Some directions are special. Multiply most vectors by a matrix and they spin off somewhere new — but a few vectors keep their direction and only stretch or shrink. Those special vectors are eigenvectors, and the stretch factors are eigenvalues.

The Eigenvalues and Eigenvectors Calculator on this page finds both for any 2×2 matrix you enter. Type in the four entries and it returns the eigenvalues, the matching normalized eigenvectors, and the trace, determinant, and discriminant it used along the way.

You do not need to remember the characteristic equation — the calculator shows it. But the article below explains what each number means, so the result reads as understanding rather than magic.

What Does the Eigenvalues and Eigenvectors Calculator Do?

The Eigenvalues and Eigenvectors Calculator solves the equation Av = λv for a 2×2 matrix A. It computes the trace and determinant, forms the characteristic polynomial, and applies the quadratic formula to get the two eigenvalues λ₁ and λ₂.

For each real eigenvalue it then finds a direction vector that the matrix only scales, normalizes it to unit length, and displays it. When the discriminant is negative it reports the complex conjugate pair and explains why no real eigenvectors exist.

How to Use the Eigenvalues and Eigenvectors Calculator

Enter the four numbers of your matrix: a and b across the top row, c and d across the bottom. Any real numbers work — integers, decimals, negatives, zeros.

Press Calculate. Each eigenvalue appears with its eigenvector beneath it, plus a line showing the trace, determinant, and discriminant used. If you see a message about complex eigenvalues instead of vectors, that is a genuine mathematical outcome, not an error — read the section on the discriminant to understand it.

What an Eigenvalue Actually Is

An eigenvalue is the factor by which a matrix stretches its eigenvector. If Av = 3v, the eigenvalue is 3: the matrix triples that direction. An eigenvalue of 1 leaves the vector untouched; 0 crushes it to nothing; a negative flips it backward.

For a 2×2 matrix there are always two eigenvalues counting multiplicity, though they may be equal or complex. They are the roots of the characteristic equation, and everything else the calculator does flows from them.

What an Eigenvector Actually Is

An eigenvector is a direction the matrix treats simply: instead of rotating it, the matrix just scales it. Finding these directions is like finding the grain of the transformation — the axes along which it behaves predictably.

The formula is:

Av = λv

Eigenvectors are never unique: any non-zero multiple of an eigenvector is also an eigenvector for the same eigenvalue. That is why the calculator normalizes each one to length 1 — it picks a standard representative so the answer is clean and comparable.

The Characteristic Equation, Step by Step

Eigenvalues come from demanding that (A − λI) squashes some non-zero vector to zero, which happens exactly when its determinant is zero. For a 2×2 matrix that condition becomes a quadratic equation.

The formula is:

λ² − (trace)λ + (determinant) = 0

The trace is a + d and the determinant is ad − bc. The calculator shows both, so you can verify: for the default matrix with a=4, b=2, c=1, d=3, the trace is 7 and the determinant is 10, giving λ² − 7λ + 10 = 0.

Reading the Discriminant Before You Start

The discriminant — trace² − 4 × determinant — predicts what kind of answer you will get before any square root is taken. Positive means two distinct real eigenvalues; zero means one repeated eigenvalue; negative means a complex pair.

Geometrically, a negative discriminant means the matrix rotates every direction — nothing stays put, so no real eigenvectors exist. A zero discriminant means the matrix scales everything almost uniformly. The calculator reports the discriminant so this diagnosis is visible, not hidden.

Worked Example: A Matrix With Two Distinct Eigenvalues

First: take a = 4, b = 2, c = 1, d = 3. The trace is 7 and the determinant is 4×3 − 2×1 = 10.

The discriminant is 49 − 40 = 9, so the eigenvalues are (7 ± 3) ÷ 2 = 5 and 2.

Then: for λ₁ = 5 the calculator builds the vector (b, λ − a) = (2, 1) and normalizes it to (0.894, 0.447).

Check it: multiplying the matrix by (0.894, 0.447) gives (5×0.894, 5×0.447) — exactly λ₁ times the vector. The second eigenvalue λ₂ = 2 gets its own direction the same way.

Worked Example: A Repeated Eigenvalue

First: take a = 3, b = 0, c = 0, d = 3 — three times the identity matrix. The trace is 6 and the determinant is 9.

The discriminant is 36 − 36 = 0, so both eigenvalues are 6 ÷ 2 = 3.

Then: since the matrix is just 3I, every direction is an eigenvector with eigenvalue 3 — the matrix triples everything uniformly.

The calculator notes this special case: when the matrix is a multiple of the identity, any non-zero vector works, so it says so instead of inventing a meaningless “the” eigenvector.

Worked Example: A Matrix With Complex Eigenvalues

First: take a = 0, b = −1, c = 1, d = 0 — a 90-degree rotation matrix. The trace is 0 and the determinant is 0×0 − (−1)×1 = 1.

The discriminant is 0 − 4 = −4, which is negative.

Then: the eigenvalues are (0 ± √−4) ÷ 2 = ±i, the purely imaginary pair i and −i.

No real vector keeps its direction under a 90-degree rotation — everything moves — so there are no real eigenvectors. The calculator reports λ₁ = 0 + 1i and λ₂ = 0 − 1i and explains exactly why the vector section is empty.

Why Eigenvectors Are Normalized

Because any multiple of an eigenvector is still an eigenvector, reporting a raw one would be arbitrary — (2, 1) and (200, 100) describe the same direction. Normalizing to unit length removes that ambiguity.

The formula is:

v̂ = v ÷ √(vx² + vy²)

Dividing each component by the vector’s length gives a direction of length exactly 1. Two people computing the same matrix now get the same answer up to a sign flip, which is the only freedom left.

Common Eigenvalue Mistakes

The classic mistake is writing the characteristic equation with the wrong sign — it is λ² − (trace)λ + (determinant), and flipping a sign gives wrong eigenvalues that still “look” plausible. Another is forgetting that eigenvectors must be non-zero; the zero vector trivially satisfies Av = λv for every λ and tells you nothing.

People also normalize incorrectly by dividing by the wrong length, or assume eigenvectors of different eigenvalues must be perpendicular — true only for symmetric matrices. Finally, many forget to check the answer by multiplying A by the reported vector.

The Zero Matrix and Identity Matrix Edge Cases

The zero matrix has eigenvalue 0 twice, and every vector is an eigenvector — it crushes everything to the origin. The identity matrix has eigenvalue 1 twice, and again every vector qualifies — it changes nothing.

These are the cases where “the eigenvector” is meaningless, and the calculator says so plainly instead of picking an arbitrary direction. In between, a diagonal matrix like diag(2, 5) has eigenvalues 2 and 5 with eigenvectors along the coordinate axes — the cleanest case of all.

Where Eigenvalues and Eigenvectors Are Useful

Google’s original PageRank was an eigenvector computation: the web’s link matrix has a dominant eigenvector that ranks pages. Engineers use eigenvalues for vibration analysis — a bridge’s resonant frequencies are eigenvalues of its stiffness matrix.

Data scientists meet them in principal component analysis, where eigenvectors of the covariance matrix reveal the directions of greatest variation. Quantum mechanics is built on them: measurable quantities are eigenvalues of operators. The 2×2 case on this page is the doorway to all of it.

Worked Example: Checking Your Answer by Multiplication

First: reuse the matrix a = 4, b = 2, c = 1, d = 3 with the reported λ₂ = 2 and its normalized eigenvector.

The calculator gives a unit vector for λ₂; call it (x, y).

Then: multiply. The top entry is 4x + 2y and the bottom is 1x + 3y.

If the vector is correct, (4x + 2y, x + 3y) equals (2x, 2y) — each entry exactly double the original. This check catches sign errors and arithmetic slips instantly, and the calculator’s own display invites you to perform it.

How to Interpret Your Result Correctly

Read each eigenvalue as a stretch factor along its eigenvector’s direction. A large positive eigenvalue means strong stretching there; a small one means compression; a negative means flipping plus scaling.

The eigenvectors themselves are directions, not destinations — only their line matters, which is why the sign is arbitrary. And when the result shows complex eigenvalues, read them as rotation-plus-scaling: the matrix has no fixed real direction at all.

Eigenvalues in One Picture

Imagine drawing arrows in every direction on a sheet of paper, then applying your matrix to each arrow. Most arrows swing to new angles — but the eigenvectors stay on their original lines, merely growing or shrinking. The eigenvalues are exactly those growth factors.

Hold that image while you read the calculator’s output: each λ with its unit vector is one of those special arrows, quantified. When the arrows all rotate and none stay put, you are looking at the complex case.

Frequently Asked Questions

1. What is an eigenvalue in simple terms?

It is the number by which a matrix stretches its eigenvector. If Av = 3v, then 3 is the eigenvalue and v is the eigenvector.

2. What is an eigenvector in simple terms?

A direction that a matrix only stretches or shrinks instead of rotating. Every other vector gets turned; eigenvectors keep their line.

3. Why does a 2×2 matrix have two eigenvalues?

Because the characteristic equation is quadratic, and a quadratic has two roots counting multiplicity. They can be distinct, repeated, or a complex pair.

4. What does a negative discriminant mean?

The eigenvalues are complex conjugates and there are no real eigenvectors. Geometrically, the matrix rotates every direction, so nothing stays fixed.

5. Why are eigenvectors not unique?

Any non-zero multiple of an eigenvector satisfies the same equation Av = λv. The calculator normalizes to unit length to give one standard answer.

6. Can an eigenvalue be zero?

Yes. It means the matrix crushes that eigenvector direction to zero — the matrix is singular and has no inverse. The determinant of such a matrix is also zero.

7. How do I verify an eigenpair by hand?

Multiply the matrix by the reported vector and check that the result equals the eigenvalue times the vector, entry by entry. This is the definitive test.

8. What are complex eigenvalues used for?

They describe rotation combined with scaling, which appears in oscillations, electrical circuits, and control systems. The pair a ± bi encodes both the growth rate and the rotation speed.

9. What is the trace of a matrix?

The sum of the diagonal entries, a + d. It equals the sum of the eigenvalues and appears in the characteristic equation.

10. What is the determinant’s role here?

The determinant ad − bc equals the product of the eigenvalues and is the constant term of the characteristic equation. A zero determinant means a zero eigenvalue.

11. Do symmetric matrices have special eigen properties?

Yes — all their eigenvalues are real and eigenvectors for different eigenvalues are perpendicular. Many physics and statistics matrices are symmetric for exactly this reason.

12. What happens with a repeated eigenvalue?

You get one eigenvalue twice. Sometimes there are still two independent eigenvectors; sometimes only one direction works. The calculator flags the identity-like case explicitly.

13. Can this calculator handle 3×3 matrices?

No — it is built for 2×2 matrices, where the quadratic formula gives exact answers. Larger matrices need numerical methods beyond a simple page calculator.

14. Why normalize eigenvectors to length 1?

To remove the arbitrariness of scaling. Unit eigenvectors give everyone the same answer up to a sign, which makes results comparable and checks easy.

15. Where do eigenvalues appear in real life?

In PageRank, bridge vibration analysis, principal component analysis, quantum mechanics, and population models — anywhere a transformation’s natural directions matter.