Skip to content

Elementary Complex Functions

In this lesson you’ll extend the exponential, logarithm, trigonometric, hyperbolic and power functions to C\mathbb{C}, see which familiar properties survive and which break, and learn to handle multi-valuedness with branches and cuts.

Define

ez=ex+iy=ex(cosy+isiny)e^z = e^{x+iy} = e^x(\cos y + i\sin y)

This is forced, not chosen: it is the only definition that is analytic and agrees with exe^x on the real axis, by the identity theorem.

Properties that survive:

  • ez1+z2=ez1ez2e^{z_1+z_2} = e^{z_1}e^{z_2}
  • ddzez=ez\frac{d}{dz}e^z = e^z
  • ez0e^z \ne 0 for all zz, since ez=ex>0|e^z| = e^x > 0
  • eze^z is entire

The property that breaks, and it is a big one:

ez+2πi=eze^{z+2\pi i} = e^z

The complex exponential is periodic, with period 2πi2\pi i. On R\mathbb{R} the exponential is injective; on C\mathbb{C} it is infinitely many-to-one. Every consequence in this lesson traces back to that.

Geometrically ez=eRez|e^z| = e^{\operatorname{Re}z} and argez=Imz\arg e^z = \operatorname{Im}z, so horizontal lines map to rays and vertical lines to circles. A horizontal strip of height 2π2\pi covers C{0}\mathbb{C}\setminus\{0\} exactly once.

We want logz\log z to invert eze^z. Since the exponential is not injective, there is no single answer:

logz=lnz+iargz=lnz+i(Argz+2πk),kZ\log z = \ln|z| + i\arg z = \ln|z| + i(\operatorname{Arg}z + 2\pi k), \quad k\in\mathbb{Z}

The complex logarithm is genuinely multi-valued, taking infinitely many values differing by 2πi2\pi i. This is not a defect to be patched but a real feature, and it is the reason the subject has branch cuts, Riemann surfaces, and - much later - the residue theorem’s factor of 2πi2\pi i.

The principal logarithm picks Argz(π,π]\operatorname{Arg}z\in(-\pi,\pi]:

Logz=lnz+iArgz\operatorname{Log} z = \ln|z| + i\operatorname{Arg}z

To make it continuous we must delete a ray from 0 to \infty. The conventional choice is the negative real axis, called the principal branch cut, and on C\mathbb{C} minus that cut, Log\operatorname{Log} is analytic with

ddzLogz=1z\frac{d}{dz}\operatorname{Log}z = \frac1z

Two points of care that cause most of the errors in this material:

  • log(z1z2)=logz1+logz2\log(z_1z_2) = \log z_1 + \log z_2 holds as sets of values, but can fail for principal values. For instance with z1=z2=1z_1=z_2=-1, Log(z1z2)=Log(1)=0\operatorname{Log}(z_1z_2)=\operatorname{Log}(1)=0, while Log(1)+Log(1)=2πi\operatorname{Log}(-1)+\operatorname{Log}(-1)=2\pi i.
  • log\log of a negative number now exists. Log(1)=iπ\operatorname{Log}(-1) = i\pi, which is the honest answer to a question you were told had none.

Rearranging Euler’s formula gives the definitions:

cosz=eiz+eiz2,sinz=eizeiz2i\cos z = \frac{e^{iz}+e^{-iz}}{2}, \qquad \sin z = \frac{e^{iz}-e^{-iz}}{2i}

Both are entire, both have the expected derivatives, and every algebraic identity - sin2z+cos2z=1\sin^2 z+\cos^2 z=1, the addition formulas, the double-angle formulas - carries over unchanged, because the identity theorem says an identity valid on R\mathbb{R} must persist.

The property that breaks is boundedness:

sin(iy)=isinhy,cos(iy)=coshy\sin(iy) = i\sinh y, \qquad \cos(iy) = \cosh y

both of which grow without bound. Complex sine and cosine are unbounded, and they must be, by Liouville’s theorem: a bounded entire function is constant. The familiar bound sinx1|\sin x|\le1 is a fact about the real axis only.

Zeros are unchanged, though: sinz=0\sin z = 0 exactly at z=nπz = n\pi, and cosz=0\cos z = 0 exactly at z=π2+nπz = \frac{\pi}{2}+n\pi, all real. Extending to C\mathbb{C} adds no new zeros.

The hyperbolic functions are the same objects rotated:

sinhz=ezez2,sin(iz)=isinhz,cos(iz)=coshz\sinh z = \frac{e^z-e^{-z}}{2}, \qquad \sin(iz) = i\sinh z, \qquad \cos(iz)=\cosh z

In C\mathbb{C} the distinction between trigonometric and hyperbolic functions essentially disappears, which retrospectively explains why their identities look so similar.

For general complex aa and z0z\ne0, define

za=ealogzz^a = e^{a\log z}

Because log\log is multi-valued, so is zaz^a in general. The number of distinct values depends on aa:

  • aa an integer: one value. Nothing new.
  • a=p/qa = p/q in lowest terms: exactly qq values, the qqth roots we already know.
  • aa irrational or complex: infinitely many values.

The famous example is

ii=eilogi=ei(iπ/2+2πik)=eπ/22πki^i = e^{i\log i} = e^{i(i\pi/2 + 2\pi ik)} = e^{-\pi/2 - 2\pi k}

Every value is real, and the principal one is eπ/20.2079e^{-\pi/2}\approx0.2079. That an imaginary number to an imaginary power is real is a good illustration of how much the exponential definition buys.

Also worth flagging: (za)b=zab(z^a)^b = z^{ab} can fail with principal branches, for the same reason as the logarithm identity. When in doubt, go back to ealogze^{a\log z} and track the branch.

These are logarithms in disguise. For instance

arcsinz=ilog(iz+1z2)\arcsin z = -i\log\left(iz+\sqrt{1-z^2}\right)

so they are multi-valued too, and need cuts. This is the cleanest explanation of why real inverse trig functions have such awkward restricted domains: those restrictions are branch choices.

Example 1: Compute e1+iπ/2e^{1+i\pi/2}.

Solution.

e1+iπ/2=e1(cosπ2+isinπ2)=e(0+i)=eie^{1+i\pi/2} = e^1\left(\cos\tfrac\pi2 + i\sin\tfrac\pi2\right) = e(0+i) = ei

So the answer is ei2.718iei \approx 2.718i. Note ez=e1=e|e^z| = e^1 = e and arg=π/2\arg = \pi/2, matching. ✓

Example 2: Find all solutions of ez=1e^z = 1.

Solution. We need ez=ex=1|e^z| = e^x = 1, so x=0x=0, and argez=y0(mod2π)\arg e^z = y \equiv 0 \pmod{2\pi}.

z=2πik,kZz = 2\pi i k, \quad k \in \mathbb{Z}

Infinitely many solutions, where the real equation ex=1e^x=1 had one. That is periodicity showing up in the simplest possible way.

Example 3: Compute all values of log(1)\log(-1) and the principal one.

Solution. 1=1|-1| = 1 so lnz=0\ln|z| = 0, and arg(1)=π+2πk\arg(-1) = \pi + 2\pi k.

log(1)=i(π+2πk),kZ\log(-1) = i(\pi+2\pi k), \quad k\in\mathbb{Z}

The principal value is Log(1)=iπ\operatorname{Log}(-1) = i\pi.

Check: eiπ=1e^{i\pi} = -1 ✓, and also e3iπ=1e^{3i\pi} = -1 ✓. Both are correct logarithms, which is exactly the point.

Example 4: Compute Log(1+i)\operatorname{Log}(1+i).

Solution. 1+i=2|1+i| = \sqrt2 and Arg(1+i)=π/4\operatorname{Arg}(1+i) = \pi/4, so

Log(1+i)=ln2+iπ4=ln22+iπ40.3466+0.7854i\operatorname{Log}(1+i) = \ln\sqrt2 + i\frac\pi4 = \frac{\ln2}{2}+\frac{i\pi}{4} \approx 0.3466+0.7854i

Example 5: Solve sinz=2\sin z = 2.

Solution. No real solution exists, but a complex one does. Write w=eizw = e^{iz}; then sinz=ww12i=2\sin z = \frac{w - w^{-1}}{2i} = 2 becomes

w24iw1=0w=2i±i3=i(2±3)w^2 - 4iw - 1 = 0 \quad\Longrightarrow\quad w = 2i \pm i\sqrt3 = i(2\pm\sqrt3)

Then iz=log(i(2±3))iz = \log\big(i(2\pm\sqrt3)\big), giving

z=π2+2πkiln(2±3)z = \frac\pi2 + 2\pi k - i\ln(2\pm\sqrt3)

Taking k=0k=0 and the plus sign, zπ21.317iz \approx \frac\pi2 - 1.317i.

Check: sin(π21.317i)=cos(1.317i)1=cosh(1.317)2.0\sin(\frac\pi2 - 1.317i) = \cos(1.317i)\cdot 1 = \cosh(1.317) \approx 2.0

Complex sine takes every complex value, which is Liouville’s theorem made concrete.

Example 6: Find all values of iii^i.

Solution. logi=i(π2+2πk)\log i = i\left(\frac\pi2+2\pi k\right), so

ii=eilogi=eii(π/2+2πk)=eπ/22πki^i = e^{i\log i} = e^{i\cdot i(\pi/2+2\pi k)} = e^{-\pi/2-2\pi k}

Every value is a positive real number, and the principal one is eπ/20.2079e^{-\pi/2}\approx0.2079. ∎

Example 7: Show log(z1z2)=logz1+logz2\log(z_1z_2) = \log z_1+\log z_2 can fail for principal values.

Solution. Take z1=z2=1z_1 = z_2 = -1. Then z1z2=1z_1z_2 = 1 and

Log(1)=0,Log(1)+Log(1)=iπ+iπ=2πi\operatorname{Log}(1) = 0, \qquad \operatorname{Log}(-1)+\operatorname{Log}(-1) = i\pi+i\pi = 2\pi i

These differ by 2πi2\pi i. ∎

The rule holds as an equality of value sets, and 2πi2\pi i is a member of the set log1=2πik\log 1 = 2\pi i k. The identity is true; it is the principal-value bookkeeping that fails, and that is the standard trap in this material.

AC steady state. The complex exponential eiωte^{i\omega t} replaces sines and cosines throughout electrical engineering, turning differential equations into algebra. Every phasor calculation is eze^z used deliberately.

Damped oscillation. A solution e(α+iω)te^{(-\alpha+i\omega)t} combines decay and oscillation in one expression, with Re\operatorname{Re} giving the decay rate and Im\operatorname{Im} the frequency. Reading behaviour off the position of an exponent in C\mathbb{C} is standard in vibration and control analysis.

Quantum phases. A wavefunction’s time evolution is multiplication by eiEt/e^{-iEt/\hbar}, a pure phase of modulus 1. That the modulus is exactly 1 is conservation of probability.

Branch cuts in physics. In scattering theory and in Feynman propagators the choice of branch cut for a square root or logarithm encodes physical boundary conditions, and choosing wrongly gives an unphysical answer.

Digital signal processing. The zz-transform uses complex powers znz^{-n}, and the unit circle in the zz-plane is the frequency axis. Evaluating a transfer function on that circle is what a frequency response is.

Logarithmic spirals. The curve z(t)=e(a+ib)tz(t)=e^{(a+ib)t} is a logarithmic spiral, the shape of nautilus shells, spiral galaxy arms, and hurricane bands. It is the orbit of a point under repeated multiplication by a fixed complex number.

What property does e^z have in ℂ that e^x does not have in ℝ?
What are all the solutions of e^z = 1?
What is the principal value of Log(−1)?
Why does the principal logarithm need a branch cut?
Which statement about sin z on ℂ is true?
How many values does z^(1/3) have for z ≠ 0?
What is the principal value of i^i?
Why can Log(z₁z₂) = Log z₁ + Log z₂ fail?