Differential Equations › Section C · Second order differential equations › Topic C.4

Repeated roots

When the characteristic equation has a single repeated root, the exponential method hands us only one solution. A factor of t supplies the missing second one, and this case turns out to mark the exact boundary between oscillating and non-oscillating behaviour.

One root, one solution

If \(b^2 - 4ac = 0\), the characteristic equation \(ar^2 + br + c = 0\) has the single root \(r = -\frac{b}{2a}\), counted twice. For \(y'' + 6y' + 9y = 0\), the equation is \((r + 3)^2 = 0\), so \(r = -3\) and \(e^{-3t}\) is a solution. The trouble is that the exponential method gives nothing else. Writing \(c_1e^{-3t} + c_2e^{-3t}\) does not help, because it collapses to \((c_1 + c_2)e^{-3t}\), a single constant times one function. A second order equation needs two genuinely different solutions, so we have to find another one by a different route.

Finding the second solution

Try multiplying the known solution by an unknown function: \(y = v(t)e^{rt}\). The product rule gives \(y' = (v' + rv)e^{rt}\) and \(y'' = (v'' + 2rv' + r^2v)e^{rt}\). Substituting and collecting terms,

ay'' + by' + cy = e^{rt}\big[av'' + (2ar + b)v' + (ar^2 + br + c)v\big]

The last bracket is zero because \(r\) is a root, and \(2ar + b = 0\) because \(r = -\frac{b}{2a}\). Only \(av''e^{rt}\) survives, so the equation requires \(v'' = 0\). Hence \(v = c_1 + c_2t\), and the new solution is \(te^{rt}\). This trick of multiplying by \(v(t)\) is developed in general in the next topic, "Reduction of order".

The general solution

For a repeated root \(r\), the general solution is

y = c_1e^{rt} + c_2te^{rt} = (c_1 + c_2t)e^{rt}

The two functions are not constant multiples of each other, because their ratio is \(t\), which is not constant. There is also an intuitive way to see where \(t\) comes from. If two distinct roots \(r\) and \(r + h\) are very close, then \(\frac{e^{(r + h)t} - e^{rt}}{h}\) solves the equation with those two roots, by superposition. As \(h \to 0\) this difference quotient becomes the derivative of \(e^{rt}\) with respect to \(r\), which is \(te^{rt}\). The simplest case is \(y'' = 0\): the root 0 is repeated, and the solution \(c_1 + c_2t\) is every straight line.

Worked example: a quick return

Solve \(y'' + 6y' + 9y = 0\) with \(y(0) = 2\) and \(y'(0) = -1\). 1. The characteristic equation is \((r + 3)^2 = 0\), so \(r = -3\) twice. 2. General solution: \(y = (c_1 + c_2t)e^{-3t}\). 3. \(y(0) = c_1 = 2\). 4. Product rule: \(y' = \big(c_2 - 3c_1 - 3c_2t\big)e^{-3t}\), so \(y'(0) = c_2 - 3c_1 = -1\), which gives \(c_2 = 5\). 5. Answer: \(y = (2 + 5t)e^{-3t}\). 6. Check: \(y'(0) = 5 - 6 = -1\). The solution decays to zero without ever crossing it, because \(2 + 5t > 0\) for all \(t \ge 0\).

Worked example: a fractional root

Solve \(4y'' - 4y' + y = 0\) with \(y(0) = 4\) and \(y'(0) = 0\). 1. The characteristic equation \(4r^2 - 4r + 1 = (2r - 1)^2 = 0\) gives \(r = \frac{1}{2}\) twice. 2. General solution: \(y = (c_1 + c_2t)e^{t/2}\). 3. \(y(0) = c_1 = 4\). 4.

y' = \big(c_2 + \tfrac{1}{2}c_1 + \tfrac{1}{2}c_2t\big)e^{t/2}

so \(y'(0) = c_2 + 2 = 0\) and \(c_2 = -2\). 5. Answer: \(y = (4 - 2t)e^{t/2}\). 6. Check: \(y'(0) = -2 + 2 = 0\). Here the root is positive, so the solution crosses zero once, at \(t = 2\), then grows ever more negative.

How these solutions behave

With a repeated root \(r < 0\), every solution tends to 0, even though the factor \(t\) grows: an exponential decay always beats a linear growth in the long run. If \(r > 0\), nonzero solutions grow without bound. Because \(c_1 + c_2t\) is a straight line, a nonzero solution can cross zero at most once, so there is no oscillation. A solution can, however, rise before it decays. With \(y(0) = 0\) and \(y'(0) = 1\), the equation \(y'' + 6y' + 9y = 0\) gives \(y = te^{-3t}\). Setting \(y' = (1 - 3t)e^{-3t} = 0\) shows a single peak at \(t = \frac{1}{3}\), of height \(\frac{1}{3e} \approx 0.12\).

The boundary case

A repeated root sits exactly between the other two cases. Hold \(a > 0\) and \(b\) fixed and increase \(c\) through \(\frac{b^2}{4a}\). Below that value the discriminant is positive and the roots are real and distinct; above it they are complex and the solutions oscillate. At the value itself the two real roots merge into one, just before they split into a complex pair. In a damped spring this boundary is called critical damping, and the topic "Mechanical vibrations" shows why engineers aim for it: a critically damped system settles quickly without overshooting back and forth.

Common mistakes to avoid

First, writing \(c_1e^{rt} + c_2e^{rt}\), which has only one effective constant. Second, putting the factor \(t\) in the wrong place, such as \(e^{rt^2}\) or \(e^{rt} + t\); the second solution is the product \(te^{rt}\). Third, forgetting the product rule, so that the derivative of \(te^{-3t}\) is taken to be \(-3te^{-3t}\) instead of \((1 - 3t)e^{-3t}\). Fourth, missing a repeated root because the leading coefficient is not 1: \(4r^2 - 4r + 1\) is a perfect square. Fifth, assuming that the factor \(t\) makes decaying solutions grow.

Key takeaways

When \(b^2 - 4ac = 0\), the root \(r = -\frac{b}{2a}\) is repeated and only \(e^{rt}\) comes from the exponential method. Substituting \(y = v(t)e^{rt}\) forces \(v'' = 0\), which gives the second solution \(te^{rt}\). The general solution is \((c_1 + c_2t)e^{rt}\). Use the product rule for \(y'\). For \(r < 0\), solutions decay with at most one zero crossing and possibly one peak; for \(r > 0\) they grow. The repeated root is the boundary between real roots and oscillating complex roots.

Flashcards

Flashcard

Repeated root

r = −b/(2a) when b² − 4ac = 0

The quadratic is a perfect square.

Test your understanding

Question 1 of 5

What is the general solution of y″ − 10y′ + 25y = 0?

Question 2 of 5

Solve y″ + 4y′ + 4y = 0 with y(0) = 1 and y′(0) = 0.

Question 3 of 5

For a repeated root r, why is c₁e^(rt) + c₂e^(rt) not the general solution?

Question 4 of 5

Which equation has a repeated root?

Question 5 of 5

The solution y = te^(−4t) starts at 0 and rises. At what time does it reach its largest value?

← Previous topicNext topic →