Unit 6, lesson 44

Complex Numbers and Complex Exponentials

How complex numbers are added, multiplied and divided, how polar form turns multiplication into rotation, why Euler's formula holds, and how complex exponentials produce real oscillating solutions.

  • Theory part 87
  • Practice part 88
  • Exact Equations and Complex Numbers

The equation y′′+y=0y'' + y = 0 describes a mass bouncing on a spring, and its solutions are cos⁡t\cos t and sin⁡t\sin t. Try the usual guess y=erty = e^{rt} and you need r2=−1r^2 = -1, which no real number satisfies. Complex numbers supply that missing rr, and Euler’s formula shows that the exponential it produces is secretly a pair of sines and cosines.

Key terms

Real and imaginary parts

Definition A complex number is an expression z=a+biz = a + bi with aa and bb real numbers, where ii is a symbol with i2=−1i^2 = -1. Two complex numbers are equal when both their parts are equal. The real number aa is the real part, written Re⁡z\operatorname{Re} z, and the real number bb is the imaginary part, written Im⁡z\operatorname{Im} z. The set of all complex numbers is C\mathbb{C}, and the real numbers are the complex numbers with b=0b = 0.

In plain words A complex number is a pair of real numbers glued together with ii. Note that the imaginary part is the real coefficient bb, not the term bibi.

Example For z=3+4iz = 3 + 4i the real part is 33 and the imaginary part is Im⁡z=4\operatorname{Im} z = 4, while 4i4i is the imaginary term. The number −2i-2i has real part 00 and imaginary part −2-2.

The complex plane

Definition The complex plane identifies z=a+biz = a + bi with the point (a,b)(a, b) of R2\mathbb{R}^2, or with the vector ⟨a,b⟩\langle a, b \rangle from the origin. The horizontal axis is the real axis and the vertical axis is the imaginary axis.

In plain words Every complex number is a point in an ordinary plane, with the real part as the first coordinate and the imaginary part as the second.

Example The number 3+4i3 + 4i is the point (3,4)(3, 4), three units right and four units up. The vector to it is the hypotenuse of a right triangle with legs 33 and 44, so it has length 55.

Arithmetic

Definition For z=a+biz = a + bi and w=c+diw = c + di,

z+w=(a+c)+(b+d)i,zw=(ac−bd)+(ad+bc)i.\begin{aligned} z + w &= (a + c) + (b + d)i, \\[0.4em] zw &= (ac - bd) + (ad + bc)i. \end{aligned}

The product rule comes from expanding and using i2=−1i^2 = -1. With these operations C\mathbb{C} obeys the same commutative, associative and distributive laws as R\mathbb{R}, and every w≠0w \ne 0 has a reciprocal, so division by nonzero numbers is defined. Unlike R\mathbb{R}, there is no ordering of C\mathbb{C} compatible with arithmetic, so inequalities like z<wz < w have no meaning.

In plain words Add parts to parts. Multiply like polynomials in ii, then replace i2i^2 by −1-1.

Example The product of 3+4i3 + 4i and 1−2i1 - 2i is

(3+4i)(1−2i)=3−6i+4i+8=11−2i.\begin{aligned} (3 + 4i)(1 - 2i) &= 3 - 6i + 4i + 8 \\[0.4em] &= 11 - 2i. \end{aligned}

Conjugates

Definition The conjugate of z=a+biz = a + bi is z‾=a−bi\overline{z} = a - bi. For all zz and ww,

z+w‾=z‾+w‾,zw‾=z‾ w‾,\overline{z + w} = \overline{z} + \overline{w}, \qquad \overline{zw} = \overline{z}\,\overline{w},

and zz‾=a2+b2z \overline{z} = a^2 + b^2, a nonnegative real number that is zero only for z=0z = 0. Also Re⁡z=12(z+z‾)\operatorname{Re} z = \frac{1}{2}(z + \overline{z}), and zz is real exactly when z‾=z\overline{z} = z. For w≠0w \ne 0, division is done by multiplying top and bottom by w‾\overline{w},

zw=z w‾w w‾.\frac{z}{w} = \frac{z\,\overline{w}}{w\,\overline{w}}.

In plain words Conjugating flips the sign of the imaginary part, which reflects the point across the real axis. Multiplying a number by its conjugate gives a real number, and that is what makes division work.

Example The conjugate of 3+4i3 + 4i is 3−4i3 - 4i, the point (3,−4)(3, -4), and (3+4i)(3−4i)=9+16=25(3 + 4i)(3 - 4i) = 9 + 16 = 25. So

13+4i=3−4i25=325−425i.\frac{1}{3 + 4i} = \frac{3 - 4i}{25} = \frac{3}{25} - \frac{4}{25}i.

Modulus and argument

Definition The modulus of z=a+biz = a + bi is

∣z∣=a2+b2=zz‾,\lvert z \rvert = \sqrt{a^2 + b^2} = \sqrt{z \overline{z}},

the distance from 00 to zz in the complex plane. It satisfies ∣zw∣=∣z∣∣w∣\lvert zw \rvert = \lvert z \rvert \lvert w \rvert. For z≠0z \ne 0, an argument of zz is any real θ\theta, in radians, with

a=∣z∣cos⁡θ,b=∣z∣sin⁡θ.a = \lvert z \rvert \cos \theta, \qquad b = \lvert z \rvert \sin \theta.

Arguments of the same zz differ by integer multiples of 2π2\pi. The principal argument Arg⁡z\operatorname{Arg} z is the one in (−π,π](-\pi, \pi]. The number 00 has no argument.

In plain words The modulus is how far the point is from the origin, and an argument is the angle its vector makes with the positive real axis.

Example For 3+4i3 + 4i the modulus is 55, and since the point is in the first quadrant Arg⁡(3+4i)=arctan⁡(4/3)≈0.9273\operatorname{Arg}(3 + 4i) = \arctan(4/3) \approx 0.9273. For −1−i-1 - i the formula arctan⁡(b/a)=arctan⁡1=π/4\arctan(b/a) = \arctan 1 = \pi/4 points to the wrong quadrant. The point (−1,−1)(-1, -1) is in the third quadrant, so Arg⁡(−1−i)=−3π/4\operatorname{Arg}(-1 - i) = -3\pi/4.

Polar form and roots

Definition Every z≠0z \ne 0 can be written in polar form

z=r(cos⁡θ+isin⁡θ),r=∣z∣>0,\begin{gathered} z = r(\cos \theta + i \sin \theta), \\[0.4em] r = \lvert z \rvert > 0, \end{gathered}

with θ\theta an argument of zz. By the angle addition formulas, if w=s(cos⁡φ+isin⁡φ)w = s(\cos \varphi + i \sin \varphi) then

zw=rs(cos⁡(θ+φ)+isin⁡(θ+φ)),\begin{aligned} zw &= rs\bigl(\cos(\theta + \varphi) \\[0.4em] &\qquad + i \sin(\theta + \varphi)\bigr), \end{aligned}

and by induction zn=rn(cos⁡nθ+isin⁡nθ)z^n = r^n(\cos n\theta + i \sin n\theta) for every positive integer nn. The same holds for n=0n = 0 and for negative nn, because 1/z=r−1(cos⁡θ−isin⁡θ)1/z = r^{-1}(\cos \theta - i \sin \theta). For a positive integer nn, a number w=R(cos⁡φ+isin⁡φ)w = R(\cos \varphi + i \sin \varphi) with R>0R > 0 has exactly nn distinct nnth roots,

R1/n(cos⁡φ+2πkn+isin⁡φ+2πkn),k=0,1,…,n−1,\begin{gathered} R^{1/n}\Bigl(\cos \tfrac{\varphi + 2\pi k}{n} + i \sin \tfrac{\varphi + 2\pi k}{n}\Bigr), \\[0.4em] k = 0, 1, \ldots, n - 1, \end{gathered}

where R1/nR^{1/n} is the positive real root.

In plain words Multiplying complex numbers multiplies the lengths and adds the angles. Taking an nnth root does the reverse, and the roots sit evenly spaced on a circle.

Example To find the cube roots of 8i8i, write 8i=8(cos⁡π2+isin⁡π2)8i = 8(\cos \frac{\pi}{2} + i \sin \frac{\pi}{2}). The roots have modulus 22 and angles π6\frac{\pi}{6}, 5π6\frac{5\pi}{6} and 3π2\frac{3\pi}{2}, namely

3+i,−3+i,−2i.\sqrt{3} + i, \qquad -\sqrt{3} + i, \qquad -2i.

As a check, (−2i)3=−8i3=8i(-2i)^3 = -8i^3 = 8i.

Euler’s formula from power series

Definition For a complex zz define

ez=∑n=0∞znn!.e^{z} = \sum_{n=0}^{\infty} \frac{z^n}{n!}.

The series converges absolutely for every zz, by comparison with the real series for e∣z∣e^{\lvert z \rvert}. For real θ\theta, the powers of iθi\theta cycle through 1,i,−1,−i1, i, -1, -i times θn\theta^n, and absolute convergence allows the even and odd terms to be collected separately,

eiθ=∑k=0∞(−1)kθ2k(2k)!+i∑k=0∞(−1)kθ2k+1(2k+1)!=cos⁡θ+isin⁡θ.\begin{aligned} e^{i\theta} &= \sum_{k=0}^{\infty} \frac{(-1)^k \theta^{2k}}{(2k)!} \\[0.4em] &\quad + i \sum_{k=0}^{\infty} \frac{(-1)^k \theta^{2k+1}}{(2k+1)!} \\[0.4em] &= \cos \theta + i \sin \theta. \end{aligned}

This is Euler’s formula, with θ\theta in radians.

In plain words Feed iθi\theta into the exponential series, and the even terms rebuild the cosine series while the odd terms rebuild ii times the sine series.

Example At θ=π\theta = \pi the formula gives eiπ=cos⁡π+isin⁡π=−1e^{i\pi} = \cos \pi + i \sin \pi = -1, and at θ=π/2\theta = \pi/2 it gives eiπ/2=ie^{i\pi/2} = i. The polar form now has the short form z=reiθz = r e^{i\theta}, so for instance 3+4i≈5e0.9273i3 + 4i \approx 5e^{0.9273 i}.

The complex exponential

Definition The series above satisfies ez+w=ezewe^{z + w} = e^z e^w for all complex zz and ww. This is a standard fact, proved by multiplying the two absolutely convergent series term by term, and here it is used without proof. Combined with Euler’s formula, for real aa and bb,

ea+bi=ea(cos⁡b+isin⁡b).e^{a + bi} = e^a(\cos b + i \sin b).

So ∣ea+bi∣=ea>0\lvert e^{a + bi} \rvert = e^a > 0, which means eze^z is never 00, and ez+2πi=eze^{z + 2\pi i} = e^z. For a complex constant rr and a real variable tt, differentiating real and imaginary parts separately gives

ddtert=rert.\frac{d}{dt} e^{rt} = r e^{rt}.

In plain words The real part of the exponent sets the size and the imaginary part sets the angle. As tt grows, erte^{rt} spins around the origin while it grows or shrinks.

Example For z=ln⁡2+iπ/3z = \ln 2 + i\pi/3,

ez=2(cos⁡π3+isin⁡π3)=1+i3.\begin{aligned} e^{z} &= 2\Bigl(\cos \frac{\pi}{3} + i \sin \frac{\pi}{3}\Bigr) \\[0.4em] &= 1 + i\sqrt{3}. \end{aligned}

The curve e(−1+2i)te^{(-1 + 2i)t} for t≥0t \ge 0 circles the origin once every π\pi units of time while its modulus e−te^{-t} shrinks, so it spirals inward.

Real sinusoidal solutions

Definition Let pp and qq be real constants and let z(t)=u(t)+iv(t)z(t) = u(t) + i v(t) be a complex valued solution of y′′+py′+qy=0y'' + p y' + q y = 0 on an interval, with uu and vv real valued. Then uu and vv are each real solutions. The reason is that

0=(u′′+pu′+qu)+i(v′′+pv′+qv),\begin{aligned} 0 &= (u'' + p u' + q u) \\[0.4em] &\quad + i(v'' + p v' + q v), \end{aligned}

and both brackets are real because pp and qq are, so both must be 00. In particular, if r=α+iβr = \alpha + i\beta with β≠0\beta \ne 0 solves r2+pr+q=0r^2 + p r + q = 0, then eαtcos⁡βte^{\alpha t} \cos \beta t and eαtsin⁡βte^{\alpha t} \sin \beta t are real solutions.

In plain words One complex exponential solution hands you two real solutions for free, its real part and its imaginary part. Real coefficients are what make this split work.

Example For y′′+4y=0y'' + 4y = 0 the number r=2ir = 2i satisfies r2+4=0r^2 + 4 = 0, so e2it=cos⁡2t+isin⁡2te^{2it} = \cos 2t + i \sin 2t is a solution and cos⁡2t\cos 2t and sin⁡2t\sin 2t are real solutions. To meet y(0)=1y(0) = 1 and y′(0)=6y'(0) = 6, try y=c1cos⁡2t+c2sin⁡2ty = c_1 \cos 2t + c_2 \sin 2t. Then y(0)=c1=1y(0) = c_1 = 1 and y′(0)=2c2=6y'(0) = 2c_2 = 6, so

y=cos⁡2t+3sin⁡2t.y = \cos 2t + 3 \sin 2t.

Checking, y′′=−4cos⁡2t−12sin⁡2t=−4yy'' = -4\cos 2t - 12 \sin 2t = -4y, as needed.

Check yourself

  1. Write 2+3i1−i\dfrac{2 + 3i}{1 - i} in the form a+bia + bi.
Show the answer

Multiply top and bottom by the conjugate 1+i1 + i. The bottom becomes 22 and the top becomes 2+2i+3i+3i2=−1+5i2 + 2i + 3i + 3i^2 = -1 + 5i, so

2+3i1−i=−12+52i.\frac{2 + 3i}{1 - i} = -\frac{1}{2} + \frac{5}{2}i.
  1. Put z=−1+i3z = -1 + i\sqrt{3} in polar form and use it to compute z3z^3.
Show the answer

The modulus is 1+3=2\sqrt{1 + 3} = 2. The point (−1,3)(-1, \sqrt{3}) is in the second quadrant, so Arg⁡z=2π/3\operatorname{Arg} z = 2\pi/3 and z=2e2πi/3z = 2e^{2\pi i/3}. Then z3=8e2πi=8z^3 = 8e^{2\pi i} = 8.

  1. Find the real solution of y′′+2y′+5y=0y'' + 2y' + 5y = 0 with y(0)=1y(0) = 1 and y′(0)=1y'(0) = 1.
Show the answer

The roots of r2+2r+5=0r^2 + 2r + 5 = 0 are r=−1±2ir = -1 \pm 2i, so the real solutions e−tcos⁡2te^{-t}\cos 2t and e−tsin⁡2te^{-t}\sin 2t are available. With y=e−t(c1cos⁡2t+c2sin⁡2t)y = e^{-t}(c_1 \cos 2t + c_2 \sin 2t), the condition y(0)=1y(0) = 1 gives c1=1c_1 = 1, and y′(0)=−c1+2c2=1y'(0) = -c_1 + 2c_2 = 1 gives c2=1c_2 = 1. The solution is

y=e−t(cos⁡2t+sin⁡2t).y = e^{-t}(\cos 2t + \sin 2t).

How this lesson connects

It also uses Real algebra and trigonometry, Power series and ordinary derivatives and The meaning of a differential-equation solution.