Differential Equations › Section C · Second order differential equations › Topic C.3

Complex roots

When the characteristic equation has no real roots, the solutions oscillate. Euler's formula turns a pair of complex exponentials into real sines and cosines, and the real part of the root decides whether the oscillation fades, grows or holds steady.

When the discriminant is negative

For \(ay'' + by' + cy = 0\) with \(b^2 - 4ac < 0\), the quadratic formula gives a complex conjugate pair of roots, \(r = \alpha \pm \beta i\), where

\alpha = -\frac{b}{2a}, \qquad \beta = \frac{\sqrt{4ac - b^2}}{2a}

and \(\beta > 0\) when \(a > 0\). For \(y'' + 2y' + 10y = 0\), the roots are \(\frac{-2 \pm \sqrt{-36}}{2}\), which is \(-1 \pm 3i\), so \(\alpha = -1\) and \(\beta = 3\). Formally, \(e^{(-1 + 3i)t}\) and \(e^{(-1 - 3i)t}\) are solutions, as the topic "Basic concepts and superposition" showed for any root. But a physical displacement is a real number, so we need a way to turn these complex functions into real ones.

Euler's formula

The bridge between exponentials and trigonometry is Euler's formula:

e^{i\theta} = \cos\theta + i\sin\theta

One way to see it is through power series: the series for \(e^{i\theta}\) splits into its real terms, which form the series for \(\cos\theta\), and its imaginary terms, which form \(i\) times the series for \(\sin\theta\). Replacing \(\theta\) by \(-\theta\) flips the sign of the sine term, giving \(\cos\theta - i\sin\theta\), because cosine is even and sine is odd. Combined with the ordinary rule \(e^{u + v} = e^ue^v\), which still holds for complex exponents, this gives

e^{(\alpha + i\beta)t} = e^{\alpha t}(\cos\beta t + i\sin\beta t)

From complex to real solutions

Call the two complex solutions \(z_1\) and \(z_2\):

z_{1,2} = e^{\alpha t}(\cos\beta t \pm i\sin\beta t)

Superposition allows any constants, even complex ones. Adding and halving gives \(e^{\alpha t}\cos\beta t\). Subtracting and dividing by \(2i\) gives \(e^{\alpha t}\sin\beta t\):

\frac{z_1 + z_2}{2} = e^{\alpha t}\cos\beta t, \qquad \frac{z_1 - z_2}{2i} = e^{\alpha t}\sin\beta t

Both results are real solutions, and neither is a constant multiple of the other, because their ratio is \(\tan\beta t\). So, for roots \(\alpha \pm \beta i\), the real general solution is

y = e^{\alpha t}(c_1\cos\beta t + c_2\sin\beta t)

Using the formula

Once the roots are known, write the answer straight away: \(\alpha\) goes in the exponential and \(\beta\) inside both trigonometric functions. For \(y'' + 2y' + 10y = 0\), the general solution is \(e^{-t}(c_1\cos 3t + c_2\sin 3t)\). If the leading coefficient is not 1, the formula handles it: for \(4y'' + 4y' + 17y = 0\),

r = \frac{-4 \pm \sqrt{16 - 272}}{8} = -\frac{1}{2} \pm 2i

so \(y = e^{-t/2}(c_1\cos 2t + c_2\sin 2t)\). When there is no \(y'\) term, as in \(y'' + \omega^2y = 0\), the roots are \(\pm\omega i\), \(\alpha = 0\), and the solution is pure oscillation:

y = c_1\cos\omega t + c_2\sin\omega t

Worked example: a growing oscillation

Solve \(y'' - 2y' + 10y = 0\) with \(y(0) = 2\) and \(y'(0) = -1\). 1. Roots: \(r = \frac{2 \pm \sqrt{4 - 40}}{2}\), that is, \(1 \pm 3i\). 2. General solution: \(y = e^{t}(c_1\cos 3t + c_2\sin 3t)\). 3. \(y(0) = c_1 = 2\). 4. By the product rule,

y' = e^{t}\big[(c_1 + 3c_2)\cos 3t + (c_2 - 3c_1)\sin 3t\big]

so \(y'(0) = c_1 + 3c_2 = -1\), giving \(c_2 = -1\). 5. Answer: \(y = e^{t}(2\cos 3t - \sin 3t)\). 6. Check: \(y'(0) = 2 - 3 = -1\). The oscillation grows, because \(\alpha = 1 > 0\).

Worked example: conditions away from zero

Solve \(y'' + 9y = 0\) with \(y(\pi/6) = 2\) and \(y'(\pi/6) = 3\). 1. Roots \(\pm 3i\), so \(y = c_1\cos 3t + c_2\sin 3t\). 2. Then \(y' = -3c_1\sin 3t + 3c_2\cos 3t\). 3. At \(t = \pi/6\), \(3t = \pi/2\), where cosine is 0 and sine is 1. 4. So \(y(\pi/6) = c_2 = 2\) and \(y'(\pi/6) = -3c_1 = 3\), giving \(c_1 = -1\). 5. Answer: \(y = -\cos 3t + 2\sin 3t\). 6. Check: at \(t = \pi/6\), \(y = 0 + 2 = 2\) and \(y' = 3 + 0 = 3\). Choosing a starting point where the angle is a multiple of \(\pi/2\) keeps the arithmetic short.

Reading the solution

The factor \(e^{\alpha t}\) acts as an envelope that the oscillation stays inside. If \(\alpha < 0\), the oscillation decays towards zero, like a plucked string falling silent. If \(\alpha > 0\), it grows without bound. If \(\alpha = 0\), it keeps a constant size forever. For \(y'' + 2y' + 10y = 0\), the envelope is \(e^{-t}\), so after 3 time units the swings are at most \(e^{-3}\), about 5%, of their starting envelope, while the solution still changes sign every \(\pi/3\). The trigonometric factor repeats every \(\frac{2\pi}{\beta}\), so the solution crosses zero at evenly spaced times \(\frac{\pi}{\beta}\) apart, even when \(\alpha \ne 0\) stops it repeating exactly. The time \(\frac{2\pi}{\beta}\) between alternate crossings is called the quasi-period, and the topic "Mechanical vibrations" uses it for damped springs.

Common mistakes to avoid

First, swapping \(\alpha\) and \(\beta\): for \(r = -3 \pm 4i\), the solution is \(e^{-3t}(c_1\cos 4t + c_2\sin 4t)\), not \(e^{4t}(\dots 3t)\). Second, putting \(\beta^2\) inside the trigonometric functions: \(y'' + 16y = 0\) gives \(\cos 4t\), not \(\cos 16t\). Third, forgetting the product rule when differentiating, which loses the \(\alpha c_1\) term in \(y'(0)\). Fourth, writing \(c_1e^{\alpha t} + c_2e^{\beta t}\), which treats the complex case like the real one. Fifth, leaving \(i\) in the final answer: the constants \(c_1\) and \(c_2\) are real for real initial conditions.

Key takeaways

A negative discriminant gives roots \(\alpha \pm \beta i\). Euler's formula, \(e^{i\theta} = \cos\theta + i\sin\theta\), turns the complex exponential solutions into the real pair \(e^{\alpha t}\cos\beta t\) and \(e^{\alpha t}\sin\beta t\). The general solution is \(e^{\alpha t}(c_1\cos\beta t + c_2\sin\beta t)\). Use the product rule for \(y'\). The sign of \(\alpha\) decides decay, growth or steady oscillation, and \(\beta\) sets how fast it oscillates, with zero crossings \(\pi/\beta\) apart.

Flashcards

Flashcard

Complex roots

r = α ± βi when b² − 4ac < 0

A conjugate pair.

Test your understanding

Question 1 of 5

What is the general solution of y″ + 6y′ + 25y = 0?

Question 2 of 5

What is the general solution of y″ + 16y = 0?

Question 3 of 5

Solve y″ + 2y′ + 2y = 0 with y(0) = 1 and y′(0) = 0.

Question 4 of 5

How do the nonzero solutions of y″ − y′ + 4y = 0 behave as t increases?

Question 5 of 5

Nonzero solutions of y″ + 2y′ + 10y = 0 cross zero repeatedly. How far apart in time are successive zero crossings?

← Previous topicNext topic →