Tensor Derivative in Curvilinear Coordinates

Sourangshu Ghosh

To cite this version:

Sourangshu Ghosh. Tensor Derivative in Curvilinear Coordinates. 2024. ⟨hal-04830269⟩

HAL Id: hal-04830269

[Link]
Preprint submitted on 11 Dec 2024

HAL is a multi-disciplinary open access archive L’archive ouverte pluridisciplinaire HAL, est des-
for the deposit and dissemination of scientific re- tinée au dépôt et à la diffusion de documents scien-
search documents, whether they are published or not. tifiques de niveau recherche, publiés ou non, émanant
The documents may come from teaching and research des établissements d’enseignement et de recherche
institutions in France or abroad, or from public or pri- français ou étrangers, des laboratoires publics ou
vate research centers. privés.

Distributed under a Creative Commons CC BY 4.0 - Attribution - International License

 Tensor Derivative in Curvilinear Coordinates
Sourangshu Ghosh*
*
sourangshug@[Link], Department of Civil Engineering, Indian Institute of Science Bangalore

Abstract
In this article, we mathematically rigorously derive the expressions for the Del Operator ∇, Divergence ∇ · ⃗v, Curl ∇ × ⃗v, Vector gradient
∇⃗v of Vector Fields ⃗v, Laplacian ∇2 f ≡ ∆f of Scalar Fields f and Divergence ∇ · T of 2nd order Tensor Fields T in both Cylindrical
and Spherical Coordinates. We also derive the Directional Derivative (A · ∇)⃗v and Vector Laplacian ∇2 ⃗v ≡ ∆⃗v of Vector Fields ⃗v using
metric coefficients in Rectangular, Cylindrical and Spherical Coordinates. We then generalized the concept of gradient, divergence and curl
to Tensor Fields in any Curvilinear Coordinates. After that we rigorously discuss the concepts of Christoffel Symbols, Parallel Transport in
Riemann Space, Covariant Derivative of Tensor Fields and Various Applications of Tensor Derivatives in Curvilinear Coordinates (Geodesic
Equation, Riemann Curvature Tensor, Ricci Tensor and Ricci Scalar).

Contents
1 Derivatives of Cylindrical Coordinate Unit Vectors with Respect to Cylindrical
Coordinates 2

2 Derivatives of Spherical Coordinate Unit Vectors with Respect to Spherical

Coordinates 5

3 Divergence of a Tensor Field 7

3.1 Cartesian coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.2 Curvilinear coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

4 Divergence ∇ · ⃗v in Cylindrical Coordinates 9

5 Divergence ∇ · ⃗v in Spherical Coordinates 11

6 Laplacian ∆f in Cylindrical Coordinates 14

7 Laplacian ∆f in Spherical Coordinates 15

8 Curl of a Tensor Field 15

8.1 Curl of a First-Order Tensor (Vector) Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
8.2 Curl of a Second-Order Tensor Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
8.3 Identities Involving the Curl of a Tensor Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

9 Curl ∇ × ⃗v in Cylindrical Coordinates 16

10 Curl ∇ × ⃗v in Spherical Coordinates 17

1
11 Gradient of a Tensor field 19
11.1 Cartesian coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
11.2 Curvilinear coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

12 Vector Gradient ∇⃗v in Cylindrical Coordinates 20

13 Vector Gradient ∇⃗v in Spherical Coordinates 22

14 Divergence ∇ · T of a Second-Order Tensor in Cylindrical Coordinates 24

15 Divergence ∇ · T of a Second-Order Tensor in Spherical Coordinates 28

16 Directional Derivative (A · ∇)⃗v in Curvilinear Coordinates 32

17 Vector Laplacian ∆⃗v in Curvilinear Coordinates 33

18 Derivatives with respect to vectors and second-order tensors 35

18.1 Derivatives of scalar and vector valued functions of vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
18.2 Derivatives of scalar and vector valued functions of second-order tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

19 Advanced Tensor Derivatives 36

19.1 Christoffel symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
19.1.1 Manifold, Coordinate System setup and the metric tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
19.1.2 Covariant Derivative of a Vector Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
19.1.3 Christoffel symbols in Euclidean Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
19.1.4 Christoffel Symbols of the Second Kind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
19.1.5 Transformation Law for Christoffel Symbols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
19.2 Parallel Transport in Riemannian Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
19.2.1 Derivation of the Relationship to Index-Free Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
19.2.2 Christoffel Symbols In Earth Surface Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
19.3 Covariant Derivative of Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
19.4 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
19.4.1 Application 1: Geodesic Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
19.4.2 Application 2: Riemann Curvature Tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
19.4.3 Application 3: Ricci Tensor and Ricci Scalar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

20 References 48

1 Derivatives of Cylindrical Coordinate Unit Vectors with Respect to Cylindrical

Coordinates
To rigorously derive the derivatives of the cylindrical coordinate unit vectors (êρ , êθ , and êz ) with respect to the cylindrical coordinates (ρ,
θ, and z), we must first recall the geometric meaning of these unit vectors and the way they vary with changes in the cylindrical coordinates.

Step 1: Cylindrical Coordinates and Unit Vectors

The Cylindrical coordinates (ρ, θ, z) are related to Cartesian coordinates (x, y, z) as follows:
p y
ρ = x2 + y 2 , θ = tan−1 , z = z.
x
The unit vectors in cylindrical coordinates are defined as:

2
 • Radial unit vector êρ :
êρ = cos θî + sin θĵ

where î and ĵ are the unit vectors in the x- and y-directions, respectively. This vector points in the direction of increasing ρ.
• Azimuthal unit vector êθ :
êθ = − sin θî + cos θĵ
which is perpendicular to êρ and points in the direction of increasing θ.

• Axial unit vector êz :

êz = k̂

where k̂ is the unit vector in the z-direction. This vector is constant and points in the direction of increasing z.

Step 2: Derivatives of Unit Vectors with Respect to ρ

Since the unit vectors êρ , êθ , and êz do not depend explicitly on ρ, their derivatives with respect to ρ are straightforward:

• Derivative of êρ with respect to ρ:

∂êρ
=0
∂ρ

• Derivative of êθ with respect to ρ:

∂êθ
=0
∂ρ

• Derivative of êz with respect to ρ:

∂êz
=0
∂ρ

Thus, none of the unit vectors in cylindrical coordinates change as a function of ρ, so all the derivatives with respect to ρ are zero.

Step 3: Derivatives of Unit Vectors with Respect to θ

The unit vectors êρ and êθ depend explicitly on the coordinate θ, as they involve trigonometric functions of θ. Therefore, we need to carefully
compute their derivatives with respect to θ.

• Derivative of êρ with respect to θ:

êρ = cos θî + sin θĵ.
Taking the derivative with respect to θ:
∂êρ
= − sin θî + cos θĵ.
∂θ
Noting that − sin θî + cos θĵ = êθ , we have:
∂êρ
= êθ
∂θ

3
 • Derivative of êθ with respect to θ:
êθ = − sin θî + cos θĵ.
Taking the derivative with respect to θ:
∂êθ
= − cos θî − sin θĵ.
∂θ
Noting that − cos θî − sin θĵ = −êρ , we have:
∂êθ
= −êρ
∂θ

• Derivative of êz with respect to θ: Since êz does not depend on θ, its derivative is zero:

∂êz
=0
∂θ

Step 4: Derivatives of Unit Vectors with Respect to z

Finally, we compute the derivatives of the unit vectors with respect to z. Since none of the unit vectors in cylindrical coordinates depend on
z, the derivatives with respect to z are zero:

• Derivative of êρ with respect to z:

∂êρ
=0
∂z

• Derivative of êθ with respect to z:

∂êθ
=0
∂z

• Derivative of êz with respect to z:

∂êz
=0
∂z

Step 5: Summary of Results

The full matrix of partial derivatives of the cylindrical unit vectors with respect to the cylindrical coordinates is:
 ∂ê ∂êρ ∂êρ

ρ  
∂ρ ∂θ ∂z 0 êθ 0
 ∂ê ∂êθ ∂êθ 
 ∂ρθ ∂θ ∂z 
= 0 −êρ 0
∂êz ∂êz ∂êz 0 0 0
∂ρ ∂θ ∂z

This is the complete set of partial derivatives of the cylindrical coordinate unit vectors with respect to the cylindrical coordinates.

Final Explanation
• êρ and êθ depend on θ, and their rates of change are captured in terms of the other unit vector (either êθ or êρ ).
• êz is independent of both ρ, θ, and z, and thus all its partial derivatives are zero.
• No unit vectors depend on ρ or z, so all partial derivatives with respect to these coordinates are zero except for θ-related derivatives.

4
2 Derivatives of Spherical Coordinate Unit Vectors with Respect to Spherical
Coordinates
The spherical coordinates (r, θ, ϕ) describe a point in three-dimensional space:
• r: radial distance from the origin,
• θ: polar angle (measured from the positive z-axis),

• ϕ: azimuthal angle (measured from the positive x-axis in the xy-plane).

In Cartesian coordinates, the point (x, y, z) can be expressed as:

x = r sin θ cos ϕ

y = r sin θ sin ϕ
z = r cos θ

Definition of Unit Vectors

In spherical coordinates, the position vector of a point is:
⃗r = r ⃗er
where ⃗er , ⃗eθ , and ⃗eϕ are the unit vectors along the r, θ, and ϕ directions, respectively. These unit vectors are related to the Cartesian unit
vectors î, ĵ, and k̂ by the following expressions:

⃗er = sin θ cos ϕ î + sin θ sin ϕ ĵ + cos θ k̂

⃗eθ = cos θ cos ϕ î + cos θ sin ϕ ĵ − sin θ k̂

⃗eϕ = − sin ϕ î + cos ϕ ĵ

Derivatives of Unit Vectors with Respect to r

Since the unit vectors ⃗er , ⃗eθ , and ⃗eϕ depend only on θ and ϕ, their derivatives with respect to r are zero:

∂⃗er ∂⃗eθ ∂⃗eϕ

= 0, = 0, =0
∂r ∂r ∂r

Derivative of ⃗er with Respect to θ

∂⃗
er
To compute ∂θ , we take the derivative of each component of ⃗er with respect to θ:

⃗er = sin θ cos ϕ î + sin θ sin ϕ ĵ + cos θ k̂

∂⃗er
= cos θ cos ϕ î + cos θ sin ϕ ĵ − sin θ k̂
∂θ
From the definition of ⃗eθ , we recognize this result as:
∂⃗er
= ⃗eθ
∂θ

5
Derivative of ⃗er with Respect to ϕ
∂⃗
er
Similarly, to compute ∂ϕ , we differentiate the components of ⃗er with respect to ϕ:

⃗er = sin θ cos ϕ î + sin θ sin ϕ ĵ + cos θ k̂

∂⃗er
= − sin θ sin ϕ î + sin θ cos ϕ ĵ
∂ϕ
This expression corresponds to:
∂⃗er
= sin θ ⃗eϕ
∂ϕ

Derivative of ⃗eθ with Respect to θ

∂⃗
eθ
Now, we compute ∂θ :
⃗eθ = cos θ cos ϕ î + cos θ sin ϕ ĵ − sin θ k̂
∂⃗eθ
= − sin θ cos ϕ î − sin θ sin ϕ ĵ − cos θ k̂
∂θ
This expression is the negative of ⃗er :
∂⃗eθ
= −⃗er
∂θ

Derivative of ⃗eθ with Respect to ϕ

∂⃗
eθ
For ∂ϕ , we differentiate the components of ⃗eθ with respect to ϕ:

⃗eθ = cos θ cos ϕ î + cos θ sin ϕ ĵ − sin θ k̂

∂⃗eθ
= − cos θ sin ϕ î + cos θ cos ϕ ĵ
∂ϕ
This is:
∂⃗eθ
= cos θ ⃗eϕ
∂ϕ

Derivative of ⃗eϕ with Respect to θ

∂⃗
eϕ
Next, we calculate ∂θ . Since ⃗eϕ = − sin ϕ î + cos ϕ ĵ, it has no θ-dependence, so:

∂⃗eϕ
=0
∂θ

6
Derivative of ⃗eϕ with Respect to ϕ
∂⃗
eϕ
Finally, we compute ∂ϕ :
⃗eϕ = − sin ϕ î + cos ϕ ĵ
∂⃗eϕ
= − cos ϕ î − sin ϕ ĵ
∂ϕ
This is the negative of ⃗eϕ :
∂⃗eϕ
= −⃗eϕ
∂ϕ

Summary of Results
The derivatives of the unit vectors in spherical coordinates are:

∂⃗er ∂⃗eθ ∂⃗eϕ

= 0, = 0, =0
∂r ∂r ∂r

∂⃗er ∂⃗er
= ⃗eθ , = sin θ ⃗eϕ
∂θ ∂ϕ

∂⃗eθ ∂⃗eθ
= −⃗er , = cos θ ⃗eϕ
∂θ ∂ϕ

∂⃗eϕ ∂⃗eϕ
= 0, = −⃗eϕ
∂θ ∂ϕ

3 Divergence of a Tensor Field

3.1 Cartesian coordinates
In a Cartesian coordinate system we have the following relations for a vector field v and a second-order tensor field S:
∂vi
∇·v = = vi,i ei
∂xi
∂Ski
∇·S= ek = Ski,i ek
∂xi
where tensor index notation for partial derivatives is used in the rightmost expressions. Note that

∇ · S ̸= ∇ · S⊤
For a symmetric second-order tensor, the divergence is also often written as:
∂Ski
∇·S= ek = Ski,i ek
∂xi
The above expression is sometimes used as the definition of ∇ · S in Cartesian component form (often also written as div S). Note that such
a definition is not consistent with the rest of this article (see the section on curvilinear coordinates).

7
The difference stems from whether the differentiation is performed with respect to the rows or columns of S, and is conventional. This
is demonstrated by an example. In a Cartesian coordinate system, the second-order tensor (matrix) S is the gradient of a vector function v:

∇ · (∇v) = ∇ · V = (vij ei ⊗ ej ) · ei = vij,i ej = (∇ · ∇ · v) = ∇2 v

∇ · [(∇v)⊤ ] = ∇ · V⊤ = (vji ei ⊗ ej ) · ei = vji,j ei = ∇2 v

The last equation is equivalent to the alternative definition/interpretation:

(∇)alt (∇v) = (∇)alt (vij ei ⊗ ej ) = vij,j ei = ∇2 v, ei = ∇2 v

3.2 Curvilinear coordinates

If g , g , g3 are the contravariant basis vectors in a curvilinear coordinate system. In curvilinear coordinates, the divergences of a vector field
1 2

v and a second-order tensor field S are:

∂v i
∇·v = + v k Γiik
∂ξ i

∂Sik
∇·S= − Slk Γlii − Sil Γlik gk
∂ξ i
More generally  
∂Sij j l i j
∇·S= − Γ S
ik ij − Γ kj il g ⊗ b
S
∂ξ i
 
∂Sij
∇·S= i m
− Γjk Sil − Γlj Slm gi ⊗ bj
∂ξ k
" #
∂Sji
= + Γiil Sjl − Γlij Sli bj
∂q i
" #
∂Sji j j
= − Γik Sl + Γkl Si g ik bj
l l
∂q k
In cylindrical polar coordinates:
 
∂vr 1 ∂vθ ∂vz
∇·v = + + vr +
∂r r ∂θ ∂z
∂Srr ∂Srθ ∂Srz
∇·S= er + eθ + ez
∂r ∂r ∂r
   
1 ∂Sθr 1 ∂Sθθ
+ + (Srr − Sθθ ) er + + (Srθ + Sθr ) eθ
r ∂θ r ∂θ
 
1 ∂Sθz
+ + Srz ez
r ∂θ
∂Szr ∂Szθ ∂Szz
+ er + eθ + ez
∂z ∂z ∂z

8
4 Divergence ∇ · ⃗v in Cylindrical Coordinates
∂ ∂ ∂ ∂ ∂ ∂
We first need to compute the partial derivatives ∂x ; ∂y and ∂z in terms of ∂ρ , ∂ϕ , and ∂z For that, let us apply the basic rule of differentiation
called the chain rule.
∂ ∂ρ ∂ ∂ϕ ∂
= +
∂x ∂x ∂ρ ∂x ∂ϕ
Note here that in the above formula I have skipped the variable z. The reason behind this is very simple. As we know for both the systems,
i.e., Cartesian and Cylindrical, z coordinate is exactly the same; only ρ and ϕ are the functions of x and y as mentioned above.
∂ρ ∂ p 2  x
= x + y2 = p
∂x ∂x x + y2
2

∂ρ ρ cos ϕ
⇒ = = cos ϕ
∂x ρ
Note the simplification in the above step. As we are going to convert into the Cylindrical coordinates from the Cartesian ones, we must
simplify to the extent so that to get cylindrical variables. Similarly,
∂ϕ ∂ y −y −y
= tan−1 =
2 = 2
∂x ∂x x 2
x + x y x + y2

∂ϕ −y − sin ϕ
⇒ = 2 =
∂x ρ ρ
So putting these values above, we have,
 
∂ ∂ − sin ϕ ∂
= cos ϕ +
∂x ∂ρ ρ ∂ϕ
Similarly we can write
 
∂ ∂ cos ϕ ∂
= sin ϕ +
∂y ∂ρ ρ ∂ϕ
Using the above 2 boxed equations, we can therefore write
∂ ∂ ∂
∇= ⃗ex + ⃗ey + ⃗ez
∂x ∂y ∂z
       
∂ − sin ϕ ∂ ∂ cos ϕ ∂ ∂
⇒ ∇ = (cos ϕ⃗eρ − sin ϕ⃗eϕ ) + + (sin ϕ⃗eρ + cos ϕ⃗eϕ ) + + ⃗ez
∂ρ ρ ∂ϕ ∂ρ ρ ∂ϕ ∂z
Collecting similar terms together we get as follows:
∂ 1 ∂ ∂
∇ = (sin2 ϕ + cos2 ϕ) ⃗eρ + (sin2 ϕ + cos2 ϕ) ⃗eϕ + ⃗ez
∂ρ ρ ∂ϕ ∂z
∂ 1 ∂ ∂
⇒ ∇ = ⃗eρ + ⃗eϕ + ⃗ez
∂ρ ρ ∂ϕ ∂z
A vector field v can be written in Cylindrical Coordinates as

⃗v = vρ eρ + vϕ eϕ + vz ez

9
We have earlier derived that the cylindrical Del operator is

∂ 1 ∂ ∂
∇ = eρ + eϕ + ez
∂ρ ρ ∂ϕ ∂z

Therefore the divergence of the vector field v shall be

∂ 1 ∂ ∂
div(⃗v) = ∇ · ⃗v = (eρ + eϕ + ez ) · (vρ eρ + vϕ eϕ + vz ez )
∂ρ ρ ∂ϕ ∂z

∂ 1 ∂ ∂
⇒ div(⃗v) = ∇ · ⃗v = eρ · (vρ eρ + vϕ eϕ + vz ez ) + eϕ · (vρ eρ + vϕ eϕ + vz ez ) + ez · (vρ eρ + vϕ eϕ + vz ez )
∂ρ ρ ∂ϕ ∂z
Now note that we have
∂ ∂ ∂
(⃗eρ ) = 0; (⃗eϕ ) = 0; (⃗ez ) = 0
∂ρ ∂ρ ∂ρ

∂ ∂ ∂
(⃗eρ ) = ⃗eϕ ; (⃗eϕ ) = −⃗eρ ; (⃗ez ) = 0
∂ϕ ∂ϕ ∂ϕ

∂ ∂ ∂
(⃗eρ ) = 0; (⃗eϕ ) = 0; (⃗ez ) = 0
∂z ∂z ∂z
Therefore the above boxed equation can be written as
 
∂ 1 ∂ ∂
∇ · ⃗v = ⃗eρ + ⃗eϕ + ⃗ez · (vρ⃗eρ + vϕ⃗eϕ + vz⃗ez )
∂ρ ρ ∂ϕ ∂z
     
∂⃗eρ ∂vρ ∂⃗eϕ ∂vϕ ∂⃗ez ∂vz
= ⃗eρ · vρ + ⃗eρ + vϕ + ⃗eϕ + vz + ⃗ez
∂ρ ∂ρ ∂ρ ∂ρ ∂ρ ∂ρ
     
1 ∂⃗eρ ∂vρ ∂⃗eϕ ∂vϕ ∂⃗ez ∂vz
+ ⃗eϕ · vρ + ⃗eρ + vϕ + ⃗eϕ + vz + ⃗ez
ρ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ
     
∂⃗ez ∂vρ ∂⃗ez ∂vϕ ∂⃗ez ∂vz
+⃗ez · vρ + ⃗eρ + vϕ + ⃗eϕ + vz + ⃗ez
∂z ∂z ∂z ∂z ∂z ∂z
     
∂vρ ∂vϕ ∂vz
= ⃗eρ · vρ (0) + ⃗eρ + vϕ (0) + ⃗eϕ + vz (0) + ⃗ez
∂ρ ∂ρ ∂ρ
     
1 ∂vρ ∂vϕ ∂vz
+ ⃗eϕ · vρ (⃗eϕ ) + ⃗eρ + vϕ (−⃗eρ ) + ⃗eϕ + vz (0) + ⃗ez
ρ ∂ϕ ∂ϕ ∂ϕ
     
∂vρ ∂vϕ ∂vz
+⃗ez · vρ (0) + ⃗eρ + vϕ (0) + ⃗eϕ + vz (0) + ⃗ez
∂z ∂z ∂z
The above expression can be simplified as
     
∂vρ ∂vϕ ∂vz
div · ⃗v = ∇ · ⃗v = (⃗eρ · ⃗eρ ) + (⃗eρ · ⃗eϕ ) + (⃗eρ · ⃗ez )
∂ρ ∂ρ ∂ρ
     
vρ 1 ∂vρ vϕ 1 ∂vϕ 1 ∂vz
+ (⃗eϕ · ⃗eϕ ) + (⃗eϕ · ⃗eρ ) − (⃗eϕ · ⃗eρ ) + (⃗eϕ · ⃗eϕ ) + (⃗eϕ · ⃗ez )
ρ ρ ∂ϕ ρ ρ ∂ϕ ρ ∂ϕ

10
      
∂vz ∂vz ∂vz
+(⃗ez · ⃗eρ ) + (⃗ez · ⃗eϕ ) + (⃗ez · ⃗ez )
∂z ∂z ∂z
Now note that we have

⃗eρ · ⃗eρ = ⃗eϕ · ⃗eϕ = ⃗ez · ⃗ez = 1

⃗eρ · ⃗eϕ = ⃗eϕ · ⃗ez = ⃗ez · ⃗eρ = 0

Therefore we can write
∂vρ vρ 1 ∂vϕ ∂vz
div · ⃗v = ∇ · ⃗v = + + +
∂ρ ρ ρ ∂ϕ ∂z
This is the expression for divergence in Cylindrical coordinates.

5 Divergence ∇ · ⃗v in Spherical Coordinates

∂ ∂ ∂ ∂ ∂ ∂
We first need to compute the partial derivatives ∂x ; ∂y and ∂z in terms of ∂r ; ∂θ and ∂ϕ . For that let us apply the basic rule of differentiation
called the chain rule.
∂ ∂ ∂r ∂ ∂θ ∂ ∂ϕ
= + +
∂x ∂r ∂x ∂θ ∂x ∂ϕ ∂x

∂ ∂ ∂r ∂ ∂θ ∂ ∂ϕ
= + +
∂y ∂r ∂y ∂θ ∂y ∂ϕ ∂y

∂ ∂ ∂r ∂ ∂θ ∂ ∂ϕ
= + +
∂z ∂r ∂z ∂θ ∂z ∂ϕ ∂z
Therefore, we will need the following derivatives:
∂r ∂θ ∂ϕ ∂r ∂θ ∂ϕ ∂r ∂θ ∂ϕ
; ; ; ; ; ; ; and
∂x ∂x ∂x ∂y ∂y ∂y ∂z ∂z ∂z
Let us calculate all the required derivatives one by one
∂r r sin θ cos ϕ
∴ = = sin θ cos ϕ
∂x r
Similarly, we also have r ! r !
∂θ ∂ x2 + y 2 −1 ∂ x2 + y 2
= tan−1 = 2
∂x ∂x z ∂x z
q
x2 +y 2
1+ z

−1 x −xz
= x2 +y 2
p p = p
1+ z
x2 + y 2 x2 + y 2 + z 2 (x2 + y 2 + z 2 ) x2 + y 2
∂θ −r sin θ cos ϕ · cos θ · r sin θ cos θ cos ϕ
∴ = p =
∂x 2 2 2 2 2 2
r r sin θ cos ϕ + r sin θ sin ϕ 2 r sin θ

∂θ cos θ cos ϕ
⇒ =
∂x r

11
Similarly, we also have
∂ϕ ∂  −1 y  −1 y −y
= tan =
2 2 = 2
∂x ∂x x 1+ x x
y x + y2

∂ϕ −r sin θ sin ϕ − sin ϕ

⇒ = 2 2 =
∂x r sin θ r sin θ
Working on the similar lines, we can get the following derivatives,

∂r ∂θ cos θ sin ϕ ∂ϕ cos ϕ

= sin θ sin ϕ; = ; =
∂y ∂y r ∂y r sin θ

∂r ∂θ − sin θ ∂ϕ
= cos θ; = ; =0
∂z ∂z r ∂z
So, now we have all the required derivatives. Let us put these into the expressions before.
∂ ∂r ∂ ∂θ ∂ ∂ϕ ∂
= + +
∂x ∂x ∂r ∂x ∂θ ∂x ∂ϕ
       
∂ ∂ ∂ cos θ cos ϕ ∂ − sin ϕ
⇒ = (sin θ cos ϕ) + +
∂x ∂r ∂θ r ∂ϕ r sin θ
∂ ∂ cos θ cos ϕ ∂ sin ϕ ∂
⇒ = sin θ cos ϕ + −
∂x ∂r r ∂θ r sin θ ∂ϕ
Similarly, we can write
∂ ∂r ∂ ∂θ ∂ ∂ϕ ∂
= + +
∂y ∂y ∂r ∂y ∂θ ∂y ∂ϕ
       
∂ ∂ ∂ cos θ sin ϕ ∂ cos ϕ
⇒ = (sin θ sin ϕ) + +
∂y ∂r ∂θ r ∂ϕ r sin θ
∂ ∂ cos θ sin ϕ ∂ cos ϕ ∂
⇒ = sin θ sin ϕ + +
∂y ∂r r ∂θ r sin θ ∂ϕ
Similarly, we can write
∂ ∂r ∂ ∂θ ∂ ∂ϕ ∂
= + +
∂z ∂z ∂r ∂z ∂θ ∂z ∂ϕ
∂ ∂ sin θ ∂
⇒ = cos θ −
∂z ∂r r ∂θ
Now note that we want to convert the below-mentioned expression of gradient given in Cartesian Coordinate System to an expression of
gradient in Spherical Coordinate System
∂ ∂ ∂
∇ · ⃗v = (vx ) + (vy ) + (vz )
∂x ∂y ∂z
∂ cos θ cos ϕ ∂ sin ϕ ∂ ∂ sin θ ∂ ∂ sin θ ∂
⇒ ∇ · ⃗v = (sin θ cos ϕ + − )vx + (cos θ − )vy + (cos θ − )vz
∂r r ∂θ r sin θ ∂ϕ ∂r r ∂θ ∂r r ∂θ
Now note that to get the vector components from the Cartesian Coordinate System to the Spherical Coordinate System, we use the
following transformation:     
vx sin θ cos ϕ cos θ cos ϕ − sin ϕ vr
vy  =  sin θ sin ϕ cos θ sin ϕ cos ϕ   vθ 
vz cos θ − sin θ 0 vϕ

12
Hence we have,

vx = (sin θ cos ϕ)vr + (cos θ cos ϕ)vθ − (sin ϕ)vϕ

vy = (sin θ sin ϕ)vr + (cos θ sin ϕ)vθ + (cos ϕ)vϕ

vz = (cos θ)vr − (sin θ)vθ

So, finally to get divergence in spherical coordinates, let us put all the terms together.
∂ ∂ ∂
∇ · ⃗v = (vx ) + (vy ) + (vz )
∂x ∂y ∂z
Using the above 4 equations, we can therefore write
 
∂ cos θ cos ϕ ∂ sin ϕ ∂
∇ · ⃗v = sin θ cos ϕ + − ((sin θ cos ϕ)vr + (cos θ cos ϕ)vθ − (sin ϕ)vϕ )
∂r r ∂θ r sin θ ∂ϕ
 
∂ cos θ sin ϕ ∂ cos ϕ ∂
+ sin θ sin ϕ + + ((sin θ sin ϕ)vr + (cos θ sin ϕ)vθ + (cos ϕ)vϕ )
∂r r ∂θ r sin θ ∂ϕ
 
∂ sin θ ∂
+ cos θ − ((cos θ)vr − (sin θ)vθ )
∂r r ∂θ
Now it’s a mechanical work. Just take the proper derivatives, club the terms and simplify. Note that in the following steps, we will use the
product rule several times. Let’s compute the Derivatives and collect the terms:
 
∂vr ∂vθ ∂vϕ
∇ · ⃗v = sin θ cos ϕ sin θ cos ϕ + cos θ cos ϕ − sin ϕ
∂r ∂r ∂r
   
cos θ cos ϕ ∂vθ ∂vϕ
+ cos θ cos ϕ vr cos θ + sin θ + cos (−vϕ sin θ) − sin ϕ
r ∂θ ∂θ
   
sin ϕ ∂vθ ∂vϕ
− sin θ cos θ −vr sin θ + cos (−vϕ cos θ) − sin ϕ + vϕ cos ϕ
r sin θ ∂θ ∂ϕ
 
∂vr ∂vθ ∂vϕ
+ sin θ sin ϕ sin θ sin ϕ + cos θ sin ϕ + cos ϕ
∂r ∂r ∂r
   
cos θ sin ϕ ∂vθ ∂vϕ
+ sin ϕ vr cos θ + sin θ + cos (−vϕ sin θ) − sin ϕ
r ∂θ ∂θ
   
cos ϕ ∂vθ ∂vϕ
+ cos θ cos ϕ vr cos θ + sin θ + cos (−vϕ cos θ) − sin ϕ
r sin θ ∂ϕ ∂ϕ
 
sin ϕ ∂vθ
− −vr sin θ + vθ cos θ +
r ∂r
Let us try to simplify by taking similar terms together and use the basic identity, sin2 θ + cos2 θ = 1 or sin2 ϕ + cos2 ϕ = 1.
∂vϕ ∂vθ ∂vθ ∂vϕ ∂vϕ
Note that the terms with derivatives ∂r , ∂θ , ∂ϕ , ∂θ , and ∂ϕ are going to vanish. Therefore we can write
∂vr ∂vr ∂vr
∇ · ⃗v = sin2 θ cos2 ϕ + sin2 θ sin2 ϕ + cos2 θ
∂r ∂r ∂r

13
 cos2 θ cos2 ϕ sin2 ϕ cos2 θ sin2 ϕ cos2 θ sin2 θ
+ vr + vr + vr + vr + vr
r r r r r

cos2 θ cos2 ϕ ∂vθ cos2 θ sin2 ϕ ∂vθ sin2 θ ∂vθ cos θ cos2 ϕ sin2 ϕ
+ + + − sin θvθ + cos θvθ − cos θ sin2 ϕ sin θvθ
r ∂θ r ∂θ r ∂θ r r

cos2 ϕ sin θ sin2 ϕ ∂vθ cos2 ϕ ∂vθ

+ cos θvθ + cos θvθ + +
r sin θ r r sin θ ∂θ r sin θ ∂ϕ
Again simplifying the terms and utilizing, sin2 θ + cos2 θ = 1 or sin2 ϕ + cos2 ϕ = 1; we can get as follows.

∂vr 2vr 1 ∂vθ cos θ 1 ∂vϕ

∇ · ⃗v = + + + vθ +
∂r r r ∂θ r sin θ r sin θ ∂ϕ
But, to have simplicity in the expression, we can rewrite the above expression as follows.

1 ∂ 1 ∂ 1 ∂vϕ
r 2 vr +


∇ · ⃗v = (vθ sin θ) +
r2 ∂r r sin θ ∂θ r sin θ ∂ϕ
This is the expression for divergence in Spherical coordinates.

6 Laplacian ∆f in Cylindrical Coordinates

Note that the Laplacian ∆f of a scalar field f can be written in terms of gradient and divergence as

∆f = ∇ · (∇f )

A vector field v can be written in Cylindrical Coordinates as

⃗v = vρ eρ + vϕ eϕ + vz ez

We earlier derived that in cylindrical coordinates the gradient of a scalar field f can be written as

∂f 1 ∂f ∂f
∇f = ⃗eρ + ⃗eϕ + ⃗ez
∂ρ ρ ∂ϕ ∂z

We earlier derived that in cylindrical coordinates the divergence of a vector field ⃗v can be written as

∂vρ vρ 1 ∂vϕ ∂vz

div · ⃗v = + + +
∂ρ ρ ρ ∂ϕ ∂z

Therefore the Laplacian of a scalar field f can be written in terms of gradient and divergence in Cylindrical Coordinates shall be
∂f 1 ∂f ∂f ∂ ∂f 1 ∂f 1 ∂ 1 ∂f ∂ ∂f
∆f = ∇ · (∇f ) = ∇ · (⃗eρ + ⃗eϕ + ⃗ez )= ( )+ ( )+ ( )+ ( )
∂ρ ρ ∂ϕ ∂z ∂ρ ∂ρ ρ ∂ρ ρ ∂ϕ ρ ∂ϕ ∂z ∂z

∂2f 1 ∂f 1 ∂2f ∂2f

⇒ ∆f = 2
+ + 2 2+ 2
∂ρ ρ ∂ρ ρ ∂ϕ ∂z

14
7 Laplacian ∆f in Spherical Coordinates
Note that the Laplacian ∆f of a scalar field f can be written in terms of gradient and divergence as

∆f = ∇ · (∇f )

A vector field v can be written in Spherical Coordinates as

⃗v = vr er + vθ eθ + vϕ eϕ

We earlier derived that in Spherical coordinates the gradient of a scalar field f can be written as

∂f 1 ∂f 1 ∂f
∇f = ⃗er + ⃗eθ + ⃗eϕ
∂r r ∂θ r sin θ ∂ϕ

We earlier derived that in Spherical coordinates the divergence of a vector field ⃗v can be written as

1 ∂ 1 ∂ 1 ∂vϕ
r 2 vr +


∇ · ⃗v = div · ⃗v = 2
(vθ sin θ) +
r ∂r r sin θ ∂θ r sin θ ∂ϕ

Therefore the Laplacian of a scalar field f can be written in terms of gradient and divergence in Spherical Coordinates shall be
   
∂f 1 ∂f 1 ∂f 1 ∂ ∂f 1 ∂ 1 ∂f 1 ∂ 1 ∂f
∆f = ∇ · (∇f ) = ∇ · (⃗er + ⃗eθ + ⃗eϕ )= 2 r2 + sin θ + ( )
∂r r ∂θ r sin θ ∂ϕ r ∂r ∂r r sin θ ∂θ r ∂θ r sin θ ∂ϕ r sin θ ∂ϕ
   
1 ∂ 2 ∂f 1 ∂ ∂f 1 ∂ ∂f
⇒ ∆f = 2 r + 2 sin θ + 2 2 ( )
r ∂r ∂r r sin θ ∂θ ∂θ r sin θ ∂ϕ ∂ϕ

8 Curl of a Tensor Field

The curl of an order-n > 1 tensor field T(x) is also defined using the recursive relation:

(∇ × T) · c = ∇ · (∇ × c · T)

(∇ × v) · c = ∇ · (v × c)
where c is an arbitrary constant vector and v is a vector field.

8.1 Curl of a First-Order Tensor (Vector) Field

Consider a vector field v and an arbitrary constant vector c. In index notation, the cross product is given by

v × c = ϵijk vj ck ei

where ϵijk is the permutation symbol, otherwise known as the Levi-Civita symbol. Then,

∇ · (v × c) = ϵijk vj,i ck = (ϵijk vj,i ek ) · c = (∇ × v) · c

Therefore,
∇ × v = ϵijk vj,i ek

15
8.2 Curl of a Second-Order Tensor Field
For a second-order tensor S,
c · S = cm Smj ej
Hence, using the definition of the curl of a first-order tensor field,

∇ × (c · S) = ϵijk cm Smj,i ek = (ϵijk Smj,i ek ⊗ em ) · c = (∇ × S) · c

Therefore, we have
∇ × S = ϵijk Smj,i ek ⊗ em

8.3 Identities Involving the Curl of a Tensor Field

The most commonly used identity involving the curl of a tensor field, T, is

∇ × (∇ × T) = 0

This identity holds for tensor fields of all orders. For the important case of a second-order tensor, S, this identity implies that

∇ × (∇ × S) = 0

9 Curl ∇ × ⃗v in Cylindrical Coordinates

A vector field v can be written in Cylindrical Coordinates as

⃗v = vρ eρ + vϕ eϕ + vz ez

We have earlier derived that the cylindrical Del operator is

∂ 1 ∂ ∂
∇ = eρ + eϕ + ez
∂ρ ρ ∂ϕ ∂z

Let us take the cross product:

 
∂ 1 ∂ ∂
∇ × ⃗v = ⃗eρ + ⃗eϕ + ⃗ez × (vρ⃗eρ + vϕ⃗aϕ + vz⃗ez )
∂ρ ρ ∂ϕ ∂z

∂ 1 ∂ ∂
⇒ ∇ × ⃗v = ⃗eρ × (vρ⃗eρ + vϕ⃗ep hi + vz⃗ez ) + ⃗eϕ × (vρ⃗eρ + vϕ⃗ep hi + vz⃗ez ) + ⃗ez × (vρ⃗eρ + vϕ⃗ep hi + vz⃗ez )
∂ρ ρ ∂ϕ ∂z
 
∂⃗eρ ∂vρ ∂⃗eϕ ∂vϕ ∂⃗ez ∂vz
⇒ ∇ × ⃗v = ⃗eρ × vρ + ⃗eρ + vϕ + ⃗eϕ + vz + ⃗ez
∂ρ ∂ρ ∂ρ ∂ρ ∂ρ ∂ρ
 
1 ∂⃗eρ ∂vρ ∂⃗eϕ ∂vϕ ∂⃗ez ∂vz
+ ⃗eϕ × vρ + ⃗eρ + vϕ + ⃗eϕ + vz + ⃗ez
ρ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ
 
∂⃗eρ ∂vρ ∂⃗eϕ ∂vϕ ∂⃗ez ∂vz
+⃗ez × vρ ⃗
+ eρ + vϕ ⃗
+ eϕ + vz ⃗
+ ez
∂z ∂z ∂z ∂z ∂z ∂z
Note that the derivatives of the unit vectors of cylindrical coordinates are

∂ ∂ ∂
(⃗eρ ) = 0; (⃗eϕ ) = 0; (⃗ez ) = 0
∂ρ ∂ρ ∂ρ

16
 ∂ ∂ ∂
(⃗eρ ) = ⃗eϕ ; (⃗eϕ ) = −⃗eρ ; (⃗ez ) = 0
∂ϕ ∂ϕ ∂ϕ

∂ ∂ ∂
(⃗eρ ) = 0; (⃗eϕ ) = 0; (⃗ez ) = 0
∂z ∂z ∂z
So putting all these derivatives in the step above, and taking the required cross product, we have,

∂vϕ ∂vz 1 ∂vρ vϕ 1 ∂vz ∂vρ ∂vϕ

∇ × ⃗v = ⃗ez − ⃗eϕ − ⃗ez + ⃗ez + ⃗eρ + ⃗eϕ − ⃗eρ
∂ρ ∂ρ ρ ∂ϕ ρ ρ ∂ϕ ∂z ∂z

Collecting together the similar terms, we get

     
1 ∂vz ∂vϕ ∂vρ ∂vz vϕ ∂vϕ 1 ∂vρ
∇ × ⃗v = − ⃗eρ + − ⃗eϕ + + − ⃗ez
ρ ∂ϕ ∂z ∂z ∂ρ ρ ∂ρ ρ ∂ϕ
This can be adjusted and rewritten as like following,
      
1 ∂vz ∂(ρvϕ ) ∂vz ∂vρ ∂(ρvϕ ) ∂vρ
∇ × ⃗v = − ⃗eρ − − ρ⃗eϕ + − ⃗ez
ρ ∂ϕ ∂z ∂ρ ∂z ∂ρ ∂ϕ

The above expression can be easily written in matrix determinant form as:

⃗eρ ρ⃗eϕ ⃗ez

1 ∂ ∂ ∂
∇ × ⃗v = ∂ρ ∂ϕ ∂z
ρ
vρ ρvϕ vz

This is the standard expression for the curl of a vector field ⃗v = vρ⃗eρ + vϕ⃗eϕ + vz⃗ez in Cylindrical coordinates.

10 Curl ∇ × ⃗v in Spherical Coordinates

A vector field v can be written in Spherical Coordinates as

⃗v = vr er + vθ eθ + vϕ eϕ

We have earlier derived that the Spherical Del operator is

∂ eθ ∂ eϕ ∂
∇ = er + +
∂r r ∂θ r sin θ ∂ϕ

Let us take the cross product:

∂ eθ ∂ eϕ ∂
∇ × ⃗v = (er + + ) × (vr er + vθ eθ + vϕ eϕ )
∂r r ∂θ r sin θ ∂ϕ
∂ eθ ∂ eϕ ∂
∇ × ⃗v = er × (vr er + vθ eθ + vϕ eϕ ) + × (vr er + vθ eθ + vϕ eϕ ) + × (vr er + vθ eθ + vϕ eϕ )
∂r r ∂θ r sin θ ∂ϕ
 
∂⃗er ∂vr ∂⃗eθ ∂vθ ∂⃗eϕ ∂vϕ
⇒ ∇ × ⃗v = ⃗er × vr + ⃗er + vθ + ⃗eθ + vϕ + ⃗eϕ
∂r ∂r ∂r ∂r ∂r ∂r

17
  
⃗eθ ∂⃗er ∂vr ∂⃗eθ ∂vθ ∂⃗eϕ ∂vϕ
+ × vr + ⃗er + vθ + ⃗eθ + vϕ + ⃗eϕ
r ∂θ ∂θ ∂θ ∂θ ∂θ ∂θ
 
⃗eϕ ∂⃗er ∂vr ∂⃗eθ ∂vθ ∂⃗eϕ ∂vϕ
+ × vr + ⃗er + vθ + ⃗eθ + vϕ + ⃗eϕ
r sin θ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ
Note that the derivatives of the unit vectors of spherical coordinates are

∂⃗er ∂⃗er ∂⃗er

= 0, = ⃗eθ , = sin θ ⃗eϕ
∂r ∂θ ∂ϕ

∂⃗eθ ∂⃗eθ ∂⃗eθ

= 0, = −⃗er , = cos θ ⃗eϕ
∂r ∂θ ∂ϕ

∂⃗eϕ ∂⃗eϕ ∂⃗eϕ

= 0, = 0, = − sin θ ⃗er − cos θ ⃗eθ
∂r ∂θ ∂ϕ
Let’s substitute these into the equation of ∇ · ⃗v mentioned above.

The First Term (Radial Derivatives) shall be

 
∂vr ∂vθ ∂vϕ
⃗er × vr · 0 + ⃗er + vθ · 0 + ⃗eθ + vϕ · 0 + ⃗eϕ
∂r ∂r ∂r
 
∂vr ∂vθ ∂vϕ
⇒ ⃗er × ⃗er + ⃗eθ + ⃗eϕ
∂r ∂r ∂r
∂vθ ∂vϕ
⇒ ⃗eϕ − ⃗eθ
∂r ∂r
The Second Term (Polar Derivatives) shall be
 
⃗eθ ∂vr ∂vθ ∂vϕ
× vr⃗eθ + ⃗er − vθ⃗er + ⃗eθ + vϕ cos θ ⃗eϕ + ⃗eϕ
r ∂θ ∂θ ∂θ
 
⃗eθ ∂vθ ∂vϕ
⇒ × ⃗eθ + ⃗eϕ + (vr − vθ )⃗er + vϕ cos θ ⃗eϕ
r ∂θ ∂θ
 
1 ∂vϕ
⇒ ⃗er − eϕ (vr − vθ ) + er vϕ cos θ
⃗ ⃗
r ∂θ

The Third Term (Azimuthal Derivatives) shall be

 
⃗eϕ ∂vr ∂vθ ∂vϕ
× sin θ ⃗eϕ · vr + ⃗er + 0 + ⃗eθ − (sin θ ⃗er + cos θ ⃗eθ )vϕ + ⃗eϕ
r sin θ ∂ϕ ∂ϕ ∂ϕ
 
1 ∂vr ∂vθ
⇒ ⃗eθ − ⃗er − vϕ (sin θ ⃗eθ − cos θ ⃗er )
r sin θ ∂ϕ ∂ϕ
    
1 ∂vr ∂vθ
⇒ ⃗eθ + vϕ sin θ − ⃗er + vϕ cos θ
r sin θ ∂ϕ ∂ϕ

18
Let’s combine all terms and simplify further. The final expression for ∇ × ⃗v is:

        
∂vθ ∂vϕ 1 ∂vϕ 1 ∂vr ∂vθ
∇ × ⃗v = ⃗eϕ − ⃗eθ + ⃗er − ⃗eϕ (vr − vθ ) + ⃗er vϕ cos θ + ⃗eθ + vϕ sin θ − ⃗er + vϕ cos θ
∂r ∂r r ∂θ r sin θ ∂ϕ ∂ϕ
Finally, let’s collect the terms for each unit vector:
  
1 ∂vϕ vϕ cos θ 1 ∂vθ
∇ × ⃗v = ⃗er + − + vϕ cos θ
r ∂θ r r sin θ ∂ϕ
  
∂vϕ 1 ∂vr
+⃗eθ − + + vϕ sin θ
∂r r sin θ ∂ϕ
 
∂vθ vr − vθ
+⃗eϕ −
∂r r
     
1 ∂ ∂vθ 1 ∂vr 1 ∂(rvϕ ) 1 ∂(rvθ ) ∂vr
⇒ ∇ × ⃗v = (vϕ sin θ) − ⃗er + − ⃗eθ + − ⃗eϕ
r sin θ ∂θ ∂ϕ r sin θ ∂ϕ r ∂r r ∂r ∂θ
This is the standard expression for the curl of a vector field ⃗v = vr⃗er + vθ⃗eθ + vϕ⃗eϕ in Spherical coordinates.

11 Gradient of a Tensor field

11.1 Cartesian coordinates
If e1 , e2 , e3 are the basis vectors in a Cartesian coordinate system, with coordinates of points denoted by (x1 , x2 , x3 ), then the gradient of the
tensor field T is given by
∂T
∇T = ⊗ ei
∂xi
The vectors x and c can be written as x = xi ei and c = ci ei . Let y := x + αc. In that case the gradient is given by

d
∇T · c = T(x1 + αc1 , x2 + αc2 , x3 + αc3 )
dα
α=0
 
d ∂T ∂y1 ∂T ∂y2 ∂T ∂y3
⇒ ∇T · c ≡ T(y1 , y2 , y3 ) = + +
dα ∂y1 ∂α ∂y2 ∂α ∂y3 ∂α α=0
α=0
 
∂T ∂T ∂T ∂T ∂T
⇒ ∇T · c = c1 + c2 + c3 = ci = ⊗ ei · c
∂x1 ∂x2 ∂x3 ∂xi ∂xi
Since the basis vectors do not vary in a Cartesian coordinate system, we have the following relations for the gradients of a scalar field ϕ, a
vector field v, and a second-order tensor field S:

∂ϕ
∇ϕ = ei = ϕ,i ei
∂xi

∂(vj ej ) ∂vj
∇v = ⊗ ei = ej ⊗ ei = vj,i ej ⊗ ei
∂xi ∂xi

∂(Sjk ej ⊗ ek ) ∂Sjk
∇S = ⊗ ei = ej ⊗ ek ⊗ ei = Sjk,i ej ⊗ ek ⊗ ei
∂xi ∂xi

19
11.2 Curvilinear coordinates
If g , g , g3 are the contravariant basis vectors in a curvilinear coordinate system, with coordinates of points denoted by (ξ 1 , ξ 2 , ξ 3 ), then the
1 2

gradient of the tensor field T is given by

∂T
∇T = ⊗ gi
∂ξ i
From this definition we have the following relations for the gradients of a scalar field ϕ, a vector field v, and a second-order tensor field S:
∂ϕ i
∇ϕ =g
∂ξ i
∂(v j gj )
 j   
i ∂v k j i ∂vj
∇v = ⊗g = + v Γik gj ⊗ g = k
− vk Γij gj ⊗ gi
∂ξ i ∂ξ i ∂ξ i

∂(Sjk gj ⊗ gk )
 
i ∂Sjk
∇S = ⊗g = l l
− Slk Γij − Sjl Γik gj ⊗ gk ⊗ gi
∂ξ i ∂ξ i
where the Christoffel symbol Γkij is defined using
∂gi ∂gi
Γkij gk = =⇒ Γkij = gk ·
∂ξ j ∂ξ j

12 Vector Gradient ∇⃗v in Cylindrical Coordinates

Cylindrical coordinates (r, ϕ, z) are defined by:
• r: Radial distance from the z-axis.
• ϕ: Azimuthal angle (angle in the xy-plane measured from the positive x-axis).
• z: Height along the z-axis, which corresponds to the Cartesian z-coordinate.
The relationship between cylindrical coordinates (r, ϕ, z) and Cartesian coordinates (x, y, z) is given by:

x = r cos ϕ, y = r sin ϕ, z=z

with inverse relations: p y

r= x2 + y 2 , ϕ = tan−1
, z=z
x
In cylindrical coordinates, the basis vectors are position-dependent. At any point in space, the unit vectors are:
• ⃗er : Unit vector in the radial direction (perpendicular to the z-axis).
• ⃗eϕ : Unit vector in the azimuthal direction (tangential to the circular path around the z-axis).
• ⃗ez : Unit vector in the z-direction (parallel to the z-axis).

These vectors are mutually orthogonal, and they can be written in terms of Cartesian unit vectors î, ĵ, k̂ as:
     
cos ϕ − sin ϕ 0
⃗er =  sin ϕ  , ⃗eϕ =  cos ϕ  , ⃗ez = 0
0 0 1

Since ⃗er and ⃗eϕ depend on ϕ, their derivatives with respect to ϕ will be non-zero, which is crucial for our derivation.

20
General Form of the Gradient of a Vector
⃗ = Vr ⃗er + Vϕ ⃗eϕ + Vz ⃗ez , the gradient tensor ∇V
For a vector field V ⃗ is obtained by applying the gradient operator to each component of the
vector field and considering the derivatives of the basis vectors. In curvilinear coordinates, the gradient of a vector field is generally written
as:
i
∇V ⃗ = ∂V ⃗ei ⊗ ⃗ej + V i ∇⃗ei
∂xj

where the first term represents the derivatives of the components of the vector field, and the second term involves the gradients of the basis
vectors.

Derivatives of the Basis Vectors

Since the basis vectors ⃗er , ⃗eϕ , and ⃗ez depend on ϕ, their derivatives must be carefully considered. Let’s compute them:

Derivative of ⃗er :
∂⃗er ∂⃗er ∂⃗er
= 0, = ⃗eϕ , =0
∂r ∂ϕ ∂z

Derivative of ⃗eϕ :
∂⃗eϕ ∂⃗eϕ ∂⃗eϕ
= 0, = −⃗er , =0
∂r ∂ϕ ∂z

Derivative of ⃗ez :
∂⃗ez ∂⃗ez ∂⃗ez
= 0, = 0, =0
∂r ∂ϕ ∂z
These derivatives are crucial when calculating the second term V i ∇⃗ei in the gradient formula.

Computing the Gradient Tensor

⃗ in cylindrical coordinates by considering each directional derivative:
Now, we will compute the gradient of the vector field V

⃗
∂V
Radial Derivative ∂r :

⃗
∂V ∂Vr ∂Vϕ ∂Vz
= ⃗er + ⃗eϕ + ⃗ez
∂r ∂r ∂r ∂r
There is no derivative of the basis vectors with respect to r, as they do not depend on r.

⃗
∂V
Azimuthal Derivative ∂ϕ :

⃗
∂V ∂Vr ∂⃗er ∂Vϕ ∂⃗eϕ ∂Vz
= ⃗er + Vr + ⃗eϕ + Vϕ + ⃗ez
∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ
Substituting the derivatives of the basis vectors:

⃗
∂V ∂Vr ∂Vϕ ∂Vz
= ⃗er + Vr ⃗eϕ + ⃗eϕ − Vϕ ⃗er + ⃗ez
∂ϕ ∂ϕ ∂ϕ ∂ϕ

21
Combining like terms:
⃗
∂V

∂Vr
 
∂Vϕ

∂Vz
= − Vϕ ⃗er + + Vr ⃗eϕ + ⃗ez
∂ϕ ∂ϕ ∂ϕ ∂ϕ

⃗
∂V
Vertical Derivative ∂z :

⃗
∂V ∂Vr ∂Vϕ ∂Vz
= ⃗er + ⃗eϕ + ⃗ez
∂z ∂z ∂z ∂z
Again, there is no derivative of the basis vectors with respect to z, as they are independent of z.

Including Metric Factors

Cylindrical coordinates have geometric scaling factors. Distances in the r-direction are straightforward, but in the ϕ-direction, the physical
length corresponding to a change in ϕ is rdϕ. This scaling factor must be included in the calculation of derivatives with respect to ϕ. Therefore,
the terms involving derivatives with respect to ϕ must be divided by r.

⃗ in Cylindrical Coordinates
Gradient of Tensor V
⃗ in cylindrical coordinates is:
Thus, the full gradient tensor ∇V
   
∂Vr 1 ∂Vr ∂Vr
− Vϕ
 ∂r r  ∂ϕ ∂z
⃗ =
 
∇V  ∂Vϕ 1 ∂Vϕ ∂Vϕ 
 ∂r r ∂ϕ + Vr ∂z 
∂Vz 1 ∂Vz ∂Vz
∂r r ∂ϕ ∂z

13 Vector Gradient ∇⃗v in Spherical Coordinates

Let the position in spherical coordinates be represented as (r, θ, ϕ), where:
• r: Radial distance.
• θ: Polar (or colatitudinal) angle (angle from the z-axis).
• ϕ: Azimuthal angle (angle in the xy-plane from the x-axis).
The basis vectors ⃗er , ⃗eθ , and ⃗eϕ in spherical coordinates are not constant but vary with position. These vectors are mutually orthogonal, and
the unit vectors are:
     
sin θ cos ϕ cos θ cos ϕ − sin ϕ
⃗er =  sin θ sin ϕ  , ⃗eθ =  cos θ sin ϕ  , ⃗eϕ =  cos ϕ 
cos θ − sin θ 0

⃗ in spherical coordinates is expressed as:

A general vector field V

⃗ = Vr (r, θ, ϕ) ⃗er + Vθ (r, θ, ϕ) ⃗eθ + Vϕ (r, θ, ϕ) ⃗eϕ

V
⃗ which is a tensor (often called the del operator applied to a vector field). To derive this rigorously, we
The goal is to find the gradient ∇V,
need to compute the derivative of each component of the vector field in the r, θ, and ϕ directions while considering the variation of the unit
vectors.

22
General Form of the Gradient of a Vector
⃗ = V i⃗ei , the gradient in curvilinear coordinates is expressed as:
In general, for a vector field V

⃗ = ∂V i
∇V ⃗ei ⊗ ⃗ej + V i ∇⃗ei
∂xj
Where V i are the components of the vector field, ⃗ei are the unit vectors, and ∇⃗ei are the gradients of the basis vectors. The term ⊗ indicates
a tensor product, resulting in a tensor. For spherical coordinates, we will first derive the derivative of the unit vectors ⃗er , ⃗eθ , and ⃗eϕ with
respect to r, θ, and ϕ.

Derivatives of the Unit Vectors

The unit vectors ⃗er , ⃗eθ , and ⃗eϕ vary with position. To compute the gradient of a vector field, we need their partial derivatives.

Derivative of ⃗er :
∂⃗er ∂⃗er ∂⃗er
= 0, = ⃗eθ , = sin θ ⃗eϕ
∂r ∂θ ∂ϕ

Derivative of ⃗eθ :
∂⃗eθ ∂⃗eθ ∂⃗eθ
= 0, = −⃗er , = cos θ ⃗eϕ
∂r ∂θ ∂ϕ

Derivative of ⃗eϕ :
∂⃗eϕ ∂⃗eϕ ∂⃗eϕ
= 0, = 0, = − sin θ ⃗er − cos θ ⃗eθ
∂r ∂θ ∂ϕ
These derivatives are essential for computing the gradient of a vector field in spherical coordinates.

⃗
Gradient of the Vector Field V
⃗ involves taking the derivative of each component in the r, θ, and ϕ directions, accounting for the variation of both the
The gradient of V
components Vr , Vθ , Vϕ and the basis vectors ⃗er , ⃗eθ , and ⃗eϕ . Let’s compute each term.

⃗
∂V
Radial Derivative ∂r :

⃗
∂V ∂Vr ∂Vθ ∂Vϕ
= ⃗er + ⃗eθ + ⃗eϕ
∂r ∂r ∂r ∂r
There is no derivative of the basis vectors with respect to r, as they are independent of r.

⃗
∂V
Polar Derivative ∂θ :

∂V⃗ ∂Vr ∂⃗er ∂Vθ ∂⃗eθ ∂Vϕ

= ⃗er + Vr + ⃗eθ + Vθ + ⃗eϕ
∂θ ∂θ ∂θ ∂θ ∂θ ∂θ
Substituting the derivatives of the basis vectors:

⃗
∂V ∂Vr ∂Vθ ∂Vϕ
= ⃗er + Vr ⃗eθ + ⃗eθ − Vθ ⃗er + ⃗eϕ
∂θ ∂θ ∂θ ∂θ

23
Combining like terms:

⃗
∂V

∂Vr
 
∂Vθ

∂Vϕ
= − Vθ ⃗er + + Vr ⃗eθ + ⃗eϕ
∂θ ∂θ ∂θ ∂θ

⃗
∂V
Azimuthal Derivative ∂ϕ :

⃗
∂V ∂Vr ∂⃗er ∂Vθ ∂⃗eθ ∂Vϕ ∂⃗eϕ
= ⃗er + Vr + ⃗eθ + Vθ + ⃗eϕ + Vϕ
∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ ∂ϕ
Using the derivatives of the unit vectors:

⃗
∂V ∂Vr ∂Vθ ∂Vϕ
= ⃗er + Vr sin θ ⃗eϕ + ⃗eθ + Vθ cos θ ⃗eϕ + ⃗eϕ − Vϕ sin θ ⃗er − Vϕ cos θ ⃗eθ
∂ϕ ∂ϕ ∂ϕ ∂ϕ
Combining terms:

⃗
∂V

∂Vr
 
∂Vθ
 
∂Vϕ

= − Vϕ sin θ ⃗er + − Vϕ cos θ ⃗eθ + + Vr sin θ + Vθ cos θ ⃗eϕ
∂ϕ ∂ϕ ∂ϕ ∂ϕ

⃗ in Spherical Coordinates
Gradient of Tensor V
⃗ in spherical coordinates is then given by:
The gradient tensor ∇V
   
∂Vr 1 ∂Vr 1 ∂Vr


− Vθ − Vϕ sin θ
 ∂r r ∂θ r sin θ
 ∂ϕ  
⃗ =  ∂Vθ 1 ∂Vθ 1 ∂Vθ


∇V + Vr − V cos θ

 ∂r ϕ
r ∂θ rsin θ ∂ϕ 
  
∂Vϕ 1 ∂Vϕ 1 ∂Vϕ
∂r r ∂θ + Vθ cos θ r sin θ ∂ϕ + Vr sin θ + Vθ cos θ

14 Divergence ∇ · T of a Second-Order Tensor in Cylindrical Coordinates

To derive the divergence of a second-order tensor in cylindrical coordinates, we begin by recalling that the divergence of a second-order tensor
field T in Cartesian coordinates is expressed as:

∂Tij
(∇ · T)i =
∂xj

In cylindrical coordinates, we have a new set of basis vectors corresponding to the radial direction ⃗er , the azimuthal direction ⃗eθ , and the
axial direction ⃗ez , and the divergence operator becomes more complicated due to the non-orthogonal coordinate system. The coordinates
(r, θ, z) are related to the Cartesian coordinates (x, y, z) by:

x = r cos θ, y = r sin θ, z = z.

24
Step 1: Coordinate Setup in Cylindrical Coordinates
The metric tensor in cylindrical coordinates is diagonal, and the components are:

grr = 1, gθθ = r2 , gzz = 1.

The volume element associated with the metric is:
q
det(gij ) = r

Step 2: General Form of the Divergence in Cylindrical Coordinates

In cylindrical coordinates, the divergence of a second-order tensor is given by:
q 
1 ∂
(∇ · T)i = p det(gij )Tij .
det(gij ) ∂xj
In cylindrical coordinates, this becomes:

1 ∂
(∇ · T)i = (rTij )
r ∂xj
where the summation is over j = r, θ, z. The divergence of the second-order tensor T in cylindrical coordinates has three components: radial
(r), azimuthal (θ), and axial (z). We will expand each component explicitly, rigorously applying the necessary derivatives.

Radial Component (∇ · T)r

The radial component is given by:

1 ∂ 1 ∂Tθr ∂Tzr Tθθ

(∇ · T)r = (rTrr ) + + −
r ∂r r ∂θ ∂z r

First Term
1 ∂
(rTrr )
r ∂r
Using the product rule:
∂ ∂Trr
(rTrr ) = Trr + r
∂r ∂r
so:
1 ∂ Trr ∂Trr
(rTrr ) = +
r ∂r r ∂r

Second Term
1 ∂Tθr
.
r ∂θ
This is a straightforward derivative of the θr-component with respect to θ, scaled by 1/r.

25
Third Term
∂Tzr
.
∂z
This is a straightforward partial derivative with respect to z.

Fourth Term
Tθθ
.−
r
This term accounts for the divergence in the azimuthal direction, involving the θθ-component, scaled by 1/r.

Final Expression for Radial Component

Thus, the radial component becomes:
Trr ∂Trr 1 ∂Tθr ∂Tzr Tθθ
(∇ · T)r = + + + −
r ∂r r ∂θ ∂z r

Azimuthal Component (∇ · T)θ

The azimuthal component is:

1 ∂ 1 ∂Tθθ ∂Tzθ Trθ

(∇ · T)θ = (rTrθ ) + + +
r ∂r r ∂θ ∂z r

First Term
1 ∂
(rTrθ ) .
r ∂r
Using the product rule:
∂ ∂Trθ
(rTrθ ) = Trθ + r ,
∂r ∂r
so:
1 ∂ Trθ ∂Trθ
(rTrθ ) = +
r ∂r r ∂r

Second Term
1 ∂Tθθ
.
r ∂θ
This is the partial derivative of Tθθ with respect to θ, scaled by 1/r.

Third Term
∂Tzθ
.
∂z
This is the straightforward partial derivative of the zθ-component with respect to z.

26
Fourth Term
Trθ
.
r
This term accounts for the contribution of the radial direction to the divergence in the azimuthal direction.

Final Expression for Azimuthal Component

Thus, the azimuthal component becomes:

Trθ ∂Trθ 1 ∂Tθθ ∂Tzθ Trθ

(∇ · T)θ = + + + +
r ∂r r ∂θ ∂z r

Axial Component (∇ · T)z

The axial component is:

1 ∂ 1 ∂Tθz ∂Tzz
(∇ · T)z = (rTrz ) + +
r ∂r r ∂θ ∂z

First Term
1 ∂
(rTrz ) .
r ∂r
Using the product rule:
∂ ∂Trz
(rTrz ) = Trz + r ,
∂r ∂r
so:
1 ∂ Trz ∂Trz
(rTrz ) = +
r ∂r r ∂r

Second Term
1 ∂Tθz
.
r ∂θ
This is the partial derivative of Tθz with respect to θ, scaled by 1/r.

Third Term
∂Tzz
.
∂z
This is the straightforward partial derivative of the zz-component with respect to z.

Final Expression for Axial Component

Thus, the axial component becomes:

Trz ∂Trz 1 ∂Tθz ∂Tzz

(∇ · T)z = + + +
r ∂r r ∂θ ∂z

27
Final Complete Answer
The derivation of the second-order tensor divergence in cylindrical coordinates results in the following expressions for the three components:

1. Radial Component

Trr ∂Trr 1 ∂Tθr ∂Tzr Tθθ

(∇ · T)r = + + + −
r ∂r r ∂θ ∂z r

2. Azimuthal Component

Trθ ∂Trθ 1 ∂Tθθ ∂Tzθ Trθ

(∇ · T)θ = + + + +
r ∂r r ∂θ ∂z r

3. Axial Component

Trz ∂Trz 1 ∂Tθz ∂Tzz

(∇ · T)z = + + +
r ∂r r ∂θ ∂z
These three components together form the complete tensor divergence in cylindrical coordinates.

15 Divergence ∇ · T of a Second-Order Tensor in Spherical Coordinates

To derive the divergence of a second-order tensor in spherical coordinates, we begin with the general expression for the divergence of a
second-order tensor field T. In Cartesian coordinates, the divergence of a second-order tensor is given by:

∂Tij
(∇ · T)i =
∂xj

However, in spherical coordinates (r, θ, ϕ), where we have the orthonormal basis vectors ⃗er , ⃗eθ , and ⃗eϕ , the tensor components must be
expressed in terms of the spherical coordinate system, and the divergence operator takes a more complex form due to the non-Cartesian
nature of the coordinates.

Step 1: Coordinate Setup in Spherical Coordinates

In spherical coordinates, the metric tensor gij is diagonal, with the following non-zero components:

grr = 1, gθθ = r2 , gϕϕ = r2 sin2 θ.

Thus, the volume element (related to the determinant of the metric tensor) is:
q
det(gij ) = r2 sin θ.

Step 2: General Form of Divergence in Spherical Coordinates

The divergence of a second-order tensor T in spherical coordinates is expressed as:
q 
1 ∂
(∇ · T)i = p det(gij )Tij
det(gij ) j
∂x

28
In spherical coordinates, for each direction i = r, θ, ϕ, we expand this expression using the appropriate scale factors and differentials in r,
θ, and ϕ. The divergence of the second-order tensor T has three components corresponding to the radial (r), polar (θ), and azimuthal (ϕ)
directions. Let’s derive each component explicitly, performing the necessary derivatives.

Radial Component (∇ · T)r

The radial component is:

1 ∂ 1 ∂ 1 ∂Tϕr
r2 Trr +


(∇ · T)r = 2
(sin θTθr ) +
r ∂r r sin θ ∂θ r sin θ ∂ϕ

First Term
1 ∂
r2 Trr .


r2 ∂r
Apply the product rule:
∂ ∂Trr
r2 Trr = 2rTrr + r2


,
∂r ∂r
so:
1 ∂
2Trr ∂Trr
2
r2 Trr = +
r ∂r r ∂r

Second Term
1 ∂
(sin θTθr )
r sin θ ∂θ
Again, apply the product rule:
∂ ∂Tθr
(sin θTθr ) = cos θTθr + sin θ
∂θ ∂θ
so:  
1 ∂ 1 ∂Tθr cos θ 1 ∂Tθr
(sin θTθr ) = cos θTθr + sin θ = Tθr +
r sin θ ∂θ r sin θ ∂θ r sin θ r ∂θ

Third Term
1 ∂Tϕr
.
r sin θ ∂ϕ
This is a straightforward partial derivative.

Final Expression for Radial Component

Thus, the radial component becomes:

2Trr ∂Trr cos θ 1 ∂Tθr 1 ∂Tϕr

(∇ · T)r = + + Tθr + +
r ∂r r sin θ r ∂θ r sin θ ∂ϕ

29
Polar Component (∇ · T)θ
The polar component is:

1 ∂ 1 ∂ 1 ∂Tϕθ Tϕϕ cot θ

r2 Trθ +


(∇ · T)θ = 2
(sin θTθθ ) + −
r ∂r r sin θ ∂θ r sin θ ∂ϕ r2

First Term
1 ∂
r2 Trθ .


r2 ∂r
Using the product rule:
∂ ∂Trθ
r2 Trθ = 2rTrθ + r2


,
∂r ∂r
so:
1 ∂
2Trθ ∂Trθ
2
r2 Trθ = +
r ∂r r ∂r

Second Term
1 ∂
(sin θTθθ )
r sin θ ∂θ
Using the product rule:
∂ ∂Tθθ
(sin θTθθ ) = cos θTθθ + sin θ
∂θ ∂θ
so:  
1 ∂ 1 ∂Tθθ cos θ 1 ∂Tθθ
(sin θTθθ ) = cos θTθθ + sin θ = Tθθ +
r sin θ ∂θ r sin θ ∂θ r sin θ r ∂θ

Third Term
1 ∂Tϕθ
.
r sin θ ∂ϕ
This is a straightforward partial derivative.

Fourth Term
Tϕϕ cot θ
− .
r2
This term arises due to the cot θ factor in the spherical coordinates, and it’s already simplified.

Final Expression for Polar Component

Thus, the polar component becomes:

2Trθ ∂Trθ cos θ 1 ∂Tθθ 1 ∂Tϕθ Tϕϕ cot θ

(∇ · T)θ = + + Tθθ + + −
r ∂r r sin θ r ∂θ r sin θ ∂ϕ r2

30
Azimuthal Component (∇ · T)ϕ
The azimuthal component is:

1 ∂ 1 ∂ 1 ∂Tϕϕ Tθϕ cot θ

r2 Trϕ +


(∇ · T)ϕ = (sin θTθϕ ) + +
r2 ∂r r sin θ ∂θ r sin θ ∂ϕ r2

First Term
1 ∂
r2 Trϕ .


2
r ∂r
Using the product rule:
∂ ∂Trϕ
r2 Trϕ = 2rTrϕ + r2


,
∂r ∂r
so:
1 ∂
2Trϕ ∂Trϕ
2
r2 Trϕ = +
r ∂r r ∂r

Second Term
1 ∂
(sin θTθϕ )
r sin θ ∂θ
Using the product rule:
∂ ∂Tθϕ
(sin θTθϕ ) = cos θTθϕ + sin θ
∂θ ∂θ
so:  
1 ∂ 1 ∂Tθϕ cos θ 1 ∂Tθϕ
(sin θTθϕ ) = cos θTθϕ + sin θ = Tθϕ +
r sin θ ∂θ r sin θ ∂θ r sin θ r ∂θ

Third Term
1 ∂Tϕϕ
.
r sin θ ∂ϕ
This is a straightforward partial derivative.

Fourth Term
Tθϕ cot θ
.
r2
This is already simplified.

Final Expression for Azimuthal Component

Thus, the azimuthal component becomes:

2Trϕ ∂Trϕ cos θ 1 ∂Tθϕ 1 ∂Tϕϕ Tθϕ cot θ

(∇ · T)ϕ = + + Tθϕ + + +
r ∂r r sin θ r ∂θ r sin θ ∂ϕ r2

31
Final Complete Answer
Therefore the second-order tensor divergence in spherical coordinates results in the following expressions for the three components:

1. Radial Component

2Trr ∂Trr cos θ 1 ∂Tθr 1 ∂Tϕr

(∇ · T)r = + + Tθr + +
r ∂r r sin θ r ∂θ r sin θ ∂ϕ

2. Polar Component

2Trθ ∂Trθ cos θ 1 ∂Tθθ 1 ∂Tϕθ Tϕϕ cot θ

(∇ · T)θ = + + Tθθ + + −
r ∂r r sin θ r ∂θ r sin θ ∂ϕ r2

3. Azimuthal Component

2Trϕ ∂Trϕ cos θ 1 ∂Tθϕ 1 ∂Tϕϕ Tθϕ cot θ

(∇ · T)ϕ = + + Tθϕ + + +
r ∂r r sin θ r ∂θ r sin θ ∂ϕ r2
These three components together form the complete tensor divergence in spherical coordinates.

16 Directional Derivative (A · ∇)⃗v in Curvilinear Coordinates

The Directional Derivative defined for a vector field A by (A · ∇), where ∇ is the gradient operator. The Directional Derivative operator
applied in arbitrary orthogonal three-dimensional coordinates to a vector field ⃗v becomes:

3   
X Ak ∂vj vk ∂hj ∂hk
[(A · ∇)⃗v]j = + Aj − Ak
hk ∂qk hk hj ∂qk ∂qj
k=1
√
where the hi ’s are related to the metric tensors by hi = gii .

Directional Derivative in Cartesian Coordinates:

In rectangular coordinates, g11 = g22 = g33 = 1, therefore the Directional Derivative defined for a vector field ⃗v in Rectangular Coordinates
shall be      
∂vx ∂vx ∂vx ∂vy ∂vy ∂vy ∂vz ∂vz ∂vz
(A · ∇)⃗v = Ax + Ay + Az x̂ + Ax + Ay + Az ŷ + Ax + Ay + Az ẑ
∂x ∂y ∂z ∂x ∂y ∂z ∂x ∂y ∂z

Directional Derivative in Cylindrical Coordinates:

In cylindrical coordinates (r, θ, z), g11 = 1, g22 = r2 , g33 = 1 , therefore the Directional Derivative defined for a vector field ⃗v in Cylindrical
Coordinates shall be
 
∂vr Aϕ ∂vr ∂vr Aϕ v ϕ
(A · ∇)⃗v = Ar + + Az − r̂
∂r r ∂ϕ ∂z r
 
∂vϕ Aϕ ∂vϕ ∂vϕ Aϕ v r
+ Ar + + Az + ϕ̂
∂r r ∂ϕ ∂z r
 
∂vz Aϕ ∂vz ∂vz
+ Ar + + Az ẑ
∂r r ∂ϕ ∂z

32
Directional Derivative in Spherical Coordinates
In spherical coordinates (r, θ, ϕ), g11 = 1, g22 = r2 , g33 = r2 sin2 θ, therefore the Directional Derivative defined for a vector field ⃗v in
Spherical Coordinates shall be
 
∂vr Aθ ∂vr Aϕ ∂vr Aθ v θ + Aϕ v ϕ
(A · ∇)⃗v = Ar + + − r̂
∂r r ∂θ r sin θ ∂ϕ r
 
∂vθ Aθ ∂vθ Aϕ ∂vθ Aθ v r Aϕ vϕ cot θ
+ Ar + + + − θ̂
∂r r ∂θ r sin θ ∂ϕ r r
 
∂vϕ Aθ ∂vϕ Aϕ ∂vϕ Aϕ v r Aϕ vθ cot θ
+ Ar + + + + ϕ̂
∂r r ∂θ r sin θ ∂ϕ r r

17 Vector Laplacian ∆⃗v in Curvilinear Coordinates

Consider the general curvilinear coordinates (u1 , u2 , u3 ). We shall limit ourselves to orthogonal coordinate systems in Euclidean 3-space, each
system being characterized by the metric coefficients g11 , g22 , g33 . The element of distance is specified by:

(ds)2 = g11 (du1 )2 + g22 (du2 )2 + g33 (du3 )2 .

Also,

1 ∂ϕ 1 ∂ϕ 1 ∂ϕ
grad ϕ = a1 1
+ a2 2
+ a3
g11 ∂u g22 ∂u g33 ∂u3
 
1 ∂ √ ∂ √ ∂ √
div ⃗v = √ ( gv1 ) + ( gv2 ) + ( gv3 )
g ∂u1 ∂u2 ∂u3
and
 
1 ∂ ∂
curl ⃗v = (g22 v3 ) − (g33 v2 )
g11 ∂u2 ∂u3
 
1 ∂ ∂
+ (g33 v1 ) − (g11 v3 )
g22 ∂u3 ∂u1
 
1 ∂ ∂
+ (g11 v2 ) − (g22 v1 )
g33 ∂u1 ∂u2
where a1 , a2 , and a3 are unit vectors, and g = g11 g22 g33 . Using the above equations, we obtain the general expression for the vector Laplacian
operating on ⃗v:
      
1 ∂ 1 ∂v1 1 ∂ v1 1 ∂ v1
∆⃗v = a1 + +
g11 ∂u1 g11 ∂u1 g22 ∂u2 g22 g33 ∂u3 g33
      
1 ∂ v2 1 ∂ 1 ∂v2 1 ∂ v2
+a2 + +
g11 ∂u1 g11 g22 ∂u2 g22 ∂u2 g33 ∂u3 g33
      
1 ∂ v3 1 ∂ v3 1 ∂ 1 ∂v3
+a3 + +
g11 ∂u1 g11 g22 ∂u2 g22 g33 ∂u3 g33 ∂u3
This is much more complex than the scalar case:

1 ∂2ϕ 1 ∂2ϕ 1 ∂2ϕ

∇2 ϕ = + +
g11 ∂(u1 )2 g22 ∂(u2 )2 g33 ∂(u3 )2

33
Vector Laplacian in Rectangular Coordinates
In rectangular coordinates, g11 = g22 = g33 = 1, therefore the Vector Laplacian ∆⃗v in Rectangular Coordinates shall be

∂ 2 vx ∂ 2 vy ∂ 2 vz
∆⃗v = ax 2
+ ay 2
+ az
∂x ∂y ∂z 2
While

∂2ϕ ∂2ϕ ∂2ϕ

∇2 ϕ = + 2 + 2
∂x2 ∂y ∂z

In this special case, each component of ∆⃗v has the same form as ∇2 ϕ.

Vector Laplacian in Cylindrical Coordinates

In cylindrical coordinates (r, θ, z), g11 = 1, g22 = r2 , g33 = 1 , therefore the Vector Laplacian ∆⃗v in Cylindrical Coordinates shall be
 2
1 ∂ 2 vr ∂ 2 vr

∂ vr 1 ∂vr vr
∆⃗v = ar + − 2+ 2 +
∂r2 r ∂r r r ∂θ2 ∂z 2
 2 2
∂ 2 vθ

∂ vθ 1 ∂vθ vθ 1 ∂ vθ
+aθ + − 2+ 2 +
∂r2 r ∂r r r ∂θ2 ∂z 2
 2 2
∂ 2 vz

∂ vz 1 ∂vz 1 ∂ vz
+az + + 2 +
∂r2 r ∂r r ∂θ2 ∂z 2
and

∂ 2 ϕ 1 ∂ϕ 1 ∂2ϕ ∂2ϕ
∇2 ϕ = 2
+ + 2 2 + 2
∂r r ∂r r ∂θ ∂z
The first two components of ∆⃗v are of different form from ∇2 ϕ, each containing two additional terms.

Vector Laplacian in Spherical Coordinates

In spherical coordinates (r, θ, ϕ), g11 = 1, g22 = r2 , g33 = r2 sin2 θ, therefore the Vector Laplacian ∆⃗v in Spherical Coordinates shall be
 2
1 ∂ 2 vr ∂ 2 vr

∂ vr 2 ∂vr 2vr cot θ ∂vr 1
∆⃗v = ar + − 2 + 2 + 2 + 2 2
∂r2 r ∂r r r ∂θ2 r ∂θ r sin θ ∂ϕ2
 2 2
∂ 2 vθ

∂ vθ 2 ∂vθ vθ 1 ∂ vθ cot θ ∂vθ 1
+aθ + − 2+ 2 + 2 + 2 2
∂r2 r ∂r r r ∂θ2 r ∂θ r sin θ ∂ϕ2
 2 2
∂ 2 vϕ

∂ vϕ 2 ∂vϕ 1 ∂ vϕ cot θ ∂vϕ 1
+aϕ + + 2 + 2 + 2 2
∂r2 r ∂r r ∂θ2 r ∂θ r sin θ ∂ϕ2
and

∂ 2 ϕ 2 ∂ϕ 1 ∂ 2 ϕ cot θ ∂ϕ 1 ∂2ϕ
∇2 ϕ = + + + + 2
∂r2 r ∂r r2 ∂θ2 r2 ∂θ r2 sin θ ∂ϕ2
Here each component of ∆⃗v contains more terms than ∇2 ϕ. Other coordinates are handled in the same manner as the foregoing three cases.
In the more complicated coordinate systems, the difference between the vector and scalar Laplacians becomes even more pronounced.

34
18 Derivatives with respect to vectors and second-order tensors
18.1 Derivatives of scalar and vector valued functions of vectors
Let f (v) be a real valued function of the vector v. Then the derivative of f (v) with respect to v (or at v) is the vector defined through its
dot product with any vector u being
 
∂f d
· u = Df (v)[u] = f (v + αu)
∂v dα α=0
for all vectors u. The above dot product yields a scalar, and if u is a unit vector, it gives the directional derivative of f at v, in the u direction.
If f (v) = f1 (v) + f2 (v) then
∂f ∂f1 ∂f2
·u= ·u+ ·u
∂v ∂v ∂v
If f (v) = f1 (v)f2 (v) then    
∂f ∂f1 ∂f2
·u= · u f2 (v) + f1 (v) ·u
∂v ∂v ∂v
If f (v) = f1 (f2 (v)) then
∂f ∂f1 ∂f2
·u= ·u
∂v ∂f2 ∂v
Let f (v) be a vector valued function of the vector v. Then the derivative of f (v) with respect to v (or at v) is the second order tensor defined
through its dot product with any vector u being
 
∂f d
· u = Df (v)[u] = f (v + αu)
∂v dα α=0
for all vectors u. The above dot product yields a vector, and if u is a unit vector, it gives the directional derivative of f at v, in the u direction.
If f (v) = f1 (v) + f2 (v) then
∂f ∂f1 ∂f2
·u= ·u+ ·u
∂v ∂v ∂v
If f (v) = f1 (v) × f2 (v) then    
∂f ∂f1 ∂f2
·u= · u × f2 (v) + f1 (v) × ·u
∂v ∂v ∂v
If f (v) = f1 (f2 (v)) then  
∂f ∂f1 ∂f2
·u= · ·u
∂v ∂f2 ∂v

18.2 Derivatives of scalar and vector valued functions of second-order tensors

Let f (S) be a real valued function of the second order tensor S. Then the derivative of f (S) with respect to S (or at S) in the direction T is
the second order tensor defined as
 
∂f d
: T = Df (S)[T] = f (S + αT)
∂S dα α=0

for all second order tensors T. If F(S) = F1 (S) + F2 (S) then

 
∂F ∂F1 ∂F2
:T= + :T
∂S ∂S ∂S

35
If F(S) = F1 (S) · F2 (S) then    
∂F ∂F1 ∂F2
:T= : T · F2 (S) + F1 (S) · :T
∂S ∂S ∂S
If F(S) = F1 (F2 (S)) then  
∂F ∂F1 ∂F2
:T= : :T
∂S ∂F2 ∂S
If f (S) = f1 (F2 (S)) then  
∂f ∂f1 ∂F2
:T= : :T
∂S ∂F2 ∂S

19 Advanced Tensor Derivatives

19.1 Christoffel symbols
This section introduces the Christoffel symbols as part of the formalism used to express the connection on a manifold in local coordinates.
These symbols help define the covariant derivative, which is critical in differential geometry, particularly in curved spaces. Here’s what we
need to rigorously derive:

• Manifold, Coordinate system setup and the metric tensor

• Covariant differentiation of a vector.

Let’s break this down step-by-step with detailed mathematical rigor.

19.1.1 Manifold, Coordinate System setup and the metric tensor

Let M be a differentiable manifold of dimension n. At each point p ∈ M, the tangent space Tp M is an n-dimensional vector space. To
specify the geometry of M, we introduce a coordinate chart (U, φ), where U ⊂ M is an open set and φ : U → Rn is a smooth, bijective map.
For each point p ∈ U , φ(p) = (x1 (p), x2 (p), . . . , xn (p)) gives the local coordinates of p. The coordinate basis vectors at each point are denoted
∂
by ∂x i , abbreviated as ∂i . These vectors span the tangent space Tp M at each point p.

To define distances and angles on the manifold, we introduce the metric tensor g, which is a smooth, symmetric, positive-definite (0,2)-
tensor field. In a coordinate chart, the metric is represented as:

g = gij dxi ⊗ dxj

∂ ∂


where gij = g ∂x i , ∂xj are the components of the metric in the chosen coordinates, and dxi are the dual basis 1-forms. The metric
components gij are smooth functions on the manifold, and gij = gji , ensuring the symmetry of the metric tensor. The inverse of the metric
tensor is denoted by g ij , where:

g ik gkj = δji
where δji is the Kronecker delta.

19.1.2 Covariant Derivative of a Vector Field

On a manifold, the covariant derivative extends the notion of differentiation to curved spaces, ensuring that differentiation is consistent with
the manifold’s geometry. To define it, we need the notion of a connection. The covariant derivative of a vector field X in the direction of
another vector field Y , denoted ∇Y X, must satisfy the following properties:

1. Linearity: ∇Y (aX + bZ) = a∇Y X + b∇Y Z for any scalar functions a and b, and vector fields X, Z.

36
 2. Leibniz Rule: ∇Y (f X) = f ∇Y X + (Y f )X for any scalar function f .
3. Compatibility with the Metric: Y (g(X, Z)) = g(∇Y X, Z) + g(X, ∇Y Z), ensuring that the covariant derivative preserves the inner
product structure defined by the metric tensor.

The Christoffel symbols Γkij are the components of the connection in a coordinate basis. They provide a way to express the covariant derivative
in terms of local coordinates. The Christoffel symbols are defined by the condition that the covariant derivative of the basis vector field ∂j is
a linear combination of the basis vectors:

∇∂i ∂j = Γkij ∂k
The Christoffel symbols Γkij are related to the metric tensor by the following expression, derived from the requirement that the covariant
derivative preserves the metric (i.e., ∇g = 0):
1 kl
Γkij = g (∂i gjl + ∂j gil − ∂l gij )
2
where g kl is the inverse of the metric tensor, and ∂i gjl denotes the partial derivative of the metric component gjl with respect to the coordinate
xi . Let’s derive the expression for the Christoffel symbols step-by-step:

1. Start from the compatibility condition for the metric tensor:

∇k gij = 0

This implies that the metric is covariantly constant.

2. Expanding this condition in terms of the Christoffel symbols:

∂k gij − Γlki glj − Γlkj gil = 0

3. Rearranging terms gives the following equation:

∂k gij = Γlki glj + Γlkj gil

4. Using the symmetry of the metric gij , cyclically permute the indices i, j, and k in the above equation, then add and subtract the
resulting equations to isolate the Christoffel symbols. After some algebra, we arrive at the expression:
1 kl
Γkij = g (∂i gjl + ∂j gil − ∂l gij )
2

Thus, the Christoffel symbols are completely determined by the metric tensor and its first derivatives. Given a vector field X = X i ∂i , the
covariant derivative of X in the direction of ∂j is:

∇j X i = ∂j X i + Γijk X k
This formula expresses the covariant derivative of a vector field in terms of the Christoffel symbols and the partial derivatives of the components
of the vector field.

19.1.3 Christoffel symbols in Euclidean Space

The section defines the Christoffel symbols in a Euclidean space, where the coordinates are typically Cartesian, but we can also consider
curvilinear coordinates (e.g., polar, cylindrical, spherical). Christoffel symbols appear when we want to express derivatives of vector fields
in a non-Cartesian coordinate system. In Euclidean space, the Christoffel symbols represent how the coordinate basis vectors change as we
move from one point to another. We will rigorously derive the following key components:
• Euclidean space and coordinate systems.

37
 • Connection between the Christoffel symbols and coordinate transformations.
• The role of Christoffel symbols in non-Cartesian coordinates.
• Detailed derivation of Christoffel symbols in Euclidean space.
We begin by considering the standard n-dimensional Euclidean space Rn . The Euclidean space is a flat space, and it can be equipped with
different coordinate systems:
• Cartesian coordinates (x1 , x2 , . . . , xn ): These coordinates are orthogonal and have a straightforward metric, which is the identity
matrix.
• Curvilinear coordinates (q 1 , q 2 , . . . , q n ): These coordinates can be functions of the Cartesian coordinates, such as polar or spherical
coordinates.
The key idea is that while Euclidean space has no intrinsic curvature, when we introduce non-Cartesian coordinates, the coordinate basis vectors
change from point to point. This change is precisely captured by the Christoffel symbols. In a general coordinate system (q 1 , q 2 , . . . , q n ), the
basis vectors ∂q∂ i span the tangent space at each point. These basis vectors are not necessarily orthogonal, and their inner products define
the components of the metric tensor gij , which in Euclidean space is:
 k l
∂x ∂x
gij = δkl
∂q i ∂q j
where δkl is the Kronecker delta, representing the flat Cartesian metric of Euclidean space. Thus, the metric tensor in curvilinear coordinates
is:
n
X ∂xk ∂xk
gij =
∂q i ∂q j
k=1
In Cartesian coordinates, the covariant derivative reduces to the partial derivative because the Christoffel symbols vanish. However, in
curvilinear coordinates, the basis vectors ∂q∂ i vary from point to point, and this variation is captured by the Christoffel symbols. The
covariant derivative ∇j of a vector field V i is given by:

∇j V i = ∂j V i + Γijk V k
where Γijk are the Christoffel symbols. The Christoffel symbols in Euclidean space are derived from the requirement that the covariant
derivative of the metric tensor is zero: ∇k gij = 0. This is the metric compatibility condition. Expanding it gives:

∂k gij − Γlki glj − Γlkj gil = 0

Rearranging this expression gives:

∂k gij = Γlki glj + Γlkj gil

After some manipulation, we obtain the Christoffel symbols as:
1 kl
Γkij =g (∂i gjl + ∂j gil − ∂l gij )
2
In Cartesian coordinates, the metric tensor gij is the identity matrix, gij = δij , and its derivatives vanish. Thus, the Christoffel symbols are:
Γkij = 0. In curvilinear coordinates, the Christoffel symbols are generally non-zero. For example, in 2D polar coordinates (r, θ), the metric
tensor is:
 
1 0
gij =
0 r2
One of the non-zero Christoffel symbols in polar coordinates is: Γθrθ = 1r .

38
19.1.4 Christoffel Symbols of the Second Kind
The Christoffel symbols of the second kind, often denoted as Γkij , represent the components of the Levi-Civita connection in a local coordinate
system on a differentiable manifold. They are used to express the covariant derivative of tensor fields and play a crucial role in general
relativity, Riemannian geometry, and differential geometry in general.

Let M be an n-dimensional differentiable manifold with a smooth Riemannian metric tensor g. The Christoffel symbols of the second
kind Γkij represent the components of the connection in a coordinate basis. The Christoffel symbols are defined by the following equation for
∂
the covariant derivative of the basis vector fields ∂x i:

∂ ∂
∇ = Γkij k
∂
∂xj∂xi ∂x
The covariant derivative on a manifold must satisfy the condition of metric compatibility, meaning that the covariant derivative of the metric
tensor g is zero ∇k gij = 0. Expanding the covariant derivative of the metric tensor gives:

∂k gij − Γlki glj − Γlkj gil = 0

Rearranging the equation from metric compatibility:

∂k gij = Γlki glj + Γlkj gil

Cyclically permuting the indices i, j, and k, and adding/subtracting the resulting equations, we obtain:

∂k gij + ∂i gjk − ∂j gki = 2Γlki glj

Multiplying by g kl , we find the Christoffel symbols:
1 kl
Γkij = g (∂i gjl + ∂j gil − ∂l gij )
2
The Christoffel symbols are symmetric in the lower two indices: Γkij = Γkji . This follows from the symmetry of the metric tensor gij = gji .
The Christoffel symbols describe how basis vectors change as we move on the manifold. In general relativity, they are used in the geodesic
equation:

d2 xi dxj dxk
2
+ Γijk =0
dτ dτ dτ
∂
 ∂ ∂

In a holonomic basis, the basis vectors ∂x i are derived directly from a coordinate system. These vectors commute: ∂xi , ∂xj = 0. In a
c c
nonholonomic basis, the basis vectors ea may not commute [ea , eb ] = Cab ec where Cab are the structure constants. The covariant derivative
in a nonholonomic basis {ea } is given by:
c
∇ea eb = ωab ec
c c c
where ωab are the connection coefficients. The structure constants Cab are defined by: [ea , eb ] = Cab ec . These constants are antisymmetric
Cab = −Cba . In a nonholonomic basis, the connection coefficients are related to the Christoffel symbols Γcab by:
c c

c 1 c
ωab = Γcab + Cab
2
We derive the connection coefficients as:
c 1 cd 1 c
ωab = g (∂a gbd + ∂b gad − ∂d gab ) + Cab
2 2
c
In general, the connection coefficients ωab are not symmetric in their lower indices due to the structure constants.

39
19.1.5 Transformation Law for Christoffel Symbols
The Christoffel symbols Γkij define the covariant derivative in a given coordinate system. In this document, we derive the transformation law
for the Christoffel symbols under a change of variables, focusing on the relationship between the original and transformed coordinates. The
Christoffel symbols in a local coordinate system (xi ) are defined through the covariant derivative of the basis vectors:

∂ ∂
∇ ∂ = Γijk i
∂xj ∂xk ∂x
Let (x1 , . . . , xn ) and (x̄1 , . . . , x̄n ) be two coordinate systems related by a smooth map x̄i = x̄i (x1 , . . . , xn ). The Jacobian matrix is given by:
∂ x̄i
∂xj . The basis vectors transform as:
∂ ∂ x̄j ∂
i
=
∂x ∂xi ∂ x̄j
The transformation law for the Christoffel symbols is:

∂ x̄m ∂xj ∂xn i ∂ x̄m ∂ 2 xp

Γ̄m
kl = i k l
Γjn +
∂x ∂ x̄ ∂ x̄ ∂xp ∂ x̄k ∂ x̄l
The Christoffel symbols are symmetric in their lower indices, and this symmetry is preserved under the transformation:

Γkij = Γkji

19.2 Parallel Transport in Riemannian Space

A Riemannian manifold (M, g) consists of a differentiable manifold equipped with a smooth, symmetric, positive-definite metric tensor g,
which defines an inner product at each point. In local coordinates, the metric tensor is:

g = gij (x) dxi ⊗ dxj

∂ ∂


where gij = g ∂xi , ∂xj . Parallel transport describes how a vector field V(t) is moved along a curve γ(t) such that its covariant derivative
vanishes:
∇γ̇(t) V(t) = 0
The covariant derivative ∇ satisfies linearity, the Leibniz rule, and metric compatibility. The Christoffel symbols Γkij are defined by:

∇j V i = ∂j V i + Γijk V k

The Christoffel symbols are determined by the metric compatibility condition: ∇k gij = 0 Expanding this, we get:

∂k gij − Γlki glj − Γlkj gil = 0

By permuting indices and solving, we find:

1 kl
Γkij = g (∂i gjl + ∂j gil − ∂l gij )
2
The Christoffel symbols describe how vectors are parallel transported along curves. For a vector V i (t) being parallel transported along a
curve γ(t), the condition is:
dV i dγ j k
+ Γijk V =0
dt dt
∂
The Christoffel symbols Γkij describe the covariant derivative of a vector field V i in the direction of ∂xj :

∇j V i = ∂j V i + Γijk V k

40
In index-free notation, the covariant derivative ∇X V of a vector field V along X satisfies linearity and the Leibniz rule:

∇X (f V ) = f ∇X V + (Xf )V

In a coordinate system, the covariant derivative is expressed in indexed notation as:

  ∂
∇X Y = X i ∂i Y j + X i Γjik Y k
∂xj
In index-free notation, it is written abstractly as:

∇X Y = abstract connection operator acting on X and Y

The Levi-Civita connection satisfies the metric compatibility condition ∇g = 0, expressed in indexed notation as: ∇k gij = 0. In index-free
notation, this is written as:
(∇X g)(Y, Z) = 0
The Christoffel symbols provide a coordinate-based description of the connection, while index-free notation describes the connection in an
abstract, coordinate-independent manner.

19.2.1 Derivation of the Relationship to Index-Free Notation

The Christoffel symbols in the indexed notation are used to represent the covariant derivative in a coordinate basis. For a Riemannian
manifold (M, g) with local coordinates (x1 , x2 , . . . , xn ), the Christoffel symbols Γkij are defined through the covariant derivative of a vector
∂
field V i in the direction of a coordinate basis vector ∂x j:

∇j V i = ∂j V i + Γijk V k
∂
Here, the covariant derivative is expressed in terms of the partial derivative ∂j = ∂xj and the Christoffel symbols Γijk , which describe how
∂
the coordinate basis vectors ∂x i change as one moves through the manifold.

In contrast to the indexed representation, the index-free notation expresses geometric objects and operations without explicit reference
to coordinates. Instead, the focus is on the intrinsic geometric meaning of these objects.

The covariant derivative of a vector field V in the direction of a vector field X is denoted by ∇X V in index-free notation. This can be
understood as a directional derivative of the vector field V along X, and is defined such that:

1. ∇X (aV + bW ) = a∇X V + b∇X W (linearity),

2. ∇X (f V ) = f ∇X V + (Xf )V (Leibniz rule).

Here, f is a smooth scalar function, and a and b are constants. In index-free notation, the Christoffel symbols do not appear explicitly.
Instead, the connection ∇ is defined abstractly as a map that takes two vector fields X and Y and returns another vector field ∇X Y . This
abstract map satisfies the linearity and Leibniz rules described above.

∂
Let {ei } be a local frame (or basis vector fields) on the manifold, such that ei = ∂x i in local coordinates. The Christoffel symbols de-
scribe how these basis vectors change with respect to each other under parallel transport, and in index-free notation, the covariant derivative
∇ei ej (in terms of the basis vectors) can be written as:

∇ei ej = Γkij ek
This relation defines the Christoffel symbols Γkij in terms of how the basis vector ej changes in the direction of the basis vector ei . In index-free
notation, we avoid explicitly writing the indices and instead describe the relationship in terms of the abstract covariant derivative and vector
fields. The Levi-Civita connection is the unique connection ∇ on a Riemannian manifold that satisfies two important properties:

41
 1. Metric compatibility: ∇g = 0, meaning that the covariant derivative of the metric tensor g is zero. In other words, the inner product
of vector fields is preserved under parallel transport.
2. Torsion-free condition: The connection is symmetric, meaning that for any two vector fields X and Y ,

∇X Y − ∇Y X = [X, Y ]

where [X, Y ] denotes the Lie bracket of the vector fields X and Y . This condition ensures that the connection is torsion-free.

In index-free notation, these two conditions provide an abstract way to define the connection without referring to the Christoffel symbols
explicitly. The connection is understood as a map that satisfies these two properties. The Christoffel symbols can be derived from the metric
compatibility condition in both indexed and index-free notation. In indexed notation, the condition ∇g = 0 is written as:

∇k gij = 0
Expanding this in terms of the Christoffel symbols gives:

∂k gij − Γlki glj − Γlkj gil = 0

This equation shows how the Christoffel symbols are related to the partial derivatives of the metric tensor.

In index-free notation, the same condition is expressed without reference to a coordinate system. The covariant derivative of the metric
tensor is written as:

(∇X g)(Y, Z) = 0
for any vector fields X, Y , and Z. This equation means that the inner product g(Y, Z) remains constant under parallel transport in the
direction of X. Using this condition, we can derive the Christoffel symbols by working with the abstract covariant derivative and the metric
in an intrinsic manner, but without introducing explicit coordinate indices.

In indexed notation, the Christoffel symbols explicitly describe the connection coefficients in terms of the coordinates. However, in index-free
notation, the connection itself is viewed as an abstract operator that satisfies the properties of linearity, the Leibniz rule, and metric com-
patibility, without direct reference to the coordinate system. To see the relationship between these two notations, we observe that in local
coordinates, the covariant derivative ∇X Y can be written as:
  ∂
∇X Y = X i ∂i Y j + X i Γjik Y k
∂xj
This is the indexed version of the covariant derivative. In index-free notation, we express this in terms of abstract vector fields and operators:

∇X Y = abstract connection operator acting on X and Y

The relationship between the two notations is that the Christoffel symbols Γjik in indexed notation provide the components of the abstract
connection ∇ in a local coordinate system. The index-free notation is more geometric and abstract, emphasizing the intrinsic properties of
the connection, while the indexed notation provides a detailed, coordinate-based description.

In many geometric applications, especially in differential geometry and general relativity, index-free notation is preferred because it ab-
stracts away the dependence on specific coordinates and emphasizes the geometric properties of objects like vectors, tensors, and connections.
This allows us to express geometric concepts such as curvature and geodesics in a coordinate-independent manner. For example, the Riemann
curvature tensor can be written in index-free notation as:

R(X, Y )Z = ∇X ∇Y Z − ∇Y ∇X Z − ∇[X,Y ] Z
This expression highlights the curvature as a map that depends on vector fields X, Y , and Z, without requiring explicit coordinate indices.

42
19.2.2 Christoffel Symbols In Earth Surface Coordinates
Given a spherical coordinate system, which describes points on the Earth surface (approximated as an ideal sphere),

x(R, θ, φ) = (R cos θ cos φ, R cos θ sin φ, R sin θ)

For a point x, R is the distance to the Earth’s core (usually approximately the Earth radius). θ and φ are the latitude and longitude. Positive
θ is the northern hemisphere. To simplify the derivatives, the angles are given in radians (where d sin(x)/dx = cos(x), the degree values
introduce an additional factor of 360/2π).

At any location, the tangent directions are eR (up), eθ (north), and eφ (east) - you can also use indices 1, 2, 3.

eR = (cos θ cos φ, cos θ sin φ, sin θ)

eθ = R · (− sin θ cos φ, − sin θ sin φ, cos θ)
eφ = R · (− cos θ sin φ, cos θ cos φ, 0)
The related metric tensor has only diagonal elements (the squared vector lengths). This is an advantage of the coordinate system and not
generally true.

gRR = 1 gθθ = R2 gφφ = R2 cos2 θ

gij = 0 else
1 1
g RR = 1 g θθ = g φφ =
R2 R2 cos2 θ
g ij = 0 else
Now the necessary quantities can be calculated. Examples:

eR = g RR eR · eR = (cos θ cos φ, cos θ sin φ, sin θ)

∂
Γφ φ
φR = e · eφ = eφ · (−R cos θ cos φ, −R cos θ sin φ, 0) = −R cos2 θ
∂R
∂e
The resulting Christoffel symbols of the second kind Γkji = ek · ∂xji then are (organized by the ”derivative” index i in a matrix):
 R
ΓRR ΓR ΓR
  
θR φR 0 0 0
 ΓθRR ΓθθR ΓθφR  =  0 0 1 
R
ΓφRR Γφ
θR Γφ
φR 0 R1 0
 R
ΓRθ ΓR ΓR 0 − R1
  
θθ φθ 0
 ΓθRθ Γθθθ Γθφθ  =  1 0 − tan θ 
R
ΓφRθ Γ φ
θθ Γ φ
φθ 0 0 0
 R
ΓRφ ΓR R 
0 0 −R cos2 θ
 
θφ Γφφ
 ΓθRφ Γθθφ Γθφφ  =  0 0 − tan θ 
φ φ φ 1 1
ΓRφ Γθφ Γφφ R R 0
These values show how the tangent directions (columns: eR , eθ , eφ ) change, seen from an outside perspective (e.g., from space), but given in
the tangent directions of the actual location (rows: R, θ, φ).

43
19.3 Covariant Derivative of Tensors
∂
We begin with the simplest case: the covariant derivative of a vector field. Let V = V i ∂x i be a vector field on a Riemannian manifold (M, g).
∂
The covariant derivative of the vector field V in the direction of another vector field X = X j ∂x j is denoted by ∇X V , and is given in local

coordinates as:

∂
∇X V = X j ∂j V i + X j Γijk V k
∂xi
where Γijk are the Christoffel symbols of the Levi-Civita connection, and ∂j V i represents the partial derivative of the component V i of the
vector field. This expression can be understood as a directional derivative that is adjusted by the Christoffel symbols to account for the
curvature of the manifold.

Next, consider a covector field (also known as a 1-form) ω = ωi dxi . The covariant derivative of a covector field is slightly different be-
cause the Christoffel symbols must be applied in a way that respects the transformation properties of covectors. The covariant derivative of
∂
the covector field ω in the direction of a vector field X = X j ∂x j is denoted by ∇X ω, and in local coordinates is given by:

∇X ω = X j ∂j ωi − X j Γkji ωk dxi

Here:
• ∂j ωi is the partial derivative of the component ωi of the covector field.
• The Christoffel symbols Γkji account for the change in the basis for the covector components.
The minus sign in the second term arises because covectors transform contragrediently to vectors.

Let’s now extend the notion of the covariant derivative to a general tensor field. A tensor field T of type (r, s) has r contravariant in-
dices (upper indices) and s covariant indices (lower indices). The components of such a tensor are denoted by Tji11ji22...j
...ir
s
, and it can be written
as:
∂ ∂
T = Tji11ji22...j
...ir
⊗ ⊗ · · · ⊗ dxj1 ⊗ dxj2 ⊗ · · ·
s
∂xi1 ∂xi2
The covariant derivative of the tensor T in the direction of a vector field X = X k ∂x∂ k is denoted ∇X T , and in local coordinates, the covariant
derivative is defined as:

∇X T = X k ∇k T
∂
where ∇k is the covariant derivative with respect to the coordinate direction ∂xk
. The components of the covariant derivative of T are given
by:
r
X s
X
∇k Tji11ji22...j
...ir
s
= ∂k Tji11ji22...j
...ir
s
+ Γiklm Tji11j...l...i
2 ...js
r
− Γlkjn Tji11...l...j
i2 ...ir
s
m=1 n=1
Pr
∂k Tji11ji22...j
...ir
is the ordinary partial derivative of the tensor components. The first summation
s
im i1 ...l...ir
m=1 Γkl Tj1 j2 ...js involves the Christoffel
symbols acting on the contravariant Pindices (upper indices). Each contravariant index contributes a positive term involving the Christoffel
s
symbol. The second summation − n=1 Γlkjn Tji11...l...j
i2 ...ir
s
involves the Christoffel symbols acting on the covariant indices (lower indices). Each
covariant index contributes a negative term involving the Christoffel symbol.

This formula ensures that the covariant derivative transforms correctly under coordinate changes, respecting the tensorial nature of T .
As an example, consider a rank-2 tensor field T i1 i2 , which is a tensor with two contravariant indices. The covariant derivative of this tensor
is given by:

44
 ∇k T i1 i2 = ∂k T i1 i2 + Γikl1 T li2 + Γikl2 T i1 l
Similarly, for a rank-2 covariant tensor Tj1 j2 , the covariant derivative is:

∇k Tj1 j2 = ∂k Tj1 j2 − Γlkj1 Tlj2 − Γlkj2 Tj1 l

These formulas are special cases of the general formula for the covariant derivative of a tensor field, with the Christoffel symbols acting
appropriately on each index.

The covariant derivative of a tensor field can be interpreted geometrically as a measure of how the tensor field changes as we move from one
point to another on the manifold, taking into account the manifold’s curvature. The Christoffel symbols adjust the ordinary derivative by
accounting for how the basis vectors and the metric change across the manifold. For example:
• In the case of a vector field, the covariant derivative measures how the vector changes as we move along the manifold while accounting
for the curvature.

• In the case of a covector field, the covariant derivative measures how the covector changes along the manifold, again incorporating the
manifold’s geometry.
The geometric importance of the covariant derivative lies in its ability to describe derivatives on curved spaces in a way that respects the
underlying geometry, which ordinary partial derivatives cannot do. Substituting equation (61) into equation (60), we get:

∂Ai
 i 
∂A k i ∂A i k
= ei + Γij A ek = + Γjk A ei , (75)
∂xj ∂xj ∂xj
where the names of the dummy indices i and k are swapped after the second equality.

Definition 4.3. The covariant derivative of a contravariant vector, Ai , is given by:

∂Ai
∇ j Ai ≡+ Γijk Ak , (76)
∂xj
where the adjective covariant refers to the fact that the index on the differentiation operator (j) is in the covariant (lower) position. Thus,
equation (75) becomes:
∂A
= ∇j Ai ei . (77)
∂xj
∂A
Thus, the i-th contravariant component of the vector ∂x j relative to the covariant basis ei is the covariant derivative of the i-th contravariant

component of the vector, A , with respect to the coordinate xj . In general, covariant derivatives are much more cumbersome than partial
i

derivatives as the covariant derivative of any one tensor component involves all tensor components for non-zero Christoffel symbols. Only for
Cartesian coordinates—where all Christoffel symbols are zero—do covariant derivatives reduce to ordinary partial derivatives.

Consider now the transformation of the covariant derivative of a contravariant vector from the coordinate system xi to x̃p . Thus, using
equations (59) and (74), we have:

∂ Ãp ∂xj ∂ x̃p ∂Ai ∂xj ∂ 2 x̃p i j k p

∂ x̃p ∂ 2 xi
  r
i ∂x ∂x ∂ x̃ ∂ x̃ l
∇q̃ Ãp = q
+ Γ̃ p
qr à r
= q i j
+ q j i
A + Γ jk q r i
+ i q r
A. (78)
∂ x̃ ∂ x̃ ∂x ∂x ∂ x̃ ∂x ∂x ∂ x̃ ∂ x̃ ∂x ∂x ∂ x̃ ∂ x̃ ∂xl
Now for some fun. Remembering we can swap the order of differentiation between linearly independent quantities, and exploiting our freedom
to rename dummy indices at whim, we rewrite the last term of equation (78) as:

∂ x̃p ∂ x̃r ∂
 i
∂ x̃p ∂
 i
∂ ∂ x̃p ∂xi ∂xi ∂
    p 
∂x l ∂x l ∂ x̃
i l r q
A = i l q
A = l i q
− q l
Al .
∂x ∂x ∂ x̃ ∂ x̃ ∂x ∂x ∂ x̃ ∂x ∂x ∂ x̃ ∂ x̃ ∂x ∂xi

45
Thus, equation (78) becomes:

∂xj ∂ x̃p
∇q̃ Ãp = ∇ j Ai ,
∂ x̃q ∂xi
and the covariant derivative of a contravariant vector is a mixed rank 2 tensor.

19.4 Applications
19.4.1 Application 1: Geodesic Equation
A geodesic is the shortest path between two points on a curved manifold, analogous to a straight line in flat Euclidean space. On a Rie-
mannian manifold, the equation that describes a geodesic is derived using the concept of parallel transport and involves the Christoffel symbols.

Let γ(t) be a curve on the manifold, parametrized by t, and let γ̇(t) represent the tangent vector to the curve at each point. The con-
dition that γ(t) is a geodesic is that the tangent vector γ̇(t) is parallel transported along the curve. This gives the geodesic equation:

∇γ̇(t) γ̇(t) = 0
In local coordinates, the components of the geodesic equation can be written as:

d2 γ i j
i dγ dγ
k
+ Γ jk =0
dt2 dt dt
This is the geodesic equation in terms of the Christoffel symbols Γijk , which account for the curvature of the manifold. Here:
d2 γ i
• dt2 represents the second derivative of the coordinates of the curve γ(t).

• Γijk are the Christoffel symbols that describe how the coordinate basis vectors change as we move along the curve.
Consider a curve γ(t) on the manifold, with tangent vector γ̇(t).The condition for γ(t) to be a geodesic is that the covariant derivative of γ̇(t)
along itself vanishes:
∇γ̇ γ̇ = 0
dγ i
In local coordinates, the tangent vector is γ̇ i = dt , and the covariant derivative of the tangent vector along itself is:

d2 γ i j
i dγ dγ
k
∇γ̇ γ̇ i = + Γ jk
dt2 dt dt
Setting this equal to zero gives the geodesic equation:
d2 γ i dγ j dγ k
2
+ Γijk =0
dt dt dt
Geodesics are the paths that particles or objects follow when moving under no external forces in curved spacetime, and the Christoffel symbols
account for the effects of curvature on these paths.

19.4.2 Application 2: Riemann Curvature Tensor

The Riemann curvature tensor is a fundamental object in differential geometry that measures the curvature of a Riemannian manifold. It
describes how the geometry of the manifold deviates from being flat and is defined in terms of the Christoffel symbols and their derivatives.

i
The Riemann curvature tensor Rjkl is given by:
i
Rjkl = ∂k Γijl − ∂l Γijk + Γikm Γm i m
jl − Γlm Γjk

46
This equation describes the curvature in terms of the change in the Christoffel symbols as we move across the manifold. ∂k Γijl − ∂l Γijk :
These terms represent how the Christoffel symbols vary from point to point on the manifold, which is related to how the geometry changes.
Γikm Γm i m
jl − Γlm Γjk : These terms account for how the Christoffel symbols interact with each other, capturing the nonlinearity of the space.
The Riemann curvature tensor measures the failure of parallel transport to be path-independent. Specifically, it quantifies how much a vector
changes when parallel transported around a small closed loop on the manifold. If the manifold is flat (like Euclidean space), the Riemann
curvature tensor vanishes.

19.4.3 Application 3: Ricci Tensor and Ricci Scalar

The Ricci tensor Rij is obtained by contracting the Riemann curvature tensor over one of its indices. It is defined as:
k
Rij = Rikj
The Ricci tensor is a trace of the Riemann curvature tensor and describes how the volume of a small geodesic ball deviates from the volume
of a similar ball in flat space due to curvature. The Ricci scalar R is obtained by further contracting the Ricci tensor:

R = g ij Rij
The Ricci scalar provides a single number at each point on the manifold that describes the degree to which the manifold is curved. Start with
i
the Riemann curvature tensor Rjkl . Contract over the first and third indices to obtain the Ricci tensor:
k
Rij = Rikj = ∂k Γkij − ∂j Γkik + Γkkl Γlij − Γkjl Γlik
Contract the Ricci tensor with the metric g ij to get the Ricci scalar:
R = g ij Rij
The Ricci tensor and Ricci scalar are important in general relativity, as they appear in the Einstein field equations that describe the relationship
between the curvature of spacetime and the distribution of matter and energy. The Einstein field equations are the fundamental equations of
general relativity, describing how matter and energy influence the curvature of spacetime. They are given by:
1
Rij − gij R = 8πGTij
2
where: Rij is the Ricci tensor, R is the Ricci scalar, gij is the metric tensor, Tij is the stress-energy tensor, which describes the distribution
of matter and energy in spacetime, G is the gravitational constant.

The Einstein field equations relate the geometry of spacetime (represented by the Ricci tensor, Ricci scalar, and metric) to the matter
and energy content of spacetime (represented by the stress-energy tensor). The term Rij − 21 gij R describes the curvature of spacetime, and
the right-hand side 8πGTij describes the matter and energy that cause this curvature. The equations are derived by combining the geometric
properties of the Ricci tensor and scalar with physical principles such as the conservation of energy and momentum.
In this extremely rigorous derivation, we have covered the following key points:
• The geodesic equation, which describes the motion of particles in curved space and involves the Christoffel symbols to account for
curvature.
• The Riemann curvature tensor, which measures the intrinsic curvature of a manifold and is expressed in terms of the Christoffel symbols
and their derivatives.
• The Ricci tensor and Ricci scalar, which are contractions of the Riemann curvature tensor and describe aspects of curvature related to
volume distortion.
• The Einstein field equations, which link the geometry of spacetime (curvature) to the matter and energy content through the stress-energy
tensor.
These applications of the Christoffel symbols are foundational to the study of Riemannian geometry and general relativity. They provide the
framework for understanding the relationship between curvature, geometry, and physical phenomena in both mathematics and physics.

47
20 References
1. Borisenko, A. I. (1968). Vector and tensor analysis with applications. Courier Corporation.
2. Salençon, J. (2012). Handbook of continuum mechanics: General concepts thermoelasticity. Springer Science and Business Media.

3. Lebedev, L. P., Cloud, M. J., and Eremeyev, V. A. (2010). Tensor analysis with applications in mechanics. World Scientific.
4. Basar, Y., and Weichert, D. (2013). Nonlinear continuum mechanics of solids: fundamental mathematical and physical concepts.
Springer Science and Business Media.
5. Papastavridis, J. G. (2018). Tensor calculus and analytical dynamics. Routledge.

6. Sahraee, S., and Wriggers, P. (2023). Representation of Tensorial Variables in Curvilinear Coordinates. In Tensor Calculus and
Differential Geometry for Engineers: With Solved Exercises (pp. 169-207). Cham: Springer Nature Switzerland.
7. Naghdi, P. M., and Wainwright, W. L. (1961). On the time derivative of tensors in mechanics of continua. Quarterly of Applied
Mathematics, 19(2), 95-109.
8. Shapiro, I. L. (2019). Derivatives of Tensors, Covariant Derivatives. In A Primer in Tensor Analysis and Relativity (pp. 55-65). Cham:
Springer International Publishing.
9. Langlois, W. E., Deville, M. O., Langlois, W. E., and Deville, M. O. (2014). Curvilinear Coordinates. Slow Viscous Flow, 81-104.
10. Grinfeld, P. (2013). Introduction to tensor analysis and the calculus of moving surfaces (pp. 135-36). New York: Springer.
11. Bishop, R. L., and Goldberg, S. I. (2012). Tensor analysis on manifolds. Courier Corporation.

12. Do Carmo, M. P. (2016). Differential geometry of curves and surfaces: revised and updated second edition. Courier Dover Publications.
13. Flügge, W. (1972). Tensor analysis and continuum mechanics (p. 153). New York: Springer.
14. Willmore, T. J. (2013). An introduction to differential geometry. Courier Corporation.

15. Boehler, J. P. (1987). Applications of tensor functions in solid mechanics (Vol. 292, pp. 13-30). J. P. Boehler (Ed.). New York:
Springer.
16. Synge, J. L., and Schild, A. (1978). Tensor calculus (Vol. 5). Courier Corporation.
17. Guggenheimer, H. W. (2012). Differential geometry. Courier Corporation.

18. Kobayashi, S., and Nomizu, K. (1996). Foundations of differential geometry, volume 2 (Vol. 61). John Wiley and Sons.
19. Lee, J. M., and Lee, J. M. (2012). Smooth manifolds (pp. 1-31). Springer New York.
20. Wald, R. M. (2010). General relativity. University of Chicago press.

48