Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix. A scalar λ∈F\lambda\in\mathbb{F} is an eigenvalue of AA if and only if there exists a non-zero vector v‾∈Fn\underline{v} \in \mathbb{F}^n such that multiplication by AA scales v‾\underline{v} by λ.\lambda.

Av‾=λv‾A\underline{v}=\lambda\underline{v}

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix and λ\lambda be an eigenvalue of A.A. A non-zero vector v‾∈Fn\underline{v} \in \mathbb{F}^n is an eigenvector of AA corresponding to λ\lambda if and only if Av‾=λv‾.A\underline{v}=\lambda\underline{v}. For an eigenvector v‾,\underline{v}, the linear transformation v‾↦Av‾\underline{v}\mapsto A\underline{v} is the scalar transformation v‾↦λv‾.\underline{v}\mapsto\lambda\underline{v}.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix. The characteristic polynomial of AA is the polynomial of degree nn in λ\lambda given by the determinant of A−λIn.A-\lambda I_n.

pA(λ):=∣A−λIn∣p_A(\lambda):=\lvert A-\lambda I_n\rvert

The eigenvalues of AA are the roots of the characteristic polynomial, since Av‾=λv‾A\underline{v}=\lambda\underline{v} has a non-zero solution v‾\underline{v} if and only if A−λInA-\lambda I_n is singular.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix. The eigen-spectrum of A,A, written σA,\sigma_A, is the set of distinct eigenvalues of A.A.

σA:={λ:∣A−λIn∣=0}\sigma_A:=\{\lambda : \lvert A-\lambda I_n\rvert=0\}

The characteristic polynomial of AA has mm distinct complex roots where 1≤m≤n,1\leq m\leq n, so taking eigenvalues over C\mathbb{C} the eigen-spectrum has cardinality m.m. A real matrix can have complex eigenvalues: the rotation matrix [0−110]\begin{bmatrix} 0 & -1 \\ 1 & 0 \end{bmatrix} has eigenvalues ±i\pm i and no real eigenvalues. The statements below that count nn eigenvalues assume that all roots of the characteristic polynomial lie in F,\mathbb{F}, which always holds for F=C.\mathbb{F}=\mathbb{C}.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix and λ∈σA.\lambda \in \sigma_A. The eigenvectors of AA corresponding to λ\lambda are the non-zero solutions v‾\underline{v} of the equation

(A−λIn)v‾=0‾.\left(A-\lambda I_n\right)\underline{v}=\underline{0}.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix and λ∈σA.\lambda \in \sigma_A. The eigen-space of AA with respect to λ,\lambda, written Eλ,\mathcal{E}_{\lambda}, is the kernel of A−λIn.A-\lambda I_n.

Eλ:={v‾∈Fn:(A−λIn)v‾=0‾}\mathcal{E}_{\lambda}:=\{\underline{v} \in \mathbb{F}^n : \left(A-\lambda I_n\right)\underline{v}=\underline{0}\}

The eigen-space is a subspace of Fn\mathbb{F}^n that contains the eigenvectors of AA corresponding to λ\lambda together with the zero vector, and it is spanned by those eigenvectors.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix and λ∈σA.\lambda \in \sigma_A. The algebraic multiplicity of λ,\lambda, written μA(λ),\mu_A\left(\lambda\right), is the number of times λ\lambda appears as a root of the characteristic polynomial of A.A. The sum of the algebraic multiplicities of all eigenvalues of AA is n.n.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix and λ∈σA.\lambda \in \sigma_A. The geometric multiplicity of λ,\lambda, written γA(λ),\gamma_A\left(\lambda\right), is the dimension of the eigen-space Eλ.\mathcal{E}_{\lambda}. The geometric multiplicity is the maximum number of linearly independent eigenvectors of AA corresponding to λ.\lambda.

The algebraic and geometric multiplicities have the following properties:

1≤γA(λ)≤μA(λ)≤n  ∀ λ∈σA1\leq\gamma_A\left(\lambda\right)\leq\mu_A\left(\lambda\right)\leq n\,\,\forall\,\lambda\in\sigma_A

Eigenvectors corresponding to distinct eigenvalues are linearly independent. A repeated eigenvalue does not necessarily correspond to multiple linearly independent eigenvectors.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix and λ1,…,λn∈F\lambda_1,\ldots,\lambda_n \in \mathbb{F} be its eigenvalues repeated according to algebraic multiplicity. The eigenvalues have the following properties:

∑i=1nλi=tr(A)\sum_{i=1}^n \lambda_i = \mathrm{tr}(A) ∏i=1nλi=∣A∣\prod_{i=1}^n \lambda_i = \lvert A\rvert

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a square matrix. The matrix AA is defective if and only if the geometric multiplicity of some eigenvalue is strictly less than its algebraic multiplicity.

γA(λ)<μA(λ)  for some  λ∈σA\gamma_A\left(\lambda\right)<\mu_A\left(\lambda\right)\,\,\mathrm{for}\,\mathrm{some}\,\,\lambda\in\sigma_A

A matrix is defective if and only if it has fewer than nn linearly independent eigenvectors. A matrix is non-defective if and only if γA(λ)=μA(λ)\gamma_A\left(\lambda\right)=\mu_A\left(\lambda\right) for every λ∈σA,\lambda\in\sigma_A, in which case it has nn linearly independent eigenvectors.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a non-defective matrix. An eigen-basis of AA is a basis of Fn\mathbb{F}^n consisting of nn linearly independent eigenvectors of A.A.

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a non-defective matrix with eigen-basis v‾1,…,v‾n\underline{v}_1,\ldots,\underline{v}_n and corresponding eigenvalues λ1,…,λn.\lambda_1,\ldots,\lambda_n. The eigenvector matrix PP is the square matrix whose columns are the eigenvectors, and the eigenvalue matrix Λ\Lambda is the diagonal matrix whose diagonal elements are the corresponding eigenvalues.

P:=[↑↑v‾1…v‾n↓↓]P:=\begin{bmatrix} \uparrow & & \uparrow \\ \underline{v}_1 & \ldots & \underline{v}_n \\ \downarrow & & \downarrow \end{bmatrix} Λ:=[λ1000⋱000λn]\Lambda:=\begin{bmatrix} \lambda_1 & 0 & 0 \\ 0 & \ddots & 0 \\ 0 & 0 & \lambda_n \end{bmatrix}

Let A∈F n×nA\in\mathbb{F}^{\,n\times n} be a non-defective matrix with eigenvector matrix PP and eigenvalue matrix Λ.\Lambda. The diagonalisation of AA is the factorisation of AA into the product of P,P, Λ\Lambda and P−1.P^{-1}.

A=PΛP−1A=P\Lambda P^{-1}

A matrix is diagonalisable if and only if it is non-defective. The diagonalisation has the property Ak=PΛkP−1A^k=P\Lambda^kP^{-1} for every k∈N,k\in\mathbb{N}, where Λk\Lambda^k is the diagonal matrix with diagonal elements λik.\lambda_i^k.