Determinants: Formula, Properties & Applications

The Determinant as a Scalar Assignment

Every square matrix

A

of size

n \times n

has a single real number (or complex number, if the entries are complex) associated with it, written

\det(A)

or

|A|

. This number is called the determinant. It is defined only for square matrices — a rectangular matrix has no determinant.

The notation

|A|

is widespread but potentially confusing, since the same vertical bars denote absolute value for real numbers and modulus for complex numbers. When there is any risk of ambiguity,

\det(A)

is preferred.

The determinant encodes a remarkable amount of structural information. Its most fundamental role is as an invertibility test:

A

is invertible if and only if

\det(A) \neq 0

. But it also measures how the linear map

x \mapsto Ax

distorts volume, determines whether that map preserves or reverses orientation, and appears in explicit formulas for eigenvalues, matrix inverses, and solutions to linear systems. The next several sections develop the determinant starting from the smallest cases and building toward the general definition.

The 2×2 Formula

For a

2 \times 2

matrix

A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}

the determinant is

\det(A) = ad - bc

One way to see where this formula comes from is to solve the system

Ax = \mathbf{v}

for a general right-hand side. Applying elimination, the solution involves dividing by

ad - bc

at every step. When

ad - bc = 0

, the system either has no solution or infinitely many, and the matrix cannot be inverted. When

ad - bc \neq 0

, there is a unique solution for every

\mathbf{v}

, and the inverse exists.

The expression

ad - bc

also carries geometric meaning. The columns of

A

are the vectors

(a, c)

and

(b, d)

in

\mathbb{R}^2

. The signed area of the parallelogram these two vectors span equals exactly

ad - bc

. A positive value means the columns are arranged counterclockwise; a negative value means clockwise; zero means the columns are parallel and the parallelogram collapses to a line segment.

Worked Examples

For

A = \begin{pmatrix} 3 & 1 \\ 2 & 5 \end{pmatrix}

, the determinant is

3 \cdot 5 - 1 \cdot 2 = 13

. The matrix is invertible and its column vectors span a parallelogram of area

13

.

For

A = \begin{pmatrix} 4 & 6 \\ 2 & 3 \end{pmatrix}

, the determinant is

4 \cdot 3 - 6 \cdot 2 = 0

. The second column is

\frac{3}{2}

times the first, so the columns are parallel and the matrix is singular.

For

A = \begin{pmatrix} 0 & -1 \\ 1 & 0 \end{pmatrix}

, the determinant is

0 \cdot 0 - (-1) \cdot 1 = 1

. This is a

90°

rotation matrix — it preserves areas and orientation.

The 3×3 Formula

For a

3 \times 3

matrix

A = \begin{pmatrix} a_{11} & a_{12} & a_{13} \\ a_{21} & a_{22} & a_{23} \\ a_{31} & a_{32} & a_{33} \end{pmatrix}

the determinant is computed by expanding along the first row:

\det(A) = a_{11}(a_{22}a_{33} - a_{23}a_{32}) - a_{12}(a_{21}a_{33} - a_{23}a_{31}) + a_{13}(a_{21}a_{32} - a_{22}a_{31})

Each first-row entry multiplies the

2 \times 2

determinant of the submatrix that remains after deleting that entry's row and column. The signs alternate:

+, -, +

.

The Sarrus Mnemonic

A shortcut for the

3 \times 3

case is the rule of Sarrus. Write the matrix and copy its first two columns to the right:

\begin{pmatrix} a_{11} & a_{12} & a_{13} \\ a_{21} & a_{22} & a_{23} \\ a_{31} & a_{32} & a_{33} \end{pmatrix} \begin{matrix} a_{11} & a_{12} \\ a_{21} & a_{22} \\ a_{31} & a_{32} \end{matrix}

Sum the three products along the downward diagonals, then subtract the three products along the upward diagonals. This gives the same result as the first-row expansion. The Sarrus rule works only for

3 \times 3

matrices — there is no analogous diagonal trick for

4 \times 4

or larger.

Worked Example

For

A = \begin{pmatrix} 2 & 3 & 1 \\ 0 & -1 & 4 \\ 5 & 2 & 3 \end{pmatrix}

, expanding along the first row:

\det(A) = 2[(-1)(3) - (4)(2)] - 3[(0)(3) - (4)(5)] + 1[(0)(2) - (-1)(5)]

= 2(-3 - 8) - 3(0 - 20) + 1(0 + 5) = 2(-11) - 3(-20) + 5 = -22 + 60 + 5 = 43

The same result follows from the Sarrus rule: downward diagonals give

2(-1)(3) + 3(4)(5) + 1(0)(2) = -6 + 60 + 0 = 54

; upward diagonals give

1(-1)(5) + 2(4)(2) + 3(0)(3) = -5 + 16 + 0 = 11

; the determinant is

54 - 11 = 43

.

The six signed products of a 3×3

Three diagonals running one way are added and three running the other are subtracted — six products in total, each using one entry from every row and every column. The pattern is a convenience for this size only: it does not extend to 4×4, where the count jumps to twenty-four. Watch the diagonals traced out on the determinant visualizer.

The general definition in the next section is what these six products are a special case of.

The General n×n Determinant

The pattern from the

3 \times 3

case extends to any dimension. For an

n \times n

matrix, expand along the first row:

\det(A) = \sum_{j=1}^{n} (-1)^{1+j} \, a_{1j} \, M_{1j}

where

M_{1j}

is the determinant of the

(n-1) \times (n-1)

submatrix obtained by deleting row

1

and column

j

. This is a recursive definition: each

n \times n

determinant reduces to

n

determinants of size

(n-1) \times (n-1)

, each of which reduces further, until reaching the base case of

1 \times 1

matrices where

\det(a) = a

.

A crucial fact is that the expansion need not use the first row. Expanding along any row or any column gives the same result. The signed sub-determinants

(-1)^{i+j} M_{ij}

are called cofactors, and the freedom to choose the expansion axis is what makes the formula practical: a row or column with many zeros dramatically reduces the number of terms.

The recursive nature of this definition means the computational cost grows factorially. An

n \times n

determinant via cofactor expansion requires on the order of

n!

arithmetic operations. For

n = 5

that is

120

operations; for

n = 10

it is over

3.6

million. This explosion is what motivates the row-reduction approach, which accomplishes the same task in

O(n^3)

operations by exploiting how row operations affect the determinant.

Singular and Nonsingular Matrices

The value

\det(A) = 0

marks a sharp dividing line. A square matrix with zero determinant is called singular; a square matrix with nonzero determinant is called nonsingular or invertible.

When

\det(A) = 0

, the columns of

A

are linearly dependent — at least one column can be written as a linear combination of the others. The homogeneous system

Ax = \mathbf{0}

has nontrivial solutions, the rank of

A

is strictly less than

n

, and the matrix maps

\mathbb{R}^n

onto a lower-dimensional subspace. Geometrically, the transformation collapses at least one dimension, flattening regions of positive volume down to zero volume.

When

\det(A) \neq 0

, everything works. The columns of

A

form a basis for

\mathbb{R}^n

. The system

Ax = \mathbf{b}

has exactly one solution for every right-hand side

\mathbf{b}

. The matrix has an inverse

A^{-1}

, and the linear map

x \mapsto Ax

is a bijection from

\mathbb{R}^n

to itself that stretches or compresses volumes by the factor

|\det(A)|

without collapsing any dimension.

The determinant thus answers the single most important structural question about a square matrix in a single number.

Computing Small Determinants

The most direct way to build intuition is to compute by hand. The

2 \times 2

formula

ad - bc

is immediate. The

3 \times 3

case requires either first-row expansion or the Sarrus shortcut. For

4 \times 4

matrices, no shortcut exists — the computation goes through cofactor expansion, but choosing a good row or column makes a significant difference.

A 4×4 Example

A = \begin{pmatrix} 1 & 0 & 2 & -1 \\ 3 & 0 & 0 & 5 \\ 2 & 1 & 4 & -3 \\ 1 & 0 & 5 & 0 \end{pmatrix}

Column

2

has three zeros and a single

1

in position

(3,2)

, making it the best expansion axis. Expanding along column

2

:

\det(A) = (-1)^{3+2} \cdot 1 \cdot M_{32}

where

M_{32}

is the

3 \times 3

determinant obtained by deleting row

3

and column

2

:

M_{32} = \det\begin{pmatrix} 1 & 2 & -1 \\ 3 & 0 & 5 \\ 1 & 5 & 0 \end{pmatrix}

Expanding this along its first row:

M_{32} = 1(0 \cdot 0 - 5 \cdot 5) - 2(3 \cdot 0 - 5 \cdot 1) + (-1)(3 \cdot 5 - 0 \cdot 1)

= 1(-25) - 2(-5) + (-1)(15) = -25 + 10 - 15 = -30

So

\det(A) = (-1)^{5} \cdot 1 \cdot (-30) = (-1)(-30) = 30

.

The choice of column

2

reduced the problem from four

3 \times 3

determinants down to one. This illustrates why scanning for zeros before expanding is the first step in any hand computation.

Expanding by Minors and Cofactors

The sub-determinants appearing in cofactor expansion have their own terminology. The

(i,j)

minor

M_{ij}

is the determinant of the submatrix formed by deleting row

i

and column

j

from

A

. The

(i,j)

cofactor is the signed version:

C_{ij} = (-1)^{i+j} M_{ij}

The sign factor

(-1)^{i+j}

follows a checkerboard pattern starting with

+

at position

(1,1)

:

\begin{pmatrix} + & - & + & - & \cdots \\ - & + & - & + & \cdots \\ + & - & + & - & \cdots \\ \vdots & & & & \ddots \end{pmatrix}

Using this notation, the Laplace expansion along row

i

takes the compact form

\det(A) = \sum_{j=1}^{n} a_{ij} \, C_{ij}

and the expansion along column

j

is

\det(A) = \sum_{i=1}^{n} a_{ij} \, C_{ij}

Both give the same result regardless of which row or column is chosen. Collecting all cofactors into a matrix and transposing produces the adjugate, which leads to an explicit formula for the matrix inverse. The full development of minors, cofactors, the adjugate, and the structural results they enable is on the cofactors page.

Expansion along a row, minors highlighted

For each entry of the chosen row, its own row and column are struck out and the determinant of what remains is taken, then signed and weighted. Any row or column gives the same answer, so the one with the most zeros is always the cheapest to pick. Choose a different row or column and compare on the determinant visualizer.

This is the definition that actually generalises, at the price of a cost that grows faster than almost anything else in the subject.

How Row Operations Affect the Determinant

The three elementary row operations interact with the determinant in simple, predictable ways. Swapping two rows multiplies the determinant by

-1

. Multiplying a row by a nonzero scalar

k

multiplies the determinant by

k

. Adding a scalar multiple of one row to a different row leaves the determinant unchanged.

These three rules turn Gaussian elimination into a determinant-computing algorithm. Reduce

A

to upper triangular form using row operations, tracking every swap and every scaling. The determinant of the resulting triangular matrix is the product of its diagonal entries. Adjusting for the tracked sign flips and scale factors gives

\det(A)

.

This approach requires on the order of

n^3

arithmetic operations — a dramatic improvement over the

n!

cost of cofactor expansion. For any matrix larger than

4 \times 4

, row reduction is the practical method.

A Quick Illustration

Starting from

A = \begin{pmatrix} 2 & 4 \\ 3 & 7 \end{pmatrix}

, subtract

\frac{3}{2}

times row

1

from row

2

(this does not change the determinant) to get

\begin{pmatrix} 2 & 4 \\ 0 & 1 \end{pmatrix}

. The product of the diagonal is

2 \cdot 1 = 2

, which matches

\det(A) = 2 \cdot 7 - 4 \cdot 3 = 2

.

The complete set of algebraic properties — including the multiplicative rule

\det(AB) = \det(A)\det(B)

, transpose invariance, and the full invertibility equivalence — is developed on the properties page. The three rules above collect into a compact reference:

Row operation	Effect on det(A)	Adjustment when reading off det from triangular form
Swap two rows	multiplies det by −1	flip the sign once per swap
Multiply a row by k ≠ 0	multiplies det by k	divide the final product by k
Add a multiple of one row to another	leaves det unchanged	no adjustment

Area, Volume, and Orientation

The determinant has a direct geometric meaning: it measures how a matrix, viewed as a linear transformation, distorts size and orientation.

In two dimensions,

|\det(A)|

equals the area of the parallelogram spanned by the columns of

A

. In three dimensions,

|\det(A)|

equals the volume of the parallelepiped spanned by the three column vectors, which also equals the scalar triple product

\mathbf{a} \cdot (\mathbf{b} \times \mathbf{c})

. In

n

dimensions,

|\det(A)|

is the factor by which the map

x \mapsto Ax

scales

n

-dimensional volumes.

The sign carries its own meaning. A positive determinant means the transformation preserves orientation — counterclockwise stays counterclockwise in

\mathbb{R}^2

, right-handed stays right-handed in

\mathbb{R}^3

. A negative determinant means orientation is reversed. A zero determinant means the image is lower-dimensional: a

3 \times 3

transformation with

\det = 0

maps all of

\mathbb{R}^3

onto a plane, a line, or a point.

Rotation matrices always have determinant

+1

. Reflection matrices always have determinant

-1

. These are the cleanest examples of orientation-preserving and orientation-reversing maps. The full geometric treatment, including the change-of-variables formula from multivariable calculus and explicit area and volume formulas, is on the geometry page.

A grid stretched, then flattened

Above, the unit square has become a parallelogram and its area is exactly the determinant. Below, the matrix has flattened the plane onto a line: the parallelogram has no area left and the determinant is zero. Singularity and zero area are the same event seen from two sides. Change the entries and watch the area respond on the linear transformation explorer.

Every geometric use of the determinant — volume, orientation, change of variables — is this one fact applied in more dimensions.

Determinant-Based Formulas

The determinant is not only a diagnostic tool — it provides closed-form expressions for quantities that might otherwise require iterative procedures.

Cramer's rule solves a linear system

Ax = \mathbf{b}

by expressing each component of the solution as a ratio of two determinants:

x_i = \det(A_i)/\det(A)

, where

A_i

is

A

with its

i

-th column replaced by

\mathbf{b}

. The adjugate formula gives the inverse of

A

as

A^{-1} = \frac{1}{\det(A)} \operatorname{adj}(A)

, writing every entry of the inverse explicitly in terms of cofactors. In the

2 \times 2

case, this reduces to the familiar swap-and-negate formula.

The cross product

\mathbf{a} \times \mathbf{b}

in

\mathbb{R}^3

can be computed as a symbolic

3 \times 3

determinant with the unit vectors

\hat{i}, \hat{j}, \hat{k}

in the first row. The characteristic polynomial

\det(A - \lambda I)

defines the eigenvalues of

A

— its roots are precisely the scalars

\lambda

for which

A - \lambda I

is singular. The Wronskian, a determinant built from functions and their derivatives, tests linear independence in the setting of differential equations.

Each of these formulas is developed with full worked examples on the applications page. The five collect into a single reference card:

Formula	Expression	What it computes
Cramer's rule	x_i = det(A_i) / det(A)	each component of the unique solution to Ax = b
Adjugate formula for A⁻¹	A⁻¹ = adj(A) / det(A)	explicit closed form for the matrix inverse
Cross product in ℝ³	a × b = symbolic 3×3 det with î, ĵ, k̂ on top	vector perpendicular to a and b with prescribed length
Characteristic polynomial	det(A − λI)	polynomial whose roots are the eigenvalues of A
Wronskian	determinant of functions and their derivatives	linear independence test for solutions of an ODE

Summary: How to Compute a Determinant

Across the methods covered above, four routes lead to

\det(A)

— the closed-form

2\times 2

formula, the closed-form

3\times 3

expansion (or Sarrus), general cofactor expansion, and row reduction. Each is the right choice in a different regime, distinguished by matrix size, sparsity of the matrix, and whether the work needs to scale to large

n

. The table below collects all four side by side on the underlying formula, the situation it suits best, and the asymptotic cost.

Method	Formula / approach	Best for	Cost
2×2 formula	ad − bc	every 2×2 case — direct, no tricks needed	O(1)
3×3 (first-row expansion or Sarrus)	sum of three 2×2 dets with alternating signs; or downward minus upward diagonal products	3×3 only; Sarrus does not generalize	O(1)
Cofactor expansion	Σ (−1)^i+j a_ij M_ij along any row or column	small matrices, especially when a row or column has many zeros	O(n!)
Row reduction	reduce to triangular form, multiply the diagonal, adjust for tracked swaps and scalings	any matrix 4×4 or larger; the standard general-purpose method	O(n³)

Determinants

A Single Number from a Square Matrix

The Determinant as a Scalar Assignment

The 2×2 Formula

Worked Examples

The 3×3 Formula

The Sarrus Mnemonic

Worked Example

The General n×n Determinant

Singular and Nonsingular Matrices

Computing Small Determinants

A 4×4 Example

Expanding by Minors and Cofactors

How Row Operations Affect the Determinant

A Quick Illustration

Area, Volume, and Orientation

Determinant-Based Formulas

Summary: How to Compute a Determinant