Skip to content

Characteristic Equation for a Differential Equation


Definition

Consider the homogeneous linear differential equation with constant real coefficients

x(n)(t)+an−1x(n−1)(t)+⋯+a1x′(t)+a0x(t)=0,x^{(n)}(t)+a_{n-1}x^{(n-1)}(t)+\dots +a_1x'(t)+a_0x(t)=0,

where x′(t)=dxdt(t)x'(t)=\dfrac{\mathrm{d}x}{\mathrm{d}t}(t) and x(k)x^{(k)} is the kk-th derivative. Its characteristic polynomial is

p(r)=rn+an−1rn−1+⋯+a1r+a0,p(r)=r^n+a_{n-1}r^{n-1}+\dots+a_1r+a_0,

and p(r)=0p(r)=0 is its characteristic equation. The equation is obtained by substituting the trial solution x(t)=ertx(t)=\mathrm{e}^{rt}: since dkdtkert=rkert\frac{\mathrm{d}^k}{\mathrm{d}t^k}\mathrm{e}^{rt}=r^k\mathrm{e}^{rt}, the left-hand side becomes p(r)ertp(r)\mathrm{e}^{rt}, which vanishes for all tt if and only if p(r)=0p(r)=0. For a second-order differential equation x′′+ax′+bx=0x''+ax'+bx=0 the characteristic equation is r2+ar+b=0r^2+ar+b=0.


Interpretation

The substitution turns a differential equation into an algebraic one: each root of the polynomial pp gives an exponential "mode" ert\mathrm{e}^{rt} of the exponential function. Writing the equation as a first-order system with a companion matrix, pp is exactly the characteristic polynomial of that matrix, so the roots are its eigenvalues.

Left, the roots -1 + 2i and -1 - 2i of r^2 + 2r + 5 = 0 in the complex plane; right, the real solution e^(-t) cos 2t oscillating inside the dashed envelope plus or minus e^(-t).

The figure follows the complex-roots example x′′+2x′+5x=0x''+2x'+5x=0. Its characteristic equation r2+2r+5=0r^2+2r+5=0 has the roots r=−1±2ir=-1\pm2\mathrm{i}, marked on the left in the complex plane. Through ert\mathrm{e}^{rt} and Euler's formula they give the real mode e−tcos⁡2t\mathrm{e}^{-t}\cos2t on the right: the real part −1-1 fixes the shrinking envelope ±e−t\pm \mathrm{e}^{-t}, and the imaginary part 22 fixes how fast the solution oscillates inside it.

Move aa and bb across a2=4ba^2=4b: two real roots give two exponentials, a repeated root needs the extra factor tt, and a complex pair α±iβ\alpha\pm\mathrm{i}\beta gives eαt(c1cos⁡βt+c2sin⁡βt)e^{\alpha t}(c_1\cos\beta t+c_2\sin\beta t).

The discrete-time analogue is the characteristic polynomial for a difference equation, where the trial solution is λt\lambda^t instead of ert\mathrm{e}^{rt}. The two are linked by λ=er\lambda=\mathrm{e}^{r}: sampling ert\mathrm{e}^{rt} at integer times gives (er)t(\mathrm{e}^{r})^t.

Common errors: forgetting that a repeated root needs the extra factor tt; writing e(α±iβ)t\mathrm{e}^{(\alpha\pm\mathrm{i}\beta)t} as the final real answer instead of converting it with Euler's formula; and applying the method to equations with non-constant coefficients, where ert\mathrm{e}^{rt} is generally not a solution.


Key Results

  • By the fundamental theorem of algebra, pp has nn complex roots counted with multiplicity.
  • A real root rr of multiplicity mm gives the mm solutions ert, tert, …, tm−1ert\mathrm{e}^{rt},\,t\mathrm{e}^{rt},\,\dots,\,t^{m-1}\mathrm{e}^{rt}.
  • Since the coefficients are real, nonreal roots come in pairs α±iβ\alpha\pm\mathrm{i}\beta (β≠0\beta\neq0) by the complex conjugate root theorem. Each pair gives the real solutions eαtcos⁡βt\mathrm{e}^{\alpha t}\cos\beta t and eαtsin⁡βt\mathrm{e}^{\alpha t}\sin\beta t, because e(α+iβ)t=eαt(cos⁡βt+isin⁡βt)\mathrm{e}^{(\alpha+\mathrm{i}\beta)t}=\mathrm{e}^{\alpha t}(\cos\beta t+\mathrm{i}\sin\beta t); a pair of multiplicity mm gives these two functions multiplied by 1,t,…,tm−11,t,\dots,t^{m-1}.
  • The nn functions obtained this way form a basis of the solution space, so the general solution is their linear combination.
  • For n=2n=2, r2+ar+b=0r^2+ar+b=0 has discriminant a2−4ba^2-4b: positive gives two real roots, zero a double root −a/2-a/2, negative a complex pair −a2±i4b−a22-\frac{a}{2}\pm\mathrm{i}\frac{\sqrt{4b-a^2}}{2}.

Examples

Distinct real roots. x′′−x′−6x=0x''-x'-6x=0: r2−r−6=(r−3)(r+2)r^2-r-6=(r-3)(r+2), so x(t)=c1e3t+c2e−2tx(t)=c_1\mathrm{e}^{3t}+c_2\mathrm{e}^{-2t}.

Complex roots. x′′+2x′+5x=0x''+2x'+5x=0: r=−2±4−202=−1±2ir=\frac{-2\pm\sqrt{4-20}}{2}=-1\pm2\mathrm{i}, so x(t)=e−t(c1cos⁡2t+c2sin⁡2t)x(t)=\mathrm{e}^{-t}(c_1\cos2t+c_2\sin2t), a damped oscillation.

Double root, checked. x′′+2x′+x=0x''+2x'+x=0: (r+1)2=0(r+1)^2=0. For x=te−tx=t\mathrm{e}^{-t}: x′=(1−t)e−tx'=(1-t)\mathrm{e}^{-t}, x′′=(t−2)e−tx''=(t-2)\mathrm{e}^{-t}, and x′′+2x′+x=(t−2+2−2t+t)e−t=0x''+2x'+x=(t-2+2-2t+t)\mathrm{e}^{-t}=0.

Triple root. x′′′−3x′′+3x′−x=0x'''-3x''+3x'-x=0: r3−3r2+3r−1=(r−1)3r^3-3r^2+3r-1=(r-1)^3, so x(t)=(c1+c2t+c3t2)etx(t)=(c_1+c_2t+c_3t^2)\mathrm{e}^{t}.


Prerequisites



Sources

Aliases: auxiliary equation, characteristic polynomial of an ODE, characteristic equation of a linear ODE Keywords: e^{rt}, characteristic roots, repeated root, complex roots, e^{alpha t} cos(beta t) Symbols: alpha, beta, lambda, square root

Last updated: 22 October 2026 19:04