Further Matrix Algebra (FP3)

Edexcel International A-Level Further Mathematics· FP3 Topic 6 (2018 Spec Issue 3)· 45 min read

1. 3D Linear Transformations & Combined Matrices★★★☆☆⏱ 10 min

✓ Calculator OK

📘 Definition

3D Linear Transformation

A transformation T of 3D column vectors such that $T (a x + b y) = a T (x) + b T (y)$ for all scalars $a, b$ and vectors $x, y$ , represented by a 3x3 matrix $M$ where $T (x) = M x$ .

This extends FP1 2D transformation work: each column of the 3x3 transformation matrix is the image of the standard basis vectors $i, j, k$ respectively. For combined transformations, the first transformation applied is the right-hand matrix in the product: performing transformation B then A is represented by $A B$ , per Edexcel specification.

📐 Worked Example

Let transformation B be a reflection in the xy-plane, and transformation A be a 90° anticlockwise rotation about the x-axis. Find the matrix representing B followed by A.

1
Matrix B (reflection in xy-plane) maps $i \to i, j \to j, k \to - k$ :
2
3
Matrix A (90° anticlockwise rotation about x-axis) maps $i \to i, j \to k, k \to - j$ :
4
5
Combined matrix for B then A is $A B$ :
6

Exam tip:

Always double-check the order of multiplication for combined transformations: the first transformation performed is the right-hand matrix in the product.

2. Transpose, 3x3 Determinants & Non-Singular Matrices★★★☆☆⏱ 10 min

✓ Calculator OK

📘 Definition

Matrix Transpose

$A^{T}$

The matrix formed by swapping the rows and columns of $A$ , so the entry at row $i$ , column $j$ of $A$ is at row $j$ , column $i$ of $A^{T}$ . The product transpose rule is $(A B)^{T} = B^{T} A^{T}$ .

To calculate a 3x3 determinant, expand along the row or column with the most zeros, applying the alternating cofactor sign matrix: $+ - + - + - + - +$ . A matrix is singular if its determinant is 0, meaning it has no inverse.

📐 Worked Example

Calculate the determinant of $A = 105216340$ and state if it is singular.

1
Expand along row 1 (has one zero to simplify calculation):
2
3
Compute each minor determinant:
4
5
$det (A) = 1 \neq = 0$ , so $A$ is non-singular.

Exam tip:

Always pick the row or column with the most zeros to expand the determinant, to minimize arithmetic errors.

3. 3x3 Matrix Inverses★★★★☆⏱ 12 min

✓ Calculator OK

📘 Definition

Adjugate Matrix

The transpose of the cofactor matrix of a square matrix, where cofactor $C_{ij} = (- 1)^{i + j} M_{ij}$ and $M_{ij}$ is the minor (determinant of the matrix with row $i$ , column $j$ removed). The inverse formula is $A^{- 1} = \frac{1}{d e t ( A )} adj (A)$ , and the product inverse rule is $(A B)^{- 1} = B^{- 1} A^{- 1}$ .

The inverse of a transformation matrix corresponds to the inverse transformation, which maps image vectors back to their original positions. Always confirm the determinant is non-zero before calculating an inverse, as singular matrices have no inverse.

📐 Worked Example

Find the inverse of $A = 105216340$ , for which $det (A) = 1$ .

1
Calculate the cofactor matrix using the alternating sign rule:
2
3
Transpose the cofactor matrix to get the adjugate:
4
5
Divide by $det (A) = 1$ to get the inverse:
6

Exam tip:

Verify your inverse calculation by checking that $A A^{- 1} = I$ , the identity matrix, to catch any transposition or arithmetic errors.

4. Eigenvalues & Eigenvectors★★★★☆⏱ 12 min

✓ Calculator OK

📘 Definition

Eigenvalue & Eigenvector

For a square matrix $A$ , an eigenvalue $λ$ is a scalar such that $A x = λ x$ for a non-zero vector $x$ , which is the corresponding eigenvector. Normalised eigenvectors have a magnitude of 1, calculated by dividing $x$ by $∣ x ∣$ .

Find eigenvalues by solving the characteristic equation $det (A - λ I) = 0$ . For each eigenvalue, solve $(A - λ I) x = 0$ to find the corresponding eigenvectors. For symmetric matrices, eigenvectors for distinct eigenvalues are always orthogonal.

📐 Worked Example

Find the eigenvalues and normalised eigenvectors of symmetric matrix $A = 101020101$ .

1
Set up and solve the characteristic equation:
2
3
Eigenvalues: $λ = 0$ , $λ = 2$ (repeated)
4
For $λ = 0$ , solve $A x = 0$ to get eigenvector $(1, 0, - 1)$ , normalise to $(\frac{1}{2}, 0, - \frac{1}{2})$
5
For $λ = 2$ , solve $(A - 2 I) x = 0$ to get orthogonal eigenvectors $(1, 0, 1)$ and $(0, 1, 0)$ , normalise to $(\frac{1}{2}, 0, \frac{1}{2})$ and $(0, 1, 0)$

Exam tip:

For repeated eigenvalues of symmetric matrices, select orthogonal eigenvectors by inspection to avoid extra orthogonalisation steps later.

5. Orthogonal Diagonalisation of Symmetric Matrices★★★★★⏱ 11 min

✓ Calculator OK

📘 Definition

Orthogonal Diagonalisation

For a symmetric matrix $A$ , the process of finding an orthogonal matrix $P$ such that $P^{T} A P = D$ , where $D$ is a diagonal matrix with the eigenvalues of $A$ on its leading diagonal.

To construct $P$ , use normalised, orthogonal eigenvectors of $A$ as its columns. Since $P$ is orthogonal, $P^{T} = P^{- 1}$ , so this process is equivalent to diagonalising $A$ using an orthonormal basis of eigenvectors.

📐 Worked Example

Find an orthogonal matrix $P$ such that $P^{T} A P$ is diagonal for $A = 101020101$ .

1
Use the normalised orthogonal eigenvectors from the previous example as columns of $P$ :
2
3
Confirm $P$ is orthogonal ( $P^{T} P = I$ ), then compute $P^{T} A P$ :
4

Exam tip:

The order of columns in $P$ corresponds to the order of eigenvalues on the leading diagonal of $D$ , so you can arrange eigenvalues in any order as long as the eigenvectors match.

6. Common Pitfalls

Wrong move:

Multiplying combined transformation matrices in order A then B for "A then B"

Why:

Edexcel specification defines the first transformation as the right-hand matrix in the product, so reverse order is required

Correct move:

Always write the first transformation as the rightmost matrix, multiply by the second transformation on the left

Wrong move:

Forgetting to transpose the cofactor matrix when calculating the adjugate for 3x3 inverses

Why:

The adjugate is the transpose of the cofactor matrix, not the cofactor matrix itself, so skipping this step gives an incorrect inverse

Correct move:

After calculating the cofactor matrix, swap its rows and columns to get the adjugate before dividing by the determinant

Wrong move:

Using the order $(A B)^{- 1} = A^{- 1} B^{- 1}$ or $(A B)^{T} = A^{T} B^{T}$ for products

Why:

Both inverse and transpose reverse the order of the matrix product, so these incorrect orders lead to wrong results

Correct move:

Always reverse the order of matrices when taking the transpose or inverse of a product: $(A B)^{T} = B^{T} A^{T}$ , $(A B)^{- 1} = B^{- 1} A^{- 1}$

Wrong move:

Failing to normalise eigenvectors when constructing an orthogonal matrix $P$ for diagonalisation

Why:

Orthogonal matrices require columns of magnitude 1, so unnormalised eigenvectors will make $P$ non-orthogonal, so $P^{T} A P$ will not be diagonal

Correct move:

Divide each eigenvector by its magnitude before using it as a column of $P$ when orthogonal diagonalisation is required

Wrong move:

Expanding a 3x3 determinant without applying the cofactor sign alternation

Why:

The cofactor sign (+/-) alternates with each row and column position, so omitting it gives an incorrect determinant value

Correct move:

Use the sign matrix $+ - + - + - + - +$ when expanding a 3x3 determinant along any row or column

7. Quick Reference Cheatsheet

Concept	Formula/Rule	Key Note
3D Transformation Matrix	Columns = images of $i, j, k$	B then A = $A B$ (right matrix first)
Transpose Product Rule	$(A B)^{T} = B^{T} A^{T}$	Swap rows and columns of $A$ to get $A^{T}$
3x3 Determinant	Expand along row/column with cofactor signs	$det (A) = 0$ → singular, no inverse
3x3 Inverse	$A^{- 1} = \frac{1}{d e t ( A )} \times adj (A)$	$(A B)^{- 1} = B^{- 1} A^{- 1}$
Eigenvalues	Solve $det (A - λ I) = 0$	$λ$ solves $A x = λ x$
Eigenvectors	Solve $(A - λ I) x = 0$	Normalise: divide by $∣ x ∣$ for orthonormal vectors
Orthogonal Diagonalisation	$P^{T} A P = D$ , $P$ orthogonal	$D$ has eigenvalues of $A$ on leading diagonal

8. Frequently Asked

Do I need to memorise 3x3 determinant and inverse formulas for FP3?

Yes, no 3x3 matrix operation formulas are provided in the FP3 formula book. You must recall the cofactor/adjugate inverse method and 3x3 determinant calculation procedure.

Why do I multiply matrices in reverse order for combined transformations?

The Edexcel specification defines that transformation B followed by transformation A is represented by the product matrix AB, as column vectors are multiplied on the left by the transformation matrix.

When do I need to normalise eigenvectors for diagonalisation?

For symmetric matrices, if you are asked for an orthogonal matrix P such that PᵀAP is diagonal, you must normalise each eigenvector to form the columns of P to ensure P is orthogonal.

Going deeper

official_specEdexcel IAL Further Maths FP3 SpecificationRefer to page 67 for FP3 matrix algebra outcomes
prereq_guideFP1 Matrix Algebra (Edexcel IAL)Recap 2x2 matrix arithmetic and 2D transformations before starting this guide

What's Next

Now that you have mastered further matrix algebra for FP3, you can apply these concepts to other FP3 topics including linear systems of equations and 3D geometry, where matrices are used to represent planes and intersections. Eigenvalues and diagonalisation are also foundational for university-level linear algebra, engineering, and data science if you plan to pursue these fields at higher education. Make sure to practice full past paper questions on this topic to build speed and accuracy, as matrix questions are often worth 10+ marks and require careful step-by-step calculation to avoid arithmetic errors. Recap FP1 matrix content if you struggled with prerequisite knowledge, as gaps in basic matrix arithmetic will slow you down significantly in FP3 exam questions.

→FP2 Further Complex Numbers