2x2 Matrix Determinant and Inverse

2×2 determinant and inverse

Take a 2×2 matrix A = [[a, b], [c, d]]. Its determinant is det(A) = ad - bc, and where an inverse exists it works out to A⁻¹ = (1/det(A)) · [[d, -b], [-c, a]]. You can invert the matrix exactly when det(A) ≠ 0. If det = 0 the matrix is singular, which is another way of saying its columns are linearly dependent.

Example: with A = [[1, 2], [3, 4]], you get det = 1·4 - 2·3 = -2, which gives A⁻¹ = (-1/2)·[[4, -2], [-3, 1]] = [[-2, 1], [1.5, -0.5]]. To check it, multiply back out: A · A⁻¹ = I.

Applications

These little matrices show up everywhere. They crack linear systems in two unknowns, whether you reach for Cramer's rule or x = A⁻¹b. In computer graphics they encode 2D linear transformations like rotations, scalings and shears. They turn up in principal component analysis (PCA) when you're working with the covariance of two variables, and the eigenvalues of a 2×2 system matrix tell you about the stability of 2D linear ODE systems.

FAQ

What does det = 0 mean geometrically? It means the columns of A line up parallel (linearly dependent), so the transformation flattens the whole plane down to a line or even a single point. Any area you started with is gone.

What does the determinant measure? It's the signed factor that areas get scaled by once the linear transformation is applied. A negative det tells you the orientation has flipped, which is a reflection.

How do I solve Ax = b? As long as det(A) ≠ 0, the answer is x = A⁻¹ · b. If you'd rather not invert, Cramer's rule gets you there with two 2×2 determinants.

Does the formula extend to larger matrices? The general version, A⁻¹ = adj(A)/det(A), holds for any n×n. That said, once n hits 3 or more, computing the adjugate gets slow and Gauss–Jordan elimination is usually the quicker route.