Complex Numbers and the Complex Plane

Why does $x^2 + 1 = 0$ have no solution among the real numbers? What happens if we insist it should? This simple demand — that every polynomial equation deserve a root — forces us to build a new number system. The system that emerges turns out to be not an ad hoc trick. Instead, it is a complete, rigid, and geometrically beautiful extension of the real line. The complex numbers $\mathbb{C}$ are a two-dimensional real vector space equipped with a multiplication that makes them a field. Remarkably, this algebraic structure is inseparable from a geometric one: multiplying by a fixed complex number is always a rotation-and-scaling of the plane. This dual nature — algebra you can compute with, geometry you can see — is why complex numbers matter far beyond satisfying curiosity about $\sqrt{-1}$ . Every complex number carries a magnitude (its modulus) and a direction (its argument). The interplay between rectangular coordinates $x+iy$ and polar coordinates $r(\cos\theta + i\sin\theta)$ is important. It turns multiplication, which is opaque in coordinates, into addition of angles, which is transparent. This single fact underlies de Moivre's formula, the existence of $n$ th roots. It eventually underlies the entire machinery of complex analysis. The big picture is that $\mathbb{C}$ is simultaneously a field, a normed vector space, and a metric space. It is isometric to $\mathbb{R}^2$ with the Euclidean metric. These are three viewpoints that must be shown to cohere. Once they do, one can speak unambiguously of distance, convergence, continuity. One can also speak of differentiability of complex-valued functions. All of this is built on the same foundations as real analysis. It is enriched by the extra structure of multiplication and rotation. This article develops the algebra of $\mathbb{C}$ from its formal definition. It establishes the modulus and triangle inequality. It proves the polar form and de Moivre's theorem. It uses these to solve $z^n = w$ completely. It closes by setting up the geometric vocabulary — lines, circles, and the topology of the plane. These are needed for Complex Functions and Limits and eventually The Complex Exponential and Logarithm.

Key takeaways

$\mathbb{C} = \mathbb{R}^2$ as a set, with addition componentwise and multiplication $(a,b)(c,d) = (ac-bd,\, ad+bc)$ .
Writing $i = (0,1)$ , every complex number is $z = x + iy$ with $i^2 = -1$ ; this makes $\mathbb{C}$ a field.
The modulus $|z| = \sqrt{x^2+y^2}$ is the Euclidean length of $z$ , and satisfies the triangle inequality $|z+w|\le |z|+|w|$ .
Every nonzero $z$ has polar form $z = r(\cos\theta + i\sin\theta)$ , and multiplication multiplies moduli and adds arguments.
De Moivre's formula $(\cos\theta+i\sin\theta)^n = \cos n\theta + i\sin n\theta$ gives all $n$ th roots of any nonzero complex number.
Key idea: $\mathbb{C}$ is a field whose multiplication is geometrically a rotation-and-scaling, unifying algebra and plane geometry.

Where this lives today

Complex numbers are the coordinate system of two-dimensional rotation and scaling, so they appear anywhere planar rigid motions combined with dilation matter: electrical engineers use phasors $Ve^{i\omega t}$ to encode the amplitude and phase of alternating currents, turning differential equations into algebra exactly because multiplication by $i$ is a $90°$ phase shift. In signal processing, the discrete and continuous Fourier transforms decompose signals into sums of complex exponentials, whose existence and uniqueness rest on the $n$ th-root corollary proved above. Quantum mechanics represents states as vectors over $\mathbb{C}$ , with the modulus-squared $|z|^2$ giving probabilities — the field structure, not just the vector space structure, is essential because complex phases interfere. In pure mathematics, the complex plane becomes the domain of Complex Functions and Limits, where continuity and differentiability are built on the metric $|z-w|$ established here, and the polar/exponential form developed here is the seed of The Complex Exponential and Logarithm. Computer graphics and robotics use complex numbers (and their quaternion generalizations) to compose 2D rotations efficiently, and number theorists use Gaussian integers $\mathbb{Z}[i]$ , a subring of $\mathbb{C}$ , to prove classical results about which integers are sums of two squares.

Ask the tutor: How does the polar form of complex multiplication foreshadow the definition of the complex exponential function?

\mathbb{C} = \mathbb{R}^2 = \{(x,y) : x, y \in \mathbb{R}\},

Intuition

Think of $\mathbb{C}$ as the ordinary Cartesian plane, but with an extra piece of structure: a way to multiply points, not just add them. Addition is nothing new — it's vector addition, following the parallelogram rule you already know from $\mathbb{R}^2$ . Multiplication is the surprise. If $z = x+iy$ sits at distance $r=|z|$ from the origin at angle $\theta$ from the positive real axis, then multiplying any other complex number $w$ by $z$ scales $|w|$ by the factor $r$ and rotates $w$ by the angle $\theta$ . This is the entire content of the polar form and it is worth internalizing as a picture before internalizing it as algebra. A second mental image worth carrying is $\mathbb{C}$ as a number system extending $\mathbb{R}$ : the real axis $\{x + i0\}$ sits inside $\mathbb{C}$ as an ordinary copy of $\mathbb{R}$ , and $i$ is a new number, off that axis, whose defining property $i^2=-1$ lets every real quadratic have roots. The imaginary axis $\{iy\}$ is the set of numbers obtained by rotating the real axis a quarter turn — indeed multiplying by $i$ is a $90°$ rotation, since $i(x+iy) = -y + ix$ , which is $(x,y)$ rotated counterclockwise by $\pi/2$ .

Figure

The complex plane: rectangular vs. polar coordinates of

z = 3+4i

z, w \in \mathbb{C}

\mathbb{C} = \mathbb{R}^2

Worked examples

Example 1 — Rationalizing and computing with an asymmetric quotient. Compute $\dfrac{(2-5i)}{(1+3i)} + (4+i)^2$ and express the result in the form $x+iy$ , then find its modulus and argument. Step 1: the quotient. Multiply numerator and denominator by the conjugate of the denominator:

\frac{2-5i}{1+3i} = \frac{(2-5i)(1-3i)}{(1+3i)(1-3i)} = \frac{2 - 6i - 5i + 15i^2}{1 - 9i^2} = \frac{2 - 11i - 15}{1+9} = \frac{-13 - 11i}{10} = -\frac{13}{10} - \frac{11}{10}i.

Step 2: the square. $(4+i)^2 = 16 + 8i + i^2 = 16+8i-1 = 15+8i$ . Step 3: sum.

\left(-\frac{13}{10} - \frac{11}{10}i\right) + (15+8i) = \left(15 - \frac{13}{10}\right) + \left(8 - \frac{11}{10}\right)i = \frac{137}{10} + \frac{69}{10}i.

Step 4: modulus.

\left|\frac{137}{10}+\frac{69}{10}i\right| = \frac{1}{10}\sqrt{137^2+69^2} = \frac{1}{10}\sqrt{18769 + 4761} = \frac{1}{10}\sqrt{23530}.

Numerically $\sqrt{23530} \approx 153.395$ , so the modulus is $\approx 15.3395$ . Step 5: argument. Since real part $137/10>0$ and imaginary part $69/10>0$ , the point lies in the first quadrant, so

\theta = \arctan\!\left(\frac{69}{137}\right) \approx \arctan(0.50365) \approx 0.4657 \text{ radians} \approx 26.68^\circ.

Example 2 — Solving $z^5 = -16 + 16\sqrt{3}\,i$ using polar form and de Moivre. Step 1: polar form of the right side. Let $w = -16+16\sqrt3 i$ . Its modulus is

|w| = \sqrt{(-16)^2 + (16\sqrt3)^2} = \sqrt{256 + 768} = \sqrt{1024} = 32.

Its argument: $\operatorname{Re}(w)<0$ , $\operatorname{Im}(w)>0$ , so $w$ is in the second quadrant. We have $\cos\phi = -16/32 = -1/2$ and $\sin\phi = 16\sqrt3/32 = \sqrt3/2$ , which identifies $\phi = 2\pi/3$ (i.e. $120^\circ$ ). So

w = 32\left(\cos\frac{2\pi}{3} + i\sin\frac{2\pi}{3}\right).

Step 2: fifth roots. By the $n$ th-root formula (proved below as a corollary of de Moivre's theorem), the solutions of $z^5=w$ are

z_k = 32^{1/5}\left(\cos\frac{\frac{2\pi}{3}+2\pi k}{5} + i \sin\frac{\frac{2\pi}{3}+2\pi k}{5}\right), \qquad k=0,1,2,3,4.

Since $32 = 2^5$ , we get $32^{1/5}=2$ exactly. The five angles are

\theta_k = \frac{2\pi/3 + 2\pi k}{5}, \quad k=0,1,2,3,4,

namely

\theta_0 = \frac{2\pi}{15},\quad \theta_1 = \frac{2\pi}{15}+\frac{2\pi}{5}=\frac{8\pi}{15},\quad \theta_2=\frac{14\pi}{15},\quad \theta_3=\frac{20\pi}{15}=\frac{4\pi}{3},\quad \theta_4=\frac{26\pi}{15}.

So the five fifth roots are $z_k = 2(\cos\theta_k + i \sin\theta_k)$ for these five explicit angles, equally spaced by $2\pi/5 = 72^\circ$ around a circle of radius $2$ — a regular pentagon of vertices, one of which sits at angle $2\pi/15 = 24^\circ$ .

Intuition, continued: polar form

z = x+iy \ne 0

z_1 = r_1(\cos\theta_1+i\sin\theta_1)

z_1z_2 = r_1r_2\big[(\cos\theta_1\cos\theta_2 - \sin\theta_1\sin\theta_2) + i(\cos\theta_1\sin\theta_2+\sin\theta_1\cos\theta_2)\big].

n

z_k

Figure

The five fifth roots of

w=-16+16\sqrt3\,i

, equally spaced by

72°

on the circle

|z|=2

\mathbb{C}

Scope, edge cases & common traps

Situation	What to watch for
Argument of $0$	Undefined — $0$ has no direction; all polar-form statements assume $z\ne0$ .
Choosing $\operatorname{Arg}$ range	Convention-dependent: $(-\pi,\pi]$ here, but $[0,2\pi)$ is equally common; state your convention.
$\sqrt{z}$ for complex $z$	Not a single-valued function without a branch choice; "the" square root is ambiguous — there are two, given by the $n=2$ case of the root corollary.
Comparing complex numbers with $<$	Meaningless: $\mathbb{C}$ has no order compatible with its field structure (see counterexample). Only $\|z\|$ , $\operatorname{Re}(z)$ , $\operatorname{Im}(z)$ can be compared as reals.
$\overline{z^n}$ vs. $(\bar z)^n$	Equal, but a common slip is forgetting conjugation distributes over powers only via the multiplicativity $\overline{zw}=\bar z\bar w$ proven above, applied repeatedly.

Common trapThe classic error

A frequent mistake is writing $\arg(z_1 z_2) = \arg(z_1) + \arg(z_2)$ as an equation of numbers rather than of cosets mod $2\pi$ . For instance with principal arguments, $z_1=z_2 = \cos(3\pi/4)+i\sin(3\pi/4)$ (so $\operatorname{Arg}(z_1)=\operatorname{Arg}(z_2) = 3\pi/4$ ), we get $z_1z_2 = \cos(3\pi/2)+i\sin(3\pi/2)$ , whose principal argument is $-\pi/2$ , not $3\pi/4+3\pi/4 = 3\pi/2$ . The correct statement is only that $3\pi/2$ and $-\pi/2$ differ by $2\pi$ : the additivity theorem holds for some choice of argument on each side, not automatically for the principal branch.

The mathematician behind it

Origins · 1 minCarl Friedrich Gauss and how this was discovered

Carl Friedrich Gauss

1777–1855 · Göttingen, German Enlightenment mathematics

Gauss gave the first widely accepted rigorous proof (1799, his doctoral dissertation) that every nonconstant polynomial with real or complex coefficients has a complex root — the Fundamental Theorem of Algebra — and later (1831) explicitly promoted the geometric representation of complex numbers as points of a plane, which we now call the Gaussian plane or Argand plane (after Jean-Robert Argand, who published a similar idea in 1806). Gauss's insistence on treating $a+bi$ as a genuine geometric point, not merely a formal symbol, was decisive in making complex numbers respectable rather than merely tolerated.

History

- 1545 — Gerolamo Cardano manipulates expressions like

\sqrt{-15}

while solving cubic equations, without believing in them as numbers.

- 1572 — Rafael Bombelli gives formal computational rules for such "impossible" square roots in his Algebra, effectively the first arithmetic of complex numbers.

- 1777 — Leonhard Euler introduces the symbol

i

for

\sqrt{-1}

and proves

e^{i\theta} = \cos\theta+i\sin\theta

- 1799 — Gauss proves the Fundamental Theorem of Algebra in his doctoral thesis.

- 1831 — Gauss publishes the geometric interpretation of complex numbers as points in a plane, cementing the Argand–Gauss plane picture.

Why do we define complex multiplication with that specific formula

(a,b)(c,d)=(ac-bd,ad+bc)

rather than componentwise?

Componentwise multiplication $(a,b)(c,d) = (ac,bd)$ does define a commutative ring on $\mathbb{R}^2$ , but it is not a field: $(1,0)$ and $(0,1)$ are both nonzero, yet their product is $(0,0)$ , so it has zero divisors and $(0,1)$ has no inverse. The specific formula in the definition is engineered precisely so that $i=(0,1)$ satisfies $i^2=-1$ and every nonzero element becomes invertible (proved in the Field Axioms theorem), which is exactly what's needed to solve $x^2+1=0$ while keeping a field structure. It is also, as the intuition section shows, exactly the formula for composing a rotation-and-scaling with another rotation-and-scaling — so it is forced by demanding both algebraic and geometric coherence simultaneously.

\mathbb{C}

literally the same thing as

\mathbb{R}^2

As sets and as topological/metric spaces, yes: the map $x+iy \mapsto (x,y)$ is a bijection that is also an isometry between $(\mathbb{C}, |{\cdot}-{\cdot}|)$ and $(\mathbb{R}^2, \text{Euclidean distance})$ , since $|z-w| = \sqrt{(x-u)^2+(y-v)^2}$ by definition of the modulus. What $\mathbb{R}^2$ lacks on its own is the multiplication; $\mathbb{C}$ is $\mathbb{R}^2$ plus the extra multiplicative structure defined above, which turns it into a field and gives meaning to complex differentiability, roots of polynomials, and rotations — none of which are native to $\mathbb{R}^2$ as a bare metric space.

Why does multiplying by

i

rotate the plane by

90°

Write $z=x+iy$ . Then $iz = i(x+iy) = ix + i^2y = -y + ix$ , i.e. $iz$ corresponds to the point $(-y,x)$ . The linear map $(x,y)\mapsto(-y,x)$ is exactly the standard counterclockwise rotation by $\pi/2$ radians on $\mathbb{R}^2$ , which one checks preserves lengths ( $\sqrt{(-y)^2+x^2}=\sqrt{x^2+y^2}$ ) and sends the positive $x$ -axis to the positive $y$ -axis. More generally, multiplying by any unit-modulus complex number $\cos\theta+i\sin\theta$ rotates by $\theta$ , by the product formula in the de Moivre theorem with $r_1=1$ .

What exactly is the "principal" argument, and why is it discontinuous?

The principal argument $\operatorname{Arg}(z)$ is the unique representative of $z$ 's argument-coset lying in $(-\pi,\pi]$ (a convention; some texts use $[0,2\pi)$ ). It must jump somewhere because argument itself only makes sense modulo $2\pi$ , and any single-valued choice function on a full circle around the origin must have a discontinuity: as shown in the counterexample block, continuously tracking the angle around the unit circle accumulates a full $2\pi$ that a bounded-range function cannot record without a jump. For $(-\pi,\pi]$ , the jump occurs on the negative real axis, where $\operatorname{Arg}$ jumps from values near $-\pi$ just below the axis to values near $\pi$ just above it.

Can I compare two complex numbers by size, like

z < w

Not in general — as the counterexample shows, $\mathbb{C}$ admits no total order compatible with its addition and multiplication, so " $z<w$ " is meaningless for complex numbers as a blanket statement. What you can do is compare their moduli $|z|$ vs. $|w|$ (both nonnegative reals), or compare real parts, or compare imaginary parts — each of these is a genuine real-number comparison, just applied to one real-valued feature of $z$ and $w$ rather than to the complex numbers themselves.

How does the polar form connect to the complex exponential

e^{i\theta}

This article develops polar form purely trigonometrically, $z=r(\cos\theta+i\sin\theta)$ , without presupposing the exponential function on complex arguments. Once $e^z$ is defined for complex $z$ (the subject of The Complex Exponential and Logarithm), Euler's formula $e^{i\theta}=\cos\theta+i\sin\theta$ identifies polar form with $z=re^{i\theta}$ , and the product rule proved here, "multiply moduli, add arguments," becomes simply the law of exponents $e^{i\theta_1}e^{i\theta_2}=e^{i(\theta_1+\theta_2)}$ . De Moivre's formula is then just $(e^{i\theta})^n = e^{in\theta}$ .

Why does every nonzero complex number have exactly

n

distinct

n

th roots, while a positive real number has only one positive

n

th root?

Over the reals, taking $n$ th roots of a positive number and requiring the root to be a specific positive real singles out one value among possibly several real roots (e.g. $\pm\sqrt{4}=\pm2$ , but only $+2$ is "the" positive square root). Over $\mathbb{C}$ there's no positivity to break ties, and the root corollary shows the full solution set of $z^n=w$ is exactly $n$ complex numbers, equally spaced by angle $2\pi/n$ around a circle of radius $\rho^{1/n}$ . Real $n$ th roots are the special case where you additionally demand the root lie back on the real axis, which happens for at most two of the $n$ complex roots (one if $n$ is odd, two if $n$ is even and $w>0$ ).