[Link].
Physics (Honours / Gen) Mathematical Physics
Topic: Tensors & Rotation Tensor Classroom Handout
UNDERSTANDING TENSORS & ROTATION
TENSORS
A Simple, Intuitive Guide Designed Specially for [Link]. Students
1. Introduction: Why Do We Need Tensors?
Until now, you have learned about two major physical quantities:
1. Scalars: Quantities that have only magnitude (value) and no direction.
Examples: Mass (m), Temperature (T), Time (t).
2. Vectors: Quantities that have both magnitude and one specific direction.
Examples: Velocity (v), Force (F), Electric Field (E).
In vectors, when you apply a cause (like a Force) in one direction, the effect (like acceleration) happens in that
same direction (F = ma). But nature is not always this simple!
Real-World Problem: Anisotropic Medium
Imagine a crystal (like quartz). If you apply an Electric Field (E) along the x-axis, the crystal gets polarized
not just along x, but also slightly along the y and z axes! Here, the cause is in one direction, but the effect
spreads across multiple directions. A single vector cannot link this cause and effect. We need a higher tool: A
Tensor.
2. What Exactly is a Tensor? (Rank of a Tensor)
A tensor is a mathematical object that generalizes the concepts of scalars and vectors. Tensors are classified by their
Rank (or Order), which tells you how many directional components or indices it requires:
S2R Physics Family • Class Notes Page 1 of 4
Tensor Number of Dimensions / Components in 3D Space
Common Name
Rank Directions (3Rank)
Rank 0 Scalar 0 Directions 30 = 1 component
Rank 1 Vector 1 Direction 31 = 3 components (x, y, z)
Dyad / Matrix
Rank 2 2 Directions combined 32 = 9 components
Tensor
Key Concept: A Rank-2 tensor relates two directions simultaneously. For example, it can map a force vector
acting on a specific surface area (Stress Tensor), where one direction belongs to the force and the second
direction belongs to the perpendicular normal of the surface area.
3. Mathematical Notation of Tensors
Instead of writing components as x, y, z, we use indices 1, 2, 3. So, x → 1, y → 2, and z → 3.
• A vector A is written as Ai (where i = 1, 2, 3). This represents three components: A1, A2, A3.
• A Rank-2 tensor is written with two indices: Tij (where both i and j can be 1, 2, 3).
This gives us 3 × 3 = 9 distinct components, which can be elegantly arranged into a square matrix format:
T11 T12 T13
Tij =
[ T21 T22 T23 ]
T31 T32 T33
Here, the first index (i) represents the row, and the second index (j) represents the column.
4. The Concept of Rotation in Physics
Suppose you have a point P in space. Its coordinates in your original frame of reference (S frame with axes x1, x2,
x3) are given by a vector x. Now, let's rotate the coordinate axes about the origin to form a new frame of reference
(S' frame with axes x'1, x'2, x'3).
The physical point P hasn't moved, but its address (coordinates) will change because our viewing perspective has
rotated. Let its new coordinates be x'.
How do we connect the old vector components to the new vector components? We use a linear transformation
operator called the Rotation Tensor (often denoted by symbol R or A).
S2R Physics Family • Class Notes Page 2 of 4
Vector in new frame: x' = R · x ⇒ In component form: x'i = Σj=13 Rij xj
5. Derivation of the Rotation Tensor (2D Case for Simplicity)
Let's derive the rotation tensor components in 2-dimensions (rotation by an angle θ in the xy-plane counter-
clockwise) so that it is absolutely easy to follow.
Let the old axes be x1, x2 with unit vectors i1, i2. Let the new axes be x'1, x'2 with unit vectors i'1, i'2.
By resolving the new unit vectors along the old directions using basic trigonometry:
i'1 = cosθ i1 + sinθ i2
i'2 = -sinθ i1 + cosθ i2
Any arbitrary position vector r can be expressed in both systems as:
r = x1 i1 + x2 i2 = x'1 i'1 + x'2 i'2
To find the new coordinate x'1, we take the dot product of r with i'1:
x'1 = r · i'1 = (x1 i1 + x2 i2) · (cosθ i1 + sinθ i2)
x'1 = x1 cosθ + x2 sinθ
Similarly, for x'2, taking dot product with i'2:
x'2 = r · i'2 = (x1 i1 + x2 i2) · (-sinθ i1 + cosθ i2)
x'2 = -x1 sinθ + x2 cosθ
Let's write these transformation equations together side-by-side:
x'1 = (cosθ) x1 + (sinθ) x2
x'2 = (-sinθ) x1 + (cosθ) x2
Now, let's put this into a neat matrix format. This matrix represents our 2D Rotation Tensor:
x'1 cosθ sinθ x1
[ x' ] [ =
-sinθ cosθ
][ x ]
2 2
The core matrix is our rotation tensor Rij:
R11 R12 cosθ sinθ
R= [ R21 R22
]=[ -sinθ cosθ
]
S2R Physics Family • Class Notes Page 3 of 4
6. General 3D Rotation Tensor
In 3D space, if we rotate the axes around the z-axis (or x3-axis) by an angle θ, the z coordinate stays identical (x'3
= x3). The full 3D Rotation Tensor matrix looks like this:
cosθ sinθ 0
Rz(θ) =
[ -sinθ cosθ 0 ]
0 0 1
7. Essential Properties of Rotation Tensors (Exam Oriented)
These properties are frequently asked in [Link]. university examinations as short questions or proofs:
1. Orthogonal Property: A rotation tensor is an orthogonal matrix. This means multiplying the tensor by its
transpose (RT) gives the Identity matrix (I).
R · RT = I or RT = R-1
2. Determinant is Unity: The determinant of any pure rotation tensor is always equal to positive one.
det(R) = |R| = +1
Note: If |R| = -1, it implies an inversion or reflection of axes, not a pure physical rotation.
3. Preservation of Length: Rotating axes doesn't shrink or stretch vectors. The magnitude of the vector remains
invariant under rotation transformation.
(x'1)2 + (x'2)2 + (x'3)2 = (x1)2 + (x2)2 + (x3)2
Summary Formula Tip for Quick Revision:
Any quantity that transforms according to the rule A'i = Rij Aj is a tensor of rank 1 (Vector).
Any quantity that transforms according to the rule T'ij = Rim Rjn Tmn is a tensor of rank 2.
S2R Physics Family • Class Notes Page 4 of 4