11.7b DeMoivre's Theorem and the nth Roots of a Complex Number

What the polar form is for. Establishes that multiplication multiplies moduli and adds arguments, that division divides and subtracts, and that powers follow by induction as DeMoivre's Theorem. Reads the product rule geometrically as a stretch followed by a rotation, then solves the root equation to show that every non-zero complex number has exactly n distinct nth roots, evenly spaced around a circle at the vertices of a regular polygon.

Subject: Trigonometry · 64 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. DeMoivre's Theorem and the nth Roots of a Complex Number

Title

Trigonometry · Section 11.7b

What the polar form was built for: powers in one line, and n roots on a circle

2. What this lesson gives you

Objectives

The previous lesson assembled the polar form of a complex number and never used it for anything. This lesson is the payoff: three rules that turn multiplication into addition of angles, and one formula that produces every root of every non-zero complex number.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, pp. 997-1004 — the pages these objectives are drawn from

The last lesson was all setup. You learned to measure a complex number by how far it is from the origin and which direction it points, and you learned to write that as a modulus times cis of an angle. Nothing you could not already do came out of it.

This lesson is where that pays. Three short rules say what multiplication, division and powers do to a modulus and an argument, and every one of them turns a hard operation into two easy ones. A fifth power that would take a binomial expansion with six terms becomes one multiplication and one addition.

Then the whole thing runs backwards. If a power multiplies the angle, a root must divide it, and the moment you write that down you discover that there is not one nth root of a complex number but exactly n of them, sitting at the corners of a regular polygon. That picture is the thing to carry out of this lesson; if you remember it, you can rebuild the formula from it.

3. Before the shortcut: multiply two of them the long way

Warm-up

Discussion prompt

Multiply 2 root 3 plus 2i by negative 1 plus i root 3, using nothing but the distributive law and the fact that i squared is negative one. Count how many separate products you had to form. Then imagine doing the fifth power of the second factor the same way.

Do the multiplication before you read on. It is not hard, just tedious: four products, two of which combine, and one that needs the fact that i squared is negative one.

Now imagine the same technique applied to the fifth power of the second factor. The Binomial Theorem would give you six terms, each with a binomial coefficient, and powers of i running from the zeroth to the fifth that all have to be simplified separately. It is perfectly possible and almost nobody gets it right first time.

Keep both of these on the page. In about ten slides you will do exactly the same two computations using the polar form, and the contrast is the entire argument for why anyone bothers with polar form at all.

4. Multiplying in polar form

Section

Part 1

5. Multiply the moduli, add the arguments

Concept

Two complex numbers in polar form carry two pieces of information each: a length and a direction. Multiplication treats the two pieces completely separately.

Theorem 11.16 — For complex numbers with polar forms z with modulus and argument alpha, and w with modulus and argument beta: the product has modulus the product of the moduli and argument the sum of the arguments; the nth power has modulus the nth power and argument n times the argument; and the quotient has modulus the quotient and argument the difference of the arguments.

\[ zw = |z||w|\,\operatorname{cis}(\alpha + \beta), \qquad \frac{z}{w} = \frac{|z|}{|w|}\,\operatorname{cis}(\alpha - \beta) \]

operationwhat happens to the moduluswhat happens to the argument
z times wthe moduli multiplythe arguments add
z divided by wthe moduli dividethe arguments subtract
z to the nth powerthe modulus is raised to nthe argument is multiplied by n
z plus wno rule at allno rule at all

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 997 — Theorem 11.16

Read the theorem as a statement about two separate channels. A complex number carries a size and a direction, and multiplication acts on each channel independently: sizes multiply, directions add. Nothing in the product rule mixes the two.

That independence is what makes it useful. Multiplying two real numbers is something you can do; adding two angles is something you can do; so the theorem reduces a genuinely awkward operation to two familiar ones.

Look at the last row of the table, because it is the row people forget. There is no rule for sums. Polar form is built out of a distance and an angle, and neither of those behaves simply when you add two numbers. When a problem asks you to add, convert back to rectangular form first.

6. Reading the product rule piece by piece

Notation

Annotate

On: \( zw = |z||w|\,\operatorname{cis}(\alpha + \beta) \)

  • A product of two ordinary positive real numbers. Nothing complex happens here: the lengths simply multiply, as they would for any scaling.
  • Shorthand for cos of the angle plus i sin of the angle, fixed in the previous lesson. It is a point on the unit circle; the modulus in front sets the distance.
  • The arguments ADD. This is the whole content of the theorem, and the reason a fifth power costs one multiplication instead of a binomial expansion.
  • Arguments are only defined up to full turns, so α + β may come out bigger than 2π. Subtract full turns at the end to get the principal argument.

Take the annotations one at a time and connect each one to the picture of an arrow from the origin. The first piece is a plain product of two positive real numbers, so it is only saying that the arrow gets longer by a factor.

The middle piece is the one that does the work. The two arguments add, and that is the entire theorem; everything else in this lesson is a consequence of it, including DeMoivre and including the root formula.

The final note is a practical one. Angles are only determined up to whole turns, so when you add two arguments you may land past a full turn. Nothing is wrong when that happens; you simply subtract full turns at the end to report a principal argument, exactly as you did with coterminal angles in Chapter 10.

7. Proving the product rule: where the sum identities pay off

Worked example

Take two complex numbers in polar form and multiply them out in full.

\[ z = |z|\,\operatorname{cis}(\alpha), \qquad w = |w|\,\operatorname{cis}(\beta) \]

Write the product and pull the moduli to the front

Why: The moduli are real numbers, so they commute with everything else.

\[ zw = |z||w|\,[\cos\alpha + i\sin\alpha][\cos\beta + i\sin\beta] \]

Expand the bracket

Why: Four products, the last of them carrying i squared.

\[ = \cos\alpha\cos\beta + i\cos\alpha\sin\beta + i\sin\alpha\cos\beta + i^2\sin\alpha\sin\beta \]

Replace i squared by negative one and group

Why: The two real terms together, the two terms with a factor of i together.

\[ = (\cos\alpha\cos\beta - \sin\alpha\sin\beta) + i(\sin\alpha\cos\beta + \cos\alpha\sin\beta) \]

Recognise the sum identities from Chapter 10

Why: Cosine of a sum in the first bracket, sine of a sum in the second.

\[ = \cos(\alpha + \beta) + i\sin(\alpha + \beta) = \operatorname{cis}(\alpha + \beta) \]

Restore the moduli

Why: They were carried along untouched.

\[ zw = |z||w|\,\operatorname{cis}(\alpha + \beta) \]

Check on a product you already know

Why: Take z and w both equal to i, whose polar form is one times cis of pi over two. The rule must reproduce the defining property of i.

\[ i\cdot i = 1\cdot 1\,\operatorname{cis}\!\left(\tfrac{\pi}{2} + \tfrac{\pi}{2}\right) = \operatorname{cis}(\pi) = -1 \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 997 — proof of Theorem 11.16, the product rule

This is a proof worth being able to reproduce, because it is short and because it is the only place in the course where the sum identities are indispensable rather than convenient.

Follow the middle three steps with your eye on the four terms of the expansion. Two of them have no factor of i and two of them do, except that one of the first pair carries i squared, which flips its sign. Group them that way and you are staring at the cosine of a sum in one bracket and the sine of a sum in the other.

The check at the end is the smallest possible test of the rule, and it is worth doing because it uses the one complex multiplication you have known since the first day: i times i is negative one. In polar terms that reads as a quarter turn followed by another quarter turn, landing on the negative real axis. If the rule failed here it would fail everywhere.

8. Where the angle of the product comes from

Picture it

Figure (svg): A unit circle with three arrows from the origin: one at thirty degrees for z, one at one hundred and twenty degrees for w, and one at one hundred and fifty degrees for the product. An arc of thirty degrees sits at the start and the same one hundred and twenty degree arc is drawn twice, once from the axis and once stacked on top of the first arc.

The same angle beta is drawn twice: once from the real axis, and once starting where alpha finished. Stacking the two arcs is what the product rule does to arguments.

The dashed arc is the argument of w measured from the real axis. The solid orange arc is the same angle again, this time starting where the argument of z finished. The product lands at the end of the stacked pair.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, pp. 999-1000 — the geometric reading of the product rule

Spend a moment tracing the arcs rather than reading the formula. The small arc at the bottom is the argument of z. The dashed arc is the argument of w, measured the ordinary way from the positive real axis.

Now look at the solid orange arc. It is the same size as the dashed one, but it starts where the first arc ended. That is what adding arguments means geometrically: you take the turn that w represents and apply it on top of the turn that z already has.

Every arrow here has length one because the picture is only about direction. Put the moduli back in and the arrows get longer or shorter, but the angle story is unchanged. That separation between length and direction is worth holding onto for the rest of the lesson.

9. Dividing: the same identities, run backwards

Concept

Division is handled by the trick that rationalises any denominator: multiply above and below by the conjugate of the bottom.

\[ \frac{z}{w} = \frac{|z|}{|w|}\cdot\frac{\cos\alpha + i\sin\alpha}{\cos\beta + i\sin\beta}\cdot\frac{\cos\beta - i\sin\beta}{\cos\beta - i\sin\beta} \]

\[ \text{top} = (\cos\alpha\cos\beta + \sin\alpha\sin\beta) + i(\sin\alpha\cos\beta - \cos\alpha\sin\beta) = \operatorname{cis}(\alpha - \beta) \]

\[ \text{bottom} = \cos^2\beta - i^2\sin^2\beta = \cos^2\beta + \sin^2\beta = 1 \]

\[ \frac{z}{w} = \frac{|z|}{|w|}\,\operatorname{cis}(\alpha - \beta) \]

The difference identities collapse the top and the Pythagorean identity collapses the bottom to the number one, which is why nothing is left over.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 998 — proof of Theorem 11.16, the quotient rule

You have met this manoeuvre before, in Chapter 3, when you divided complex numbers in rectangular form: multiply the top and the bottom by the conjugate of the bottom. The only difference is that everything is written with cosines and sines now.

Watch the bottom first, because it is the surprise. The conjugate product leaves cosine squared minus i squared times sine squared, and since i squared is negative one that becomes cosine squared plus sine squared, which the Pythagorean identity collapses to the number one. The denominator simply disappears.

The top is the sum identities again, run with a minus sign in the middle, which makes them the difference identities. So the quotient rule is not really a new proof at all. If you can reproduce the product rule you can reproduce this one, and you should expect the arguments to subtract for exactly the reason they added before.

10. Example 11.7.3, parts 1 and 3: a product and a quotient

Worked example

Both numbers arrive in rectangular form, so the first job is to convert them.

\[ z = 2\sqrt3 + 2i, \qquad w = -1 + i\sqrt3 \]

Find the modulus of each

Why: Square, add, take the square root.

\[ |z| = \sqrt{(2\sqrt3)^2 + 2^2} = \sqrt{16} = 4, \qquad |w| = \sqrt{(-1)^2 + (\sqrt3)^2} = 2 \]

Find an argument of each from the tangent and the quadrant

Why: z is in Quadrant I, w is in Quadrant II, so the arctangent has to be adjusted for w.

\[ \tan\alpha = \frac{2}{2\sqrt3} = \frac{\sqrt3}{3} \Rightarrow \alpha = \frac{\pi}{6}, \qquad \tan\beta = -\sqrt3 \Rightarrow \beta = \frac{2\pi}{3} \]

Write both in polar form

Why: Modulus and argument, side by side.

\[ z = 4\,\operatorname{cis}\!\left(\tfrac{\pi}{6}\right), \qquad w = 2\,\operatorname{cis}\!\left(\tfrac{2\pi}{3}\right) \]

Multiply the moduli and add the arguments

Why: Four times two, and a sixth of pi plus two thirds of pi.

\[ zw = 8\,\operatorname{cis}\!\left(\tfrac{\pi}{6} + \tfrac{2\pi}{3}\right) = 8\,\operatorname{cis}\!\left(\tfrac{5\pi}{6}\right) \]

Convert the product back to rectangular form

Why: Cosine of five sixths of pi is negative root three over two; the sine is one half.

\[ zw = 8\!\left(-\tfrac{\sqrt3}{2} + \tfrac12 i\right) = -4\sqrt3 + 4i \]

Divide the moduli and subtract the arguments

Why: Four over two, and a sixth of pi minus two thirds of pi.

\[ \frac{z}{w} = 2\,\operatorname{cis}\!\left(\tfrac{\pi}{6} - \tfrac{2\pi}{3}\right) = 2\,\operatorname{cis}\!\left(-\tfrac{\pi}{2}\right) = -2i \]

Check both answers against the long multiplication

Why: The warmup multiplied the same two numbers out by hand and also got negative four root three plus four i; and multiplying the quotient back by w must return z.

\[ (-2i)(-1 + i\sqrt3) = 2i - 2i^2\sqrt3 = 2\sqrt3 + 2i = z \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 999 — Example 11.7.3, parts 1 and 3

This example is honest about the cost of the method. Both numbers arrive in rectangular form, so you have to convert them first, and for a single product that conversion is more work than just multiplying the brackets out. The conversion is the price of admission.

The conversion itself is last lesson's skill: the modulus by Pythagoras, the argument by the tangent and the quadrant. Notice that for w the quadrant matters, because the inverse tangent of negative root three hands you an angle in Quadrant IV, and w is in Quadrant II. That is the single most common slip in this whole topic.

Once both are in polar form the product and the quotient are each one line. The check is worth copying as a habit: multiply your quotient back by w and you should recover z. Any slip in the arithmetic shows up immediately, and it costs you about fifteen seconds.

11. Finish the product and the quotient

Fill the middle

\[ z = 3\,\operatorname{cis}\!\left(\tfrac{\pi}{4}\right), \qquad w = 5\,\operatorname{cis}\!\left(\tfrac{\pi}{6}\right) \]

Fill in the blanks

zw = 15\,\operatorname5\!\left(\frac1\pi}___\right), \qquad \frac______ = \frac______\,\operatorname___\!\left(\frac___\pi}___\right)

Why: The moduli multiply to fifteen and divide to three fifths. Over a common denominator of twelve the arguments are three twelfths and two twelfths of pi, so they add to five twelfths and subtract to one twelfth.

Fill both lines in before you reveal. The moduli are the easy half: they multiply for a product and divide for a quotient, so fifteen and three fifths.

The arguments need a common denominator. A quarter is three twelfths and a sixth is two twelfths, so the sum is five twelfths of pi and the difference is one twelfth of pi. Putting everything over twelve before adding or subtracting is a small discipline that prevents most of the arithmetic errors in this section.

Notice how little work this was compared with multiplying two rectangular numbers whose angles are a quarter and a sixth of pi. Those have irrational real and imaginary parts, and the rectangular multiplication would be genuinely unpleasant.

12. Trap: expecting a rule for sums

Trap

The trap

The three rules are so tidy that a fourth gets invented:

\[ z + w = (|z| + |w|)\,\operatorname{cis}(\alpha + \beta)\,? \]

Wrong. Try it with i and negative i.

\[ i + (-i) = 0 \]

\[ (1+1)\,\operatorname{cis}(0) = 2 \]

The fix

Addition is the operation rectangular form is good at: add the real parts, add the imaginary parts. Multiplication is the operation polar form is good at. Neither form is good at both, which is exactly why you convert back and forth.

\[ z + w: \text{ rectangular} \]

\[ zw, \; z^n, \; \sqrt[n]{z}: \text{ polar} \]

The three rules are so symmetrical that the brain wants to complete the set with a fourth. There is no fourth. The counterexample on the left settles it in one line: i and its negative add to zero, but the invented rule would give two.

The deeper reason is worth understanding rather than memorising. The modulus of a sum depends on the angle between the two numbers, not just on their two moduli, which is the triangle inequality in disguise. No rule of the form on the left could ever capture that.

The practical rule is simple. If the problem is an addition or a subtraction, work in rectangular form. If it is a multiplication, a division, a power or a root, work in polar form. When a problem mixes them, convert at the boundary between the two kinds of operation.

13. Two arguments that cancel

Prediction

\[ z = 5\,\operatorname{cis}\!\left(\tfrac{\pi}{3}\right), \qquad w = 4\,\operatorname{cis}\!\left(-\tfrac{\pi}{3}\right) \]

Predict first

Without computing anything in rectangular form, where does the product zw lie?

  • On the positive real axis, at 20
  • On the negative real axis, at −20
  • On the imaginary axis, at 20i
  • At 9 cis(0)

Correct: On the positive real axis, at 20

Why: Arguments add, and these two are negatives of each other, so the product has argument zero and lies on the positive real axis. Moduli multiply, so the distance is five times four, which is twenty. The tempting wrong answer adds the moduli instead of multiplying them.

Commit to an option before revealing, and try to answer it without writing anything down. Both numbers have the same argument up to sign, so the product rule tells you the answer's direction immediately.

The arguments are a third of pi and minus a third of pi, and they add to zero. A complex number with argument zero lies on the positive real axis, so the product is a positive real number even though neither factor was real.

The distance is the only thing left to get right, and the trap is to add the moduli. They multiply: five times four is twenty. If you chose the option with nine in it, you added when you should have multiplied, which is exactly the confusion the previous trap slide was about.

14. Which form for which job?

Sorting

Sort into buckets

Sort each task by the form that makes it easy.

Rectangular form
adding two complex numbers; subtracting two complex numbers; reading off the real part
Polar form
raising a number to the twentieth power; finding all six sixth roots; dividing one number by another
rect
Addition, subtraction and reading off a coordinate all act on the real and imaginary parts one at a time, and that is exactly what rectangular form displays.
pol
Multiplication, division, powers and roots all act on the modulus and the argument separately, and that is exactly what polar form displays.

Sort all six before revealing. The test for each one is simple: does the operation act on the real and imaginary parts separately, or on the size and the direction separately?

Addition, subtraction and reading off a coordinate all work on the parts, so they belong to rectangular form. Multiplication, division, powers and roots all work on size and direction, so they belong to polar form.

This is the single most useful piece of strategy in the section, and it is worth saying out loud before you start any problem. Choosing the wrong form does not make a problem impossible, but it usually multiplies the work by five and the chance of a slip by more than that.

15. DeMoivre's Theorem

Section

Part 2

16. One factor at a time gives the power rule

Concept

A power is a repeated product, so apply the product rule to a number times itself, then again, and watch what accumulates.

\[ z^2 = |z||z|\,\operatorname{cis}(\theta + \theta) = |z|^2\operatorname{cis}(2\theta) \]

\[ z^3 = z^2 z = |z|^2|z|\,\operatorname{cis}(2\theta + \theta) = |z|^3\operatorname{cis}(3\theta) \]

DeMoivre's Theorem — For a complex number in polar form and any natural number n, the nth power has modulus the nth power of the modulus and argument n times the argument.

\[ z^n = |z|^n\operatorname{cis}(n\theta) \]

The pattern is clear after two steps, but a pattern is not a proof; the next slide runs the induction that Stitz and Zeager give.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 997 — Theorem 11.16, the power rule

Watch how the exponents build. Squaring multiplies the modulus by itself and adds the argument to itself; cubing does it once more. After two steps the pattern is unmistakable: the modulus is raised to the power and the argument is multiplied by it.

Notice that these are two completely different operations on two completely different quantities, which is exactly why this is the one thing people get wrong. The length gets an exponent; the angle gets a multiplication.

A pattern seen twice is not a proof, though, and the book does not settle for one. The next slide is the induction, and it is short enough to reproduce on an exam if you are ever asked to justify the theorem rather than use it.

17. Proving DeMoivre's Theorem by induction

Worked example

Let the sentence to be proved for every natural number n be the power rule itself.

\[ P(n): \quad z^n = |z|^n\operatorname{cis}(n\theta) \]

Check the base case

Why: The first power is the number itself, and its polar form is given.

\[ z^1 = z = |z|\,\operatorname{cis}(\theta) = |z|^1\operatorname{cis}(1\cdot\theta) \;\checkmark \]

Assume the statement for some k

Why: This is the induction hypothesis, assumed for one unspecified k at least 1.

\[ z^{k} = |z|^{k}\operatorname{cis}(k\theta) \]

Split off one factor from the next power

Why: Properties of exponents, nothing more.

\[ z^{k+1} = z^{k}z \]

Substitute the hypothesis and the polar form of z

Why: Both factors are now in polar form.

\[ = \left[|z|^{k}\operatorname{cis}(k\theta)\right]\left[|z|\,\operatorname{cis}(\theta)\right] \]

Apply the product rule

Why: Moduli multiply, arguments add.

\[ = |z|^{k}|z|\,\operatorname{cis}(k\theta + \theta) = |z|^{k+1}\operatorname{cis}((k+1)\theta) \]

Check that the conclusion is the statement for k plus one

Why: It is P of k plus 1 exactly, so by the Principle of Mathematical Induction the rule holds for every natural number; and at n equal to 2 it agrees with the direct computation on the previous slide.

\[ z^2 = |z|^2\operatorname{cis}(2\theta) \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, pp. 997-998 — proof of the Power Rule

If you have seen induction before, this is a model example: a base case in one line and an inductive step that is a single application of a rule you have already proved. If you have not, the shape is worth learning here, where the content is easy.

The base case just says that the first power of a number is itself, which is true by the definition of the polar form. The inductive step splits the next power into the previous power times one more factor, substitutes the assumption for the previous power, and then applies the product rule to the two polar forms.

Note what the induction is really doing. It is saying that the product rule can be applied as many times as you like without anything going wrong, so you never have to write out a chain of n multiplications. Compare it with the proof of the power rule for the modulus in the previous lesson; the two proofs have the same skeleton, which is not a coincidence.

18. Reading DeMoivre's Theorem piece by piece

Notation

Annotate

On: \( z^{n} = |z|^{n}\operatorname{cis}(n\theta) \)

  • n is a natural number. The theorem as stated says nothing about a fractional exponent, and the last part of this lesson shows why: a fractional power has several values, not one.
  • An ordinary real power of a positive real number. This is the part a calculator does without complaining.
  • The angle is MULTIPLIED by n, never raised to the power n. Almost every wrong answer in this part of the section comes from mixing these two up.
  • nθ often exceeds a full turn. Subtract multiples of 2π until the angle is coterminal and in the range you want, then evaluate the cosine and the sine.

Read the second and third annotations side by side, because together they are the mistake that costs the most marks in this section. The modulus is raised to the power. The argument is multiplied by it. Saying the sentence out loud a few times is a genuinely effective defence.

The first annotation quietly plants something you will need at the end of the lesson. The theorem is stated for natural numbers only. It does not say what a one-third power means, and that gap is exactly where the root formula will go, because a one-third power turns out to have three values rather than one.

The last annotation is the routine bookkeeping. Multiplying the argument by a large n usually pushes it past a full turn, sometimes several, and you should reduce it before evaluating any cosine or sine. Reduce first, evaluate second; doing it the other way round is where sign errors come from.

19. Example 11.7.3, part 2: a fifth power without the Binomial Theorem

Worked example

This is the computation the warmup abandoned.

\[ w = -1 + i\sqrt3, \qquad \text{find } w^5 \]

Convert to polar form

Why: Modulus two, and Quadrant II puts the argument at two thirds of pi.

\[ w = 2\,\operatorname{cis}\!\left(\tfrac{2\pi}{3}\right) \]

Apply DeMoivre

Why: Raise the modulus to the fifth, multiply the argument by five.

\[ w^5 = 2^5\operatorname{cis}\!\left(5\cdot\tfrac{2\pi}{3}\right) = 32\,\operatorname{cis}\!\left(\tfrac{10\pi}{3}\right) \]

Reduce the argument by a full turn

Why: Ten thirds of pi is more than two pi, so subtract six thirds of pi.

\[ \tfrac{10\pi}{3} - 2\pi = \tfrac{4\pi}{3} \;\Longrightarrow\; w^5 = 32\,\operatorname{cis}\!\left(\tfrac{4\pi}{3}\right) \]

Convert back to rectangular form

Why: Cosine of four thirds of pi is minus one half; the sine is minus root three over two.

\[ w^5 = 32\!\left(-\tfrac12 - \tfrac{\sqrt3}{2}i\right) = -16 - 16i\sqrt3 \]

Check the modulus of the answer

Why: It must be two to the fifth, namely thirty-two, and a modulus check catches any slip in the final evaluation.

\[ \sqrt{(-16)^2 + (-16\sqrt3)^2} = \sqrt{256 + 768} = \sqrt{1024} = 32 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 999 — Example 11.7.3, part 2

This is the computation you abandoned in the warmup, and it now takes four lines. Convert, apply the theorem, reduce the angle, convert back. That is the entire method for every power problem in this section.

The reduction step deserves attention. Five times two thirds of pi is ten thirds of pi, which is more than a full turn, so subtract two pi, which is six thirds of pi, and you are left with four thirds of pi. That is an angle in Quadrant III whose cosine and sine you know exactly.

The check is the one you should run on every power problem. The modulus of the answer has to be the modulus of the original raised to the same power, here two to the fifth, which is thirty-two. Computing the modulus of your rectangular answer takes ten seconds and catches almost every arithmetic slip, including sign errors in the final evaluation.

20. What repeated multiplication looks like

Picture it

Figure (svg): Three complex planes side by side. In the first the powers of a number of modulus 1.18 spiral outward; in the second, modulus one, they stay on a circle; in the third, modulus 0.85, they spiral inward toward the origin.

Every power turns by the same angle, so the dots are equally spaced around; the modulus alone decides whether they run out to infinity, stay put, or collapse to zero.

Each panel plots the first twelve powers of a number whose argument is a twelfth of a full turn. The angle steps round by the same amount every time in all three panels; only the modulus changes what happens to the distance.

Look at the middle panel first. When the modulus is one, every power stays on the unit circle and the dots simply march round it, because multiplying by a number of modulus one is a pure rotation. This is the case that produces the roots of unity later in the lesson.

Now compare the outer panels. The angles do exactly the same thing in all three: each power is a further thirty degrees round. Only the distance behaves differently, and it is governed entirely by whether the modulus is above or below one.

That is the whole geometry of DeMoivre's Theorem in one picture. Repeated multiplication is repeated stretching and repeated turning, and since both happen by the same amount each time, the result is a spiral, or a circle when the stretching factor happens to be one.

21. Find the error: the modulus was multiplied by n

Error analysis

Annotate

On: \( \left[2\,\operatorname{cis}\!\left(\tfrac{2\pi}{3}\right)\right]^{5} = 10\,\operatorname{cis}\!\left(\tfrac{32\pi}{3}\right) \)

  • The modulus was multiplied by 5 instead of raised to the power 5. It should be 2 to the fifth, which is 32, not 2 times 5.
  • And the argument was raised to the power instead of multiplied: two thirds of pi to the fifth power is not what the theorem says. It should be 5 times two thirds of pi, which is ten thirds of pi.
  • This is one single mistake, made twice: the operations on the modulus and on the argument have been exchanged. Say the theorem aloud as 'power the length, multiply the angle'.
  • The modulus of the answer must be 32, and a quick look at the size of the first few powers (2, 4, 8, 16, 32) would have shown that 10 is far too small.

Read the line and try to name both errors before you reveal the annotations. There is a hint in the fact that the answer looks too small: the fifth power of something two units from the origin ought to be a long way out.

The two errors are really one error made twice. The operation that belongs to the modulus has been applied to the argument and the operation that belongs to the argument has been applied to the modulus. They have been swapped.

The defence is the modulus check from the previous slide. If you simply ask what the modulus of the answer should be, you get thirty-two, and ten is nowhere near it. Getting into the habit of predicting the modulus before you compute anything makes this class of error nearly impossible to commit unnoticed.

22. Exercise 64: the eighth power of 1 minus i

Worked example

From the section's exercise set. Eight is a high enough power that no one would expand it.

\[ (1 - i)^8 \]

Find the modulus and an argument

Why: The point is in Quadrant IV, one unit right and one unit down.

\[ |1-i| = \sqrt{1+1} = \sqrt2, \qquad \theta = -\frac{\pi}{4} \]

Write the polar form

Why: Root two at minus a quarter turn of a right angle.

\[ 1 - i = \sqrt2\,\operatorname{cis}\!\left(-\tfrac{\pi}{4}\right) \]

Apply DeMoivre

Why: Root two to the eighth, and eight times the argument.

\[ (1-i)^8 = (\sqrt2)^8\operatorname{cis}\!\left(8\cdot\!\left(-\tfrac{\pi}{4}\right)\right) = 16\,\operatorname{cis}(-2\pi) \]

Reduce the argument

Why: Minus two pi is one full turn backwards, which is the same direction as zero.

\[ 16\,\operatorname{cis}(-2\pi) = 16\,\operatorname{cis}(0) = 16 \]

Check by squaring in stages

Why: Square once in rectangular form and then take the fourth power of the result, which is real and easy.

\[ (1-i)^2 = -2i, \quad (-2i)^4 = 16i^4 = 16 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1005 — Exercise 64

Try this one on your own before reading the steps. It is the cleanest possible illustration of the method, because everything comes out exactly.

The conversion is easy: one right and one down, so the modulus is root two and the argument is minus a quarter of pi. DeMoivre then gives root two to the eighth power, which is two to the fourth, or sixteen, and eight times minus a quarter of pi, which is minus two pi, exactly one full turn backwards.

An angle of minus a full turn points along the positive real axis, so the answer is the real number sixteen. The check runs the calculation a different way entirely, squaring once in rectangular form to get negative two i and then taking the fourth power of that. When two independent routes agree, you are done.

23. Slide the modulus and the angle

Tweak it

Parameter explorer

The curve is the real part of z raised to the power x, for a number z with the modulus and angle you choose. Set the modulus to 1 and watch the curve become a pure wave; push it above 1 and the wave grows. What does the angle do to the curve?

\[ \operatorname{Re}(z^{x}) = ({r})^{x}\cos({t}x) \]

  • r — from 0.7 to 1.3: modulus of z
  • t — from 0.2 to 1.6: argument of z (radians)

This slider plots the real part of a power as the exponent varies continuously, which lets you see the whole family at once instead of one power at a time. The integer values of the horizontal variable are the actual powers.

Start by setting the modulus to one. The curve becomes a pure cosine wave of constant height, which is the middle panel of the spiral picture: pure rotation, no growth. Now nudge the modulus above one and the wave grows; drop it below one and the wave dies away.

Then leave the modulus alone and move the angle. A larger angle means each power turns further, so fewer powers fit into one full turn, and the wave oscillates faster. A small angle means many powers before the number comes back round. Both effects are visible in the spiral picture too, as the tightness of the spiral.

24. Multiplication as a motion

Section

Part 3

25. Stretch, then turn

Concept

Fix w and ask what multiplication by w does to every point z of the plane. The product rule answers it in two words.

\[ z \;\longmapsto\; zw = \underbrace{|w|}_{\text{stretch}}\cdot|z|\;\operatorname{cis}(\alpha + \underbrace{\beta}_{\text{turn}}) \]

Every length is multiplied by the modulus of w, and every direction is turned by the argument of w. The plane is magnified and rotated, both about the origin, and nothing else happens to it.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 999 — the geometry hidden in Theorem 11.16

Here the point of view changes. Instead of asking what the product of two particular numbers is, fix one of them and ask what multiplying by it does to the whole plane. The answer is remarkably rigid: the plane is magnified about the origin and rotated about the origin, and that is all.

That rigidity is worth appreciating. Multiplication by a complex number cannot bend, shear or flip the plane, and it cannot move the origin, because zero times anything is zero. Circles stay circles, straight lines stay straight lines, and angles between lines are preserved.

The four bullet cases are the ones you should be able to recognise instantly. A real multiplier only stretches. A multiplier of modulus one only turns. Anything else does both, and you can read off which by looking at the modulus and the argument separately.

26. Building the product in two moves

Picture it

Figure (svg): Two complex planes. On the left, z at radius 4 and an arrow along the same ray out to radius 8. On the right, an arc of one hundred and twenty degrees carrying the radius-8 point round to the product at one hundred and fifty degrees.

Multiplying by w never bends the picture: it scales every length by the modulus of w and turns the whole plane by the argument of w.

The book's own pair of diagrams for the numbers of Example 11.7.3. First the arrow to z is stretched to twice its length, because the modulus of w is two; then the stretched arrow is turned two thirds of pi counter-clockwise, because that is the argument of w.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1000 — visualising the product

These are the book's own diagrams for the numbers you worked with a few slides ago, and they are worth walking through slowly, with a finger on the arrow.

In the left panel nothing turns. The arrow to z is stretched along its own ray until it is twice as long, because the modulus of w is two. In the right panel nothing stretches. The stretched arrow swings counter-clockwise by two thirds of a full half turn, because that is the argument of w.

The order does not matter, incidentally. Stretching first and turning second gives the same point as turning first and stretching second, because the two operations act on different channels. That is another way of saying that the modulus and the argument really are independent.

27. Multiplying by i is a quarter turn

Concept

Figure (svg): One point at 3 plus 2i and the three points obtained by multiplying by i again and again, all on a circle of radius root 13, each a quarter turn from the last, with arcs marking the turns.

Multiplying by i keeps the modulus and adds a quarter turn, so four multiplications bring every point exactly back to where it started.

\[ i = 1\,\operatorname{cis}\!\left(\tfrac{\pi}{2}\right) \]

The modulus of i is one, so multiplying by i changes no distances at all; its argument is a right angle, so it turns everything a quarter turn counter-clockwise. Four quarter turns is a full turn, which is the geometric reason the fourth power of i is one.

\[ i(3+2i) = 3i + 2i^2 = -2 + 3i \]

\[ i^4 = \operatorname{cis}\!\left(4\cdot\tfrac{\pi}{2}\right) = \operatorname{cis}(2\pi) = 1 \]

This is the single most useful special case in the topic, and it explains something you have taken on faith since you first met complex numbers. The imaginary unit has modulus one and argument a right angle, so multiplying by it never changes a distance and always adds a quarter turn.

Trace the figure round. The starting point is three right and two up. Multiply by i and you land two left and three up, which is the same distance from the origin, a quarter turn round. Two more multiplications bring you to the opposite point and then to its reflection, and the fourth brings you home.

Now the identity at the bottom is obvious rather than memorised. Four quarter turns is a full turn, and a full turn leaves every point exactly where it was, so the fourth power of i has to be one. If you can see that, you can also see instantly that the seventeenth power of i is i again.

28. Dividing runs the motion backwards

Picture it

Figure (svg): Two complex planes. On the left, z at radius 4 and a short arrow back along the ray to radius 2. On the right, an arc running clockwise from thirty degrees down to minus ninety degrees, landing on the point 2 units below the origin.

Division undoes both moves of multiplication: the modulus is divided rather than multiplied, and the turn goes clockwise rather than counter-clockwise.

Dividing by w shrinks each length by the modulus of w and turns the plane clockwise by its argument. For the numbers of Example 11.7.3 that means halving the arrow and then turning it back two thirds of pi, landing two units below the origin.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1000 — visualising the quotient

Division is the inverse motion, so it does the opposite on both channels: it shrinks instead of stretching, and it turns clockwise instead of counter-clockwise.

Follow the same two panels as before, but with the arrows reversed. The arrow to z is pulled back to half its length, because you are dividing by a modulus of two, and then swung clockwise by the argument of w, landing on the negative imaginary axis two units below the origin.

There is a useful sanity check hiding here. If you divide and then multiply by the same number you should be back where you started, so the two pictures should undo each other exactly. Whenever you are unsure whether a turn should go clockwise or counter-clockwise, ask which direction would undo the multiplication.

29. Match each multiplier to its motion

Matching

Match the pairs

  • a. multiply by 3
  • b. multiply by i
  • c. multiply by −1
  • d. multiply by (1/2) cis(−π/3)
  • e. multiply by cis(π/4)
  • v. triple every distance, no turn
  • w. a quarter turn counter-clockwise, same size
  • x. a half turn, same size
  • y. halve every distance and turn sixty degrees clockwise
  • z. an eighth of a turn counter-clockwise, same size

Why: Read the modulus for the stretch and the argument for the turn, every time. A positive real multiplier has argument zero, so it never turns; a multiplier of modulus one never resizes; and negative one has modulus one with argument pi, which is a half turn.

Pair all five before revealing. For each multiplier, read its modulus to get the stretch and its argument to get the turn, and do not let the rectangular appearance of a number distract you.

Three of them are pure motions. Multiplying by three is a pure stretch, because the argument of a positive real number is zero. Multiplying by i or by cis of anything is a pure turn, because the modulus is one. Multiplying by negative one is a half turn, which is why negating a number reflects it through the origin.

The remaining one does both, and its argument is negative, so the turn is clockwise. If you can read a multiplier and say its motion out loud in under five seconds, you have understood the geometry of this section.

30. Exercises 41 and 42: a product and a quotient in principal form

Worked example

The exercise set asks for the answers in polar form using the principal argument, which means the angle must be brought inside the half-open interval from minus pi to pi.

\[ z = -\frac{3\sqrt3}{2} + \frac32 i, \qquad w = 3\sqrt2 - 3i\sqrt2 \]

Convert z

Why: Its modulus is three, and it sits in Quadrant II with a tangent of minus one over root three.

\[ |z| = \sqrt{\tfrac{27}{4} + \tfrac94} = 3, \quad z = 3\,\operatorname{cis}\!\left(\tfrac{5\pi}{6}\right) \]

Convert w

Why: Its modulus is six, and it sits in Quadrant IV with a tangent of minus one.

\[ |w| = \sqrt{18 + 18} = 6, \quad w = 6\,\operatorname{cis}\!\left(-\tfrac{\pi}{4}\right) \]

Multiply

Why: Three times six, and the two arguments added over a common denominator of twelve.

\[ zw = 18\,\operatorname{cis}\!\left(\tfrac{10\pi}{12} - \tfrac{3\pi}{12}\right) = 18\,\operatorname{cis}\!\left(\tfrac{7\pi}{12}\right) \]

Divide

Why: Three over six, and the arguments subtracted.

\[ \frac{z}{w} = \frac12\operatorname{cis}\!\left(\tfrac{10\pi}{12} + \tfrac{3\pi}{12}\right) = \frac12\operatorname{cis}\!\left(\tfrac{13\pi}{12}\right) \]

Bring the quotient's argument into principal range

Why: Thirteen twelfths of pi is past pi, so subtract a full turn.

\[ \tfrac{13\pi}{12} - 2\pi = -\tfrac{11\pi}{12} \;\Longrightarrow\; \frac{z}{w} = \frac12\operatorname{cis}\!\left(-\tfrac{11\pi}{12}\right) \]

Check numerically

Why: Evaluating both answers in rectangular form to four places gives the same numbers as multiplying and dividing the original expressions directly.

\[ zw \approx -4.6587 + 17.3867i, \qquad \frac{z}{w} \approx -0.4830 - 0.1294i \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1004 — Exercises 41 and 42

The exercise set adds one requirement that the worked examples did not: the answers have to be in principal form, with the argument between minus pi and pi. That is a small extra step at the end, and it is easy to forget.

Both conversions need the quadrant. The first number is in Quadrant II and the second is in Quadrant IV, and in both cases the inverse tangent alone would give you the wrong angle. Draw the point before you trust the calculator.

The product's argument, seven twelfths of pi, is already inside the principal range, so nothing more is needed there. The quotient's argument comes out as thirteen twelfths of pi, which is just past pi, so subtract a full turn to get minus eleven twelfths. Notice that this does not change the number at all, only the name you give its direction.

31. Trap: reading multiplication as a shift

Trap

The trap

Vectors are added by sliding one along the other, so multiplication gets the same treatment:

\[ z\cdot w \overset{?}{=} \text{slide } z \text{ by } w \]

Wrong. Test it on z equal to 0.

\[ 0\cdot w = 0 \]

A slide would have moved the origin somewhere else.

The fix

Sliding the plane is what addition does. Multiplication fixes the origin and changes every length and every direction in proportion, which is a magnification together with a rotation.

\[ z + w: \text{ slide} \]

\[ zw: \text{ stretch and turn} \]

This confusion comes from vectors. You have spent time adding arrows by sliding one to the tip of the other, and it is natural to assume that multiplying does something similar.

The one-line refutation is on the left. Multiplying anything by zero gives zero, so the origin cannot move, and no slide except the zero slide leaves a point fixed. Multiplication pins the origin in place and moves everything else in proportion to how far out it is.

Keep the two operations in separate mental boxes. Adding slides the whole plane rigidly. Multiplying stretches and turns it about the fixed origin. Section 11.8 will pick up the sliding picture again under the name of vectors, and it is worth knowing which of the two is which before you get there.

32. The nth roots

Section

Part 4

33. An nth root, and how many of them there are

Concept

nth root (Definition 11.4) — For complex numbers z and w, and a natural number n, the number w is an nth root of z whenever the nth power of w equals z.

The article is deliberately indefinite. There is no single principal root here, and the reason is visible the moment you try to find all the cube roots of eight by factoring.

\[ w^3 = 8 \;\Longleftrightarrow\; w^3 - 8 = 0 \;\Longleftrightarrow\; (w - 2)(w^2 + 2w + 4) = 0 \]

\[ w = 2 \quad\text{or}\quad w = \frac{-2 \pm \sqrt{4 - 16}}{2} = -1 \pm i\sqrt3 \]

Three roots, of which one is real. The Complex Factorization Theorem from Chapter 3 says a cubic has exactly three complex zeros counting multiplicity, so this list is complete.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, pp. 1000-1001 — Definition 11.4 and the cube roots of 8 by factoring

Read the definition carefully and notice the indefinite article. It says an nth root, not the nth root, and that word choice is the whole point. With real numbers you learned to write a radical sign and get one answer; that convention does not survive the move to complex numbers.

The factoring argument shows why. The equation for the cube roots of eight rearranges into a cubic with a difference of cubes, and the quadratic factor contributes two more solutions that the radical sign never mentions. Both of them are genuine cube roots: cube either one and you get eight.

Chapter 3 told you to expect exactly this. A polynomial equation of degree n has exactly n complex solutions counting multiplicity, so a cube root equation must have three solutions. The surprise is not that there are three; it is that the radical sign has been hiding two of them from you for years.

34. The cube roots of 8, the polar way

Worked example

Now solve the same equation with the machinery of this section, because factoring will not survive a fifth root.

\[ w^3 = 8, \qquad w = |w|\,\operatorname{cis}(\alpha) \]

Write both sides in polar form

Why: Eight lies eight units along the positive real axis, so its argument is zero.

\[ |w|^3\operatorname{cis}(3\alpha) = 8\,\operatorname{cis}(0) \]

Match the moduli

Why: Both moduli are non-negative real numbers, so they must be equal; extract the principal cube root.

\[ |w|^3 = 8 \;\Longrightarrow\; |w| = \sqrt[3]{8} = 2 \]

Match the arguments up to full turns

Why: Two polar forms name the same number when their arguments differ by a whole number of turns.

\[ 3\alpha = 0 + 2\pi k \;\Longrightarrow\; \alpha = \frac{2\pi k}{3}, \quad k \in \mathbb{Z} \]

Run k through consecutive integers and stop when the angles repeat

Why: At k equal to 3 the angle is two pi, coterminal with k equal to 0, so only three of them are different.

\[ \alpha = 0, \; \frac{2\pi}{3}, \; \frac{4\pi}{3}, \; \underbrace{2\pi}_{\text{same as } k=0} \]

Write the three roots in both forms

Why: Modulus two at each of the three angles.

\[ w_0 = 2, \quad w_1 = 2\,\operatorname{cis}\!\left(\tfrac{2\pi}{3}\right) = -1 + i\sqrt3, \quad w_2 = 2\,\operatorname{cis}\!\left(\tfrac{4\pi}{3}\right) = -1 - i\sqrt3 \]

Figure (svg): A circle of radius 2 in the complex plane carrying three dots at zero, one hundred and twenty and two hundred and forty degrees, joined into an equilateral triangle, with the real cube root 2 marked on the positive real axis.

Only one of the three cube roots of 8 is real. The other two are the remaining corners of an equilateral triangle on the circle of radius 2.

Check against the factoring

Why: The three roots are exactly the zeros the quadratic formula produced, so the two methods agree; and cubing the middle one returns eight.

\[ (-1 + i\sqrt3)^3 = 2^3\operatorname{cis}(2\pi) = 8 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1001 — the cube roots of 8 in polar form

Factoring worked here, but it will not work for a fifth root, and it certainly will not work when the number on the right is not real. So do the easy case again with the method that scales.

The key move is the third step. Two polar forms name the same complex number exactly when their moduli agree and their arguments differ by a whole number of full turns. That converts one complex equation into two real ones, one for the modulus and one for the argument, and the argument equation carries an unknown integer.

That integer is where the extra roots come from. Let it run and the angles come out a third of a full turn apart, and after three of them you are back where you started. Compare the three answers with the ones the factoring produced: identical. Two methods, same three numbers, and only one of them will still be usable in a moment.

35. Theorem 11.17: one formula for every root

Concept

Nothing in that calculation used the number eight or the exponent three. Running it on a general non-zero z gives the theorem.

Theorem 11.17 — Let z be a non-zero complex number with polar form modulus r and argument theta. For each natural number n, the number z has exactly n distinct nth roots, obtained by taking the principal nth root of the modulus, dividing the argument by n, and then adding one nth of a full turn repeatedly.

\[ w_k = \sqrt[n]{r}\;\operatorname{cis}\!\left(\frac{\theta}{n} + \frac{2\pi}{n}k\right), \qquad k = 0, 1, \ldots, n-1 \]

Two claims are packed in there: that each of these really is an nth root, and that there are exactly n of them, no more and no fewer. Both are proved next.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1001 — Theorem 11.17

Nothing in the cube-root calculation depended on the eight or on the three. Replace them by a general modulus and a general exponent and the same three steps give the formula on this slide.

Read the formula as an instruction rather than a statement. Take the ordinary nth root of the distance. Divide the angle by n. Then step round by one nth of a full turn as many times as you can before repeating.

Two things are being claimed at once, and it is worth separating them. First, that every one of these numbers really is an nth root, which is a computation with DeMoivre. Second, that there are exactly n of them, no more and no fewer, which needs an argument about when two angles name the same direction. The next two slides prove them in that order.

36. Reading the root formula piece by piece

Notation

Annotate

On: \( w_k = \sqrt[n]{r}\;\operatorname{cis}\!\left(\frac{\theta}{n} + \frac{2\pi}{n}k\right) \)

  • The ordinary principal nth root of a positive real number, the one your calculator gives. Every root has this same modulus, so all n of them lie on one circle.
  • The argument of z divided by n. This alone gives the first root, and it is the only place the original angle appears.
  • One nth of a full turn, added k times. This is what generates the other roots: each is the previous one rotated by the same small angle.
  • Exactly n values of k. Taking k = n adds a whole turn and returns you to the first root, so the list closes up rather than continuing.

Work through the four annotations with a picture in mind. The first two build one specific number: a circle of a certain radius, and a starting direction on it. Everything else in the formula is the same for every root.

The third annotation is the generator. Adding one nth of a full turn k times is what produces the other roots, and because it is the same angle every time, the roots come out equally spaced. That is the sentence that turns the formula into a polygon.

The fourth annotation is the one that answers the question everyone asks: why stop at n minus one? Because taking k equal to n adds a complete turn to the first root's argument, which brings you back to the first root itself. The list is a loop, and you stop just before it repeats.

37. Proving that each formula really is an nth root

Worked example

Raise the kth candidate to the nth power and see whether z comes back.

\[ w_k = \sqrt[n]{r}\;\operatorname{cis}\!\left(\frac{\theta}{n} + \frac{2\pi}{n}k\right) \]

Apply DeMoivre to the candidate

Why: Modulus to the nth power, argument multiplied by n.

\[ (w_k)^n = \left(\sqrt[n]{r}\right)^{n}\operatorname{cis}\!\left(n\!\left[\frac{\theta}{n} + \frac{2\pi}{n}k\right]\right) \]

Simplify the modulus

Why: The nth power undoes the principal nth root of a positive real number.

\[ \left(\sqrt[n]{r}\right)^{n} = r \]

Simplify the argument

Why: The n cancels both denominators.

\[ n\!\left[\frac{\theta}{n} + \frac{2\pi}{n}k\right] = \theta + 2\pi k \]

Discard the whole turns

Why: Cosine and sine are unchanged by adding a whole number of full turns, so cis is too.

\[ \operatorname{cis}(\theta + 2\pi k) = \operatorname{cis}(\theta) \]

Read off the result

Why: The nth power of the kth candidate is the polar form of z itself.

\[ (w_k)^n = r\,\operatorname{cis}(\theta) = z \]

Check on the case already computed

Why: With r equal to 8, theta equal to 0 and n equal to 3, the formula returns the three cube roots found by factoring.

\[ w_k = 2\,\operatorname{cis}\!\left(\tfrac{2\pi}{3}k\right): \quad 2, \; -1+i\sqrt3, \; -1-i\sqrt3 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, pp. 1001-1002 — first half of the proof of Theorem 11.17

This proof has exactly one idea in it: apply DeMoivre to the candidate and watch everything cancel. Read it once for the shape and once for the details.

The modulus part is immediate, since raising an nth root to the nth power undoes it. The argument part is nicer: multiplying by n cancels both denominators at once, leaving the original angle plus a whole number of full turns.

And a whole number of full turns is invisible to cosine and sine, so cis does not notice it. That is the step where the extra roots earn their keep, because it is precisely the freedom to add whole turns before dividing that produced them in the first place. Check the result against the cube roots of eight and you will find the formula reproduces the list you already have.

38. Why the list stops at k equal to n minus 1

Concept

Every root in the list has the same modulus, so two of them can only be the same number if their arguments differ by a whole number of full turns. Subtract one argument from another and see what that would need.

\[ \left(\frac{\theta}{n} + \frac{2\pi}{n}k\right) - \left(\frac{\theta}{n} + \frac{2\pi}{n}j\right) = 2\pi\!\left(\frac{k-j}{n}\right) \]

\[ \text{coterminal} \iff \frac{k-j}{n} \in \mathbb{Z} \iff n \mid (k-j) \]

\[ 0 < k - j \le n-1 \;\Longrightarrow\; n \nmid (k-j) \]

So no two entries of the list coincide: there are at least n distinct roots. And the equation has degree n, so by the Complex Factorization Theorem it has at most n solutions. Between them the two facts pin the count at exactly n.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1002 — second half of the proof of Theorem 11.17

This is the fussy half of the proof, and it is worth reading slowly once, because it is the only part of the section that explains why the count is exactly n rather than some other number.

Begin with the observation that all the candidates have the same modulus. That means two of them can only be the same number if they point in the same direction, and two directions are the same exactly when their angles differ by a whole number of full turns.

Subtract two of the arguments and everything cancels except a full turn times the difference of the two indices over n. For that to be a whole number of turns, n would have to divide the difference of the indices. But the indices are both between zero and n minus one, so their difference is at least one and at most n minus one, and nothing in that range is a multiple of n. So all n are different, and since a degree n equation has at most n solutions, you have found all of them.

39. Example 11.7.4(1): both square roots of negative 2 plus 2i root 3

Worked example

\[ z = -2 + 2i\sqrt3 \]

Convert z to polar form

Why: Modulus four, and Quadrant II gives an argument of two thirds of pi.

\[ z = 4\,\operatorname{cis}\!\left(\tfrac{2\pi}{3}\right), \qquad r = 4, \; \theta = \tfrac{2\pi}{3}, \; n = 2 \]

Take the principal square root of the modulus

Why: This is the modulus every root will have.

\[ \sqrt{4} = 2 \]

Divide the argument by n for the first root

Why: Two thirds of pi, halved.

\[ w_0 = 2\,\operatorname{cis}\!\left(\frac{2\pi/3}{2}\right) = 2\,\operatorname{cis}\!\left(\tfrac{\pi}{3}\right) = 1 + i\sqrt3 \]

Add one half of a full turn for the second

Why: With n equal to 2 the step between roots is pi, a half turn.

\[ w_1 = 2\,\operatorname{cis}\!\left(\tfrac{\pi}{3} + \pi\right) = 2\,\operatorname{cis}\!\left(\tfrac{4\pi}{3}\right) = -1 - i\sqrt3 \]

Notice the relationship between the two

Why: A half turn is multiplication by negative one, so the two square roots are negatives of each other, exactly as for real numbers.

\[ w_1 = -w_0 \]

Check by squaring

Why: Square the first root in rectangular form and recover z.

\[ (1 + i\sqrt3)^2 = 1 + 2i\sqrt3 + 3i^2 = -2 + 2i\sqrt3 = z \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1002 — Example 11.7.4, part 1

Start by noting how mechanical this is once the formula is in hand. Convert, take a root of the modulus, divide the angle, then step round.

With n equal to two the step between roots is half a full turn, which is multiplication by negative one. So the two square roots of a complex number are always negatives of each other, exactly as with real numbers where the two square roots of nine are three and minus three. The structure you already knew is a special case of this one.

Do run the check yourself rather than trusting the formula. Squaring one plus i root three by hand takes one line and returns the original number, and having verified one root you get the other free, since squaring a negative gives the same answer.

40. Example 11.7.4(2): the four fourth roots of negative 16

Worked example

\[ z = -16 \]

Convert to polar form

Why: Sixteen units along the negative real axis, so the argument is pi.

\[ z = 16\,\operatorname{cis}(\pi), \qquad r = 16, \; \theta = \pi, \; n = 4 \]

Find the common modulus

Why: The principal fourth root of sixteen.

\[ \sqrt[4]{16} = 2 \]

Find the first argument and the step

Why: The argument divided by four, and a quarter of a full turn between consecutive roots.

\[ \frac{\theta}{4} = \frac{\pi}{4}, \qquad \frac{2\pi}{4} = \frac{\pi}{2} \]

List the four roots in polar form

Why: Start at a quarter of pi and step by a half of pi three times.

\[ w_0 = 2\,\operatorname{cis}\!\left(\tfrac{\pi}{4}\right), \; w_1 = 2\,\operatorname{cis}\!\left(\tfrac{3\pi}{4}\right), \; w_2 = 2\,\operatorname{cis}\!\left(\tfrac{5\pi}{4}\right), \; w_3 = 2\,\operatorname{cis}\!\left(\tfrac{7\pi}{4}\right) \]

Convert them to rectangular form

Why: Each angle is an odd multiple of a quarter of pi, where cosine and sine are both root two over two.

\[ w_0 = \sqrt2 + i\sqrt2, \; w_1 = -\sqrt2 + i\sqrt2, \; w_2 = -\sqrt2 - i\sqrt2, \; w_3 = \sqrt2 - i\sqrt2 \]

Figure (svg): A circle of radius 2 with four dots at forty-five, one hundred and thirty-five, two hundred and twenty-five and three hundred and fifteen degrees, joined into a square, and the point negative 16 is noted as lying far out on the negative real axis.

The four fourth roots of a negative real number sit at the corners of a square, tilted so that none of them is real.

Check one of them by squaring twice

Why: Squaring the first root gives 4i, and squaring that gives negative 16, so the fourth power is z.

\[ (\sqrt2 + i\sqrt2)^2 = 4i, \quad (4i)^2 = -16 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, pp. 1002-1003 — Example 11.7.4, part 2

Notice the thing that should feel impossible at first glance. A negative real number has no real fourth root at all, and yet it has four complex ones. The reason is that the argument of a negative number is pi, and a quarter of pi is not an angle on the real axis.

The arithmetic is as easy as it gets: the modulus is the fourth root of sixteen, which is two, and the step is a quarter turn. Every argument comes out as an odd multiple of a quarter of pi, where the cosine and the sine are both root two over two up to sign, so all four convert to exact rectangular form.

Look at the picture once you have the answers. The four roots are the corners of a square, tilted forty-five degrees so that no corner touches either axis. That tilt is where the negative sign in the original number went.

41. Finish the list of fourth roots

Fill the middle

The four fourth roots of the number sixteen itself, rather than of its negative. The modulus is the same as in the example; only the starting angle changes.

Fill in the blanks

16 = 16\,\operatorname2(0) \;\Longrightarrow\; w_k = 4\,\operatorname2i\!\left(\frac______}k\right): \quad 2, \; ___, \; -2, \; -2i

Why: The principal fourth root of sixteen is two, the step between roots is a quarter of a full turn, and the first argument is zero because sixteen lies on the positive real axis. The four roots are therefore two, two i, negative two and negative two i.

This is the previous example with one change: the number is now positive sixteen instead of negative sixteen, so its argument is zero instead of pi.

The modulus of the roots does not change at all, because the modulus of the number did not change. What changes is the starting angle, which is now zero, so the first root is the real number two, and the square of roots is untilted: two, two i, negative two and negative two i.

Compare the two pictures in your head. Both are squares of the same size on the same circle; one is rotated by half a step relative to the other. Changing the argument of the number rotates the whole polygon and nothing else, which is a useful thing to know when you want to sanity-check an answer.

42. Example 11.7.4(3): three cube roots with awkward angles

Worked example

\[ z = \sqrt2 + i\sqrt2 \]

Convert to polar form

Why: Both parts are root two, so the modulus is two and the argument is a quarter of pi.

\[ z = 2\,\operatorname{cis}\!\left(\tfrac{\pi}{4}\right), \qquad r = 2, \; \theta = \tfrac{\pi}{4}, \; n = 3 \]

Find the common modulus and the step

Why: The principal cube root of two, and a third of a full turn.

\[ \sqrt[3]{2} \approx 1.2599, \qquad \frac{2\pi}{3} \]

Divide the argument by three for the first root

Why: A quarter of pi, divided by three.

\[ w_0 = \sqrt[3]{2}\,\operatorname{cis}\!\left(\tfrac{\pi}{12}\right) \]

Add a third of a turn twice

Why: One twelfth of pi plus eight twelfths, then plus eight twelfths again.

\[ w_1 = \sqrt[3]{2}\,\operatorname{cis}\!\left(\tfrac{9\pi}{12}\right) = \sqrt[3]{2}\,\operatorname{cis}\!\left(\tfrac{3\pi}{4}\right), \qquad w_2 = \sqrt[3]{2}\,\operatorname{cis}\!\left(\tfrac{17\pi}{12}\right) \]

Say why the book stops here

Why: One twelfth of pi is not one of the special angles, so turning the first and last roots into exact rectangular form needs the difference or half-angle identities of Chapter 10, which Exercise 77 asks for.

\[ \cos\!\left(\tfrac{\pi}{12}\right) = \frac{\sqrt6 + \sqrt2}{4}, \qquad \sin\!\left(\tfrac{\pi}{12}\right) = \frac{\sqrt6 - \sqrt2}{4} \]

Figure (svg): A circle of radius the cube root of 2 carrying three dots, the first just above the positive real axis at fifteen degrees and the others a third of a turn on, joined into a tilted equilateral triangle.

Dividing the argument by three tilts the whole triangle: the first root sits at one third of the original angle, and everything else follows from there.

Check the first root numerically

Why: Its rectangular form to four places, cubed, returns root two plus i root two to four places.

\[ w_0 \approx 1.2170 + 0.3261i, \quad w_0^3 \approx 1.4142 + 1.4142i = \sqrt2 + i\sqrt2 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1003 — Example 11.7.4, part 3, and Exercise 77

This example exists to make an honest point: the method always works, but the final conversion to rectangular form does not always come out nicely.

The polar answers are as clean as ever. Cube root of two for the modulus, a twelfth of pi for the first argument, and a third of a turn between consecutive roots. There is nothing hard here, and you should be able to produce the three polar forms in under a minute.

The rectangular forms are another matter. A twelfth of pi is not one of the special angles, so you would need a difference identity or a half-angle identity from Chapter 10 to get exact values, which is what Exercise 77 asks you to do. The book leaves it, and so should you unless you are asked. Notice that the middle root, at three quarters of pi, is an exception and converts easily.

43. Example 11.7.4(4): the five fifth roots of 1

Worked example

\[ z = 1 \]

Convert to polar form

Why: One unit along the positive real axis.

\[ 1 = 1\,\operatorname{cis}(0), \qquad r = 1, \; \theta = 0, \; n = 5 \]

Find the common modulus

Why: The fifth root of one is one, so every root lies on the unit circle.

\[ \sqrt[5]{1} = 1 \]

Find the first root and the step

Why: The argument zero divided by five is zero, and the step is a fifth of a full turn.

\[ w_0 = \operatorname{cis}(0) = 1, \qquad \frac{2\pi}{5} \]

List the other four

Why: Add two fifths of pi each time.

\[ w_1 = \operatorname{cis}\!\left(\tfrac{2\pi}{5}\right), \; w_2 = \operatorname{cis}\!\left(\tfrac{4\pi}{5}\right), \; w_3 = \operatorname{cis}\!\left(\tfrac{6\pi}{5}\right), \; w_4 = \operatorname{cis}\!\left(\tfrac{8\pi}{5}\right) \]

Approximate them, as Exercise 78 asks

Why: No identity of this course evaluates the cosine of two fifths of pi exactly, so a calculator is the honest tool here.

\[ w_1 \approx 0.3090 + 0.9511i, \qquad w_2 \approx -0.8090 + 0.5878i \]

Figure (svg): The unit circle with five dots at zero, seventy-two, one hundred and forty-four, two hundred and sixteen and two hundred and eighty-eight degrees, joined into a regular pentagon with one vertex at the number 1.

The fifth roots of 1 are the corners of a regular pentagon inscribed in the unit circle, with one corner nailed down at the number 1 itself.

Check the fifth power of one of them

Why: Raising the approximation to the fifth power returns one to four decimal places, and the modulus of every root is exactly one.

\[ (0.3090 + 0.9511i)^5 \approx 1.0000 + 0.0000i \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1003 — Example 11.7.4, part 4, and Exercise 78

The roots of the number one have a special name, the roots of unity, and they turn up everywhere from signal processing to number theory. This is their first appearance.

Because the modulus of one is one, every root has modulus one, so they all sit on the unit circle. And because the argument of one is zero, the first root is the number one itself. Everything else is just five equal steps round the circle.

Here again exact rectangular values are out of reach with the identities you have, because two fifths of pi is seventy-two degrees and nothing in this course evaluates its cosine exactly. A calculator is the right tool, and Exercise 78 asks for exactly that. What you can say exactly, and what matters more, is that these five points are the corners of a regular pentagon.

44. Trap: stopping at one root

Trap

The trap

The habit of a lifetime with real numbers:

\[ w^4 = 16 \;\Longrightarrow\; w = 2\,? \]

Incomplete. Three more numbers satisfy it.

\[ (-2)^4 = 16 \]

\[ (2i)^4 = 16 \]

\[ (-2i)^4 = 16 \]

The fix

An equation of degree n has n complex solutions, and the root formula produces all of them at once. The radical sign picks one preferred root of a positive real number; the phrase an nth root does not.

\[ \sqrt[4]{16} = 2: \text{ one number} \]

\[ w^4 = 16: \text{ four numbers} \]

Years of writing a radical sign and getting one number make this the most persistent error in the topic. The equation on the left has four solutions; two is only the one that happens to be positive and real.

The distinction to keep is between a notation and a question. The radical sign is a notation, and it is defined to pick out one preferred value for a positive real number. An equation is a question, and its answer is every number that satisfies it.

So read the wording of a problem carefully. Find the fourth root of sixteen and find all fourth roots of sixteen are different requests, and in this section it is nearly always the second one. If the problem says all, produce the whole polygon.

45. How many roots, and how many are real?

Sorting

Sort into buckets

Each equation has as many complex solutions as its degree. Sort them by how many of those solutions are real numbers.

No real solutions
w⁴ = −16; w² = −1
Exactly one real solution
w³ = 8; w⁵ = 1
Exactly two real solutions
w⁶ = 64; w⁴ = 16
none
A negative number has no real square or fourth root: every root of a negative number under an even exponent lands off the real axis, at an odd multiple of a small angle.
one
An odd exponent always gives exactly one real root, because a polygon with an odd number of corners and one corner on the positive real axis has no other corner on that line.
two
A positive number under an even exponent has a real root at each end of a diameter, one positive and one negative, and the rest come in conjugate pairs.

Sort all six before revealing, and do it by picturing the polygon rather than by computing anything. The question is just how many corners of the polygon land on the horizontal axis.

An odd number of corners with one on the positive real axis can never have a second one on the axis, because the corner opposite a vertex of an odd polygon is the middle of an edge. So odd exponents applied to positive numbers give exactly one real root.

For even exponents everything depends on the sign of the number. If it is positive, the polygon has a corner at each end of a diameter along the real axis, so two real roots. If it is negative, the whole polygon is rotated by half a step and no corner touches the axis at all, so none. Being able to say that without calculating is the goal of this slide.

46. The geometry of the roots

Section

Part 5

47. One root, then a sequence of equal turns

Concept

Figure (svg): A complex plane with z at radius 4 and angle one hundred and thirty-five degrees, a dashed circle of radius 4, and inside it a smaller circle of radius the cube root of 4 carrying three roots, the first at forty-five degrees and the other two a third of a turn apart.

Taking a cube root pulls the modulus in toward one and divides the angle by three; after that the other roots are free, each a third of a turn from the last.

Read the root formula as a recipe rather than a formula. The first root is found by shrinking the modulus and dividing the angle; after that no more arithmetic on z is needed, because each further root is the previous one turned by the same fixed angle.

\[ w_0 = \sqrt[n]{r}\,\operatorname{cis}\!\left(\frac{\theta}{n}\right), \qquad w_{k+1} = w_k\cdot\operatorname{cis}\!\left(\frac{2\pi}{n}\right) \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1003 — the geometric reading of Theorem 11.17

Read the formula as a two-phase recipe and the work drops sharply. Phase one uses z: shrink the modulus, divide the angle. Phase two never looks at z again; it just turns.

The picture shows the scale change clearly. The dashed outer circle is where z lives, and the inner circle is where all three of its cube roots live, much closer to the origin. Taking a root of a number bigger than one always pulls it in toward the unit circle, and taking a root of a number smaller than one pushes it out toward the unit circle.

The second line of the formula is the one to remember for speed. Each root is the previous one multiplied by cis of one nth of a full turn, so once you have the first root, you get the rest by rotating, with no further division and no further root extraction.

48. Always a regular polygon

Picture it

Figure (svg): Four unit circles side by side carrying 2, 3, 6 and 12 equally spaced roots, joined up into a diameter, a triangle, a hexagon and a twelve-sided polygon.

Whatever n is, the roots are the corners of a regular n-sided polygon on one circle. Larger n means more corners and a smaller step between them.

Every root has the same modulus, so they all sit on one circle; consecutive arguments differ by the same angle, so they are equally spaced round it. Equally spaced points on a circle are the corners of a regular polygon, for every n at once.

This is the best check you have. Plot your answers: if they are not on one circle, a modulus is wrong, and if they are not evenly spaced, an argument is wrong.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1003 — the four fourth roots equally spaced around the plane

This is the picture the lesson exists for. Look across all four panels and notice that only one thing changes: the number of corners. The circle is the same, the first corner is in the same place, and the spacing adjusts to fit.

The two halves of the formula produce the two halves of the picture. A common modulus puts every root on one circle. A common angular step spaces them evenly round it. Equally spaced points on a circle is the definition of a regular polygon, so the polygon is not a coincidence or a pretty pattern; it is a restatement of the formula.

Use it as your check on every root problem, every time. Plot the answers roughly by hand. If one of them is closer to the origin than the others, you have made a mistake in a modulus. If the gaps are uneven, you have made a mistake in an argument. It takes fifteen seconds and it catches almost everything.

49. Exercise 75: the six sixth roots of 64

Worked example

From the exercise set, and a case where every answer comes out exactly.

\[ w^6 = 64 \]

Convert 64 to polar form

Why: It lies on the positive real axis, so its argument is zero.

\[ 64 = 64\,\operatorname{cis}(0), \qquad r = 64, \; \theta = 0, \; n = 6 \]

Find the common modulus

Why: Sixty-four is two to the sixth, so its principal sixth root is exactly two.

\[ \sqrt[6]{64} = 2 \]

Find the step between roots

Why: A full turn divided by six.

\[ \frac{2\pi}{6} = \frac{\pi}{3} \]

List the six arguments

Why: Starting from zero and stepping by a third of pi.

\[ 0, \; \frac{\pi}{3}, \; \frac{2\pi}{3}, \; \pi, \; \frac{4\pi}{3}, \; \frac{5\pi}{3} \]

Convert to rectangular form

Why: Each angle is a multiple of a third of pi, where the cosine and sine are the familiar halves and root three over two.

\[ 2, \; 1 + i\sqrt3, \; -1 + i\sqrt3, \; -2, \; -1 - i\sqrt3, \; 1 - i\sqrt3 \]

Figure (svg): A circle of radius 2 with six dots a sixth of a turn apart joined into a regular hexagon, two of them on the real axis at 2 and negative 2.

An even number of roots of a positive real number always includes a real pair, one at each end of a diameter.

Check a non-real one and count the real ones

Why: Cubing one plus i root three gives negative eight, and squaring that gives sixty-four; and the hexagon has exactly two corners on the real axis, at 2 and negative 2.

\[ (1 + i\sqrt3)^6 = 2^6\operatorname{cis}\!\left(6\cdot\tfrac{\pi}{3}\right) = 64\,\operatorname{cis}(2\pi) = 64 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1005 — Exercise 75

Try this one before reading, because it is the case where everything is exact and you can do the whole thing without a calculator.

The one piece of arithmetic worth pausing on is the modulus. The sixth root of sixty-four is two, because sixty-four is two to the sixth, and spotting that saves you from an ugly decimal. Whenever the number in a root problem is a perfect power, look for it.

The six arguments step by a third of pi, and every one of them is a familiar angle, so all six rectangular forms come out exactly. Look at the hexagon afterwards: two corners on the real axis at two and negative two, and the other four in conjugate pairs above and below. Conjugate pairs are what you should always expect when the original number is real.

50. Finish the three cube roots of i

Faded example

Exercise 71. The modulus of i is one and its argument is a right angle, so this one can be done entirely in your head once the pattern is familiar.

Figure (svg): The unit circle with i marked at the top and three cube roots at thirty, one hundred and fifty and two hundred and seventy degrees, joined into an equilateral triangle with one vertex at negative i.

The roots all have modulus 1 because the modulus of i is 1, and the first of them sits at one third of a right angle.

Fill in the blanks

i = \operatorname6\!\left(\tfrac-i___\right) \Rightarrow \; w_0 = \operatorname___\!\left(\frac______}\right), \; w_1 = \operatorname___\!\left(\frac______\right), \; w_2 = \operatorname___\!\left(\frac______\right) = ___

Why: Dividing a right angle by three gives a sixth of pi for the first root. Adding two thirds of pi twice gives five sixths and then three halves of pi, and the last of those points straight down, at negative i.

Try to fill both blanks without writing anything down. The modulus of i is one, so every cube root has modulus one and the only question is the angles.

The argument of i is a right angle, and a third of a right angle is a sixth of pi, which is thirty degrees. That is the first root. Add a third of a full turn twice and you land at five sixths of pi and then at three halves of pi.

The last one is the nice surprise. Three halves of pi points straight down, so the third cube root of i is exactly negative i. Check it directly: negative i cubed is negative i times negative i times negative i, which works out to i. Whenever a root lands on an axis you should verify it by hand, because it costs nothing and confirms the whole set.

51. Slide n and watch the spacing close up

Tweak it

Parameter explorer

The nth roots of 1 have real parts equal to the cosine of two pi k over n. The curve shows that expression for every real k, and the roots themselves are the values at the whole numbers k equal to 0 up to n minus 1. Slide n and watch how many whole numbers fit inside one period.

\[ \operatorname{Re}(w_k) = \cos\!\left(\frac{2\pi k}{{n}}\right) \]

  • n — from 2 to 14: how many roots, n

A slider draws an ordinary graph, so it cannot show you the circle directly. Instead it plots the real part of a root as a function of the index, and the actual roots are the values at the whole numbers.

Start at n equal to two. One period of the curve is only two units wide, and the two whole numbers inside it give plus one and minus one, the two square roots of one. Now slide n upward and watch the period stretch: more whole numbers fit inside it, which is the graphical version of more corners fitting round the circle.

The thing to take away is that the spacing between roots shrinks as n grows, and it shrinks in a completely predictable way, as one full turn split n ways. For very large n the roots crowd together until the polygon is almost indistinguishable from the circle itself.

52. Add all n of the roots of unity

Prediction

Figure (svg): The five fifth roots of unity drawn as arrows laid nose to tail: the chain starts at the origin, walks round through five equal steps and comes back exactly to the origin, tracing a closed pentagon.

Laid nose to tail the roots close into a polygon, which is the geometric reason their sum is zero for every n above one.

Predict first

Take all five fifth roots of 1 and add them together as complex numbers. What do you get?

  • 5
  • 1
  • 0
  • It depends on which root you start from

Correct: 0

Why: The five roots are arrows of equal length pointing in five equally spaced directions, so laying them nose to tail walks a closed regular pentagon and returns to the start. The same symmetry argument works for every n above one, and it is the reason the coefficient of the second-highest power in the polynomial is zero.

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1003 — the roots as the vertices of a regular polygon

Guess before you reveal, and try to guess from the picture rather than from arithmetic. Each root is an arrow of the same length, and they point in equally spaced directions.

The figure lays them nose to tail, which is how you add arrows. Because the directions are equally spaced, the chain turns by the same amount at every step, and after n steps it has turned all the way round and arrived back at its starting point. A closed chain means the total is zero.

This is one of those facts that is much easier to see than to compute. Adding five cosines of awkward angles is unpleasant; noticing the symmetry takes a second. And the same argument works for every n above one, which is why the second coefficient of the polynomial whose roots these are is always zero.

53. Complete the table of root sets

Comparison

Comparison matrix

equationmodulus of each rootargument of the first rootangle between roots
w³ = 8202π/3
w⁴ = −162π/4π/2
w³ = √2 + i√2∛2π/122π/3
w⁵ = 1102π/5
w⁶ = 6420π/3

Fill in the four blanks before checking. Every one of them can be read straight off the root formula without computing anything from scratch.

The angle between roots is always a full turn divided by n, so it depends only on the exponent, never on the number. The modulus of each root is always the principal nth root of the modulus of the number, so it depends only on the size of the number, never on its direction. The first argument is the only entry that mixes them.

Read the table down the columns rather than across the rows and you will see the structure clearly: three independent quantities, each controlled by exactly one feature of the problem. That separation is what makes root problems quick once you trust it.

54. Where a set of equally spaced roots turns up

Real world

Discussion prompt

A piano octave is divided into twelve equally spaced steps, and each step multiplies the frequency by the same factor. What number is that factor a root of, and how is it related to the twelve twelfth roots of 1?

Write your answer before revealing. The clue is that twelve equal multiplicative steps take you from a note to the note an octave above, whose frequency is exactly double.

So the step factor is a number whose twelfth power is two, which makes it a twelfth root of two. Of the twelve twelfth roots of two, the tuning uses the one on the positive real axis, because frequencies are positive real numbers. The other eleven are perfectly good roots of the same equation; they simply do not describe frequencies.

That is the general shape of things. In an application you usually want one root, the real positive one, but the equation always has n of them, and knowing where the others sit is what lets you factor a polynomial completely, analyse an alternating current with several phases, or run a discrete Fourier transform. The polygon is not decoration; it is the structure the applications use.

55. Exercise 79: factoring a twelfth-degree polynomial

Worked example

Roots of a complex number are zeros of a polynomial, so the root formula factors things that no other method here can touch.

\[ p(x) = x^{12} - 4096 \]

Recognise the root equation

Why: The zeros of p are exactly the twelfth roots of 4096.

\[ p(x) = 0 \iff x^{12} = 4096 \]

Convert the constant to polar form

Why: Four thousand and ninety-six is two to the twelfth, and it lies on the positive real axis.

\[ 4096 = 4096\,\operatorname{cis}(0), \quad \sqrt[12]{4096} = 2 \]

Write the twelve zeros

Why: Modulus two, arguments a sixth of pi apart, starting at zero.

\[ x_k = 2\,\operatorname{cis}\!\left(\frac{\pi k}{6}\right), \qquad k = 0, 1, \ldots, 11 \]

Separate the real zeros

Why: At k equal to 0 and k equal to 6 the argument is 0 and pi, both on the real axis.

\[ x_0 = 2, \qquad x_6 = -2 \]

Pair the rest as conjugates

Why: The remaining ten come in five conjugate pairs; each pair multiplies to an irreducible real quadratic, using that the sum of a conjugate pair is twice the real part and the product is the modulus squared.

\[ (x - x_k)(x - \overline{x_k}) = x^2 - 4\cos\!\left(\tfrac{\pi k}{6}\right)x + 4 \]

Write the factorisation

Why: Two linear factors and five quadratics, with the cosines evaluated at a sixth, a third, a half, two thirds and five sixths of pi.

\[ p(x) = (x-2)(x+2)(x^2 + 4)\;\times \]

\[ (x^2 - 2\sqrt3x + 4)(x^2 - 2x + 4)(x^2 + 2x + 4)(x^2 + 2\sqrt3x + 4) \]

Check the degree and the constant term

Why: Two linear factors and five quadratics give degree twelve, and the constant terms multiply to the right value.

\[ (-2)(2)(4)^5 = -4\cdot 1024 = -4096 \;\checkmark \]

Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers §11.7, p. 1005 — Exercise 79

This is the capstone problem of the section, and it shows why anyone would care about finding all the roots rather than just one. A twelfth-degree polynomial is hopeless by any method from Chapter 3 alone.

The chain of reasoning is short. The zeros of the polynomial are the twelfth roots of four thousand and ninety-six, the root formula gives all twelve of them at once, and the Complex Factorization Theorem then says the polynomial is the product of the twelve corresponding linear factors.

The last move is the elegant one. Complex zeros of a real polynomial come in conjugate pairs, and multiplying a conjugate pair of linear factors gives a real quadratic with no complex numbers left in it. Ten of the zeros pair off into five quadratics, the two real zeros stay as linear factors, and you have factored a twelfth-degree polynomial over the real numbers with no guessing at all.

56. Putting it together

Section

Part 6

57. Pattern: a power or a root, start to finish

Pattern

Figure (svg): A flow diagram: a complex number goes into a polar-form box, then splits into a power branch using DeMoivre and a root branch using the root formula, and both branches end in a rectangular-form box after a check on the modulus.

Both branches begin with the same conversion and end with the same cheap check on the modulus; only the middle box differs.
  1. Convert first. Find the modulus and one argument, using the quadrant to fix the angle.
  2. For a power, raise the modulus to n and multiply the argument by n; then subtract full turns.
  3. For a root, take the principal nth root of the modulus and divide the argument by n; that is the first root only.
  4. Then rotate, adding one nth of a full turn n minus 1 times, and stop: the next one repeats.
  5. Convert back if rectangular form is asked for, and check the modulus of every answer.
  6. Plot the roots. One circle, equal spacing, a regular polygon, or something is wrong.

This is the order to work in for every problem in this section, and the diagram shows where the two branches share their work. Conversion at the start and a modulus check at the end are common to both.

The branch point is the only place you need to think. For a power there is one answer, and the work is raising and multiplying. For a root there are n answers, and the work is extracting and dividing, followed by n minus one rotations.

The last two items on the list are the ones students skip and should not. Checking the modulus catches arithmetic slips in seconds, and plotting the roots catches structural errors that arithmetic checks miss. Both together cost less than half a minute on any problem in this section.

58. Put the root procedure in order

Ranking

Put in order

Order the steps of a complete solution to the equation w to the n equals z.

  1. Find the modulus of z and an argument of z from its quadrant
  2. Take the principal nth root of the modulus
  3. Divide the argument by n to get the first root
  4. Add one nth of a full turn repeatedly, stopping after n roots in all
  5. Convert each root to rectangular form where the angle allows it
  6. Plot the roots and confirm one circle and equal spacing

Why: Conversion always comes first, because the formula reads the modulus and the argument. The modulus and the first argument can be done in either order, but both must precede the rotations, and the plot is the last thing because it checks everything before it.

Drag the steps into order before checking. Only one ordering makes sense, and the reason is that each step consumes the output of the one before it.

The conversion has to come first, because the formula reads a modulus and an argument and you have neither until you convert. The rotations have to come after the first root, because they rotate it. And the plot has to come last, because its whole job is to check the work that precedes it.

The one genuinely optional step is the conversion back to rectangular form, which some problems ask for and some do not. Do not spend time on messy half-angle work unless the question actually demands it, as the cube roots with a twelfth of pi in them showed.

59. Check: the product and quotient rules

Check

Check your understanding

With z of modulus 6 and argument two thirds of pi, and w of modulus 3 and argument pi over 6, what is z divided by w?

  • A. 2 cis(π/2) (correct)
  • B. 2 cis(5π/6)
  • C. 18 cis(π/2)
  • D. 3 cis(π/2)

Answer: A

Why: Moduli divide, so six over three is two; arguments subtract, so two thirds of pi minus one sixth of pi is four sixths minus one sixth, which is three sixths, a half of pi. The answer is two cis of pi over two, which is 2i.

Why B tempts people
That adds the arguments instead of subtracting them; addition is the product rule, not the quotient rule.
Why C tempts people
That multiplies the moduli. Division divides them, so the modulus should be two, not eighteen.
Why D tempts people
That subtracts the moduli instead of dividing them. There is no subtraction rule for moduli.

The question is about which operation belongs to which channel, and three of the four options are built from plausible confusions of exactly that.

Moduli divide and arguments subtract. Six over three is two, and two thirds of pi minus one sixth of pi is a half of pi. If you added the arguments you were using the product rule; if you subtracted the moduli you were using no rule at all.

Worth noticing: the answer, two cis of pi over two, is the purely imaginary number two i. Quotients of two messy complex numbers can be extremely clean, which is one of the reasons polar form is worth the conversion cost.

60. Check: DeMoivre's Theorem

Check

Check your understanding

What is the sixth power of 2 cis(π/3), written with its principal argument?

  • A. 12 cis(2π)
  • B. 64 cis(0) (correct)
  • C. 64 cis(2π/3)
  • D. 12 cis(0)

Answer: B

Why: The modulus two is raised to the sixth power, giving 64, and the argument is multiplied by six, giving six times a third of pi, which is two pi. A full turn is coterminal with zero, so the principal form is 64 cis(0), the real number 64.

Why A tempts people
The modulus was multiplied by six instead of raised to the sixth power, and a full turn should be reduced to zero.
Why C tempts people
That divides the argument by something or leaves a stray third of pi; six times a third of pi is exactly two pi.
Why D tempts people
The argument is right after reduction, but the modulus was multiplied by six rather than raised to the sixth power.

Two of the wrong options here multiply the modulus by the exponent instead of raising it, which is the error the annotation slide warned about. If you chose one of those, say the theorem out loud a few times: power the length, multiply the angle.

The other thing being tested is the reduction. Six times a third of pi is exactly two pi, a full turn, which is coterminal with zero. So the answer is a positive real number, sixty-four, even though the number you started from was nowhere near the real axis.

That is a pattern worth spotting. Whenever the argument divides evenly into a full turn, some power of the number will land on the positive real axis, and the polar method finds it instantly.

61. Check: counting and placing the roots

Check

Check your understanding

The cube roots of a complex number z include the number 3 cis(π/9). Which of these is another of them?

  • A. 3 cis(π/3)
  • B. 3 cis(7π/9) (correct)
  • C. 9 cis(π/9)
  • D. 1 cis(π/9 + 2π/3)

Answer: B

Why: All three cube roots share the modulus 3, and consecutive arguments differ by a third of a full turn, which is two thirds of pi, or six ninths of pi. One ninth plus six ninths is seven ninths of pi.

Why A tempts people
The modulus is right, but a third of pi is only three ninths of pi away from the given root, which is half the required step.
Why C tempts people
The roots all have the same modulus as each other, not the cube of it. Nine is the modulus of a square of the root, not of another root.
Why D tempts people
The argument is right but the modulus was changed to one. Every root of the same number has the same modulus.

This question gives you one root and asks for another, which is the fastest possible test of whether you understand the structure rather than the formula.

Two things are fixed for all the roots of a number: the modulus, which is shared, and the angular step, which is a full turn divided by n. Here n is three, so the step is two thirds of pi, and one ninth of pi plus six ninths of pi gives seven ninths of pi.

Each wrong option changes something that is not allowed to change. One keeps the modulus but uses half the correct step. One cubes the modulus. One keeps the correct argument but alters the modulus to one. If you can say which of those three mistakes an option makes, you understand the theorem.

62. Explain the polygon

Explain it to yourself

Discussion prompt

In two sentences, explain why the n distinct nth roots of any non-zero complex number must be the corners of a regular polygon, referring to the two parts of the root formula separately.

Write your two sentences before revealing, and make sure each one is about a different part of the formula. This is the explanation that lets you rebuild the theorem if you ever forget it.

The first sentence should be about the modulus: it is the same for every root, so they all lie on one circle. The second should be about the argument: consecutive roots differ by the same angle, so they are spread evenly round that circle.

If your explanation does not mention both, it is not complete. Points on a circle need not be a polygon, and equally spaced angles alone say nothing about distance. It is the two conditions together that force the regular polygon, which is why the plot check catches both kinds of mistake.

63. Exit ticket

Exit ticket

\[ \text{Find all four fourth roots of } z = -8 + 8i\sqrt3. \]

Discussion prompt

Convert z, give the four roots in polar form, say what polygon they form, and name one check you could run on your answers without a calculator.

This one problem uses everything: conversion with attention to the quadrant, the root formula, the polygon, and a check.

The number is in Quadrant II with modulus sixteen and argument two thirds of pi. The fourth root of sixteen is two, two thirds of pi divided by four is a sixth of pi, and the step between roots is a quarter of a full turn. That gives the four arguments on the answer slide.

For the check, the cheapest thing is the modulus: two to the fourth is sixteen, which matches. The next cheapest is the polygon: four roots of equal modulus a quarter turn apart form a square. And if you want a third, multiply each argument by four and confirm that all four results are coterminal with two thirds of pi, which is the defining property being tested.

64. Recap

Recap

rulewhat it does to the moduluswhat it does to the argument
productmultipliesadds
quotientdividessubtracts
power (DeMoivre)raises to the power nmultiplies by n
nth roottakes the principal nth rootdivides by n, then adds multiples of one nth of a turn

\[ w_k = \sqrt[n]{r}\;\operatorname{cis}\!\left(\frac{\theta}{n} + \frac{2\pi}{n}k\right), \qquad k = 0, 1, \ldots, n-1 \]

One picture carries the whole lesson: multiplication stretches and turns, and the n roots of any non-zero number are the corners of a regular polygon on one circle. Section 11.8 leaves complex numbers behind and starts on vectors, where the same arrows reappear without the multiplication.

OpenStax Algebra and Trigonometry 2e, §10.5 Polar Form of Complex Numbers §10.5 — the same material in OpenStax Algebra and Trigonometry

Four rules in the table, and each of them does something separate to the modulus and to the argument. If you only keep one sentence from this lesson, keep that one, because every result here is an instance of it.

The root formula is the only thing worth memorising in detail, and even that is recoverable from the picture. A circle whose radius is the nth root of the distance, a first corner at one nth of the original angle, and n corners equally spaced round it. If you can draw that, you can write the formula.

Section 11.8 starts vectors, which look like the arrows you have been drawing all lesson and behave like them under addition and scaling. What vectors in the plane do not have is multiplication, and the loss of it is exactly what makes complex numbers a richer object than a pair of coordinates.

Sources

  1. Stitz & Zeager, College Trigonometry, Ch. 11 Applications of Trigonometry — §11.7 Polar Form of Complex Numbers — Stitz & Zeager, College Trigonometry, Version 3 Corrected Edition, July 4 2013, pp. 997-1004
  2. OpenStax Algebra and Trigonometry 2e, §10.5 Polar Form of Complex Numbers

Want this taught 1-on-1? Alexander tutors Trigonometry — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108