In short
The nth roots of unity are the n solutions of z^n = 1. They are z_k = \cos\frac{2k\pi}{n} + i\sin\frac{2k\pi}{n} for k = 0, 1, \dots, n-1, equally spaced around the unit circle like the vertices of a regular n-gon. For n = 3, the non-real roots are called cube roots of unity and written \omega and \omega^2, where \omega = \frac{-1 + i\sqrt{3}}{2}. A key property: the nth roots of unity always sum to zero.
Take any regular polygon — a triangle, a square, a pentagon — and place it on the complex plane so that one vertex sits at 1 on the real axis and the centre is at the origin. The vertices of that polygon are the nth roots of unity.
This is a striking claim. Why should the solutions of a purely algebraic equation z^n = 1 arrange themselves into a perfectly symmetric geometric figure? The answer comes from De Moivre's theorem and the polar form of complex numbers.
Finding the nth roots of unity
You want all complex numbers z such that z^n = 1. Write z in polar form: z = \cos\theta + i\sin\theta (the modulus must be 1, since |z^n| = |z|^n = 1 forces |z| = 1). By De Moivre's theorem:
For this to equal 1 = \cos 0 + i\sin 0, you need n\theta = 2k\pi for some integer k. So:
As k runs through 0, 1, 2, \dots, n-1, you get n distinct angles equally spaced around the circle, each separated by \frac{2\pi}{n} radians. At k = n the angle is 2\pi, which is the same point as k = 0, so the cycle repeats.
$n$th roots of unity
The n solutions of z^n = 1 are
In exponential notation, z_k = e^{2\pi i k/n}. These n points are equally spaced on the unit circle, forming a regular n-gon with one vertex at 1.
A single symbol captures all the roots. Let \omega = e^{2\pi i/n} = \cos\frac{2\pi}{n} + i\sin\frac{2\pi}{n}. This is the root at k = 1 — the first one after 1 in the counterclockwise direction. Then the n roots are simply:
Each root is the previous one multiplied by \omega, which rotates the point by 2\pi/n around the circle. The root \omega is called a primitive nth root of unity because its successive powers generate all the others.
The cube roots of unity
The case n = 3 is the most heavily tested in exams and the richest in algebraic properties. The three cube roots of unity solve z^3 = 1.
Set \omega = e^{2\pi i/3} = \cos\frac{2\pi}{3} + i\sin\frac{2\pi}{3}. Compute the real and imaginary parts:
So:
The three cube roots of unity are 1, \omega, and \omega^2. What is \omega^2?
Notice that \omega^2 is the complex conjugate of \omega: \omega^2 = \bar{\omega}. This is because \omega and \omega^2 are symmetric about the real axis — they sit at angles 120° and 240° (= -120°), which are reflections of each other.
Properties of cube roots of unity
These properties appear constantly in JEE problems. Each one follows from \omega = e^{2\pi i/3}.
Property 1: \omega^3 = 1. By definition — \omega is a cube root of unity.
Property 2: 1 + \omega + \omega^2 = 0. Factor: z^3 - 1 = (z - 1)(z^2 + z + 1). The roots of z^3 - 1 = 0 are 1, \omega, \omega^2. Since \omega and \omega^2 are roots of z^2 + z + 1 = 0, the sum of all three roots is 1 + \omega + \omega^2 = 0 (by Vieta's formulas on z^3 - 1, or by direct substitution into z^2 + z + 1 = 0 and adding 1).
This property is the workhorse: whenever 1 + \omega + \omega^2 appears in a calculation, replace it with 0.
Property 3: \omega^2 = \bar{\omega} and \omega = \bar{\omega^2}. As computed above, \omega and \omega^2 are complex conjugates. This means |\omega| = |\omega^2| = 1 and \omega \cdot \omega^2 = |\omega|^2 = 1 (consistent with \omega^3 = 1).
Property 4: Powers cycle with period 3. \omega^3 = 1, so \omega^4 = \omega, \omega^5 = \omega^2, \omega^6 = 1, and in general \omega^k = \omega^{k \bmod 3}. When simplifying \omega to a large power, just take the exponent modulo 3.
Property 5: \omega^2 is also a primitive cube root of unity. Its powers are \omega^2, \omega^4 = \omega, \omega^6 = 1 — all three cube roots. So \omega^2 generates the same group as \omega. In fact, for any n, e^{2\pi i k/n} is a primitive nth root of unity if and only if \gcd(k, n) = 1.
Sum of nth roots of unity
The property 1 + \omega + \omega^2 = 0 for cube roots generalises beautifully. For any n \ge 2:
where \omega = e^{2\pi i/n}.
The proof is a one-line application of the geometric series formula. Since \omega \ne 1:
Geometrically, this makes perfect sense: the n roots of unity are equally spaced around the unit circle, so their vector sum — the "centre of mass" — is the origin. If you placed equal weights at each vertex of a regular n-gon centred at the origin, the balance point would be at 0.
More generally, for any integer m that is not a multiple of n:
because \omega^m is also a primitive nth root of unity (or at least another root of unity whose powers cycle through all nth roots). And when m is a multiple of n, each \omega^{mk} = (\omega^n)^{m/n \cdot k}... no — more directly: \omega^{mk} = 1 for all k, so the sum is n. This "orthogonality" property is the engine behind the discrete Fourier transform.
Interactive: roots of unity on the unit circle
Drag the red point along the curve below to change the angle \theta. The readouts show \theta, \cos\theta, and \sin\theta. When the point passes through a cube root of unity (\theta = 0, 2\pi/3, 4\pi/3), notice how the coordinates match the values 1, \omega, and \omega^2.
Worked examples
Example 1: Simplify $(1 + \omega - \omega^2)^3$ where $\omega$ is a primitive cube root of unity
A pure algebraic manipulation using the properties of cube roots of unity.
Step 1. Use 1 + \omega + \omega^2 = 0 to replace \omega^2.
From 1 + \omega + \omega^2 = 0, you get \omega^2 = -1 - \omega.
Why: this identity lets you eliminate \omega^2 everywhere, reducing all expressions to linear combinations of 1 and \omega.
Step 2. Substitute into 1 + \omega - \omega^2.
Why: opening the bracket and collecting terms. The expression simplifies to 2(1 + \omega).
Step 3. Simplify 1 + \omega using the same identity.
From 1 + \omega + \omega^2 = 0: 1 + \omega = -\omega^2.
So 2(1 + \omega) = -2\omega^2.
Why: the identity 1 + \omega + \omega^2 = 0 can be rearranged in three ways — use whichever form is most convenient.
Step 4. Cube the result.
Why: \omega^6 = (\omega^3)^2 = 1^2 = 1 since \omega^3 = 1. The cube root property collapses the high power instantly.
Result: (1 + \omega - \omega^2)^3 = -8.
The answer is a real integer, which is typical of symmetric expressions in \omega. Whenever you see \omega in a JEE problem, the strategy is almost always the same: use 1 + \omega + \omega^2 = 0 and \omega^3 = 1 to reduce everything.
Example 2: Show that the fourth roots of unity sum to zero, and find their product
A geometric verification using the explicit coordinates of the roots.
Step 1. List the four roots of z^4 = 1.
With n = 4, the roots are at angles 0, \pi/2, \pi, 3\pi/2:
Why: the spacing is 2\pi/4 = \pi/2 = 90°, so the roots sit at the four compass points on the unit circle.
Step 2. Compute the sum.
Why: real parts cancel (1 + (-1) = 0), imaginary parts cancel (i + (-i) = 0). This is the general property: nth roots of unity sum to zero for n \ge 2.
Step 3. Compute the product.
In polar form, each root has modulus 1, so the product has modulus 1. The arguments add: 0 + \pi/2 + \pi + 3\pi/2 = 3\pi. Reducing modulo 2\pi: 3\pi \equiv \pi. So the product is e^{i\pi} = -1.
Why: the product of all nth roots of unity is (-1)^{n+1} when n \ge 2. For n = 4: (-1)^5 = -1. (More precisely, the product equals (-1)^{n-1} — both give -1 for n = 4.)
Step 4. Verify using Vieta's formulas.
The polynomial z^4 - 1 = 0 has roots 1, i, -1, -i. The product of all roots equals (-1)^4 \cdot \frac{(-1)}{1} = -1 (constant term divided by leading coefficient, with sign (-1)^n).
Why: Vieta's formula for the product of roots of z^n + \cdots + c_0 = 0 is (-1)^n \cdot c_0. For z^4 - 1, c_0 = -1 and n = 4.
Result: The four fourth roots of unity sum to 0 and their product is -1.
The geometric picture makes the sum-to-zero property obvious — opposite vertices of the square are negatives of each other, so they cancel in pairs. For the product, the polar form argument addition gives the answer directly. This is the power of working on the unit circle: sums are visible as balancing vectors, and products are visible as accumulated rotations.
Common confusions
-
"Every nth root of unity is a primitive root." Only the roots e^{2\pi i k/n} with \gcd(k, n) = 1 are primitive. For n = 6, the root e^{2\pi i \cdot 2/6} = e^{2\pi i/3} is a primitive cube root but not a primitive sixth root, because its order is 3, not 6.
-
"\omega always means e^{2\pi i/3}." In most JEE problems, \omega refers to a primitive cube root of unity. But in general discussion, \omega can denote a primitive nth root for any n. Always check the context.
-
"The sum of the nth roots of unity is 1." The sum is 0, not 1. The confusion comes from including z_0 = 1 and mentally treating it as "the answer" rather than one term in a sum of n terms.
-
"1 + \omega + \omega^2 = 0 means \omega is negative." \omega is complex, not real. The equation says that the three vectors 1, \omega, \omega^2 form a closed triangle when placed head-to-tail, not that any single one is negative.
-
"The product of the nth roots of unity is always 1." It depends on n. The product is (-1)^{n-1}: it equals 1 when n is odd and -1 when n is even. For instance, the product of the cube roots is 1 \cdot \omega \cdot \omega^2 = \omega^3 = 1, but the product of the fourth roots is -1 as computed in Example 2.
Going deeper
If you can list the nth roots of unity, use the properties \omega^n = 1 and 1 + \omega + \cdots + \omega^{n-1} = 0 in calculations, and handle cube root problems fluently, you have the core. The material below is for readers who want to see the algebraic structure underneath.
Cyclotomic polynomials
The polynomial z^n - 1 factors over the integers. For n = 6:
The factor z^2 + z + 1 has roots \omega, \omega^2 (the primitive cube roots). The factor z^2 - z + 1 has roots e^{i\pi/3}, e^{5i\pi/3} (the primitive sixth roots). These irreducible factors are called cyclotomic polynomials — from the Greek for "circle-cutting," because they cut the circle into arcs.
The nth cyclotomic polynomial \Phi_n(z) has degree \phi(n) (Euler's totient function) and its roots are exactly the primitive nth roots of unity. These polynomials are always irreducible over the rationals — a deep result in algebra.
The discrete Fourier transform
The "orthogonality" property of roots of unity — \sum_{k=0}^{n-1} \omega^{mk} = 0 when n \nmid m and = n when n \mid m — is the mathematical heart of the Fourier transform applied to finite sequences. If you have a sequence a_0, a_1, \dots, a_{n-1}, its discrete Fourier transform is A_j = \sum_k a_k \omega^{jk}. The inverse transform uses the conjugate roots \omega^{-jk}. This is the algorithm behind signal processing, image compression, and the fast multiplication of large numbers.
Gauss and the regular 17-gon
In 1796, a 19-year-old mathematician showed that the regular 17-gon can be constructed with compass and straightedge. The proof uses the fact that 17 is a Fermat prime (17 = 2^{2^2} + 1), and the key step is factoring the 16th-degree cyclotomic polynomial \Phi_{17}(z) into quadratic factors using nested square roots. The constructibility of regular polygons is entirely determined by the factorisation of roots of unity — geometry and algebra meeting through the unit circle.
Where this leads next
- nth Roots of Complex Numbers — finding z^n = w for any complex w, not just w = 1. The roots of unity appear as the multiplicative "offsets" between the n roots.
- De Moivre's Theorem — the power formula that makes roots of unity computable: (\cos\theta + i\sin\theta)^n = \cos(n\theta) + i\sin(n\theta).
- Polar Form of Complex Numbers — the representation that reveals why roots of unity sit on the unit circle.
- Complex Numbers — Introduction — the foundation: i, the Argand plane, and arithmetic of complex numbers.
- Number Theory — Basics — divisibility and \gcd, which determine which roots of unity are primitive.