Learn faster with bite‑sized practice that actually sticks.

Second-Order Linear ODEs (Constant Coefficients): your friendly pattern-recognition lesson

Second-order linear ODEs with constant coefficients look intimidating… until you realize they’re basically a matching game.

We’ll focus on equations like:
$ay'' + by' + cy = 0$
where $a,b,c$ are constants.

The big idea: try solutions shaped like exponentials. Exponentials are “shape-stable” under derivatives, which makes them perfect for this job.

1) The characteristic equation: turning calculus into algebra

Try:
$y = e^{rt}$
Then:
$y' = re^{rt}, \quad y'' = r^2 e^{rt}$
Plug into:
$ay''+by'+cy=0$
You get:
$a r^2 e^{rt} + b r e^{rt} + c e^{rt} = 0$
Factor out $e^{rt}$ (never zero):
$a r^2 + b r + c = 0$
That’s the characteristic equation.

So the calculus problem becomes: solve a quadratic.

Three root “personalities”

Depending on the discriminant $\Delta = b^2 - 4ac,$ you get three behaviors:

A) Two real distinct roots: $r_1 \neq r_2$

$y(t)=C_1 e^{r_1 t} + C_2 e^{r_2 t}$
Intuition: two independent exponential modes (growth/decay at different rates).

B) One repeated real root: $r_1 = r_2 = r$

$y(t) = (C_1 + C_2 t)e^{rt}$
Why the extra $t$? Because you need two linearly independent solutions for a second-order equation, and $e^{rt}$ alone can’t do it twice.

C) Complex conjugate roots: $r = \alpha \pm i\beta$

$y(t)=e^{\alpha t}\left(C_1\cos(\beta t)+C_2\sin(\beta t)\right)$
Intuition: oscillation ($\cos,\sin$) with an envelope $e^{\alpha t}$.

• If $\alpha<0$: decaying oscillation
• If $\alpha=0$: pure oscillation
• If $\alpha>0$: growing oscillation

2) Complex exponentials: the ultimate shortcut (and why it’s not “cheating”)

When roots are complex, people often ask: “Where did sine and cosine come from?”

They come from this identity (Euler’s formula):
$e^{i\beta t} = \cos(\beta t) + i\sin(\beta t)$
So if the characteristic roots are $\alpha \pm i\beta$, the exponential solutions are:
$e^{(\alpha+i\beta)t} = e^{\alpha t}(\cos \beta t + i\sin \beta t)$
$e^{(\alpha-i\beta)t} = e^{\alpha t}(\cos \beta t - i\sin \beta t)$

Now here’s the magic: real solutions are made by taking real and imaginary parts.
So instead of guessing trig forms, you can:

1. solve the characteristic equation normally,
2. write down $e^{rt}$ even if $r$ is complex,
3. convert to real sines/cosines at the end.

Why this is useful: differentiating $e^{rt}$ is always easy, even when $r$ is complex. Trig shows up automatically via Euler.

3) Superposition: why we’re allowed to add solutions

These ODEs are linear and homogeneous.
That means if $y_1$ solves the equation and $y_2$ solves the equation, then:
$y = C_1 y_1 + C_2 y_2$
also solves it.

That’s the superposition principle: solutions form a “mix-and-match” family.

The key requirement: linear independence

A second-order ODE needs two independent solution shapes.

• $e^{r_1 t}$ and $e^{r_2 t}$ are independent if $r_1\neq r_2$.
• $e^{rt}$ and $t e^{rt}$ are independent when roots repeat.
• $\cos(\beta t)$ and $\sin(\beta t)$ are independent.

Intuition: you need two “knobs” ($C_1, C_2$) to match two initial conditions like $y(0)$ and $y'(0)$.

4) A bridge to QM-flavored forms: oscillatory vs decaying

In quantum mechanics (and waves generally), two ultra-famous ODEs appear again and again.

A) Oscillatory: $y'' + k^2 y = 0$

Characteristic equation:
$r^2 + k^2 = 0 \quad\Rightarrow\quad r = \pm i k$
So:
$y(t) = C_1\cos(kt) + C_2\sin(kt)$

Intuition: derivatives of sine/cosine keep cycling—nothing “pushes” the solution to blow up or die out.

ASCII sketch (pure oscillation):

In QM language, oscillations often correspond to traveling/standing waves in classically allowed regions.

B) Decaying/growing: $y'' - \kappa^2 y = 0$

Characteristic equation:
$r^2 - \kappa^2 = 0 \quad\Rightarrow\quad r = \pm \kappa$
So:
$y(t) = C_1 e^{\kappa t} + C_2 e^{-\kappa t}$

Intuition: one exponential wants to explode, the other wants to fade away.

ASCII sketch (decay):

In QM, decaying exponentials show up in classically forbidden regions (like tunneling), where the wavefunction often must stay finite, so the growing exponential gets rejected by boundary conditions.

Misconceptions (tiny potholes to avoid)

“Is it $k$ or $k^2$? I keep mixing them up.”

In $y'' + k^2 y = 0,$ the characteristic equation is $r^2 + k^2 = 0,$ so roots are $r=\pm ik.$
That means the oscillation frequency is $k$, not $k^2$.

“I found one solution. Isn’t that enough?”

Not for a second-order ODE. You need two linearly independent solutions to cover all initial conditions.

“I forgot the constants $C_1, C_2$.”

Those constants aren’t decoration—they’re the dials that let you fit the initial conditions.

“Complex solutions aren’t physical, so I shouldn’t use them.”

Complex exponentials are often just a computational tool. Real solutions come from taking real/imaginary parts (or linear combinations) at the end.

Takeaway (the comforting pattern)

Second-order constant-coefficient ODEs are all about one move:

1. assume $y=e^{rt}$
2. solve the characteristic quadratic for $r$
3. translate roots into shapes:

• real $r$ → exponentials
• repeated $r$ → add a $t$
• complex $\alpha\pm i\beta$ → sines/cosines with envelope $e^{\alpha t}$

And when you meet $y''+k^2 y=0$ versus $y''-\kappa^2 y=0,$ you can immediately think:
oscillate vs decay/grow—a theme that shows up everywhere from springs to quantum waves.