Mathematics · Ch 2 — De Moivre's Theorem
De Moivre's Theorem for Integral and Rational Indices
De Moivre's Theorem for Integral and Rational Indices
We already know that multiplying two complex numbers written in the form is easy on the angles: , because and satisfy the angle-addition formulas. Squaring a single is just a special case of this: . The natural question is what happens for a general power — does always collapse to , for any integer , not just ? The answer is yes, and this fact is De Moivre's theorem, named after the French mathematician Abraham de Moivre.
Theorem (integral index). For any real and any integer ,
The cleanest way to see why this must be true is to build it up in three stages, matching the three flavours of integer: positive, zero, and negative.
Positive integers — induction. For the two sides are identical, so the base case is free. Assume the formula already holds for some positive integer , i.e. . Multiply both sides by one more factor of and expand the right-hand product using ; grouping real and imaginary parts turns the product into purely by the ordinary cosine/sine addition formulas. That is exactly the statement for . So the claim, once true for one positive integer, is automatically true for the next, and by induction it holds for every positive integer.
Zero. , so the formula holds trivially at .
Negative integers. Write with a positive integer. Then is , and the positive-integer case already tells us the denominator is . Rationalising this fraction (multiply top and bottom by , and use ) leaves exactly , which is . So the negative case follows from the positive case for free.
Putting the three cases together proves the theorem for every integer , positive, zero, or negative — nothing is assumed about the sign of anywhere the theorem is actually used.
It is standard shorthand to write for ; in that notation De Moivre's theorem is simply . A few immediate consequences are worth keeping handy: (apply the theorem to , since and ); and because , each of and is literally the reciprocal of the other.
Rational indices. The theorem as stated is only about integer powers, because 'raising to a rational power' isn't single-valued the way integer powers are — a rational power like genuinely has two possible values, and has three. What survives for a rational exponent (with ) is a weaker, one-valued statement: is one of the values of . This follows by raising to the power : using the integral-index theorem twice, , which exactly says is a root of — i.e. a legitimate value of . This rational-index version is what powers the root-finding in the next section.
Worked example. Simplify . …