Diagonalize a 2×2 or 3×3 matrix online. Find eigenvalues, eigenvectors, matrices P and D where A = PDP⁻¹, determinant, trace, and characteristic polynomial with step-by-step results.
Matrix diagonalization is a foundational technique in linear algebra that decomposes a square matrix A into the product PDP⁻¹, where D is a diagonal matrix of eigenvalues and P is a matrix whose columns are the corresponding eigenvectors. This decomposition dramatically simplifies computing matrix powers, exponentials, and solving systems of differential equations. Our Diagonalize Matrix Calculator supports both 2×2 and 3×3 matrices, walking you through each step: forming the characteristic polynomial det(A − λI) = 0, solving for eigenvalues, computing eigenvectors, and assembling the P and D matrices. Understanding whether a matrix is diagonalizable also reveals important structural properties — for example, a matrix with n distinct eigenvalues is always diagonalizable, while repeated eigenvalues may or may not allow full diagonalization depending on the dimension of their eigenspaces. Engineers use diagonalization for vibration analysis in mechanical systems, physicists rely on it for quantum state representations, and data scientists apply it in Principal Component Analysis (PCA) for dimensionality reduction. Whether you are a student solving homework problems or a professional verifying computation, this tool provides instant, accurate results with clear visual comparisons of eigenvalue magnitudes and a complete step-by-step summary table.
Diagonalization becomes valuable the moment you need more than a determinant or a one-off eigenvalue check. In linear algebra coursework, it helps verify whether a repeated eigenvalue still produces enough independent eigenvectors. In applied math, it is a fast way to decide whether a matrix power, recurrence, or linear dynamical system can be simplified by working in an eigenbasis instead of multiplying the original matrix again and again.
This calculator is especially useful because it keeps the decision path together: trace, determinant, characteristic polynomial, eigenvalues, and the final diagonalizable verdict are shown in one place. For 2×2 examples it also exposes the eigenvector columns used to build P, making it easier to compare your handwritten work to the exact matrix decomposition the interface is using.
Characteristic polynomial: det(A − λI) = 0. For 2×2: λ² − tr(A)λ + det(A) = 0. Eigenvalues are roots; eigenvectors solve (A − λI)x = 0. A = PDP⁻¹ when P is invertible.
Result: det(A) = 3, trace(A) = 4, eigenvalues = 3 and 1, diagonalizable = Yes
For A = [[2, 1], [1, 2]], the calculator forms λ² − 4λ + 3 = 0, so the eigenvalues are 3 and 1. A valid eigenvector matrix is P = [[-1, -1], [-1, 1]] using the displayed column vectors, and D = diag(3, 1). Because P is invertible, A can be written as PDP⁻¹.
To diagonalize a matrix, you are looking for a change of basis that turns a full square matrix into a diagonal one. The diagonal entries are the eigenvalues, and the change-of-basis matrix is built from independent eigenvectors. When this works, matrix powers and many repeated transformations become much easier to compute because a diagonal matrix can be raised to powers entry by entry.
The output order reflects the real workflow used in class and in applied work. First check the trace and determinant as quick invariants. Next inspect the characteristic polynomial and eigenvalues. After that, the key question is whether the matrix has enough independent eigenvectors to build an invertible P matrix. The diagonalizable status card summarizes that decision, while the step table shows the same logic in a more explicit sequence.
Diagonalization appears in vibration problems, repeated linear transformations, Markov-style models, systems of differential equations, and PCA-style reasoning about directions of stretch. It is also a good diagnostic tool: if a matrix has distinct eigenvalues, diagonalization is usually straightforward; if an eigenvalue repeats, the decomposition may fail unless the eigenspace is large enough. That is why this calculator is useful not just for answers, but for quickly testing examples and counterexamples.
Diagonalization rewrites a matrix A as PDP⁻¹, where D is a diagonal matrix of eigenvalues and P contains the eigenvectors as columns. This makes computing Aⁿ trivial: Aⁿ = PDⁿP⁻¹.
A matrix is not diagonalizable when it has repeated eigenvalues whose eigenspaces have dimension less than their algebraic multiplicity — for example, the matrix [[2, 1], [0, 2]]. This provides clearer practical guidance for reliable use.
This calculator focuses on real eigenvalues. If the characteristic polynomial has no real roots (negative discriminant), the matrix cannot be diagonalized over the reals.
It is det(A − λI), a polynomial in λ whose roots are the eigenvalues. For a 2×2 matrix it is quadratic; for 3×3 it is cubic.
It simplifies matrix exponentiation, solving differential equations, and performing PCA in data science. Raising a diagonal matrix to a power is trivially just raising each element.
The trace (sum of diagonal entries) always equals the sum of eigenvalues, and the determinant equals the product of eigenvalues. These identities hold for any square matrix.