You already met the machine z = r e i θ on the parent note . This page is the firing range : we walk through every kind of case that a complex number can hand you when you try to write it as r e i θ — every quadrant, the axes, the degenerate zero, huge powers, division, and a real-world twist. If a scenario exists, it lives in the matrix below and gets a fully worked example.
Intuition Read this first — what "a case" even means
To write z = x + i y in the form r e i θ you must find two numbers:
the modulus r = x 2 + y 2 — always a length, so never negative;
the argument θ — the anticlockwise angle from the positive real axis.
Finding r is easy and never breaks. Finding θ is where cases appear, because the naive formula arctan ( y / x ) only ever hands back an angle between − 9 0 ∘ and + 9 0 ∘ (that is − 2 π to + 2 π ). Points living to the left of the imaginary axis, or on an axis, need a correction. Each correction is a "case".
Below, x is the real part, y the imaginary part. "Ref" means the reference angle α = arctan x y — the acute angle the arrow makes with the real axis, always positive.
Cell
Signs of ( x , y )
Where the arrow points
Correct θ
Example
A Quadrant I
x > 0 , y > 0
up-right
α
Ex 1
B Quadrant II
x < 0 , y > 0
up-left
π − α
Ex 2
C Quadrant III
x < 0 , y < 0
down-left
α − π
Ex 3
D Quadrant IV
x > 0 , y < 0
down-right
− α
Ex 4
E On an axis
one of x , y = 0
along an axis
0 , 2 π , π , − 2 π
Ex 5
F Degenerate
x = 0 , y = 0
nowhere
undefined
Ex 5
G Big power
any
—
De Moivre
Ex 6
H Division
any
—
subtract angles
Ex 7
I Word problem
any
AC circuit phasor
modulus & phase
Ex 8
J Exam twist
negative "radius"
flip by π
Ex 9
The map below shows exactly which region each quadrant cell owns and how the reference angle α (yellow) turns into the true angle θ (blue).
Prerequisites you can lean on: Modulus and argument of a complex number , Polar form (trigonometric) r(cosθ + i sinθ) , De Moivre's theorem , Complex multiplication as rotation and scaling .
Worked example Ex 1 — Convert
z = 3 + 3 i
Forecast: Both parts positive → arrow points up-right → first quadrant. Guess: is θ exactly 4 5 ∘ ? Write your guess before reading on.
Step 1. r = 3 2 + 3 2 = 18 = 3 2 .
Why this step? Modulus is the length of the arrow, x 2 + y 2 — the Pythagorean distance from the origin to the point ( 3 , 3 ) .
Step 2. Reference angle α = arctan 3 3 = arctan ( 1 ) = 4 π .
Why this step? tan is "opposite over adjacent" on the right triangle whose legs are x = 3 and y = 3 . The ratio y / x = 1 tells us the arrow rises as fast as it runs — a perfect 4 5 ∘ slope.
Step 3. x > 0 , y > 0 is Cell A, so θ = α = 4 π — no correction.
Why this step? arctan already returns first-quadrant angles, so its raw output is right here.
Answer: z = 3 2 e iπ /4 .
Verify: r cos θ = 3 2 ⋅ 2 1 = 3 = x ✓ and r sin θ = 3 2 ⋅ 2 1 = 3 = y ✓.
Worked example Ex 2 — Convert
z = − 2 + 2 3 i
Forecast: x < 0 , y > 0 → up-left → second quadrant. The naive arctan ( y / x ) = arctan ( − 3 ) will hand back a negative angle. That is wrong; predict the true angle before checking.
Step 1. r = ( − 2 ) 2 + ( 2 3 ) 2 = 4 + 12 = 16 = 4 .
Why this step? Squaring kills the sign, so the modulus doesn't care that x is negative.
Step 2. Reference angle α = arctan − 2 2 3 = arctan ( 3 ) = 3 π .
Why this step? We use the absolute ratio so α is the honest acute angle the arrow makes with the real axis — 6 0 ∘ here.
Step 3. Cell B correction: θ = π − α = π − 3 π = 3 2 π .
Why this step? Look at figure s01. In quadrant II the arrow is a mirror of the reference triangle across the imaginary axis, so its true angle is 18 0 ∘ minus the reference angle. Blindly trusting arctan ( − 3 ) = − 3 π would place the arrow in quadrant IV — the wrong half of the plane.
Answer: z = 4 e i 2 π /3 .
Verify: 4 cos 3 2 π = 4 ⋅ ( − 2 1 ) = − 2 = x ✓ and 4 sin 3 2 π = 4 ⋅ 2 3 = 2 3 = y ✓.
Worked example Ex 3 — Convert
z = − 1 − i
Forecast: Both parts negative → down-left → third quadrant. Here y / x = − 1 − 1 = + 1 , so arctan gives + 4 π — pointing up-right , the opposite corner! Predict how far off it is.
Step 1. r = ( − 1 ) 2 + ( − 1 ) 2 = 2 .
Step 2. Reference angle α = arctan − 1 − 1 = arctan ( 1 ) = 4 π .
Why this step? Same acute 4 5 ∘ triangle as Ex 1 — the magnitude of the slope is identical; only the direction differs.
Step 3. Cell C correction: θ = α − π = 4 π − π = − 4 3 π .
Why this step? In quadrant III the arrow points into the down-left corner. Subtracting π from the reference angle rotates it a half-turn into the correct quadrant. (You could equally write θ = π + α = 4 5 π ; the two differ by a full turn 2 π and name the same direction. We keep arguments in ( − π , π ] , so − 4 3 π is the principal value.)
Answer: z = 2 e − i 3 π /4 .
Verify: 2 cos ( − 4 3 π ) = 2 ⋅ ( − 2 1 ) = − 1 = x ✓ and 2 sin ( − 4 3 π ) = 2 ⋅ ( − 2 1 ) = − 1 = y ✓.
Worked example Ex 4 — Convert
z = 1 − i 3
Forecast: x > 0 , y < 0 → down-right → fourth quadrant. This is the one quadrant where raw arctan ( y / x ) is already correct — but let's see why.
Step 1. r = 1 2 + ( − 3 ) 2 = 1 + 3 = 2 .
Step 2. Reference angle α = arctan 1 − 3 = arctan ( 3 ) = 3 π .
Step 3. Cell D correction: θ = − α = − 3 π .
Why this step? The arrow dips below the real axis by 6 0 ∘ , so its anticlockwise angle is negative. Here arctan ( − 3 ) = − 3 π directly , because arctan 's range ( − 2 π , 2 π ) exactly covers quadrants I and IV. That's the whole reason cells A and D need no fix.
Answer: z = 2 e − iπ /3 .
Verify: 2 cos ( − 3 π ) = 2 ⋅ 2 1 = 1 = x ✓ and 2 sin ( − 3 π ) = 2 ⋅ ( − 2 3 ) = − 3 = y ✓.
Worked example Ex 5 — Four points on axes, plus
z = 0
Forecast: On an axis one component is zero, so arctan either gives 0 or divides by zero . Guess each angle from the picture before computing.
Look at figure s02: four arrows lying flat along the axes, plus a dot at the origin.
(i) z = 5 (positive real axis). r = 5 , arrow points along + real, so θ = 0 . 5 e i 0 = 5 .
Why? No imaginary part means no turn from the reference direction.
(ii) z = 4 i (positive imaginary axis). r = 4 , arrow points straight up, so θ = 2 π . 4 e iπ /2 .
Why this step matters? arctan ( y / x ) = arctan ( 4/0 ) is undefined (division by zero) — you must read the angle off the picture, not the formula.
(iii) z = − 7 (negative real axis). r = 7 , arrow points left, so θ = π . 7 e iπ .
Why? A half-turn from the positive real axis. (This is the seed of Euler's identity e^(iπ)+1=0 : e iπ = − 1 .)
(iv) z = − 3 i (negative imaginary axis). r = 3 , arrow points down, so θ = − 2 π . 3 e − iπ /2 .
(v) z = 0 (Cell F, degenerate). r = 0 . There is no arrow , so no direction — arg ( 0 ) is undefined . We simply write z = 0 ; it has no exponential form because you can't ask "which way does a point of zero length face?"
Verify: 5 cos 0 = 5 ✓; 4 sin 2 π = 4 ✓; 7 cos π = − 7 ✓; 3 sin ( − 2 π ) = − 3 ✓.
Common mistake "The origin has argument
0 ."
Why it feels right: 0 is a natural default and r = 0 anyway.
The trap: The exponential form r e i θ with r = 0 gives 0 no matter what θ is , so the angle carries zero information — it isn't 0 , it's genuinely undefined. Treat z = 0 as a special case and never feed it into an argument formula.
Worked example Ex 6 — Compute
( − 1 − i ) 10
Forecast: Expanding a 10th power by binomial is brutal. Exponential form turns it into "power the length, multiply the angle." Guess: will the answer be real, imaginary, or neither?
Step 1. From Ex 3, − 1 − i = 2 e − i 3 π /4 .
Why this step? Once in exponential form, the power rule from De Moivre's theorem applies: ( r e i θ ) n = r n e in θ .
Step 2. ( − 1 − i ) 10 = ( 2 ) 10 e − i ⋅ 10 ⋅ 3 π /4 = 2 5 e − i 30 π /4 = 32 e − i 15 π /2 .
Why this step? ( 2 ) 10 = ( 2 1/2 ) 10 = 2 5 = 32 and the angle simply scales by 10 .
Step 3. Reduce the angle mod 2 π : − 2 15 π = − 2 15 π + 8 π = − 2 15 π + 2 16 π = 2 π .
Why this step? Adding whole turns (2 π = 2 4 π each; we added four turns = 8 π ) doesn't move the arrow but lands θ in the principal range.
Answer: 32 e iπ /2 = 32 i — purely imaginary.
Verify: 32 cos 2 π = 0 (real part) and 32 sin 2 π = 32 (imag part), so 32 i ✓. Cross-check numerically: ( − 1 − i ) 2 = 2 i , so ( − 1 − i ) 10 = (( − 1 − i ) 2 ) 5 = ( 2 i ) 5 = 32 i 5 = 32 i ✓.
Worked example Ex 7 — Compute
1 + i 2 + 2 3 i
Forecast: Division means "divide the lengths, subtract the angles." Convert each, then subtract. Guess the modulus of the result.
Step 1. Numerator: 2 + 2 3 i has r = 4 + 12 = 4 , quadrant I, α = arctan 3 = 3 π , so θ = 3 π . Thus 4 e iπ /3 .
Step 2. Denominator: 1 + i = 2 e iπ /4 (Cell A).
Why these steps? Both must be in exponential form so we can use r 2 e i θ 2 r 1 e i θ 1 = r 2 r 1 e i ( θ 1 − θ 2 ) .
Step 3. 2 e iπ /4 4 e iπ /3 = 2 4 e i ( π /3 − π /4 ) = 2 2 e iπ /12 .
Why this step? 2 4 = 2 4 2 = 2 2 ; and 3 π − 4 π = 12 4 π − 3 π = 12 π .
Answer: 2 2 e iπ /12 .
Verify: Multiply back: 2 2 e iπ /12 ⋅ 2 e iπ /4 = 4 e i ( π /12 + 3 π /12 ) = 4 e iπ /3 , which is the original numerator ✓. Modulus check: 2 4 = 2 2 ≈ 2.828 ✓.
Worked example Ex 8 — Two AC voltages combine
Two voltage sources at the same frequency are written as phasors (complex numbers whose modulus is amplitude and argument is phase):
V 1 = 10 e i 0 , V 2 = 10 e iπ /2 volts.
The total voltage is V = V 1 + V 2 . Find its amplitude and phase.
Forecast: V 1 is 10 V "in phase", V 2 is 10 V "9 0 ∘ ahead". Adding two equal, perpendicular arrows — guess the total amplitude (bigger or smaller than 20 ?).
Step 1. Turn each phasor back to rectangular form to add: V 1 = 10 ( cos 0 + i sin 0 ) = 10 ; V 2 = 10 ( cos 2 π + i sin 2 π ) = 10 i .
Why this step? Exponential form multiplies beautifully but does not add directly — to sum arrows you need their x , y components, exactly like adding vectors.
Step 2. V = 10 + 10 i .
Step 3. Amplitude r = 1 0 2 + 1 0 2 = 10 2 ≈ 14.14 V. Phase: quadrant I, θ = arctan 10 10 = 4 π = 4 5 ∘ .
Why this step? The resultant of two equal perpendicular phasors is 2 times one of them, tilted halfway between — not the naive 20 V, because they are out of phase.
Answer: V = 10 2 e iπ /4 V — amplitude ≈ 14.14 V, phase 4 5 ∘ .
Verify (units + geometry): Both inputs are volts; V 2 + V 2 = V ✓. Sanity: 10 2 ≈ 14.1 < 20 , correct because perpendicular arrows partly "waste" each other's contribution ✓.
Worked example Ex 9 — Simplify
z = − 3 e iπ /6 into standard r e i θ with r > 0
Forecast: A modulus can never be negative, yet here it looks like r = − 3 . What's really going on? Guess the true angle.
Step 1. Factor the minus sign as a rotation: − 1 = e iπ .
Why this step? A leading − 1 is a half-turn of the plane (18 0 ∘ ), i.e. e iπ (Cell E, point (iii)). Absorbing it into the angle keeps the modulus honest and positive.
Step 2. z = e iπ ⋅ 3 e iπ /6 = 3 e i ( π + π /6 ) = 3 e i 7 π /6 .
Why this step? Exponents of the same base add.
Step 3. Put in principal range ( − π , π ] : 6 7 π − 2 π = 6 7 π − 12 π = − 6 5 π .
Why this step? 6 7 π > π , so subtract one full turn to name the same arrow with the standard argument.
Answer: z = 3 e − i 5 π /6 (modulus 3 , argument − 6 5 π , quadrant III).
Verify: Compare rectangular parts. Original: − 3 ( cos 6 π + i sin 6 π ) = − 3 ( 2 3 + 2 1 i ) = − 2 3 3 − 2 3 i . New: 3 ( cos ( − 6 5 π ) + i sin ( − 6 5 π )) = 3 ( − 2 3 − 2 1 i ) = − 2 3 3 − 2 3 i ✓ — identical.
Common mistake "Leave the answer as
− 3 e iπ /6 ."
Why it feels right: Algebraically it's a valid expression.
The fix: Exponential form by convention has r = ∣ z ∣ ≥ 0 . A negative coefficient hides a π -rotation; always fold it in so the modulus is a genuine length.
Recall Which cell needs which correction?
Quadrant I correction? ::: none, θ = α
Quadrant II correction? ::: θ = π − α
Quadrant III principal angle? ::: θ = α − π
Quadrant IV correction? ::: θ = − α
Argument of z = 0 ? ::: undefined (no direction)
How to remove a leading minus sign? ::: multiply by e iπ (add π to the angle)
"A: as-is, B: π minus, C: minus π , D: dip negative." The two quadrants touching the positive real axis (I and IV) need no or only a sign flip; the two on the left (II and III) need a π shift.