GraphMath

Complex eigenvalues and eigenvectors

Rotation and scaling hidden inside a real 2×2 matrix

What geometric structure replaces real eigendirections?

A real 2×2 matrix with complex eigenvalues has no nonzero real eigenvector direction. Instead, one complex eigenvector provides two real vectors, Re(x) and Im(x), that form a basis in which the matrix becomes a rotation-scaling block. This gives the factorization A = XSX⁻¹ and makes powers and exponentials of A easier to understand.

Key ideas

Complex eigenvalues of a real matrix occur in conjugate pairs and encode rotation together with scaling.

  • For a real 2×2 matrix, non-real eigenvalues occur as λ₁ = α + iβ and λ₂ = α − iβ
  • If Ax₁ = λ₁x₁, then x₁ᶜ is an eigenvector associated with λ₂
  • A complex eigenvector can be written uniquely as Re(x) + i Im(x)
  • Different free-variable choices change the displayed real and imaginary parts without changing the eigenvector relation
  • For a rotation-scaling matrix, Re(x) and Im(x) are orthogonal and equal in length
  • With X = [Re(x) | Im(x)], a real matrix with complex eigenvalues factors as A = XSX⁻¹
  • The block S rotates by θ = atan2(β, α) and scales by ρ = √(α² + β²) = |λ|
  • Powers satisfy Aᵗ = XSᵗX⁻¹, and matrix exponentials satisfy eᵗᴬ = XeᵗˢX⁻¹

The chapter follows these ideas through symbolic derivations, basis-change diagrams and orbit visualizations.

Why does one complex eigenvector produce a real basis?

Write x = Re(x) + i Im(x) and λ = α + iβ. Separating the real and imaginary parts of Ax = λx shows that A maps the pair [Re(x) | Im(x)] according to a real rotation-scaling matrix S. When these two real vectors are linearly independent, they form the columns of X and give A = XSX⁻¹.

Related chapters

Chapter contents

How do complex eigenvalues control powers of A?

In A = XSX⁻¹, the matrix S is a rotation-scaling block determined by λ = α + iβ. Its modulus |λ| controls contraction or expansion, while θ = atan2(β, α) controls rotation. Therefore Sᵗ rotates by tθ and scales by |λ|ᵗ, and Aᵗ = XSᵗX⁻¹ carries the same behavior back to the original coordinates. The orbit may appear elliptical in the standard basis even though the underlying action in the S basis is circular.

Was this chapter helpful?

Quick feedback helps us improve the site.