is an odd permutation of (123); (3.1)
0, otherwise.

It has 27 components, of which only 6 are nonzero. It follows directly from this definition
that ijk changes sign if any two of its indices are exchanged,

ijk = jki = kij = −jik = −ikj = −kji . (3.2)

The Levi-Civita symbol is convenient for expressing cross products and curls in tensor
notation. For example, if A and B are two vectors, then

(A×B)i = ijk Aj Bk , (3.3)

and
∂Bk
(∇×B)i = ijk . (3.4)
∂xj
Any combination of an even number of Levi-Civita symbols (or an even number of cross
products and curls) can be reduced to dot products with the following system of identities.
Similarly, any combination of an odd number of Levi-Civita symbols (or an odd number of
cross products and curls) can be reduced to a single Levi-Civita symbol (or a cross product
or a curl) plus dot products. The first is the most general:

δi` δim δin


ijk `mn = δj` δjm δjn . (3.5)
δk` δkm δkn

Notice that the indices (ijk) label the rows, while (`mn) label the columns. If this is
contracted on i and l, we obtain
δjm δjn
ijk imn = = δjm δkn − δjn δkm . (3.6)
δkm δkn
–2–

This identity is the one used most often, for boiling down two cross products that have one
index in common, such as ∇×(A×B). By contracting Eq. (3.6) in j and m we obtain

ijk ijn = 2δkn . (3.7)

Finally, contracting on k and n we obtain

ijk ijk = 6. (3.8)

It should be clear how to generalize these identities to higher dimensions.


If Aij = −Aji is an antisymmetric, 3 × 3 tensor, it has 3 independent components that
we can associate with a 3-vector A, as follows:
 
0 A3 −A2
Aij =  −A3 0 A1  = ijk Ak . (3.9)
A2 −A1 0
The inverse of this is
1
Aij = ijk Ak . (3.10)
2
Using thes identities, the multiplication of an antisymmetric matrix times a vector can be
reexpressed in terms of a cross product. That is, if

Xi = Aij Yj (3.11)

then
X = Y×A. (3.12)
Similarly, if A and B are two vectors, then

Ai Bj − Aj Bi = ijk (A×B)k , (3.13)

and
∂Bj ∂Bi
− = ijk (∇×B)k . (3.14)
∂xi ∂xj
Finally, if Mij is a 3 × 3 matrix (or tensor), then
1
det M = ijk M1i M2j M3k = ijk `mn Mi` Mjm Mkn . (3.15)
6
The Levi-Civita symbol has been defined here only on 3 , but most of the properties
above are easily generalized to n (including the case n = 2). It only transforms as a
tensor under proper orthogonal changes of coordinates, which is why we are calling it a
“symbol” instead of a “tensor.” It can, however, be used to create so-called tensor densities
on arbitrary manifolds with a metric, and has fascinating applications in Hodge-de Rham
theory in differential geometry.

You might also like