Complex Eigenvalues & Real Solutions — Complex Eigenvalues and Real Solutions (Intro) - Linear Algebra

SECTION 1

Historical Context & Motivation

For centuries, mathematicians struggled with equations that had no "real" answers. When you try to solve x² = −1, no ordinary number works. In the 1500s, Italian mathematicians began using a strange new idea: imaginary numbers (numbers involving the square root of −1). At first, people thought these were useless tricks. But over time, imaginary and complex numbers (which combine a real part and an imaginary part) turned out to be incredibly powerful — especially in understanding how things spin, oscillate, and vibrate.

Meanwhile, in the 1800s, mathematicians like Cauchy and Cayley developed the theory of matrices — rectangular grids of numbers that can describe transformations such as stretching, flipping, and rotating shapes. When mathematicians asked, "What are the special directions of a matrix?" they discovered eigenvalues and eigenvectors. Sometimes these eigenvalues are ordinary real numbers. But sometimes, especially when the matrix involves rotation, the eigenvalues turn out to be complex numbers. This lesson introduces that fascinating case.

ℹ

Prerequisites

Before starting, you should be comfortable with: (1) multiplying 2×2 matrices, (2) computing a 2×2 determinant, (3) complex numbers of the form a + bi, (4) the quadratic formula, and (5) a basic idea of what a differential equation is. If any of these topics are unfamiliar, review them before continuing.

1545 • Cardano Uses Imaginary Numbers

Gerolamo Cardano first uses the square root of negative numbers to solve cubic equations, calling them "as subtle as they are useless."

1748 • Euler's Formula

Leonhard Euler discovers that e^iθ = cos θ + i sin θ, linking complex numbers to circles and rotation.

1829 • Cauchy and Eigenvalue Theory

Augustin-Louis Cauchy develops the theory of eigenvalues for matrices, showing how they reveal a matrix's fundamental behavior.

1855 • Cayley Formalizes Matrix Algebra

Arthur Cayley publishes a memoir formalizing matrix multiplication, making it easier to study eigenvalues systematically.

1900s • Modern Applications Emerge

Complex eigenvalues become central to quantum mechanics, electrical engineering, and computer graphics — anywhere rotation and oscillation appear.

So here is the central question this lesson addresses: When a matrix has complex eigenvalues, what does that mean, and how do we still get real-number solutions? You will see that complex eigenvalues always come in pairs and that they describe rotation. Even though the eigenvalues themselves involve imaginary numbers, the final answer for a real-world system stays completely real.

SECTION 2

Core Principles & Definitions

Before diving in, let's build up the key ideas one at a time. This lesson requires fluency with matrix determinants, complex number arithmetic, the quadratic formula applied to characteristic equations, and a basic understanding of differential equations. We will define each concept as it appears, but prior exposure to these ideas is strongly recommended.

Complex Numbers

A complex number has the form a + bi, where a is the real part, b is the imaginary part, and i = √(−1). Example: 3 + 2i.

Eigenvalues (Quick Review)

An eigenvalue λ of a matrix A is a number such that Av = λv for some nonzero vector v. We find eigenvalues by solving det(A − λI) = 0.

Conjugate Pairs

When a matrix has all real entries and one eigenvalue is a + bi, the other eigenvalue is always a − bi. These are called complex conjugates.

Rotation Connection

Complex eigenvalues signal that the matrix rotates vectors rather than simply stretching them. The angle of rotation and the amount of stretching are both encoded in the eigenvalue.

Real Solutions from Complex Eigenvalues

Even with complex eigenvalues, we can extract real-valued solutions by using the real and imaginary parts of the complex eigenvector. The imaginary pieces cancel out. This technique is central to solving systems of differential equations.

✦

KEY TAKEAWAY

Think of complex eigenvalues like the settings on a merry-go-round. One number (the real part) controls how fast the ride speeds up or slows down, and the other number (the imaginary part) controls how fast it spins. Together, they describe spiraling motion — and even though "imaginary" sounds fake, the spiral you see in real life is very real!

SECTION 3

Visual Explanation — The Geometry of Complex Eigenvalues

When a 2×2 matrix has real eigenvalues, it stretches or compresses vectors along certain fixed directions. But when the eigenvalues are complex, something fundamentally different happens: the matrix rotates every vector. No vector stays pointing in the same direction after the transformation.

Left: With real eigenvalues, the vector v gets stretched into Av along the same line. Right: With complex eigenvalues, the vector v gets rotated by an angle θ into Av. The dashed circle shows the path of rotation. No vector stays pointing in the same direction.

In the left panel above, you see a matrix with real eigenvalues. The pink vector v gets mapped to the cyan dashed vector Av, which points in the same direction — just longer. In the right panel, the matrix has complex eigenvalues. The pink vector v gets rotated to the green vector Av. Notice the angle θ between them. The dashed purple circle shows the "orbit" that vectors follow. This rotation is exactly what complex eigenvalues describe.

ℹ

Why "No Fixed Direction"?

When eigenvalues are complex, no real vector satisfies Av = λv. That's because the eigenvectors themselves have complex entries. To get real solutions, we split the complex eigenvector into its real and imaginary parts.

SECTION 4

Mathematical Framework

Let's see how complex eigenvalues show up mathematically. Consider a 2×2 matrix A with real entries. To find its eigenvalues, we solve the characteristic equation. This technique is standard in Linear Algebra and assumes familiarity with matrix determinants.

CHARACTERISTIC EQUATION

det(A − λI) = 0

A is the matrix, λ (lambda) is the eigenvalue we want to find, and I is the identity matrix. This equation expands into a polynomial in λ.

For a 2×2 matrix, the characteristic equation is always a quadratic: λ² − (trace)λ + (determinant) = 0. The trace is the sum of the diagonal entries, and the determinant is ad − bc for a matrix [[a, b], [c, d]]. We use the quadratic formula to solve it.

QUADRATIC FORMULA FOR EIGENVALUES

λ = (T ± √(T² − 4D)) / 2

T = trace of A, D = determinant of A. When T² − 4D < 0, the square root involves a negative number, producing complex eigenvalues.

When T² − 4D is negative, we get eigenvalues of the form λ = a ± bi. Here, a = T/2 is the real part, and b = √(4D − T²)/2 is the imaginary part. The two eigenvalues are complex conjugates of each other.

COMPLEX EIGENVALUE PAIR

λ₁ = a + bi, λ₂ = a − bi

a is the real part (controls growth or decay), b is the imaginary part (controls the speed of rotation). Together, they describe a spiral.

EXTRACTING REAL SOLUTIONS (Differential Equations)

x(t) = e^(at) [c₁ cos(bt) + c₂ sin(bt)]

In systems of differential equations, this is the general form of the real-valued solution when eigenvalues are a ± bi. The e^at factor handles growth or decay over time, while cos(bt) and sin(bt) produce oscillation. This formula is derived by applying Euler's formula (e^iθ = cos θ + i sin θ) to the complex exponential solution and then separating real and imaginary parts.

SECTION 5

Detailed Breakdown — From Complex Eigenvectors to Real Solutions

Let's walk through the process of turning complex eigenvalues into real solutions step by step. When you find a complex eigenvalue λ = a + bi, you also find a complex eigenvector. That eigenvector will have entries like (1, 2 + 3i) — numbers with real and imaginary parts mixed together. The trick is to split that eigenvector into two pieces: its real part and its imaginary part. Each piece becomes one of the two independent real solutions.

This flowchart shows the five steps to go from a matrix with complex eigenvalues to two real-valued solutions. Notice how the complex eigenvector gets split into its real part p and imaginary part q, then combined with cosine and sine to produce fully real answers.

The key insight is in Step 4. When you write the complex eigenvector as v = p + iq, the vector p is what you get by taking only the real numbers, and q is what you get by taking only the coefficients of i. These two real vectors become the building blocks of your final answer. The cosine and sine functions in the final formulas are what replace the complex exponential — they are the "real-world version" of imaginary number arithmetic.

💡

Remember

You only need to find one complex eigenvalue and its eigenvector. The conjugate pair automatically gives you the second solution. This saves you half the work!

SECTION 6

Worked Example

Let's work through a complete example. Consider the matrix A = [[1, −2], [2, 1]]. We will find its eigenvalues, discover they are complex, and then extract real solutions.

Finding Real Solutions from Complex Eigenvalues

1
Step 1 — Write the Characteristic Equation

We compute det(A − λI) = 0. The matrix A − λI = [[(1−λ), −2], [2, (1−λ)]]. The determinant is (1−λ)(1−λ) − (−2)(2) = (1−λ)² + 4.

Characteristic equation: (1−λ)² + 4 = 0
2
Step 2 — Solve for λ

Expanding: (1−λ)² = −4. Taking the square root: 1−λ = ±√(−4) = ±2i. So λ = 1 ∓ 2i. The two eigenvalues are λ₁ = 1 + 2i and λ₂ = 1 − 2i.

λ₁ = 1 + 2i, λ₂ = 1 − 2i (complex conjugate pair, a = 1, b = 2)
3
Step 3 — Find the Eigenvector for λ₁ = 1 + 2i

Solve (A − λ₁I)v = 0. The matrix A − (1+2i)I = [[−2i, −2], [2, −2i]]. From the first row: −2i·v₁ − 2·v₂ = 0, so v₂ = −i·v₁. Setting v₁ = 1, we get v = (1, −i).

Complex eigenvector: v = (1, −i)
4
Step 4 — Split into Real and Imaginary Parts

Write v = p + iq. Here v = (1, −i) = (1, 0) + i(0, −1). So the real part is p = (1, 0) and the imaginary part is q = (0, −1).

p = (1, 0), q = (0, −1)
5
Step 5 — Write the Two Real Solutions

Using a = 1, b = 2, p = (1, 0), and q = (0, −1), the two real solutions are: x₁(t) = e^t [(1, 0) cos 2t − (0, −1) sin 2t] = e^t (cos 2t, sin 2t). And x₂(t) = e^t [(1, 0) sin 2t + (0, −1) cos 2t] = e^t (sin 2t, −cos 2t).

x₁(t) = eᵗ(cos 2t, sin 2t) and x₂(t) = eᵗ(sin 2t, −cos 2t) — both entirely real!

Notice what happened: we started with imaginary numbers but ended with solutions made of cosines and sines multiplied by an exponential. The solutions describe a spiral that grows outward because a = 1 > 0 (the e^t part grows). If a were negative, the spiral would shrink inward. If a were zero, the solutions would trace a perfect circle.

SECTION 7

Real vs. Complex Eigenvalues — Comparisons

Let's compare the two cases side by side so you can see exactly when and why each type of eigenvalue appears, and what the resulting behavior looks like.

Feature	Real Eigenvalues	Complex Eigenvalues
Discriminant	T² − 4D ≥ 0	T² − 4D < 0
Eigenvalue form	λ₁ and λ₂ are real numbers	λ = a ± bi (conjugate pair)
Geometric effect	Stretching and/or flipping along fixed directions	Rotation combined with scaling
Solution involves	Pure exponentials: eλt	Exponential × trig: eᵃᵗ cos(bt), eᵃᵗ sin(bt)
Physical example	Exponential growth/decay, population models	Oscillation, vibrating springs, AC circuits
Fixed directions?	Yes — eigenvectors are real directions	No — every vector gets rotated

Comparison of real and complex eigenvalue cases for 2×2 matrices

✦

KEY TAKEAWAY

Think of it like two different types of music. Real eigenvalues are like a single sustained note that gets louder or softer — the pitch stays the same. Complex eigenvalues are like a note that wobbles up and down (vibrato) while also getting louder or softer. The real part is the volume knob, and the imaginary part controls the wobble speed.

SECTION 8

Connection to Advanced Theory

Complex eigenvalues are not just a curiosity — they connect to some of the most powerful ideas in mathematics and science. Once you are comfortable with the introductory version, here is where the road leads. All of the advanced topics in the table below are standard subjects in more advanced mathematics and engineering courses.

This Lesson (Intro)	Advanced Topic
2×2 matrices with complex eigenvalues	n×n matrices can have multiple pairs of complex eigenvalues, creating coupled oscillations
eᵃᵗ cos(bt) and eᵃᵗ sin(bt) solutions	Matrix exponential $e^{At}$ provides a unified formula for all cases (studied in advanced Linear Algebra)
Spirals growing, shrinking, or staying circular	Stability theory: the sign of a determines if a system is stable, unstable, or neutrally stable (studied in Differential Equations and Control Theory)
Splitting eigenvector into real and imaginary parts	Jordan normal form and generalized eigenvectors for repeated complex eigenvalues (advanced Linear Algebra)

How introductory concepts connect to more advanced linear algebra and differential equations topics

In courses on differential equations and control theory, you will use complex eigenvalues constantly. Engineers analyze whether bridges vibrate dangerously, whether electrical circuits resonate, and whether autopilot systems stay stable — all by checking the eigenvalues of a matrix. If the real part a is negative, the system calms down. If a is positive, the system blows up. If a is exactly zero, the system oscillates forever. This makes eigenvalue analysis one of the most practical tools in all of applied mathematics.

📝

Looking Ahead

In future lessons, you will learn how to sketch phase portraits (pictures of all possible solution trajectories) for systems with complex eigenvalues. You will see that spirals and circles emerge naturally from the math — a beautiful connection between algebra and geometry.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

A 2×2 matrix with real entries has eigenvalues λ₁ = 3 + 4i and λ₂ = 3 − 4i. Without doing any calculations, what type of motion do these eigenvalues describe — pure growth, pure rotation, or a growing spiral? Explain your reasoning.

Show answer

These eigenvalues describe a growing spiral. The real part a = 3 is positive, so the exponential factor e^3t causes the solution to grow over time. The imaginary part b = 4 means the solution also oscillates (rotates) with angular frequency 4. If the real part were zero, we would have pure rotation. If the imaginary part were zero, we would have pure growth. Since both are nonzero, we get a spiral that expands outward.

Question 2: BASIC CALCULATION

Find the eigenvalues of the matrix A = [[0, −1], [1, 0]]. Are they real or complex?

Show answer

The characteristic equation is det(A − λI) = (0 − λ)(0 − λ) − (−1)(1) = λ² + 1 = 0. Solving gives λ² = −1, so λ = ±i. The eigenvalues are λ₁ = i and λ₂ = −i. These are purely imaginary (the real part a = 0 and b = 1). Since the real part is zero, this matrix describes a pure rotation by 90° with no growth or decay. In fact, this is exactly the standard rotation-by-90° matrix!

Question 3: INTERMEDIATE

The matrix B = [[2, −5], [1, −2]] has eigenvalues λ = ±i. Find the eigenvector for λ = i, then split it into its real part p and imaginary part q.

Show answer

Solve (B − iI)v = 0. We get [[(2−i), −5], [1, (−2−i)]] v = 0. From the second row: v₁ + (−2−i)v₂ = 0, so v₁ = (2+i)v₂. Setting v₂ = 1: v = (2+i, 1). Splitting: p = (2, 1) and q = (1, 0). Since the real part of λ is 0, the general solution involves pure cosines and sines with no exponential growth: x(t) = c₁[(2,1) cos t − (1,0) sin t] + c₂[(2,1) sin t + (1,0) cos t]. The trajectories are closed ellipses around the origin.

Question 4: APPLIED

A simple electrical circuit is modeled by the system x' = Ax where A = [[−1, −4], [1, −1]]. Find the eigenvalues and determine whether the circuit's current oscillations grow, decay, or stay constant over time.

Show answer

The trace T = −1 + (−1) = −2 and the determinant D = (−1)(−1) − (−4)(1) = 1 + 4 = 5. The discriminant is T² − 4D = 4 − 20 = −16 < 0, so the eigenvalues are complex: λ = (−2 ± √(−16))/2 = −1 ± 2i. The real part is a = −1, which is negative. This means e^−t shrinks toward zero. The circuit's oscillations decay over time — the current spirals inward to zero. The oscillation frequency is b = 2. Physically, this describes a damped oscillation like a ringing that fades away.

Question 5: CRITICAL THINKING

Suppose a 2×2 matrix A has real entries and its eigenvalues are λ = a ± bi with b ≠ 0. Explain why the eigenvalues must always come in conjugate pairs (i.e., why you can never have just one complex eigenvalue without the other). Hint: think about the characteristic polynomial and its coefficients.

Show answer

The characteristic polynomial of a 2×2 matrix with real entries is λ² − Tλ + D = 0, where T (trace) and D (determinant) are both real numbers. By the quadratic formula, the roots are λ = (T ± √(T² − 4D))/2. When T² − 4D < 0, the square root produces an imaginary number, and the ± sign creates two roots that differ only in the sign of the imaginary part: a + bi and a − bi. This is actually a general theorem: complex roots of any polynomial with real coefficients always come in conjugate pairs. This is because if you substitute a + bi into a polynomial with real coefficients and get zero, then substituting a − bi also gives zero (the imaginary parts cancel in conjugate fashion). This guarantees you can never have a lone complex eigenvalue for a real matrix.

SUMMARY

Summary

When a 2×2 matrix with real entries has a negative discriminant (T² − 4D < 0), its eigenvalues are complex conjugates of the form λ = a ± bi. The real part a controls whether solutions grow (a > 0), decay (a < 0), or stay constant (a = 0). The imaginary part b controls the speed of rotation and oscillation. Together, they describe spiraling behavior.

To obtain real-valued solutions from complex eigenvalues, find one complex eigenvector and split it into its real part p and imaginary part q. The two independent real solutions combine exponentials with cosine and sine functions: x₁ = eᵃᵗ(p cos bt − q sin bt) and x₂ = eᵃᵗ(p sin bt + q cos bt). Complex eigenvalues always come in conjugate pairs for matrices with real entries, guaranteeing that the imaginary parts cancel out in the final answer.