Skip to main content

De Moivre's theorem and roots of unity in HSC Mathematics Extension 2: powers in polar form, nth roots, the roots of unity and their geometry

Syllabus dot point

“Use de Moivre's theorem to find powers and nth roots of complex numbers and to derive the roots of unity and their geometric arrangement”

HSCMaths Extension 2Using Complex Numbers (MEX-N2)14 min read

Quick answer

De Moivre's theorem states that for any integer nn, [r(cos⁡θ+isin⁡θ)]n=rn(cos⁡nθ+isin⁡nθ)\left[r(\cos\theta + i\sin\theta)\right]^n = r^n(\cos n\theta + i\sin n\theta). To find the nn distinct nnth roots of a complex number, take the real nnth root of the modulus and divide the argument by nn, then add 2πkn\dfrac{2\pi k}{n} for k=0,1,…,n−1k = 0, 1, \dots, n-1. The nnth roots of unity are nn equally spaced points on the unit circle and sum to zero.

Jump to a section
  1. What this dot point is asking
  2. De Moivre's theorem
  3. Finding nth roots
  4. Roots of unity
  5. Why the roots of unity sum to zero
  6. Multiple-angle formulae from de Moivre
  7. Roots of a general complex number
  8. Reducing arguments
  9. How exam questions ask about de Moivre and roots of unity
  10. Exam-style questions

What this dot point is asking

NESA wants you to use de Moivre's theorem to raise complex numbers to powers and to find all the nnth roots of a complex number. You must derive the nnth roots of unity, plot them as equally spaced points on the unit circle, and use their structure (for example, that they sum to zero). The geometry, nn points evenly spread round a circle forming a regular polygon, is the picture that makes every roots question quick, so this page builds it up carefully.

De Moivre's theorem

If z=r(cos⁡θ+isin⁡θ)z = r(\cos\theta + i\sin\theta) then for every integer nn,

zn=rn(cos⁡nθ+isin⁡nθ).z^n = r^n(\cos n\theta + i\sin n\theta).

This follows from the fact that multiplication in polar form multiplies moduli and adds arguments; raising to the nnth power multiplies the modulus by itself nn times and adds the argument nn times. The theorem can be proved formally by induction for positive integers, then extended to negative integers via reciprocals.

A typical use is computing high powers. To find (1+i)8(1 + i)^8, write 1+i=2(cos⁡π4+isin⁡π4)1 + i = \sqrt{2}\left(\cos\frac{\pi}{4} + i\sin\frac{\pi}{4}\right), so

(1+i)8=(2)8(cos⁡2π+isin⁡2π)=16(1+0i)=16.(1 + i)^8 = (\sqrt{2})^8\left(\cos 2\pi + i\sin 2\pi\right) = 16(1 + 0i) = 16.

Computing (1+i)8(1+i)^8 by repeated Cartesian multiplication is laborious and error-prone; in polar form it is two short lines. That contrast is the whole reason de Moivre's theorem is examined.

Finding nth roots

To solve zn=wz^n = w where w=R(cos⁡ϕ+isin⁡ϕ)w = R(\cos\phi + i\sin\phi), write the unknown as z=r(cos⁡θ+isin⁡θ)z = r(\cos\theta + i\sin\theta). Then rn=Rr^n = R gives r=R1/nr = R^{1/n} (the positive real root), and nθ=ϕ+2πkn\theta = \phi + 2\pi k for integer kk. The distinct solutions are

zk=R1/n(cos⁡ϕ+2πkn+isin⁡ϕ+2πkn),k=0,1,…,n−1.z_k = R^{1/n}\left(\cos\frac{\phi + 2\pi k}{n} + i\sin\frac{\phi + 2\pi k}{n}\right), \qquad k = 0, 1, \dots, n-1.

Values of kk outside this range repeat earlier roots. Geometrically the nn roots lie on a circle of radius R1/nR^{1/n}, equally spaced by angle 2πn\dfrac{2\pi}{n}.

Roots of unity

The nnth roots of unity solve zn=1z^n = 1. Since 1=cos⁡0+isin⁡01 = \cos 0 + i\sin 0, they are

ωk=cos⁡2πkn+isin⁡2πkn,k=0,1,…,n−1.\omega_k = \cos\frac{2\pi k}{n} + i\sin\frac{2\pi k}{n}, \qquad k = 0, 1, \dots, n-1.

Writing ω=cos⁡2πn+isin⁡2πn\omega = \cos\frac{2\pi}{n} + i\sin\frac{2\pi}{n}, every root is a power ωk\omega^k. They sit at the vertices of a regular nn-gon inscribed in the unit circle, with one vertex at 11.

The figure builds the sixth roots of unity (z6=1z^6 = 1) stage by stage, so you can see the regular polygon emerge.

Stage 1, start on the unit circle. Every root of z6=1z^6 = 1 has modulus 11, so all six lie on the unit circle. The simplest root is k=0k = 0: ω0=cos⁡0+isin⁡0=1\omega_0 = \cos 0 + i\sin 0 = 1, sitting on the positive real axis.

The sixth roots of unity, stage 1 The sixth roots of unity are six equally spaced points on the unit circle, 60 degrees apart, with one root at 1. Built up stage by stage they form the vertices of a regular hexagon. Re Im -1 1 Start on the unit circle: the root k=0 is z = 1.

Stage 2, step round by 2π6=60∘\tfrac{2\pi}{6} = 60^\circ. Each successive root advances the argument by 2πn=60∘\dfrac{2\pi}{n} = 60^\circ. So ω1=cis⁡60∘\omega_1 = \operatorname{cis}60^\circ and ω2=cis⁡120∘\omega_2 = \operatorname{cis}120^\circ. Writing ω=ω1\omega = \omega_1, the roots are the powers 1,ω,ω2,…1, \omega, \omega^2, \ldots

The sixth roots of unity, stage 2 The sixth roots of unity are six equally spaced points on the unit circle, 60 degrees apart, with one root at 1. Built up stage by stage they form the vertices of a regular hexagon. Re Im -1 1 60° ω Step round by 60° each time: roots at 0°, 60°, 120°.

Stage 3, all six roots. Continuing around, the six arguments are 0∘,60∘,120∘,180∘,240∘,300∘0^\circ, 60^\circ, 120^\circ, 180^\circ, 240^\circ, 300^\circ, equally spaced 360∘6=60∘\dfrac{360^\circ}{6} = 60^\circ apart. The root at 180∘180^\circ is −1-1, as expected for an even nn.

The sixth roots of unity, stage 3 The sixth roots of unity are six equally spaced points on the unit circle, 60 degrees apart, with one root at 1. Built up stage by stage they form the vertices of a regular hexagon. Re Im -1 1 60° ω All six roots: equally spaced 360°/6 = 60° apart.

Stage 4, join them into a regular polygon. The six roots are the vertices of a regular hexagon inscribed in the unit circle. By symmetry the vertices balance about the origin, which is the geometric reason the roots of unity sum to zero.

The sixth roots of unity, stage 4 The sixth roots of unity are six equally spaced points on the unit circle, 60 degrees apart, with one root at 1. Built up stage by stage they form the vertices of a regular hexagon. Re Im -1 1 60° ω Joined up, they are the vertices of a regular hexagon.

Why the roots of unity sum to zero

A key property: the roots of unity sum to zero for n≥2n \ge 2. There are two ways to see it. Algebraically, zn−1=(z−1)(zn−1+⋯+z+1)z^n - 1 = (z - 1)(z^{n-1} + \dots + z + 1), and the sum of all roots equals the negative of the coefficient of zn−1z^{n-1} in zn−1z^n - 1, which is 00. Geometrically, the roots are the vertices of a regular polygon centred at the origin, so the position vectors cancel in symmetric pairs (or triples), leaving a resultant of zero, exactly the balance you can see in the hexagon above.

Multiple-angle formulae from de Moivre

One of the most powerful uses of de Moivre's theorem is deriving identities for cos⁡nθ\cos n\theta and sin⁡nθ\sin n\theta. Expand (cos⁡θ+isin⁡θ)n(\cos\theta + i\sin\theta)^n with the binomial theorem, then equate the real part to cos⁡nθ\cos n\theta and the imaginary part to sin⁡nθ\sin n\theta. Replacing sin⁡2θ=1−cos⁡2θ\sin^2\theta = 1 - \cos^2\theta converts the cosine identity to a polynomial in cos⁡θ\cos\theta alone, as in the cos⁡5θ\cos 5\theta question above. This technique turns a trigonometric identity that would be hard to prove directly into a routine expansion.

Roots of a general complex number

The same method finds the roots of any complex number, not just 11. To solve zn=wz^n = w, first write ww in modulus-argument form w=Rcis⁡ϕw = R\operatorname{cis}\phi. The nn roots then sit on a circle of radius R1/nR^{1/n}, equally spaced by 2πn\frac{2\pi}{n}, with the first root at argument ϕn\frac{\phi}{n}. Geometrically the roots of unity are the special case w=1w = 1; multiplying every root of unity by one particular root of zn=wz^n = w rotates and scales the regular polygon onto the roots of the general equation. This is why drawing the roots of unity well prepares you for every roots question.

Reducing arguments

After raising to a high power the argument often exceeds 2π2\pi or falls below −2π-2\pi. Always reduce it into a single revolution by adding or subtracting multiples of 2π2\pi before reading off the Cartesian value, since cos⁡\cos and sin⁡\sin are periodic with period 2π2\pi. Forgetting this step leaves the answer correct in principle but in an awkward, unmarked form.

Key fact

De Moivre: [rcis⁡θ]n=rncis⁡(nθ)[r\operatorname{cis}\theta]^n = r^n\operatorname{cis}(n\theta) for integer nn. The nn solutions of zn=Rcis⁡ϕz^n = R\operatorname{cis}\phi are

zk=R1/ncis⁡ ⁣(ϕ+2πkn),k=0,1,…,n−1,z_k = R^{1/n}\operatorname{cis}\!\left(\frac{\phi + 2\pi k}{n}\right), \qquad k = 0, 1, \dots, n - 1,

equally spaced by 2πn\dfrac{2\pi}{n} on a circle of radius R1/nR^{1/n}. The nnth roots of unity are the vertices of a regular nn-gon on the unit circle with one vertex at 11, and they sum to 00.

How exam questions ask about de Moivre and roots of unity

  • "Find znz^n" or "evaluate ( … )n(\,\ldots\,)^n in the form x+iyx + iy": convert to polar form, apply de Moivre, reduce the argument, then convert back.
  • "Solve zn=wz^n = w" or "find the nnth roots of ...": use the nnth-roots formula for k=0,…,n−1k = 0, \dots, n-1; expect to plot them.
  • "Show that the roots ... form a regular polygon" or "sketch on an Argand diagram": state the common modulus R1/nR^{1/n} and the equal spacing 2πn\tfrac{2\pi}{n}.
  • "Show that the roots sum to zero" or "1+ω+ω2+⋯=01 + \omega + \omega^2 + \dots = 0": use the factorisation of zn−1z^n - 1 or the symmetry of the polygon.
  • "Show that cos⁡nθ=…\cos n\theta = \ldots": expand (cos⁡θ+isin⁡θ)n(\cos\theta + i\sin\theta)^n by the binomial theorem and take the real part.
  • "It is known that u=z+1zu = z + \tfrac{1}{z} ...": recognise z+1z=2cos⁡θz + \tfrac1z = 2\cos\theta for a root on the unit circle, then solve the resulting real equation.
Worked example

Find the three cube roots of unity and verify they sum to zero

Find the three cube roots of unity and verify they sum to zero.

The cube roots of unity are the three solutions of z3=1z^3 = 1. We use the general formula for nnth roots, then check the sum using the key property that the roots of unity always sum to zero.

Apply the nnth roots formula: Since 1=cos⁡0+isin⁡01 = \cos 0 + i\sin 0, we have R=1R = 1, ϕ=0\phi = 0, and n=3n = 3. The general formula gives three roots, one for each value of kk from 00 to n−1n - 1:

zk=cos⁡2πk3+isin⁡2πk3,k=0,1,2.z_k = \cos\frac{2\pi k}{3} + i\sin\frac{2\pi k}{3}, \qquad k = 0, 1, 2.

The three roots are equally spaced around the unit circle, 120∘120^\circ apart.

Evaluate each root explicitly: Substitute k=0k = 0, 11, 22 in turn, reading off the cosine and sine values at each angle:

  • k=0k = 0: z0=cos⁡0+isin⁡0=1z_0 = \cos 0 + i\sin 0 = 1.
  • k=1k = 1: z1=cos⁡2π3+isin⁡2π3=−12+32iz_1 = \cos\frac{2\pi}{3} + i\sin\frac{2\pi}{3} = -\frac{1}{2} + \frac{\sqrt{3}}{2}i.
  • k=2k = 2: z2=cos⁡4π3+isin⁡4π3=−12−32iz_2 = \cos\frac{4\pi}{3} + i\sin\frac{4\pi}{3} = -\frac{1}{2} - \frac{\sqrt{3}}{2}i.

Notice that z1z_1 and z2z_2 are complex conjugates, as expected for a real polynomial.

Verify that the three roots sum to zero: Add the three roots, grouping real and imaginary parts separately:

1+(−12+32i)+(−12−32i)=1−1+0i=0.1 + \left(-\frac{1}{2} + \frac{\sqrt{3}}{2}i\right) + \left(-\frac{1}{2} - \frac{\sqrt{3}}{2}i\right) = 1 - 1 + 0i = 0.

The imaginary parts cancel and the real parts sum to zero, confirming the result.

Check that z1z_1 is genuinely a cube root: Since ∣z1∣=1|z_1| = 1 and arg⁡z1=2π3\arg z_1 = \frac{2\pi}{3}, de Moivre's theorem gives z13z_1^3 modulus 13=11^3 = 1 and argument 3×2π3=2π3 \times \frac{2\pi}{3} = 2\pi, which equals the argument of 11. So z13=1z_1^3 = 1. Correct.

Final answer: the three cube roots of unity are 11, −12+32i-\frac{1}{2} + \frac{\sqrt{3}}{2}i, and −12−32i-\frac{1}{2} - \frac{\sqrt{3}}{2}i, and their sum is 00.

Worked example: the fourth roots of a complex number

Solve z4=−8+83 iz^4 = -8 + 8\sqrt{3}\,i for the fourth roots

Solve z4=−8+83 iz^4 = -8 + 8\sqrt{3}\,i, giving the roots in modulus-argument form.

Write the right-hand side in polar form: Compute the modulus and argument of w=−8+83 iw = -8 + 8\sqrt{3}\,i. The modulus is R=(−8)2+(83)2=64+192=256=16R = \sqrt{(-8)^2 + (8\sqrt3)^2} = \sqrt{64 + 192} = \sqrt{256} = 16. The point is in the second quadrant with reference angle tan⁡−1838=tan⁡−13=π3\tan^{-1}\frac{8\sqrt3}{8} = \tan^{-1}\sqrt3 = \frac{\pi}{3}, so ϕ=π−π3=2π3\phi = \pi - \frac{\pi}{3} = \frac{2\pi}{3}. Hence w=16cis⁡2π3w = 16\operatorname{cis}\frac{2\pi}{3}.

Apply the nth-roots formula: With n=4n = 4, the modulus of each root is R1/4=161/4=2R^{1/4} = 16^{1/4} = 2, and the arguments are ϕ+2πk4\dfrac{\phi + 2\pi k}{4} for k=0,1,2,3k = 0, 1, 2, 3:

θk=2π3+2πk4=π6+πk2.\theta_k = \frac{\frac{2\pi}{3} + 2\pi k}{4} = \frac{\pi}{6} + \frac{\pi k}{2}.

List the four roots: Substituting k=0,1,2,3k = 0, 1, 2, 3 and keeping each argument in (−π,π](-\pi, \pi]:

  • k=0k = 0: 2cis⁡π62\operatorname{cis}\dfrac{\pi}{6}.
  • k=1k = 1: 2cis⁡2π32\operatorname{cis}\dfrac{2\pi}{3}.
  • k=2k = 2: 2cis⁡7π6=2cis⁡ ⁣(−5π6)2\operatorname{cis}\dfrac{7\pi}{6} = 2\operatorname{cis}\!\left(-\dfrac{5\pi}{6}\right).
  • k=3k = 3: 2cis⁡5π3=2cis⁡ ⁣(−π3)2\operatorname{cis}\dfrac{5\pi}{3} = 2\operatorname{cis}\!\left(-\dfrac{\pi}{3}\right).

Sense-check the geometry: All four roots have modulus 22 and are spaced 2π4=π2\dfrac{2\pi}{4} = \dfrac{\pi}{2} apart, so they are the vertices of a square centred at the origin. That even spacing confirms the arguments are correct.

Final answer: the four roots are 2cis⁡π62\operatorname{cis}\dfrac{\pi}{6}, 2cis⁡2π32\operatorname{cis}\dfrac{2\pi}{3}, 2cis⁡ ⁣(−5π6)2\operatorname{cis}\!\left(-\dfrac{5\pi}{6}\right) and 2cis⁡ ⁣(−π3)2\operatorname{cis}\!\left(-\dfrac{\pi}{3}\right).

Common traps
Using only one value of kk
A degree-nn equation has nn roots. You must list k=0k = 0 through k=n−1k = n-1, not just the principal root.
Taking a complex nnth root of the modulus
The modulus is a positive real number; its nnth root is the ordinary positive real root. Do not over-complicate it.
Dropping the 2πk2\pi k term
Without adding 2πk2\pi k before dividing the argument by nn, you get only one root instead of all nn.
Misplacing roots on the diagram
The roots are equally spaced by 2πn\frac{2\pi}{n}; an uneven sketch signals an arithmetic slip in the arguments.
Forgetting to reduce a high-power argument
After de Moivre on a large power, subtract whole multiples of 2π2\pi before reading off cos⁡\cos and sin⁡\sin.
Exam technique

For a roots question, draw the circle first, mark one root, then step round by 2πn\tfrac{2\pi}{n}; an evenly spaced sketch is itself worth marks and catches arithmetic slips. State the common modulus and the equal angular spacing explicitly. Give every root, k=0k = 0 to n−1n - 1, and put each argument in the principal range (−π,π](-\pi, \pi] if asked. For "show cos⁡nθ=…\cos n\theta = \ldots" expand (cos⁡θ+isin⁡θ)n(\cos\theta + i\sin\theta)^n with Pascal's triangle, take the real part, and substitute sin⁡2θ=1−cos⁡2θ\sin^2\theta = 1 - \cos^2\theta. Keep 2\sqrt2, π6\tfrac{\pi}{6} and the like exact throughout.

Exam-style questions

Questions in the style of NESA exam questions on this dot point, each with a worked answer. They are written by ExamExplained unless tagged "Past paper"; the year shows the paper a question is modelled on.

HSC 2021-style3 marks
Using de Moivre's theorem and the binomial expansion of (cos⁡θ+isin⁡θ)5(\cos\theta + i\sin\theta)^5, or otherwise, show that cos⁡5θ=16cos⁡5θ−20cos⁡3θ+5cos⁡θ\cos 5\theta = 16\cos^5\theta - 20\cos^3\theta + 5\cos\theta.
Show worked answer →

By de Moivre's theorem, (cos⁡θ+isin⁡θ)5=cos⁡5θ+isin⁡5θ(\cos\theta + i\sin\theta)^5 = \cos 5\theta + i\sin 5\theta.

Expand the left side with the binomial theorem (writing c=cos⁡θc = \cos\theta, s=sin⁡θs = \sin\theta):

(c+is)5=c5+5c4(is)+10c3(is)2+10c2(is)3+5c(is)4+(is)5(c + is)^5 = c^5 + 5c^4(is) + 10c^3(is)^2 + 10c^2(is)^3 + 5c(is)^4 + (is)^5.

Using i2=−1i^2 = -1, i3=−ii^3 = -i, i4=1i^4 = 1, i5=ii^5 = i, the real part is c5−10c3s2+5cs4c^5 - 10c^3 s^2 + 5c s^4.

Equating real parts: cos⁡5θ=c5−10c3s2+5cs4\cos 5\theta = c^5 - 10 c^3 s^2 + 5 c s^4.

Now replace s2=1−c2s^2 = 1 - c^2 and s4=(1−c2)2=1−2c2+c4s^4 = (1 - c^2)^2 = 1 - 2c^2 + c^4:

cos⁡5θ=c5−10c3(1−c2)+5c(1−2c2+c4)=16c5−20c3+5c\cos 5\theta = c^5 - 10 c^3 (1 - c^2) + 5 c (1 - 2c^2 + c^4) = 16 c^5 - 20 c^3 + 5 c.

Hence cos⁡5θ=16cos⁡5θ−20cos⁡3θ+5cos⁡θ\cos 5\theta = 16\cos^5\theta - 20\cos^3\theta + 5\cos\theta.

Mark notes: 1 mark for applying de Moivre and the binomial expansion, 1 mark for taking the real part correctly, 1 mark for eliminating sin⁡2θ\sin^2\theta to reach the required cosine-only form.

HSC 2022-style2 marks
Write the complex number −3+i-\sqrt{3} + i in exponential form. Hence, find the exact value of (−3+i)10(-\sqrt{3} + i)^{10}, giving your answer in the form x+iyx + iy.
Show worked answer →

Modulus: ∣−3+i∣=3+1=2|-\sqrt{3} + i| = \sqrt{3 + 1} = 2.

Argument: the point is in the second quadrant, with reference angle arctan⁡13=π6\arctan\frac{1}{\sqrt{3}} = \frac{\pi}{6}, so arg⁡=π−π6=5π6\arg = \pi - \frac{\pi}{6} = \frac{5\pi}{6}.

Exponential form: −3+i=2ei5π/6-\sqrt{3} + i = 2 e^{i 5\pi/6}.

Now apply de Moivre to the tenth power:

(−3+i)10=210ei50π/6=1024 ei25π/3(-\sqrt{3} + i)^{10} = 2^{10} e^{i 50\pi/6} = 1024\, e^{i 25\pi/3}.

Reduce the argument modulo 2π2\pi: 25π3−8π=25π−24π3=π3\frac{25\pi}{3} - 8\pi = \frac{25\pi - 24\pi}{3} = \frac{\pi}{3}.

So the value is 1024(cos⁡π3+isin⁡π3)=1024(12+i32)=512+5123 i1024\left(\cos\frac{\pi}{3} + i\sin\frac{\pi}{3}\right) = 1024\left(\frac{1}{2} + i\frac{\sqrt{3}}{2}\right) = 512 + 512\sqrt{3}\, i.

Mark notes: 1 mark for the exponential form 2ei5π/62 e^{i 5\pi/6}, 1 mark for raising to the power and reducing the argument to give 512+5123 i512 + 512\sqrt{3}\, i.

HSC 2022-style3 marks
Consider the equation z5+1=0z^5 + 1 = 0. It is known that if zz is a solution with z≠−1z \neq -1, then u=z+1zu = z + \frac{1}{z} is a solution of u2−u−1=0u^2 - u - 1 = 0. Hence find the exact value of cos⁡3π5\cos\frac{3\pi}{5}.
Show worked answer →

The non-real solutions of z5+1=0z^5 + 1 = 0 lie on the unit circle at z=eiθz = e^{i\theta}. For such zz, 1z=e−iθ\frac{1}{z} = e^{-i\theta}, so

u=z+1z=eiθ+e−iθ=2cos⁡θu = z + \frac{1}{z} = e^{i\theta} + e^{-i\theta} = 2\cos\theta.

The fifth roots of −1-1 (excluding z=−1z = -1) occur at θ=π5,3π5,7π5,9π5\theta = \frac{\pi}{5}, \frac{3\pi}{5}, \frac{7\pi}{5}, \frac{9\pi}{5}, which pair into conjugates giving the two distinct real values u=2cos⁡π5u = 2\cos\frac{\pi}{5} and u=2cos⁡3π5u = 2\cos\frac{3\pi}{5}.

Solving u2−u−1=0u^2 - u - 1 = 0 by the quadratic formula: u=1±1+42=1±52u = \frac{1 \pm \sqrt{1 + 4}}{2} = \frac{1 \pm \sqrt{5}}{2}.

Since 3π5\frac{3\pi}{5} is between π2\frac{\pi}{2} and π\pi, cos⁡3π5\cos\frac{3\pi}{5} is negative, so we take the negative root:

2cos⁡3π5=1−522\cos\frac{3\pi}{5} = \frac{1 - \sqrt{5}}{2}, hence cos⁡3π5=1−54\cos\frac{3\pi}{5} = \frac{1 - \sqrt{5}}{4}.

Mark notes: 1 mark for recognising u=2cos⁡θu = 2\cos\theta for roots on the unit circle, 1 mark for solving the quadratic, 1 mark for selecting the negative root to give cos⁡3π5=1−54\cos\frac{3\pi}{5} = \frac{1 - \sqrt{5}}{4}.

Practise this

Sources & how we know this

ExamExplained