Second-Order Homogeneous Equations

Characteristic equations and the nature of solutions

This is where physics truly lives. Newton's second law, F=maF = ma, relates force to acceleration, which is the second derivative of position. Springs, pendulums, electrical circuits, vibrating strings: the equations that govern them all involve second derivatives. Understanding second-order differential equations unlocks the mathematics of motion and oscillation.

The Standard Form

A second-order linear homogeneous equation with constant coefficients has the form:

ay+by+cy=0ay'' + by' + cy = 0

where aa, bb, and cc are constants, and a0a \neq 0. The term homogeneous means the right side is zero: there is no external forcing function. The term linear means yy, yy', and yy'' appear only to the first power, with no products like yyy \cdot y'.

This seemingly simple form describes an enormous range of physical phenomena. A mass on a spring, a swinging pendulum with friction, the current in an RLC circuit: all are modeled by equations of exactly this type.

The Exponential Guess

From first-order equations, we know that y=ekty = e^{kt} solves y=kyy' = ky. The exponential function has a remarkable property: its derivative is proportional to itself. This suggests trying y=erty = e^{rt} as a solution to our second-order equation.

If y=erty = e^{rt}, then:

  • y=rerty' = re^{rt}
  • y=r2erty'' = r^2 e^{rt}

Substituting into ay+by+cy=0ay'' + by' + cy = 0:

ar2ert+brert+cert=0ar^2 e^{rt} + br e^{rt} + c e^{rt} = 0

Factor out erte^{rt}, which is never zero:

ert(ar2+br+c)=0e^{rt}(ar^2 + br + c) = 0

Since ert0e^{rt} \neq 0, we need:

ar2+br+c=0ar^2 + br + c = 0

This is the characteristic equation. The values of rr that satisfy it determine our solutions. What started as a differential equation has become a simple quadratic.

The Characteristic Equation

The characteristic equation ar2+br+c=0ar^2 + br + c = 0 is solved using the quadratic formula:

r=b±b24ac2ar = \frac{-b \pm \sqrt{b^2 - 4ac}}{2a}

The expression under the square root, b24acb^2 - 4ac, is the discriminant. It determines the nature of the roots and, consequently, the character of the solutions.

Interactive: Explore the Characteristic Equation

Characteristic equation:

1.0r2+(3.0)r+(2.0)=01.0r^2 + (-3.0)r + (2.0) = 0

Discriminant: b24ac=1.00b^2 - 4ac = 1.00 (two distinct real roots)

General solution:

y=C1e2.00t+C2e1.00ty = C_1 e^{2.00t} + C_2 e^{1.00t}

Adjust the coefficients aa, bb, and cc to see how the roots move in the complex plane. When the discriminant is positive, both roots lie on the real axis. When negative, the roots become a complex conjugate pair. When the discriminant equals zero, the two roots coincide.

Case 1: Distinct Real Roots

When b24ac>0b^2 - 4ac > 0, we get two distinct real roots r1r_1 and r2r_2. Each gives a solution:

y1=er1tandy2=er2ty_1 = e^{r_1 t} \quad \text{and} \quad y_2 = e^{r_2 t}

These solutions are linearly independent: neither is a constant multiple of the other. This means they span a two-dimensional solution space. The general solution is any linear combination:

y=C1er1t+C2er2ty = C_1 e^{r_1 t} + C_2 e^{r_2 t}

The constants C1C_1 and C2C_2 are determined by initial conditions.

Interactive: Distinct Real Roots

Roots

Constants

y(t)=2.0e0.5t+1.0e2.0ty(t) = 2.0 e^{-0.5t} + -1.0 e^{-2.0t}

The purple curve is the sum of the blue and red exponential components.

Watch how the solution is built from two exponential components. When both roots are negative, solutions decay to zero. When one root is positive, that component eventually dominates, causing the solution to grow without bound.

Case 2: Repeated Roots

When b24ac=0b^2 - 4ac = 0, the quadratic has a single repeated root:

r=b2ar = -\frac{b}{2a}

Now we have a problem. The function y1=erty_1 = e^{rt} is a solution, but where is the second independent solution? Simply using erte^{rt} twice does not give us two independent functions.

The answer is unexpected: the second solution is y2=terty_2 = te^{rt}.

Let us verify this. If y=terty = te^{rt}, then:

  • y=ert+rtert=ert(1+rt)y' = e^{rt} + rte^{rt} = e^{rt}(1 + rt)
  • y=rert(1+rt)+rert=rert(2+rt)y'' = re^{rt}(1 + rt) + re^{rt} = re^{rt}(2 + rt)

Let us verify this works. Substituting into ay+by+cyay'' + by' + cy:

arert(2+rt)+bert(1+rt)+cterta \cdot re^{rt}(2 + rt) + b \cdot e^{rt}(1 + rt) + c \cdot te^{rt}

Factoring out erte^{rt}:

ert[2ar+b+t(ar2+br+c)]e^{rt}[2ar + b + t(ar^2 + br + c)]

Since rr satisfies the characteristic equation, ar2+br+c=0ar^2 + br + c = 0. And since r=b/(2a)r = -b/(2a), we have 2ar+b=2a(b/(2a))+b=02ar + b = 2a(-b/(2a)) + b = 0. Both terms vanish, confirming tertte^{rt} is indeed a solution.

The general solution for repeated roots is:

y=(C1+C2t)erty = (C_1 + C_2 t)e^{rt}

Interactive: Repeated Root Solutions

y(t)=(1.0+1.0t)e1.0ty(t) = (1.0 + 1.0t) e^{-1.0t}

Blue: 1.0e1.0t1.0e^{-1.0t} Red: 1.0te1.0t1.0t \cdot e^{-1.0t}

With a repeated root, the second independent solution is tertte^{rt}, not just another exponential.

The factor tertte^{rt} starts at zero and initially grows linearly (when r<0r < 0, it eventually decays as the exponential takes over). This polynomial-times-exponential form appears whenever roots repeat.

Why the tertte^{rt} Solution?

There is an elegant way to understand where tertte^{rt} comes from. Consider what happens as two distinct roots r1r_1 and r2r_2 approach each other.

The general solution is:

y=C1er1t+C2er2ty = C_1 e^{r_1 t} + C_2 e^{r_2 t}

Suppose we want a solution with y(0)=0y(0) = 0 and y(0)=1y'(0) = 1. This gives C1=C2=1/(r1r2)C_1 = -C_2 = 1/(r_1 - r_2), so:

y=er1ter2tr1r2y = \frac{e^{r_1 t} - e^{r_2 t}}{r_1 - r_2}

Now let r1rr_1 \to r and r2rr_2 \to r. This is a 0/00/0 indeterminate form. By L'Hopital's rule (differentiating with respect to r1r_1):

limr1rer1tertr1r=tert\lim_{r_1 \to r} \frac{e^{r_1 t} - e^{r t}}{r_1 - r} = te^{rt}

The tertte^{rt} solution emerges naturally as the limit when two independent exponential solutions merge into one.

Linear Independence and the General Solution

A second-order equation has a two-dimensional solution space. To span this space, we need two linearly independent solutions. Two functions y1y_1 and y2y_2 are linearly independent if the only constants satisfying C1y1+C2y2=0C_1 y_1 + C_2 y_2 = 0 for all tt are C1=C2=0C_1 = C_2 = 0.

For distinct real roots, er1te^{r_1 t} and er2te^{r_2 t} are independent because exponentials with different exponents cannot cancel each other.

For a repeated root, erte^{rt} and tertte^{rt} are independent. If C1ert+C2tert=0C_1 e^{rt} + C_2 te^{rt} = 0 for all tt, then at t=0t = 0 we get C1=0C_1 = 0, and then C2tert=0C_2 te^{rt} = 0 for all tt implies C2=0C_2 = 0.

This concept connects to linear algebra: the solutions form a vector space, and any two independent solutions form a basis for that space.

The Wronskian: A systematic way to test linear independence uses the Wronskian, defined for two functions y1y_1 and y2y_2 as:

W(y1,y2)=y1y2y1y2=y1y2y1y2W(y_1, y_2) = y_1 y_2' - y_1' y_2 = \begin{vmatrix} y_1 & y_2 \\ y_1' & y_2' \end{vmatrix}

Two solutions to a second-order equation are linearly independent if and only if their Wronskian is nonzero. For example, with y1=er1ty_1 = e^{r_1 t} and y2=er2ty_2 = e^{r_2 t}:

W=er1tr2er2tr1er1ter2t=(r2r1)e(r1+r2)t0W = e^{r_1 t} \cdot r_2 e^{r_2 t} - r_1 e^{r_1 t} \cdot e^{r_2 t} = (r_2 - r_1)e^{(r_1 + r_2)t} \neq 0

The Wronskian will appear again when we study nonhomogeneous equations and variation of parameters.

Why Two Initial Conditions?

For a first-order equation, specifying y(0)y(0) determines the solution uniquely. For a second-order equation, specifying y(0)y(0) is not enough. We also need y(0)y'(0).

The reason is geometric. Knowing the position of a particle tells you where it is, but not where it is going. Two particles at the same position but moving with different velocities will follow different trajectories.

Interactive: The Effect of Initial Conditions

Equation: y+3y+2y=0y'' + 3y' + 2y = 0 with roots r1=1,r2=2r_1 = -1, r_2 = -2

Initial conditions: y(0)=2.0,  y(0)=0.0y(0) = 2.0, \; y'(0) = 0.0

Constants: C1=4.00,  C2=2.00C_1 = 4.00, \; C_2 = -2.00

Each choice of initial conditions picks exactly one solution from the family. The red arrow shows the initial slope.

The general solution has two free constants, C1C_1 and C2C_2. Two initial conditions provide two equations to solve for these constants. Specifying both y(0)y(0) and y(0)y'(0) picks out exactly one solution from the infinite family.

Interactive: Why Two Conditions Are Needed

Knowing only y(0)=1.0y(0) = 1.0 leaves infinitely many solutions passing through that point. The gray curves all satisfy the same initial position.

Knowing only the initial position leaves an entire family of solutions passing through that point. Adding the initial velocity narrows it down to exactly one.

Solving Initial Value Problems

Here is the procedure for solving ay+by+cy=0ay'' + by' + cy = 0 with y(0)=y0y(0) = y_0 and y(0)=v0y'(0) = v_0:

  1. Write the characteristic equation: ar2+br+c=0ar^2 + br + c = 0
  2. Find the roots using the quadratic formula
  3. Write the general solution based on the root type
  4. Apply initial conditions to find C1C_1 and C2C_2

Example: Solve y+3y+2y=0y'' + 3y' + 2y = 0 with y(0)=1y(0) = 1 and y(0)=0y'(0) = 0.

The characteristic equation is r2+3r+2=0r^2 + 3r + 2 = 0, which factors as (r+1)(r+2)=0(r + 1)(r + 2) = 0.

Roots: r1=1r_1 = -1, r2=2r_2 = -2 (distinct).

General solution: y=C1et+C2e2ty = C_1 e^{-t} + C_2 e^{-2t}

Apply initial conditions:

  • y(0)=C1+C2=1y(0) = C_1 + C_2 = 1
  • y(0)=C12C2=0y'(0) = -C_1 - 2C_2 = 0

Solving: C1=2C_1 = 2, C2=1C_2 = -1

Final answer: y=2ete2ty = 2e^{-t} - e^{-2t}

Physical Interpretation

The sign of the roots determines long-term behavior:

  • Both roots negative: Solutions decay to zero. This is overdamped motion, like a mass in thick honey returning to equilibrium.
  • Both roots positive: Solutions grow without bound. This is unstable, like a pencil balanced on its tip.
  • One positive, one negative: The positive root dominates at large tt, causing growth.
  • Repeated negative root: Solutions still decay, but the tertte^{rt} term causes them to initially grow before decaying. This is critically damped motion.

The discriminant tells us whether the system is overdamped (distinct roots) or critically damped (repeated root). In the next chapter, we will see what happens when the discriminant is negative: the roots become complex, and solutions oscillate.

Looking Ahead

We have handled the cases where the characteristic equation has real roots. But what happens when b24ac<0b^2 - 4ac < 0? The quadratic formula gives complex numbers. In the next chapter, Complex Roots and Oscillations, we will see that complex roots produce sinusoidal oscillations, connecting second-order equations to waves, vibrations, and the rhythms of the physical world.

Key Takeaways

  • Second-order homogeneous equations ay+by+cy=0ay'' + by' + cy = 0 arise throughout physics, particularly in mechanics and circuits
  • The characteristic equation ar2+br+c=0ar^2 + br + c = 0 determines the form of solutions
  • Distinct real roots r1r2r_1 \neq r_2 give y=C1er1t+C2er2ty = C_1 e^{r_1 t} + C_2 e^{r_2 t}
  • Repeated roots give y=(C1+C2t)erty = (C_1 + C_2 t)e^{rt}, where tertte^{rt} is the second independent solution
  • Two initial conditions (y(0)y(0) and y(0)y'(0)) are needed because the solution space is two-dimensional
  • The discriminant b24acb^2 - 4ac classifies solutions: positive for distinct roots, zero for repeated roots, negative for complex roots (next chapter)