This is a companion drill page for the parent note . There we built the master formula from scratch. Here we hunt it into every corner — every quadrant of the input angle, the degenerate cases (zero, real, negative real, pure imaginary), a limiting-behaviour case, a word problem, and an exam twist. If you meet a root problem in the wild, it belongs to one of the cells below.
Recall The one formula we keep reusing
If w = r ( cos θ + i sin θ ) (that is, r = ∣ w ∣ is the length and θ = arg w is the angle), then the n solutions of z n = w are
z k = r 1/ n [ cos ( n θ + 2 π k ) + i sin ( n θ + 2 π k ) ] , k = 0 , 1 , … , n − 1.
Recipe = "R oot the length, A dd 2 π k , D ivide by n , run k from 0 to n − 1 ." Built from De Moivre's Theorem on top of Polar Form of Complex Numbers .
Here cis α is just shorthand for cos α + i sin α — a unit arrow pointing at angle α . We write i for the number with i 2 = − 1 , drawn as one step straight up on the Argand Diagram .
Intuition A word on units — radians vs degrees
The master formula is written in radians (that is why you see 2 π k , one full turn = 2 π ). But radians like 6 5 π are hard to see , so after computing each angle we convert to degrees for the picture, using 180° = π rad (so 6 π = 30° , 2 π = 90° , π = 180° , 2 π = 360° ). Both units name the same arrow direction; degrees are just easier to draw. Whichever you use, be consistent: add 2 π k in radians or 360° k in degrees, never a mix.
Every root problem is fixed by two inputs : where the number w sits (its angle/quadrant, or a degenerate position) and which root n we want. The table lists every distinct case class, and the example that nails it.
Cell
Input class w
What is tricky
Example
A
Positive real (θ = 0 )
"Real ⇒ one root?" trap
Ex 1 (z 3 = 8 revisited fully)
B
Pure imaginary, up (θ = 2 π , boundary)
odd n , spread of 120°
Ex 2 (z 3 = i )
B′
Pure imaginary, down (θ = − 2 π , boundary)
negative axis angle
Ex 2b (z 2 = − i )
C
Negative real (θ = π )
angle exactly π
Ex 3 (z 4 = − 16 check)
D
Quadrant I input
general non-nice angle
Ex 4 (z 2 = 8 + 8 i 3 )
E
Quadrant II input (real part negative)
naive arctan gives wrong quadrant
Ex 5 (z 2 = − 1 + i 3 )
F
Quadrant III input (both parts negative)
naive arctan gives wrong quadrant
Ex 6 (z 3 = − 4 − 4 i 3 )
G
Quadrant IV input (imag part negative)
negative arctan , add 360°
Ex 7 (z 2 = 1 − i 3 )
H
Degenerate w = 0
how many roots now?
Ex 8 (z 5 = 0 )
I
Limiting / large n
roots crowd a circle
Ex 9 (z 12 = 1 , spacing)
J
Word problem (rotation)
translate physics → root
Ex 10 (three equal spins)
K
Exam twist (equation, not z n = w )
rearrange first
Ex 11 (z 4 + z 2 + 1 = 0 )
z 3 = 8 , find ALL roots.
Forecast: 8 is a plain real number. Guess how many roots — one, or three? Where do they sit?
Step 1. Write 8 in polar form: 8 = 8 ( cos 0 + i sin 0 ) , so r = 8 , θ = 0 .
Why this step? The formula eats r and θ ; a positive real number lies on the positive x -axis, angle 0 .
Step 2. Root the length: r 1/3 = 8 1/3 = 2 . All three roots have length 2 .
Why? Only the modulus gets the ordinary real cube root; the angle is handled separately.
Step 3. Angles: ϕ k = 3 0 + 2 π k = 3 2 π k for k = 0 , 1 , 2 .
Why k = 0 , 1 , 2 ? An n -th root has exactly n distinct angles; here n = 3 .
k = 0 : z 0 = 2 ( cos 0 + i sin 0 ) = 2.
k = 1 : z 1 = 2 cis 3 2 π = 2 ( − 2 1 + i 2 3 ) = − 1 + i 3 .
k = 2 : z 2 = 2 cis 3 4 π = − 1 − i 3 .
Verify: ( − 1 + i 3 ) 3 : length 2 cubed = 8 , angle 120° × 3 = 360° = 0° , so it equals 8 . ✓ Three trees planted 120° apart on a circle of radius 2 .
Intuition Figure 1 — the roots of
z 3 = 8
The cyan circle has radius 2 (the rooted length). The three amber arrows are z 0 , z 1 , z 2 ; the white dots at their tips are the roots. The amber triangle joining them is a perfect equilateral triangle — read off the equal 120° gaps between consecutive arrows.
Common mistake "Real number ⇒ single real root."
A cubic over C must have 3 roots (Fundamental Theorem of Algebra ). Two of them are the complex conjugate pair. Real value ≠ one root.
z 3 = i .
Forecast: i points straight up. Cube roots of an "upward" number — will one of them be real?
Step 1. i = 1 ⋅ cis 2 π , so r = 1 , θ = 2 π (that is 90° ).
Why? i has length 1 and points at 90° on the Argand Diagram .
Step 2. Length of roots: 1 1/3 = 1 . They sit on the unit circle.
Step 3. Angles: ϕ k = 3 π /2 + 2 π k = 6 π + 3 2 π k .
Why this step? We apply the recipe with r = 1 , θ = 2 π , n = 3 : Add 2 π k to the angle and Divide by 3 , giving 6 π + 3 2 π k — each step in k turns the root by 3 2 π = 120° .
k = 0 : ϕ = 6 π = 30° ⇒ z 0 = cos 30° + i sin 30° = 2 3 + 2 1 i .
k = 1 : ϕ = 6 π + 3 2 π = 6 5 π = 150° ⇒ z 1 = − 2 3 + 2 1 i .
k = 2 : ϕ = 6 π + 3 4 π = 2 3 π = 270° ⇒ z 2 = − i .
Why does z 2 have zero real part and equal − i ? An angle of 270° (equivalently 2 3 π rad) points straight down the imaginary axis, where cos 270° = 0 and sin 270° = − 1 ; a unit arrow there is exactly the number − i .
Verify: ( − i ) 3 = − i 3 = − ( − i ) = i . ✓ And each step in k adds 120° , spacing 3 2 π — a regular triangle.
z 2 = − i .
Forecast: − i points straight down . That axis is a boundary just like + i , but the angle is negative. Guess where the two square roots land — will they be the same as the roots of z 2 = i ?
Step 1 — polar form. − i has length 1 and points along the negative imaginary axis, so θ = − 2 π = − 90° (equivalently 2 3 π = 270° ; same arrow).
Why the negative angle? On the boundary "straight down" is − 90° ; either name works, but we must pick one and keep it — here − 2 π keeps the arithmetic light.
Step 2 — root length. 1 1/2 = 1 . Both roots sit on the unit circle.
Step 3 — angles. ϕ k = 2 − π /2 + 2 π k = − 4 π + π k for k = 0 , 1 .
Why this step? Recipe with r = 1 , θ = − 2 π , n = 2 : Add 2 π k , Divide by 2 , giving − 4 π + π k — the two square roots sit 180° apart, as square roots always do.
k = 0 : ϕ = − 45° ⇒ z 0 = cos ( − 45° ) + i sin ( − 45° ) = 2 1 ( 1 − i ) .
k = 1 : ϕ = 135° ⇒ z 1 = − 2 1 ( 1 − i ) .
Verify: ( 2 1 − i ) 2 = 2 ( 1 − i ) 2 = 2 1 − 2 i − 1 = − i . ✓ Note these are the mirror image (across the real axis) of the roots of z 2 = i — because − i is the reflection of i .
z 4 = − 16 .
Forecast: − 16 points along the negative x -axis. Four fourth-roots — will any land on an axis, or all off-axis?
Step 1. − 16 = 16 cis π , so r = 16 , θ = π (that is 180° ).
Why θ = π ? A negative real number lies to the left of the origin, angle 180° .
Step 2. 1 6 1/4 = 2 . Circle radius 2 .
Step 3. ϕ k = 4 π + 2 π k for k = 0 , 1 , 2 , 3 : gives 4 π , 4 3 π , 4 5 π , 4 7 π — i.e. 45° , 135° , 225° , 315° .
z 0 = 2 cis 45° = 2 + i 2 .
z 1 = − 2 + i 2 , z 2 = − 2 − i 2 , z 3 = 2 − i 2 .
Why none on an axis? Because θ = π is odd-ish w.r.t. the quarter-turns; the first root already sits at 45° .
Verify: ( 2 + i 2 ) 4 : length 2 4 = 16 , angle 45° × 4 = 180° , giving 16 cis 180° = − 16 . ✓
Intuition Figure 2 — the roots of
z 4 = − 16
Four amber arrows, all length 2 , land at 45° , 135° , 225° , 315° . The white dots form a perfect square (an amber outline connects them), rotated 45° from the axes. Each quarter-turn of 90° takes you to the next root.
z 2 = 8 + 8 i 3 .
Forecast: Both parts positive ⇒ w sits in Quadrant I. Guess its angle before computing.
Step 1 — modulus. r = 8 2 + ( 8 3 ) 2 = 64 + 192 = 256 = 16.
Why? ∣ w ∣ = x 2 + y 2 is the arrow's length by Pythagoras.
Step 2 — argument. tan θ = 8 8 3 = 3 . Since both parts are positive we are in Quadrant I, so θ = 3 π = 60° (no adjustment needed).
Why check the quadrant? arctan 3 could be 60° or 240° ; the positive signs pin it to 60° .
Step 3. Root length: 1 6 1/2 = 4 ; angles ϕ k = 2 π /3 + 2 π k = 6 π + π k .
Why this step? We now feed r = 16 and θ = 3 π into the master recipe: Root the length (1 6 1/2 = 4 , so both roots have length 4 ), then Add 2 π k to the angle and Divide by n = 2 , which is exactly 2 π /3 + 2 π k = 6 π + π k — each step in k turning the root by π = 180° .
k = 0 : ϕ = 6 π = 30° ⇒ z 0 = 4 ( cos 30° + i sin 30° ) = 2 3 + 2 i .
k = 1 : ϕ = 6 π + π = 210° ⇒ z 1 = − 2 3 − 2 i .
Verify: ( 2 3 + 2 i ) 2 = 12 + 2 ⋅ 2 3 ⋅ 2 i + ( 2 i ) 2 = 12 + 8 3 i − 4 = 8 + 8 3 i . ✓ Two square roots, negatives of each other.
z 2 = − 1 + i 3 .
Forecast: Real part negative, imaginary part positive ⇒ Quadrant II. Your calculator's arctan ( − 1 3 ) returns − 60° — pointing into Quadrant IV. Guess the true angle first.
Step 1 — modulus. r = ( − 1 ) 2 + ( 3 ) 2 = 1 + 3 = 2.
Why? ∣ w ∣ = x 2 + y 2 , the arrow's length.
Step 2 — argument (careful). tan θ = − 1 3 = − 3 , base angle 60° . But x < 0 , y > 0 ⇒ Quadrant II, so measure from the negative real axis: θ = 180° − 60° = 120° = 3 2 π .
Why 180° − 60° ? tan repeats every 180° and cannot tell Q II from Q IV; the signs (x < 0 , y > 0 ) force us into the upper-left, angle 120° .
Step 3. Root length 2 ; angles ϕ k = 2 2 π /3 + 2 π k = 3 π + π k .
Why this step? Feed r = 2 , θ = 3 2 π into the recipe: Root length 2 1/2 = 2 ; Add 2 π k and Divide by n = 2 ⇒ 3 π + π k , each k turning by 180° .
k = 0 : ϕ = 60° ⇒ z 0 = 2 ( cos 60° + i sin 60° ) = 2 2 + i 2 6 .
k = 1 : ϕ = 240° ⇒ z 1 = − 2 2 − i 2 6 .
Verify: z 0 2 : length ( 2 ) 2 = 2 , angle 60° × 2 = 120° , i.e. 2 cis 120° = 2 ( − 2 1 + i 2 3 ) = − 1 + i 3 . ✓
z 3 = − 4 − 4 i 3 .
Forecast: Both parts negative ⇒ Quadrant III. If you blindly type arctan ( − 4 − 4 3 ) your calculator gives 60° — but that points into Quadrant I. Guess the true angle.
Step 1 — modulus. r = ( − 4 ) 2 + ( − 4 3 ) 2 = 16 + 48 = 64 = 8.
Step 2 — argument (the careful bit). tan θ = − 4 − 4 3 = 3 , whose base angle is 60° . But x < 0 and y < 0 ⇒ Quadrant III, so add 180° :
θ = 180° + 60° = 240° = 3 4 π .
Why add 180° ? tan repeats every 180° , so it cannot tell Quadrant I from III. The signs of the real and imaginary parts break the tie — see the two arrows in the figure.
Step 3. Root length: 8 1/3 = 2 . Angles: ϕ k = 3 4 π /3 + 2 π k = 9 4 π + 3 2 π k .
Why this step? Root length 8 1/3 = 2 ; Add 2 π k and Divide by n = 3 , giving 9 4 π + 3 2 π k — each k turning by 3 2 π = 120° .
k = 0 : ϕ = 9 4 π = 80°.
k = 1 : ϕ = 80° + 120° = 200°.
k = 2 : ϕ = 80° + 240° = 320°.
So z k = 2 cis ( 80° ) , 2 cis ( 200° ) , 2 cis ( 320° ) .
Verify: Take z 0 = 2 cis 80° : cube ⇒ length 8 , angle 240° , i.e. 8 cis 240° = 8 ( − 2 1 − i 2 3 ) = − 4 − 4 i 3 . ✓
Intuition Figure 3 — signs fix the quadrant
The solid amber arrow is the true w in Quadrant III at 240° . The dashed cyan arrow is the wrong Q I direction (60° ) a bare arctan hands you. The white arc marks the + 180° correction that carries one into the other — always let the signs of x and y decide.
arctan without checking signs.
arctan only returns angles in ( − 90° , 90° ) . For Quadrant II subtract from 180° ; for Quadrant III add 180° ; for Quadrant IV either leave negative or add 360° . Locate w first, then trust the formula.
z 2 = 1 − i 3 .
Forecast: Real part positive, imaginary part negative ⇒ Quadrant IV. Here the naive arctan actually does land in the right quadrant (as a negative angle) — but we want a tidy positive angle. Guess it.
Step 1 — modulus. r = 1 2 + ( − 3 ) 2 = 1 + 3 = 2.
Step 2 — argument. tan θ = 1 − 3 = − 3 , base angle 60° . Since x > 0 , y < 0 we are in Quadrant IV, so θ = − 60° , or equivalently θ = 360° − 60° = 300° = 3 5 π .
Why add 360° ? − 60° and 300° point at the exact same arrow (a full turn apart); we choose the positive 300° so the root angles come out positive. Both are fine so long as you stay consistent.
Step 3. Root length 2 ; angles ϕ k = 2 5 π /3 + 2 π k = 6 5 π + π k .
Why this step? Root 2 1/2 = 2 ; Add 2 π k , Divide by n = 2 ⇒ 6 5 π + π k , each k turning by 180° .
k = 0 : ϕ = 150° ⇒ z 0 = 2 cis 150° = − 2 6 + i 2 2 .
k = 1 : ϕ = 330° ⇒ z 1 = 2 6 − i 2 2 .
Verify: z 1 2 : length ( 2 ) 2 = 2 , angle 330° × 2 = 660° ≡ 300° , i.e. 2 cis 300° = 2 ( 2 1 − i 2 3 ) = 1 − i 3 . ✓
z 5 = 0 .
Forecast: The master formula assumed w = 0 . What happens when w = 0 — five roots, or something else?
Step 1. w = 0 has modulus r = 0 and no defined argument (a zero-length arrow points nowhere).
Why this breaks the formula: the recipe's "spread the angle" step needs a starting direction θ = arg w , but arg 0 is undefined — there is no arrow to rotate. The formula simply does not apply.
Step 2. Even if we tried, the root length would be 0 1/5 = 0 . Every one of the five "spokes" has length zero, so they all collapse onto the single point at the origin — the regular polygon shrinks to a dot.
Why? A polygon of radius 0 has all its vertices piled at the centre; the angles no longer separate anything.
Step 3 — solve directly instead. z 5 = 0 ⟺ z ⋅ z ⋅ z ⋅ z ⋅ z = 0 ⟺ z = 0 . So there is exactly one distinct root , z = 0 .
Why still call it "five roots"? The Fundamental Theorem of Algebra counts roots with multiplicity : z 5 = ( z − 0 ) 5 has the root 0 repeated 5 times, so z = 0 has multiplicity 5 . The "n distinct roots" statement of the master formula holds only for w = 0 ; the w = 0 case is exactly where the n separate points merge into one point of multiplicity n .
Verify: z 5 = 0 factors as ( z − 0 ) 5 ; the only solution is z = 0 , counted 5 times. ✓ Degenerate: n roots become one point of multiplicity n .
Worked example The 12th roots of unity:
z 12 = 1 . How far apart are neighbours, and what happens as n grows?
Forecast: With n = 12 , is the spacing 30° or 12° ? And if n → ∞ ?
Step 1. 1 = 1 cis 0 , so r = 1 , θ = 0 . Root length 1 1/12 = 1 — all on the unit circle.
Why this step? Feed r = 1 , θ = 0 , n = 12 into the recipe; the rooted length is 1 , so every root lands on the unit circle.
Step 2. z k = cis 12 2 π k = cis ( 30° k ) , k = 0 , … , 11 .
Why this step? Add 2 π k (with θ = 0 ) and Divide by 12 gives angle 12 2 π k = 30° k . These are the Roots of Unity : powers of ω = e 2 π i /12 , a Geometric Progression 1 , ω , … , ω 11 .
Step 3 — spacing. Consecutive roots differ by 12 2 π = 6 π = 30° .
Why this is a limit story: spacing = n 2 π → 0 as n → ∞ ; the n -gon's vertices get denser and the polygon tends to the full circle.
Verify (sum): ∑ k = 0 11 ω k = ω − 1 ω 12 − 1 = ω − 1 1 − 1 = 0. ✓ By symmetry the 12 arrows cancel exactly.
Intuition Figure 4 — twelve roots on the unit circle
Twelve white dots sit evenly on the cyan unit circle, joined by an amber 12-gon. Each amber arrow is separated from its neighbour by exactly 30° . Imagine doubling n : the dots halve their spacing and the polygon hugs the circle ever more tightly.
Worked example A robot arm at position
8 (i.e. the point 8 + 0 i , 8 units right) is reached by applying the same rotation-and-scaling three times starting from the point 1 . Each application multiplies the position by a fixed complex number z . Find all possible single moves z .
Forecast: "Same move three times, landing on 8 " — this is secretly z 3 = 8 . How many valid moves?
Step 1 — translate. Three identical steps from 1 : 1 ⋅ z ⋅ z ⋅ z = z 3 . Requiring the end point 8 gives z 3 = 8 .
Why a root problem? "Repeat a multiplication n times to reach w " is exactly z n = w .
Step 2 — reuse Ex 1. Roots: z = 2 , − 1 + i 3 , − 1 − i 3 .
Why? This is the same equation as Ex 1; the three cube roots of 8 are the three admissible moves.
Interpretation: each move scales length by 2 (= 8 1/3 ) and rotates by 0° , 120° , or 240° — all three end at 8 after three repeats.
Step 3 — physical read. The rotating moves (120° , 240° ) still land on 8 because 3 × 120° = 360° ≡ 0° .
Why? Three turns of 120° complete exactly one full loop, cancelling the rotation and leaving only the length scaling to 8 .
Verify: 2 3 = 8 ; and ( − 1 + i 3 ) 3 = 8 (length 2 3 = 8 , angle 3 × 120° = 360° = 0° ). ✓ Three admissible moves.
z 4 + z 2 + 1 = 0 .
Forecast: This is not in the form z n = w . Can we still use roots? Try substituting.
Step 1 — substitute u = z 2 . Then u 2 + u + 1 = 0 .
Why? It turns a quartic into a quadratic we can solve, then peel back.
Step 2 — solve the quadratic. u = 2 − 1 ± 1 − 4 = 2 − 1 ± i 3 . In polar: both have modulus 1 , and
u + = cis 120° , u − = cis 240°.
Why polar? We now need square roots of u , and roots love polar form.
Step 3 — take square roots of each u (that is z 2 = u ).
Why the formula ϕ = 2 θ + 180° k ? Taking a square root divides the argument by 2 (that is the 2 θ ), and the recipe's "+ 2 π k before dividing by n = 2 " becomes 2 2 π k = π k = 180° k — the two square roots are always 180° apart because arg is only fixed up to full turns.
z 2 = cis 120° : angles 2 120° + 180° k = 60° , 240° . So z = cis 60° , cis 240° .
z 2 = cis 240° : angles 120° , 300° . So z = cis 120° , cis 300° .
Step 4 — collect. Four roots: cis 60° , cis 120° , cis 240° , cis 300° — the primitive 6th roots of unity.
Why 6th? z 4 + z 2 + 1 = z 2 − 1 z 6 − 1 , so its roots are 6th-roots of unity that are not square-roots of unity.
Verify: For z = cis 60° : z 2 = cis 120° , z 4 = cis 240° , and cis 240° + cis 120° + 1 = ( − 2 1 − i 2 3 ) + ( − 2 1 + i 2 3 ) + 1 = 0. ✓
Recall One-line summary of the matrix
Every case reduces to: find r = ∣ w ∣ and θ = arg w with the correct quadrant , then run "Root, Add 2 π k , Divide, k=0..n−1." Degenerate w = 0 gives a single root of multiplicity n ; equations not of the form z n = w get a substitution first.
Which two facts fully determine a root problem's answer? The polar data of w (r and correctly-quadranted θ ) and the index n .
What is arg 0 , and how many distinct roots does z n = 0 have? Undefined; exactly one distinct root z = 0 , of multiplicity n .
Why must you subtract 60° from 180° for a Quadrant II input? Because tan has period 180° ; the signs (x < 0 , y > 0 ) put the arrow upper-left, so θ = 180° − base angle .
For a Quadrant IV input why can you add 360° to the raw negative arctan ? A full turn names the same direction; 300° and − 60° are the same arrow, and + 300° keeps root angles positive.
For an equation like z 4 + z 2 + 1 = 0 , first move? Substitute u = z 2 to get a quadratic, solve, then take square roots.