Characteristic equations, real and complex roots, general solutions, initial value problems, and the algebraic structure that determines whether a system oscillates, decays, or grows
This chapter introduces second-order differential equations, with a focus on linear equations having constant coefficients. A first-order equation describes how a quantity changes based on its current value or the current independent variable. A second-order equation goes one level deeper: it describes how the rate of change itself is changing. This is the mathematics of acceleration, oscillation, circuits, beams, and dynamic systems with inertia.
Students who have studied first-order equations know how to work with equations of the form dy/dx = f(x, y). This chapter asks a new question: what happens when the equation involves y'' alongside y' and y? The answer leads to richer behavior, including oscillation and exponential decay, and to a powerful connection between algebraic roots and the shape of solutions.
By the end of this chapter, students should be able to classify second-order linear equations, form and solve characteristic equations, write general solutions for all root types, apply initial conditions to find particular solutions, and interpret what each solution form means physically and graphically. This chapter is foundational for everything that follows in Part III.
A mass hangs from a spring. You pull the mass downward and release it. Without air resistance or friction, the mass bobs up and down forever. With damping, it gradually slows. If you push on it periodically, it might oscillate more and more violently. All of this behavior, the oscillation, the decay, the resonance, is captured by a single second-order differential equation. Why does a spring system produce oscillation? Why does adding damping produce exponential decay? Why does the type of root in an algebraic equation determine which kind of behavior appears? This chapter answers those questions.
Second-order differential equations are among the most important equations in all of applied mathematics. They appear wherever a system has inertia, wherever a second derivative represents a physically meaningful quantity such as acceleration, curvature, or flux change rate. Newton's second law F = ma gives my'' = F, a second-order equation. The behavior of an RLC circuit involves both inductance and capacitance, giving a second-order equation for charge or current. The deflection of a beam under load satisfies a second-order equation. The Schrödinger equation in quantum mechanics is second order in space.
Students who understand second-order linear equations with constant coefficients will be prepared for all of these applications, and for the deeper study of mechanical vibrations, resonance, and forcing that follows in Chapters 7 and 8.
A second-order differential equation is an equation that involves the second derivative of an unknown function. More precisely, it is an equation in which the highest-order derivative appearing is the second derivative. If y is a function of x, then a second-order equation involves y'', and may also involve y', y, and the independent variable x.
| Equation | Second-Order? | Reason |
|---|---|---|
y'' + 3y' + 2y = 0 | Yes | Highest derivative is y'' |
y'' − 4y = 0 | Yes | Highest derivative is y'' |
y' + y = 0 | No | Highest derivative is y' (first-order) |
y''' + y = 0 | No | Highest derivative is y''' (third-order) |
Compare this to a first-order equation, which involves y' but not y''. A first-order equation describes how a quantity changes. A second-order equation describes how the rate of change itself changes, which is the description of acceleration. The order of a differential equation determines how many arbitrary constants appear in its general solution, and therefore how many initial or boundary conditions are needed to select a unique solution. A first-order equation has one arbitrary constant and needs one condition. A second-order equation has two arbitrary constants and needs two conditions.
The most natural physical interpretation of a second derivative is acceleration. If x(t) describes the position of an object at time t, then x'(t) is velocity and x''(t) is acceleration. Newton's second law states that force equals mass times acceleration: F = ma = m·x''(t). If the force F depends on position x, velocity x', and possibly time t, then Newton's law becomes a second-order differential equation for x(t).
| System | Differential Equation | Variables |
|---|---|---|
| Spring (Hooke's Law) | mx'' + kx = 0 | m = mass, k = spring constant |
| RLC Circuit | Lq'' + Rq' + (1/C)q = E(t) | L = inductance, R = resistance, C = capacitance |
| Pendulum (small angle) | θ'' + (g/L)θ = 0 | g = gravity, L = length |
The spring equation and the RLC circuit equation have the same mathematical structure. One involves a mass and a spring constant; the other involves inductance and capacitance. Because the equations are mathematically identical, they share the same solution methods and the same behavior types. This is one of the deep lessons of differential equations: different physical systems can share the same mathematical form.
A second-order differential equation is linear if it can be written in the form a₂(x)·y'' + a₁(x)·y' + a₀(x)·y = g(x), where the coefficient functions depend only on x, and the unknown function y and its derivatives appear only to the first power and are not multiplied together. The standard form divides through by a₂(x): y'' + P(x)·y' + Q(x)·y = f(x).
| Equation | Linear? | Reason |
|---|---|---|
y'' + 5y' − 6y = cos(x) | Yes | y and its derivatives appear to first power only |
y'' + y² = 0 | No | y appears squared |
y'' + yy' = 0 | No | y and y' are multiplied together |
y'' = sin(y) | No | y appears inside a nonlinear function |
For a linear homogeneous equation, if y₁ and y₂ are both solutions, then any linear combination c₁y₁ + c₂y₂ is also a solution. This principle is the foundation for the general solution structure. It applies only to linear equations — for nonlinear equations, the sum of two solutions is generally not a solution.
A second-order linear equation is homogeneous if the right-hand side is zero: y'' + P(x)·y' + Q(x)·y = 0. It is nonhomogeneous if the right-hand side is a nonzero function. This chapter focuses on homogeneous equations. Chapter 8 will address nonhomogeneous equations. A homogeneous equation describes a system free from external input or forcing, where behavior depends entirely on the system's own properties and initial conditions.
| Equation | Classification |
|---|---|
y'' + 4y' + 4y = 0 | Homogeneous — right side is zero |
y'' − y = eˣ | Nonhomogeneous — right side is eˣ |
y'' + 9y = 0 | Homogeneous |
y'' + 3y' = sin(x) | Nonhomogeneous — right side is sin(x) |
The most tractable second-order linear equations are those with constant coefficients: ay'' + by' + cy = 0, where a, b, and c are real constants with a ≠ 0. When coefficients are constant, there is a complete and elegant method for finding the general solution based on the characteristic equation, the main subject of this chapter.
Many real systems, to a first approximation, have properties that do not change over time or position. The spring constant of a steel spring is essentially the same whether the spring is at position 0 or 5. In such systems, the governing differential equation has constant coefficients. Beyond physical justification, constant-coefficient equations admit exact, explicit solutions, making them ideal for learning the conceptual structure of second-order equations before approaching the variable-coefficient case.
The key insight for solving ay'' + by' + cy = 0 is to look for solutions of the form y = eʳˣ, where r is a constant to be determined. Exponential functions have the remarkable property that their derivatives are proportional to themselves: if y = eʳˣ, then y' = reʳˣ and y'' = r²eʳˣ. Substituting into the equation and factoring out eʳˣ (which is never zero) gives the characteristic equation:
Finding a solution to the differential equation reduces to finding the roots of the characteristic equation, an ordinary quadratic in r, solvable by factoring or the quadratic formula: r = (−b ± √(b² − 4ac)) / 2a. The nature of the roots depends on the discriminant Δ = b² − 4ac.
| Discriminant | Root Type | General Solution Form |
|---|---|---|
Δ > 0 | Two distinct real roots r₁ ≠ r₂ | c₁eʳ¹ˣ + c₂eʳ²ˣ |
Δ = 0 | One repeated real root r | (c₁ + c₂x)eʳˣ |
Δ < 0 | Complex conjugate roots α ± βi | eᵅˣ(c₁cos βx + c₂sin βx) |
The characteristic equation does not involve x or y at all. It is purely algebraic. This is one of the beautiful features of constant-coefficient equations: the solution structure is determined entirely by the algebraic roots of a quadratic.
When Δ > 0, the characteristic equation has two distinct real roots r₁ and r₂. Each root gives a solution: y₁ = eʳ¹ˣ and y₂ = eʳ²ˣ. These two solutions are linearly independent; neither is a constant multiple of the other. By the superposition principle, the general solution is:
Characteristic equation: r² − 5r + 6 = 0
Factor: (r − 2)(r − 3) = 0
Roots: r₁ = 2, r₂ = 3
General solution: y = c₁e²ˣ + c₂e³ˣ
r² + 3r + 2 = (r + 1)(r + 2) = 0
Roots: r₁ = −1, r₂ = −2
y = c₁e⁻ˣ + c₂e⁻²ˣ
r² − 1 = (r − 1)(r + 1) = 0
Roots: r₁ = 1, r₂ = −1
y = c₁eˣ + c₂e⁻ˣ
When Δ = 0, the characteristic equation has exactly one real root of multiplicity two: r = −b/(2a). The exponential function y₁ = eʳˣ is a solution, but a second-order equation needs two linearly independent solutions. Using the method of reduction of order, it can be shown that a second independent solution is y₂ = xeʳˣ. The factor of x is not arbitrary — it emerges necessarily from the structure of the equation when there is a repeated root.
When you find a repeated root r, do not write y = c₁eʳˣ + c₂eʳˣ. This simplifies to (c₁ + c₂)eʳˣ, a solution with only one effective arbitrary constant. The second solution must be xeʳˣ, not another copy of eʳˣ.
r² − 4r + 4 = (r − 2)² = 0
Repeated root: r = 2
y = (c₁ + c₂x)e²ˣ
r² + 6r + 9 = (r + 3)² = 0
Repeated root: r = −3
y = (c₁ + c₂x)e⁻³ˣ
When Δ < 0, the characteristic equation has two complex conjugate roots: r = α ± βi, where α = −b/(2a) and β = √(4ac − b²)/(2a) > 0. The formal exponential solutions e^((α+βi)x) and e^((α−βi)x) are complex-valued. To obtain real-valued solutions, we apply Euler's formula: e^(iθ) = cos θ + i sin θ. Taking real and imaginary parts of e^((α+βi)x) = e^(αx)(cos βx + i sin βx) yields two real-valued, linearly independent solutions.
The factor e^(αx) controls the amplitude. The cosine and sine terms produce oscillation. Together: pure oscillation if α = 0, decaying oscillation if α < 0, and growing oscillation if α > 0. These three behaviors correspond directly to undamped, underdamped, and unstable oscillation.
r² + 9 = 0 → r² = −9 → r = ±3i
α = 0, β = 3
y = c₁cos(3x) + c₂sin(3x)
r² + 2r + 5 = 0
r = (−2 ± √(4 − 20)) / 2 = (−2 ± √(−16)) / 2
r = (−2 ± 4i) / 2 = −1 ± 2i
α = −1, β = 2
y = e⁻ˣ(c₁cos(2x) + c₂sin(2x))
r² − 2r + 5 = 0
r = (2 ± √(4 − 20)) / 2 = 1 ± 2i
α = 1, β = 2
y = eˣ(c₁cos(2x) + c₂sin(2x))
When roots are complex α ± βi, do not write y = c₁e^((α+β)x) + c₂e^((α−β)x). Complex roots produce cosine and sine solutions, not a sum of real exponentials. Always use Euler's formula to convert: the solution is y = e^(αx)(c₁cos βx + c₂sin βx). If α ≠ 0, the factor e^(αx) must appear in front of both the cosine and sine terms.
The three root cases give a complete classification of solutions for ay'' + by' + cy = 0.
| Root Type | Discriminant | Roots | General Solution |
|---|---|---|---|
| Real, distinct | b² − 4ac > 0 | r₁, r₂ real, r₁ ≠ r₂ | c₁eʳ¹ˣ + c₂eʳ²ˣ |
| Real, repeated | b² − 4ac = 0 | r = r₁ = r₂ | (c₁ + c₂x)eʳˣ |
| Complex conjugates | b² − 4ac < 0 | α ± βi | eᵅˣ(c₁cos βx + c₂sin βx) |
In each case, the two functions forming the general solution are linearly independent: neither is a constant multiple of the other. Linearly independent solutions form a fundamental set of solutions. A theorem in the theory of second-order linear equations guarantees that every solution can be written as some linear combination c₁y₁ + c₂y₂ of any fundamental set {y₁, y₂}. Knowing just two independent solutions is enough to capture all solutions.
The Wronskian of two functions y₁ and y₂ is W(y₁, y₂) = y₁y₂' − y₂y₁'. Two solutions are linearly independent if and only if their Wronskian is not identically zero on the interval of interest. Use the Wronskian to confirm that two solutions are genuinely independent.
W = y₁y₂' − y₂y₁'
= e²ˣ · 3e³ˣ − e³ˣ · 2e²ˣ
= 3e⁵ˣ − 2e⁵ˣ
= e⁵ˣ ≠ 0 for all x ✓
One of the most important skills in differential equations is interpreting what a solution tells us about the underlying system. For ay'' + by' + cy = 0, the roots of the characteristic equation determine the qualitative behavior of all solutions.
| Root Description | Long-Term Behavior | Physical Analogy |
|---|---|---|
| Real, both negative | Decays to zero; no oscillation | Overdamped spring |
| Real, both positive | Grows without bound | Unstable system |
| Real, mixed signs | Grows without bound (unless c₂ = 0) | Saddle equilibrium |
| Repeated, negative | Decays to zero | Critically damped spring |
| Repeated, positive | Grows without bound | Unstable critical |
| Complex, α < 0 | Decaying oscillation | Underdamped spring |
| Complex, α = 0 | Sustained oscillation | Ideal spring (no friction) |
| Complex, α > 0 | Growing oscillation | Unstable resonance |
(a) r² + 5r + 4 = 0 → Δ = 25 − 16 = 9 > 0
Roots: r = (−5 ± 3)/2 → r₁ = −1, r₂ = −4
Both negative. → Decays to zero. Overdamped.
(b) r² + 4 = 0 → r = ±2i, α = 0
→ Sustained oscillation. Simple harmonic motion.
(c) r² − 3r + 2 = 0 → Δ = 9 − 8 = 1 > 0
Roots: r = (3 ± 1)/2 → r₁ = 1, r₂ = 2
Both positive. → Grows without bound. Unstable.
(d) r² + 2r + 10 = 0 → Δ = 4 − 40 = −36 < 0
Roots: r = −1 ± 3i, α = −1 < 0
→ Decaying oscillation. Underdamped.
The general solution y = c₁y₁ + c₂y₂ contains two arbitrary constants. To find a unique solution, two conditions must be specified. For an initial value problem, these are typically y(x₀) = y₀ and y'(x₀) = y₀'. The procedure is: (1) find the general solution, (2) apply the first condition to produce one equation in c₁ and c₂, (3) differentiate and apply the second condition to produce a second equation, (4) solve the system, and (5) write the particular solution.
Apply y(0) = 2:
c₁e⁰ + c₂e⁰ = 2 → c₁ + c₂ = 2 ... (i)
Differentiate: y' = 2c₁e²ˣ + 3c₂e³ˣ
Apply y'(0) = 7:
2c₁ + 3c₂ = 7 ... (ii)
From (i): c₁ = 2 − c₂
Substitute into (ii): 2(2 − c₂) + 3c₂ = 7
4 + c₂ = 7 → c₂ = 3
c₁ = 2 − 3 = −1
Particular solution: y = −e²ˣ + 3e³ˣ
r² + 4 = 0 → r = ±2i
α = 0, β = 2
General solution: y = c₁cos(2x) + c₂sin(2x)
Apply y(0) = 3:
c₁cos(0) + c₂sin(0) = 3 → c₁ = 3
Differentiate: y' = −2c₁sin(2x) + 2c₂cos(2x)
Apply y'(0) = −2:
2c₂ = −2 → c₂ = −1
Particular solution: y = 3cos(2x) − sin(2x)
r² + 4r + 4 = (r + 2)² = 0 → r = −2 (repeated)
General solution: y = (c₁ + c₂x)e⁻²ˣ
Apply y(0) = 1:
(c₁ + 0)e⁰ = 1 → c₁ = 1
Differentiate: y' = c₂e⁻²ˣ − 2(c₁ + c₂x)e⁻²ˣ = [c₂ − 2c₁ − 2c₂x]e⁻²ˣ
Apply y'(0) = 0:
c₂ − 2c₁ = 0 → c₂ = 2c₁ = 2
Particular solution: y = (1 + 2x)e⁻²ˣ
This simplifies to (c₁ + c₂)e³ˣ, which has only one effective arbitrary constant. The second solution must be xe^(rx), not another copy of e^(rx). Correct form: y = (c₁ + c₂x)e³ˣ.
Complex roots produce oscillatory solutions involving cos and sin, not a sum of real exponentials. Correct form: y = e^(αx)(c₁cos βx + c₂sin βx).
A second-order equation needs two conditions. Always differentiate the general solution and apply y'(x₀) = y₀' to produce a second equation. Then solve the resulting 2×2 system for both constants.
r² + 4 = 0 gives r² = −4, so r = ±2i. These are purely imaginary roots with α = 0 and β = 2, giving y = c₁cos(2x) + c₂sin(2x), not y = c₁e²ˣ + c₂e⁻²ˣ.
If the real part α is nonzero, the factor e^(αx) must appear in front. Correct form for roots −1 ± 2i: y = e⁻ˣ(c₁cos(2x) + c₂sin(2x)).
The characteristic equation is formed by replacing y'' with r², y' with r, y with 1, and preserving all coefficient signs. The correct characteristic equation is r² − 4r + 4 = 0.
After computing c₁ and c₂, always substitute back into both y(x₀) and y'(x₀) to verify. Sign errors in the linear system are common, and checking catches them immediately.
This equation is second-order but nonlinear because (y')² is a nonlinear term. Check: do y, y', and y'' appear only to the first power and not multiplied together? If not, the equation is nonlinear.
Classify each equation as second-order or not, linear or nonlinear, and homogeneous or nonhomogeneous.
y'' + 3y' + 2y = 0y''' + y = 0y'' + y² = 0y'' − 4y = eˣy'' + yy' = 0y'' + 5y' − 6y = cos(x)y'' = y(y'')² + y = 0y'' + p(x)y' + q(x)y = 0y'' − 3y' = 0Form the characteristic equation, find all roots, and identify the root type.
y'' + 5y' + 6y = 0y'' − 4y = 0y'' + 4y' + 4y = 0y'' + y = 0y'' − 2y' − 3y = 0y'' + 4y' + 13y = 0y'' + 2y' + y = 0y'' − 6y' + 9y = 0y'' − y' − 6y = 0y'' + 6y' + 10y = 04y'' − 4y' + y = 0y'' + 2y' + 5y = 0y'' − 4y' + 4y = 02y'' + 3y' − 2y = 0y'' + 9y = 0Find the general solution of each equation.
y'' + 3y' + 2y = 0y'' − 4y = 0y'' + 4y' + 4y = 0y'' + 4y = 0y'' − 2y' − 8y = 0y'' + 2y' + 5y = 0y'' − 6y' + 9y = 0y'' + 6y' + 8y = 0y'' + 6y' + 10y = 0y'' − 5y' + 6y = 0y'' + 4y' + 13y = 09y'' − 6y' + y = 0y'' − 4y' + 5y = 0y'' + y = 04y'' + 12y' + 9y = 0Solve each initial value problem. Verify both initial conditions after finding constants.
y'' − 5y' + 6y = 0, y(0) = 1, y'(0) = 0y'' + 4y = 0, y(0) = 2, y'(0) = 4y'' + 4y' + 4y = 0, y(0) = 3, y'(0) = 1y'' + 2y' + 5y = 0, y(0) = 0, y'(0) = 3y'' − 4y' + 4y = 0, y(0) = 2, y'(0) = −1y'' + y' − 6y = 0, y(0) = 1, y'(0) = 1y'' + 9y = 0, y(0) = 0, y'(0) = 6y'' − 2y' − 3y = 0, y(0) = 1, y'(0) = −1y'' + 6y' + 9y = 0, y(0) = 2, y'(0) = 0y'' − 4y' + 13y = 0, y(0) = 1, y'(0) = 2For each root description, state the qualitative long-term behavior of all solutions.
Each of the following contains an error. Identify the mistake and write the correct solution.
Identify the error. Write the correct general solution.
Check the computation of β carefully. Is the final answer correct?
What step did the student skip? Complete the solution correctly.
Explain the error and write the correct general solution.
What is missing from the general solution? Write the correct form.
y'' + 3yy' = 0 linear or nonlinear? Explain why.y'' + 5y' + 6y = eˣ homogeneous or nonhomogeneous? Why?y'' + y' = 0? How many arbitrary constants does its general solution contain?2y'' + y' − y = 0 and find its roots. Identify the root type.y'' + 6y' + 9y = 0 and identify the root type.y'' + 2y' + 10y = 0 and find the roots in the form α ± βi.y'' − y' − 6y = 0.y'' + 10y' + 25y = 0.y'' − 4y' + 8y = 0.y'' − 3y' + 2y = 0, y(0) = 0, y'(0) = 1. Verify both initial conditions.y'' + 4y = 0, y(0) = 3, y'(0) = 0. What is the amplitude of the oscillation?y'' + 4y' + 4y = 0, y(0) = 2, y'(0) = 0. Describe the long-term behavior.y'' + 8y' + 15y = 0 without solving. Justify using the characteristic roots.Chapter 7 uses every idea developed in this chapter to analyze real physical systems. The equation for a mass-spring-dashpot system is mx'' + bx' + kx = F(t). With F(t) = 0, this is a homogeneous constant-coefficient equation whose characteristic equation is mr² + br + k = 0. The discriminant b² − 4mk determines whether the system is overdamped (b² > 4mk), critically damped (b² = 4mk), or underdamped (b² < 4mk). Students who understand the three root cases from Chapter 6 will immediately see which physical behavior each case corresponds to. Chapter 7 will also introduce amplitude, period, and frequency, the relationship between complex roots and oscillation rate, and the phenomenon of resonance, when forcing frequency matches natural frequency. All of these depend on the solution framework built here.