COMPREHENSIVE TENSOR

ANALYSIS
Step-by-Step Mathematical Derivations & Explanations

Course Code / Level: [Link]. Physics (Mathematical Physics)

Target Audience: Designed specially for clear foundational understanding

and simplified step-by-step master derivations.

Core Focus: 10 Mandatory University Examination Derivations with

full algebraic expansions.

S2R PHYSICS FAMILY

Empowering Physics Education & Conceptual Depth

S2R Physics Family • Classroom Notes Page 1 of 22

Prerequisites & Concept Boosters

Before jumping directly into derivations, let us understand 3 golden rules of tensor algebra in an extremely easy
format. If you master these 3 concepts, every single proof in this document will become effortless!

Rule 1: Einstein Summation Convention

Whenever an index is repeated twice in a single term (one as a subscript/lower index, and one as a
superscript/upper index), it implies a automatic summation over all possible values of that index (usually 1, 2,
3 in 3D space).

Ai Bi = Σi=13 Ai Bi = A1B1 + A2B2 + A3B3

A repeated index is called a Dummy Index. You can change its letter freely (e.g., AiBi = AkBk). An index
appearing only once is called a Free Index, and its letter cannot be changed arbitrarily without changing it on
both sides of an equation.

Rule 2: The Kronecker Delta (δij or δij)

Think of it as a substitution machine or a switch:

• If the two indices are equal (i = j), its value is 1.

• If the two indices are different (i ≠ j), its value is 0.

The Substitution Property: When δij multiplies a vector component Aj, it changes the index j into i.

δij Aj = Ai (because only the term where j = i survives!).

Rule 3: Alternating / Levi-Civita Tensor (εijk)

This is a 3-index symbol used to represent cross products and determinants:

• ε123 = ε231 = ε312 = +1 (Cyclic order: 1 → 2 → 3 → 1)

• ε132 = ε321 = ε213 = -1 (Anti-cyclic / reversed order)

• εijk = 0 if ANY two indices are equal (e.g., ε112 = 0).

S2R Physics Family • Classroom Notes Page 2 of 22

1. Relation Between Alternating Tensor (ε) and Kronecker Delta
(δ)

This is the most important identity in tensor analysis. It connects the 3-index alternating symbol with 2-index
Kronecker deltas. It is widely used to solve vector identity proofs like ∇ × (∇ × A).

The Master Formula

εijk εimn = δjm δkn − δjn δkm

Trick to remember: Notice that the first index i is repeated (dummy index). For the remaining indices, it follows
the inner-outer product rule: (First with First) × (Second with Second) MINUS (First with Second) × (Second with
First).

Step-by-Step Analytical Proof

Let us prove this identity by testing different possible configurations of the free indices j, k, m, n. Since these
indices take values from {1, 2, 3}, we check the main logical cases:

Case I: Let j = m and k = n, with j ≠ k.

Let us assign real numbers to understand easily: Let j = 1, k = 2. This means m = 1, n = 2.
Substitute these into Left Hand Side (LHS):

LHS = εi12 εi12 = Σi=13 (εi12)2

Since indices must be distinct for a non-zero value, i can only be equal to 3.
Therefore, LHS = (ε312)2 = (+1)2 = 1.

Now, check Right Hand Side (RHS) using the definition of δ:

RHS = δ11 δ22 − δ12 δ21 = (1 × 1) − (0 × 0) = 1 − 0 = 1

Thus, LHS = RHS = 1.

S2R Physics Family • Classroom Notes Page 3 of 22

 Case II: Let j = n and k = m, with j ≠ k.
Let us assign numbers: Let j = 1, k = 2, which means n = 1, m = 2.
Substitute into LHS:

LHS = εi12 εi21

Again, i must be 3 to avoid zero values:

LHS = ε312 ε321 = (+1) × (−1) = −1

Now, evaluate the RHS with these values:

RHS = δ12 δ21 − δ11 δ22 = (0 × 0) − (1 × 1) = −1

Thus, LHS = RHS = −1.

Case III: If any two indices on the same tensor match (e.g., j = k).
If j = k, then εijk = 0, making LHS = 0.
On the RHS, if j = k, the expression becomes δjm δjn − δjn δjm = 0. Thus, LHS = RHS = 0 holds perfectly
for all permutations.

Hence, the epsilon-delta relation is universally proven. ■

S2R Physics Family • Classroom Notes Page 4 of 22

2. Prove Kronecker Delta (δij) is a Rank-2 Mixed/Covariant
Tensor

To prove any quantity is a tensor of a specific rank, we must show that its components transform from an old
coordinate system (x) to a new primed coordinate system (x') according to the standard tensor transformation laws.

Transformation Law Expectation

For a rank-2 tensor, we require two transformation coefficients (partial derivative terms):

δ'μν = δij Aiμ Ajν = δij (∂xi / ∂x'μ) (∂xj / ∂x'ν)

Step-by-Step Derivation

Step 1: Start with the partial derivative definition.

By using the standard multi-variable chain rule of partial differentiation, we can write the derivative of a
primed coordinate with respect to another primed coordinate as:

∂x'μ / ∂x'ν = δμν

(Since if μ = ν, ∂x'μ / ∂x'μ = 1, and if μ ≠ ν, they are independent coordinates, so the derivative is 0).

Step 2: Introduce unprimed coordinates using the chain rule.

Let us insert the unprimed intermediate coordinate system variables xi and xj into the differentiation chain:

∂x'μ / ∂x'ν = (∂x'μ / ∂xi) · (∂xi / ∂x'ν)

S2R Physics Family • Classroom Notes Page 5 of 22

 Step 3: Insert the Kronecker delta identity artificially.
We know that ∂xi / ∂xj = δij. Let us utilize this to structure our expression symmetrically:

Using the property that δij (∂xj / ∂x'ν) = ∂xi / ∂x'ν, we can substitute this back into equation (2.2):

δμν = (∂x'μ / ∂xi) (∂xj / ∂x'ν) δij

Conclusion

Look carefully at equation (2.3). It states that the components in the primed system (δμν) are obtained by

multiplying the components in the unprimed system (δij) by two transformation matrices—one contravariant index

rule (∂x'μ/∂xi) and one covariant index rule (∂xj/∂x'ν).

This is exactly the transformation rule for a Mixed Tensor of Rank 2 (1 covariant and 1 contravariant index).
Hence, Kronecker Delta is a tensor of rank 2. ■

S2R Physics Family • Classroom Notes Page 6 of 22

3. Prove Alternating / Levi-Civita Tensor (εijk) is a Rank-3 Tensor

To prove that the alternating symbol εijk transforms as a covariant tensor of rank 3, we must establish that under
coordinate transformation:

ε'μνρ = (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ) εijk

Step-by-Step Mathematical Proof

Step 1: Recall the definition of a Matrix Determinant.

Let J be a transformation matrix whose elements are given by Jiμ = ∂xi / ∂x'μ. By definition of determinants
using the permutation symbol, we write:

|J| εμνρ = εijk (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ)

Where |J| is the Jacobian determinant of the transformation matrix.

Step 2: Restrict to Orthogonal Transformations.

In standard physical systems and Cartesian/orthogonal spaces, the coordinate transformation preserves length
scales, meaning the Jacobian determinant satisfies:

|J| = 1

(For proper rotations/orthogonal frameworks, the determinant of the transformation matrix is always unity).

S2R Physics Family • Classroom Notes Page 7 of 22

 Step 3: Direct substitution.
Substitute |J| = 1 directly into our structural definition in Equation (3.1):

εμνρ = (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ) εijk

Since the components are identical in numerical value across reference systems (ε'μνρ = εμνρ), we can
explicitly re-write this as:

ε'μνρ = (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ) εijk

Conclusion

The expression matches the standard coordinate transformation equation requiring three distinct partial derivative
coefficients for lower indices. This establishes definitively that εijk is an isotropic Covariant Tensor of Rank 3. ■

S2R Physics Family • Classroom Notes Page 8 of 22

4. The Quotient Law of Tensors

Testing if an unknown multi-indexed quantity is a tensor using the fundamental definition can sometimes be highly
tedious. The Quotient Law is a shortcut rule: If an inner product of an unknown set of components with an
arbitrary known tensor always results in a known tensor, then the unknown components form a tensor.

Statement of the Problem

Let A(i,j,k) be a set of quantities. Suppose we are given that for an arbitrary contravariant vector Xk, the
combination:

A(i,j,k) Xk = Bij

where Bij is known to be a covariant tensor of rank 2. We must prove that A(i,j,k) must be a covariant tensor of
rank 3, i.e., Aijk.

Detailed Step-by-Step Derivation

Step 1: Write down transformations of the known quantities.

Since Bij is a rank-2 covariant tensor, its expression in a primed coordinate frame is:

B'μν = (∂xi / ∂x'μ) (∂xj / ∂x'ν) Bij

We also know that in the primed frame, the relation holds true as:

B'μν = A'(μ,ν,ρ) X'ρ

Step 2: Equate the two expressions and substitute.

Equating equations (4.1) and (4.2) gives:

A'(μ,ν,ρ) X'ρ = (∂xi / ∂x'μ) (∂xj / ∂x'ν) Bij

Since Bij = A(i,j,k) Xk, substitute this directly into the right-hand side:

S2R Physics Family • Classroom Notes Page 9 of 22

 A'(μ,ν,ρ) X'ρ = (∂xi / ∂x'μ) (∂xj / ∂x'ν) A(i,j,k) Xk

Step 3: Convert the arbitrary vector Xk into the primed frame.

Using the standard contravariant transformation vector rule: Xk = (∂xk / ∂x'ρ) X'ρ. Plug this into equation
(4.3):

A'(μ,ν,ρ) X'ρ = (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ) A(i,j,k) X'ρ

Bring all terms to the left hand side and factor out the common vector X'ρ:

[ A'(μ,ν,ρ) − (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ) A(i,j,k) ] X'ρ = 0

Step 4: Use the property of arbitrariness.

Since Xk (and hence X'ρ) is completely arbitrary, it can be non-zero. For the equation to hold true for any
choice of vector components, the bracketed quantity itself must be equal to zero:

A'(μ,ν,ρ) = (∂xi / ∂x'μ) (∂xj / ∂x'ν) (∂xk / ∂x'ρ) A(i,j,k)

This is exactly the transformation matrix requirement for a rank-3 covariant tensor. Hence, A(i,j,k) = Aijk is proven
to be a true tensor. ■

S2R Physics Family • Classroom Notes Page 10 of 22

5. Proof that the Stress Tensor (σij) is Symmetric

The state of stress at a point in a continuous deformable body is described by a rank-2 tensor σij. To prove it is
symmetric, we must show that σij = σji based on the physical law of conservation of angular momentum.

Step-by-Step Physical Derivation

Step 1: Set up the Total Torque Equation.

Consider a small volume element V bounded by a surface S. In the absence of distributed internal body
torques, the total torque acting on the body must equal zero under static equilibrium. Torque is given by the
cross product of position vector and force:

τi = ∫S εijk xj Tk dS + ∫V εijk xj fk dV = 0

Where:
• Tk is the surface traction force vector acting on surface element dS.

• fk is the body force vector per unit volume acting inside dV.

Step 2: Express Traction in terms of Stress Tensor components.

By Cauchy's Stress Principle, the traction vector is related to the stress tensor through the surface normal unit
vector nl: Tk = σlk nl. Substitute this into the surface integral:

∫S εijk xj σlk nl dS + ∫V εijk xj fk dV = 0

Step 3: Apply Gauss Divergence Theorem.

Convert the surface integral into a volume integral using the Gauss Divergence Theorem: ∫S (...) nl dS = ∫V
∂l(...) dV.

∂l ( εijk xj σlk ) = εijk [ (∂xj / ∂xl) σlk + xj ∂l(σlk) ]

We know that ∂xj / ∂xl = δjl. Thus, the expression becomes:

S2R Physics Family • Classroom Notes Page 11 of 22

 = εijk δjl σlk + εijk xj ∂l σlk = εijk σjk + εijk xj (∂l σlk)

Step 4: Combine into a single Volume Integral.

Substitute this expanded derivative format back into Equation (5.2):

∫V [ εijk σjk + εijk xj (∂l σlk + fk) ] dV = 0

From linear momentum equilibrium (Navier's equation), we know that ∂l σlk + fk = 0. Therefore, the second
term in the bracket vanishes entirely! We are left with:

∫V εijk σjk dV = 0 ⇒ εijk σjk = 0

Step 5: Expand the remaining terms explicitly.

Let's take i = 1 as a clear example:

ε123 σ23 + ε132 σ32 = 0 ⇒ (1)σ23 + (−1)σ32 = 0 ⇒ σ23 = σ32

Repeating for all coordinates yields σij = σji. Thus, the stress tensor is proved to be symmetric. ■

S2R Physics Family • Classroom Notes Page 12 of 22

6. Stress-Strain Generalized Hooke's Law Relation

In linear elasticity, Generalized Hooke's Law states that each component of stress is linearly proportional to every
component of strain. Since both stress (σij) and strain (ekl) are rank-2 tensors, the proportionality constant
connecting them must be a rank-4 tensor, denoted by Cijkl.

The Structural Equation

σij = Cijkl ekl

Reduction of Independent Elastic Constants (Step-by-Step Explanation)

Initially, since i, j, k, l can each take 3 values, Cijkl has 3 × 3 × 3 × 3 = 81 independent components. This is too
large for practical tracking. Let's simplify this step by step using physical tensor symmetries:

Reduction 1: Symmetry of the Stress Tensor (σij = σji)

Since the left-hand side of Equation (6.1) is symmetric with respect to swapping i and j, the right-hand side
components must mirror this symmetry:

Cijkl = Cjikl

This single constraint reduces the number of independent variables from 81 down to 54 constants.

Reduction 2: Symmetry of the Strain Tensor (ekl = elk)

Because the strain tensor components are inherently symmetric due to their definition in displacement fields,
swapping the inner indices k and l does not change the result:

Cijkl = Cijlk

This reduction brings the total number of independent constants down to 36 variables.

Reduction 3: Thermodynamic Potential / Strain Energy Density

S2R Physics Family • Classroom Notes Page 13 of 22

 From thermodynamics, elastic deformation energy density W can be formulated as a exact differential
quadratic form: W = ½ Cijkl eij ekl. Differentiating with respect to strain components shows that the order of
partial differentiation is commutative, leading to:

Cijkl = Cklij

This final symmetry requirement brings the 36 parameters down to exactly 21 independent constants for a
completely general anisotropic elastic material (like triclinic crystals).

Summary Note for Exams: Anisotropic material has 21 constants. If the material properties further exhibit
isotropic behavior (same properties in all directions), these 21 reduce dramatically to just 2 independent
parameters (Lamé's constants)!

S2R Physics Family • Classroom Notes Page 14 of 22

7. Derivation of Isotropic Form of Elastic Stiffness Tensor

An isotropic material has no preferred directional properties. Therefore, its elastic tensor Cijkl must be an isotropic
tensor. An isotropic tensor of rank 4 can only be expressed using linear combinations of products of Kronecker
deltas.

The Most General Fourth-Rank Isotropic Form

Cijkl = λ δij δkl + μ1 δik δjl + μ2 δil δjk

Where λ, μ1, μ2 are arbitrary scalar scalars.

Step-by-Step Derivation of Constitutive Equations

Step 1: Enforce the internal symmetry requirement.

We already know from Topic 6 that Cijkl = Cijlk (strain symmetry). Let us swap k and l in equation (7.1):

Cijlk = λ δij δlk + μ1 δil δjk + μ2 δik δjl

Comparing this term by term to equation (7.1), we can see that for the terms to match exactly, we must have:

μ1 = μ2 = μ

Thus, the equation simplifies with only two independent constants remaining:

Cijkl = λ δij δkl + μ ( δik δjl + δil δjk )

Here, λ and μ are known as the Lamé Constants of the material (where μ corresponds to the Shear Modulus).

Step 2: Derive the Stress-Strain Hooke's relation.

Substitute our simplified isotropic expression for Cijkl back into Hooke's Law equation (σij = Cijkl ekl):

σij = [ λ δij δkl + μ ( δik δjl + δil δjk ) ] ekl

Distribute the strain components across each Kronecker delta term individually:

S2R Physics Family • Classroom Notes Page 15 of 22

 σij = λ δij (δkl ekl) + μ (δik δjl ekl) + μ (δil δjk ekl)

Step 3: Simplify via the Index Substitution Trick.

Let's apply our Kronecker Delta substitution property here:
• δkl ekl = ekk (This represents the volumetric dilation trace, e11+e22+e33).

• For the second term, δik turns k into i, and then δjl turns l into j. Thus, δik δjl ekl = eij.

• For the third term, δil turns l into i, and δjk turns k into j. Thus, δil δjk ekl = eji = eij (since strain is
symmetric).

Combining these yields:

σij = λ δij ekk + μ eij + μ eij

Combining the identical shear terms results in the final form:

σij = λ ekk δij + 2μ eij

This is the final isotropic stress-strain relationship used everywhere in solid mechanics. ■

S2R Physics Family • Classroom Notes Page 16 of 22

8. Lorentz Transformation Equation in Matrix/Tensor Form

In Special Relativity, we represent a coordinate point in 4D spacetime as a position 4-vector, denoted by xμ = (x0,
x1, x2, x3) = (ct, x, y, z).

Standard Algebraic Form

For a reference frame moving at a constant speed v along the positive x-axis relative to a stationary frame, the
standard coordinate relationships are:

x' = γ(x − vt) ⇒ x'1 = γ(x1 − βx0)

y' = y ⇒ x'2 = x2
z' = z ⇒ x'3 = x3
ct' = γ(ct − βx) ⇒ x'0 = γ(x0 − βx1)

Where we define the velocity ratio β = v/c and the Lorentz scale factor γ = 1 / √(1 − β2).

Tensor/Matrix Representation

We can write this system of linear equations as a single tensor transformation equation:

x'μ = Λμν xν

Where Λμν represents the Lorentz transformation matrix. Writing out the full matrix system gives:

| x'0 | | γ -γβ 0 0 | | x0 |
| x'1 | = | -γβ γ 0 0 | × | x1 |
| x'2 | | 0 0 1 0 | | x2 |
| x'3 | | 0 0 0 1 | | x3 |

Explicit Element Index Breakdown

To ensure perfect clarity for the exam, let's explicitly list the individual component values of the tensor matrix Λμν
(where row index corresponds to μ and column index corresponds to ν):

S2R Physics Family • Classroom Notes Page 17 of 22

 Component Value Physical Meaning / Verification

Λ 00 γ Time dilation factor connecting ct' to ct

Λ 01 −γβ Relativistic cross-term between space position and new time coordinate

Λ 10 −γβ Spatial translation component dependent on frame elapsed time

Λ 11 γ Lorentz contraction parameter along the primary axis of travel

Λ 22 , Λ33 1 Perpendicular coordinates show zero change under axial boost

All other unlisted elements are equal to zero. This matrix framework forms the mathematical foundation of
relativistic mechanics. ■

S2R Physics Family • Classroom Notes Page 18 of 22

9. Invariance of the Spacetime Integral / Volume Element

We want to prove that the four-dimensional spacetime volume element integral, denoted by dV4 = dx0 dx1 dx2 dx3
= c dt dx dy dz, is an invariant quantity. This means its value remains identical across all inertial reference frames:
d4x' = d4x.

Step-by-Step Mathematical Derivation

Step 1: Express volume transformations using the Jacobian.

From standard multi-dimensional calculus, when we transform coordinates from an old system to a new
system, the differential volume components scale according to the absolute value of the Jacobian determinant:

dx'0 dx'1 dx'2 dx'3 = |J| dx0 dx1 dx2 dx3

Where |J| is the determinant of the partial derivative matrix mapping the transformation:

J = det( ∂x'μ / ∂xν ) = det( Λμν )

Step 2: Evaluate the determinant of the Lorentz Matrix.

Let us calculate the determinant of the transformation matrix Λ that we defined explicitly in Topic 8:

|J| = det Σ = γ [ γ(1) − 0 ] − (−γβ) [ −γβ(1) − 0 ] + 0

Simplifying this algebraically yields:

|J| = γ2 − γ2β2 = γ2 (1 − β2)

Step 3: Substitute the definition of the Lorentz scale factor.

Recall that by definition, γ = 1 / √(1 − β2), which means γ2 = 1 / (1 − β2). Let us substitute this back into our
expression for the determinant:

|J| = [ 1 / (1 − β2) ] × (1 − β2) = 1

S2R Physics Family • Classroom Notes Page 19 of 22

 Thus, the Jacobian determinant of a Lorentz boost is exactly equal to 1.

Step 4: Final Substitution.

Substitute |J| = 1 back into our volume scaling relationship in Equation (9.1):

dx'0 dx'1 dx'2 dx'3 = (1) · dx0 dx1 dx2 dx3

Integrating over the respective spacetime domains yields the final result:

∫ d4x' = ∫ d4x

This proves that the spacetime volume element is invariant under Lorentz transformations. This invariance is
crucial for formulating relativistic physical laws and action integrals in quantum field theory. ■

S2R Physics Family • Classroom Notes Page 20 of 22

10. Christoffel Symbols and Metric Tensor Relations

In General Relativity and curved spaces, the Christoffel symbols of the second kind, denoted by Γλμν, act as
gravitational force fields. They are not true tensors, but they are constructed directly from the derivatives of the
metric tensor gμν.

The Master Definition

Γλμν = ½ gλσ ( ∂ν gμσ + ∂μ gνσ − ∂σ gμν )

Where we use the short-hand comma/index notation for partial derivatives: ∂ν gμσ = ∂gμσ / ∂xν.

Step-by-Step Identity Derivations

Property A: Symmetry with respect to lower index configurations

Let us swap the positions of the lower indices μ and ν in the definition formula:

Γλνμ = ½ gλσ ( ∂μ gνσ + ∂ν gμσ − ∂σ gνμ )

Since the metric tensor is symmetric (gνμ = gμν), the terms inside the parentheses are identical to those in
Equation (10.1), just rearranged. Therefore:

Γλμν = Γλνμ

This proves that Christoffel symbols are always symmetric with respect to their lower indices in a torsion-free
space (Riemannian geometry).

Property B: Derivation of Metric Compatibility Constraint

Let us compute the total partial derivative of a metric component ∂gμν / ∂xρ using Christoffel symbols. Let's
add two cyclic permutations of the definition equation:

∂ρ gμν = gμσ Γσνρ + gνσ Γσμρ

S2R Physics Family • Classroom Notes Page 21 of 22

 Detailed Proof:
Let's substitute the definition of the Christoffel symbols into the right-hand side of the proposed identity:

gμσ Γσνρ = ½ gμσ gσα ( ∂ρ gνα + ∂ν gρα − ∂α gνρ )

Using the property that gμσ gσα = δμα, the index α is replaced by μ:

= ½ ( ∂ρ gνμ + ∂ν gρμ − ∂μ gνρ )

Now do the same for the second term:

gνσ Γσμρ = ½ ( ∂ρ gμν + ∂μ gρν − ∂ν gμρ )

Adding these two expressions together, we see that the terms containing ∂μ and ∂ν cancel out completely due
to opposite signs, leaving:

= ½ ∂ρ gνμ + ½ ∂ρ gμν = ∂ρ gμν

This matches the left-hand side perfectly! This property is equivalent to saying that the covariant derivative of
the metric tensor is zero (∇ρ gμν = 0), meaning the metric is compatible with the connection. ■

S2R Physics Family • Classroom Notes Page 22 of 22