You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
If De Moivre's theorem tells you how to raise a complex number to a power, running it backwards tells you how to take roots — and the answer is beautifully geometric. Every non-zero complex number has exactly n distinct nth roots, and they sit at the vertices of a regular n-gon on a circle. The simplest and most important case is zn=1: the nth roots of unity, equally spaced points on the unit circle whose remarkable algebra (they multiply like a clock, and they sum to zero) underpins everything from factorising zn−1 to the discrete Fourier transform. This lesson is one of the most elegant in the whole A-Level.
This is compulsory pure content for Papers 1 & 2, the direct sequel to De Moivre's theorem. It tests AO1 (write down the n roots systematically), AO2 (justify why there are exactly n, prove the sum-is-zero and the 1+ω+⋯+ωn−1=0 relation, and reason about the regular-polygon geometry) and AO3 (use roots of unity to factorise, to solve zn=w efficiently, or to handle a multi-step geometric problem). It connects backwards to De Moivre and conjugate roots, and forwards to the further-algebra strand's work on roots and coefficients.
To solve zn=w where w=R(cosϕ+isinϕ) (with R=∣w∣>0), write z=r(cosθ+isinθ). By De Moivre, zn=rn(cosnθ+isinnθ), so matching with w:
The crucial point is the +2kπ: adding whole turns to ϕ before dividing by n gives genuinely different roots. Taking k=0,1,2,…,n−1 yields exactly n distinct values (then k=n repeats k=0, since it adds n2nπ=2π to the argument):
zk=R1/n(cosnϕ+2kπ+isinnϕ+2kπ),k=0,1,…,n−1.
Consecutive roots differ in argument by n2π and share the modulus R1/n, so they are equally spaced on a circle of radius R1/n — the vertices of a regular n-gon.
Taking w=1=cos0+isin0 (so R=1, ϕ=0) gives the nth roots of unity:
zk=cosn2kπ+isinn2kπ,k=0,1,…,n−1.
Write ω=cosn2π+isinn2π=e2πi/n, the primitive nth root of unity (the first one anticlockwise from 1). Then zk=ωk, and the complete set of roots is
1,ω,ω2,…,ωn−1.
They all lie on the unit circle, are equally spaced by n2π, form a regular n-gon, and always include z=1. Multiplying by ω just rotates one root to the next — the roots behave like the hours on a clock face. Note too that ωn=1 and ω−1=ωn−1=ω, so the powers of ω cycle with period n and reciprocals coincide with conjugates — both facts used constantly when manipulating roots.
For n≥2, the nth roots of unity sum to zero:
1+ω+ω2+⋯+ωn−1=∑k=0n−1ωk=1−ω1−ωn=1−ω1−1=0,
a finite geometric series with ωn=1 and ω=1. Geometrically this says the position vectors to the vertices of a regular polygon centred at the origin cancel — a tidy AO2 result examiners like to ask you to prove.
Once you have any single solution z0 of zn=w, the full set is
z0,z0ω,z0ω2,…,z0ωn−1,
because multiplying by ωk keeps the modulus (∣ω∣=1) and rotates by n2kπ, stepping between the equally spaced roots. This is often the fastest route in the exam.
Since the roots of unity are exactly the roots of zn−1=0,
zn−1=(z−1)(z−ω)(z−ω2)⋯(z−ωn−1),
and dividing by (z−1) gives the cyclotomic-style factor
zn−1+zn−2+⋯+z+1=(z−ω)(z−ω2)⋯(z−ωn−1).
Two consequences are worth noting. First, setting z=1 in the right-hand product gives ∏k=1n−1(1−ωk)=n (the left side is 1+1+⋯+1=n) — a tidy identity that appears in harder problems. Second, when n is even the real-coefficient factorisation pairs each non-real root ωk with its conjugate ωn−k=ωk to give real quadratic factors z2−2cosn2kπz+1, since each pair has sum 2cosn2kπ and product 1. This is exactly the route to factorising zn−1 (or zn+1) over the reals.
Solve z3=1.
With R=1, ϕ=0, n=3: zk=cos32kπ+isin32kπ, k=0,1,2. (M1 set up the formula.)
z0=1;z1=cos32π+isin32π=−21+23i;z2=cos34π+isin34π=−21−23i.(A1 all three)
The roots form an equilateral triangle on the unit circle, and z2=zˉ1 (the non-real roots are a conjugate pair). Writing ω=−21+23i, the roots are 1,ω,ω2, with ω3=1, 1+ω+ω2=0 and ω2=ωˉ. (M1 formula; A1 three correct roots in exact form.)
Solve z4=1.
zk=cos42kπ+isin42kπ=cos2kπ+isin2kπ,k=0,1,2,3⟹z=1,i,−1,−i.(M1 A1)
These are the vertices of a square on the unit circle, lying on the axes. (A simple sanity check: their sum 1+i−1−i=0 ✓.)
Find all cube roots of 8i.
Write 8i=8(cos2π+isin2π), so R=8, ϕ=2π, and R1/3=2. (M1 polar form and R1/3.)
zk=2(cos32π+2kπ+isin32π+2kπ),k=0,1,2.(M1 add 2kπ before dividing)
z0=2(cos6π+isin6π)=3+i; z1=2(cos65π+isin65π)=−3+i; z2=2(cos23π+isin23π)=−2i.(A1 all three)
Check: (3+i) has modulus 2 and argument 6π, so by De Moivre (3+i)3=23(cos2π+isin2π)=8i ✓. (M1 polar form; M1 the +2kπ step; A1 three roots — with a verification that one cubes to 8i.)
Find the fourth roots of −16, and hence factorise z4+16 into real quadratics.
−16=16(cosπ+isinπ), R1/4=2:
zk=2(cos4π+2kπ+isin4π+2kπ),k=0,1,2,3⇒2(1+i),2(−1+i),2(−1−i),2(1−i).
The roots pair into conjugates 2(±1+i) and 2(±1−i). Each conjugate pair gives a real quadratic; e.g. 2(1±i) has sum 22 and product 2(1+1)=4, giving z2−22z+4; the other pair gives z2+22z+4. Hence
z4+16=(z2−22z+4)(z2+22z+4).
Given that z0=3+i is one cube root of 8i, write down the other two without re-deriving from scratch.
The other roots are z0ω and z0ω2, where ω=e2πi/3=−21+23i is the primitive cube root of unity. (M1 multiply by ω,ω2.) Geometrically each multiplication rotates by 32π and keeps the modulus 2, so:
z0ω=2cis(6π+32π)=2cis65π=−3+i;z0ω2=2cis(6π+34π)=2cis23π=−2i.(A1)
These match Example 3 exactly. (M1 the rotation idea; A1 the two roots. This "one root times the roots of unity" route is often the quickest in the exam once you have any single root.)
If ω=e2πi/6 is a primitive sixth root of unity, evaluate 1+ω2+ω4.
The powers 1,ω2,ω4 are themselves the three cube roots of unity (since (ω2)3=ω6=1, and they are distinct). (M1 recognise as cube roots of unity.) Therefore their sum is 0:
1+ω2+ω4=0.(A1)
(M1 spotting that {1,ω2,ω4} are the cube roots of unity; A1 for 0. Selecting every other root of unity reproduces a smaller set of roots of unity — a recurring exam idea, and the seed of the "roots-of-unity filter" in the stretch section.)
(specimen-style — not from any past paper)
(a) Solve z4=−8+83i, giving the roots in the form r(cosθ+isinθ) with −π<θ≤π. (b) State the geometric figure formed by the four roots on an Argand diagram.
(a) Modulus: ∣−8+83i∣=64+192=256=16. The point is in the second quadrant (Re<0, Im>0), so arg=π−arctan883=π−3π=32π. With R1/4=161/4=2:
zk=2(cos432π+2kπ+isin432π+2kπ)=2(cos6π+3kπ+isin6π+3kπ),k=0,1,2,3.
Evaluating the arguments 6π,64π=32π,67π,610π=35π and reducing the last two into (−π,π] (subtract 2π): 67π→−65π, 35π→−3π. So
z=2cis6π,2cis32π,2cis(−65π),2cis(−3π).
(b) The four roots have equal modulus 2 and arguments spaced by 2π, so they are the vertices of a square inscribed in the circle ∣z∣=2.
Subscribe to continue reading
Get full access to this lesson and all 10 lessons in this course.