Introduction to Complex Numbers in Physics/Engineering

Reference: Mary L. Boas, Mathematical Methods in the Physical Sciences

Chapter 2 & 14
George Arfken, Mathematical Methods for Physicists
Chapter 6

The real numbers (denoted R) are incomplete (not closed) in the sense that standard op-
erations
√ applied to some real numbers do not yield a real number result (e.g., square root:
−1). It is surprisingly easy to enlarge the set of real numbers producing√ a set of numbers
that is closed under standard operations: one simply needs to include −1 (and linear
combinations of it). This enlarged field of numbers,
√ called the complex numbers (denoted
C), consists of numbers of the form: z = a + b −1 where a√and b are real numbers. There
are lots of notations for theses numbers. In mathematics, −1 is called √ i (so z = a + bi),
whereas in electrical engineering i is frequently used for current, so −1 is called j (so
z = a + bj). In Mathematica complex numbers are constructed using I for i. Since com-
plex numbers require two real numbers to specify them they can also be represented as an
ordered pair: z = (a, b). In any case a is called the real part of z: a = Re(z) and b is
called the imaginary part of z: b = Im(z). Note that the imaginary part of any complex
number is real and the imaginary part of any real number is zero. Finally there is a polar
notation which reports the radius (a.k.a. absolute value or magnitude) and angle (a.k.a.
phase or argument) of the complex number in the form: r∠θ. The polar notation can
be converted to an algebraic expression because of a surprising relationship between the
exponential function and the trigonometric functions:

ejθ = cos θ + j sin θ (1)

Thus there is a simple formula for the complex number z1 in terms of its magnitude and
angle:
p
|z1 | ≡ a2 + b2 = r (2)
a = r cos θ = |z1 | cos θ (3)
b = r sin θ = |z1 | sin θ (4)
jθ
z1 = a + bj = |z1 |(cos θ + j sin θ) = |z1 |e (5)

For example, we have the following notations for the complex number 1 + i:
√ √
1 + i = 1 + j = 1 + I = (1, 1) = 2∠45◦ = 2ejπ/4 (6)

Note that Equation 1 can be used to express the usual trigonometric functions in terms of
complex exponentials:

ejθ + e−jθ ejθ − e−jθ

cos θ = Re(ejθ ) = sin θ = Im(ejθ ) = (7)
2 2j
Since complex numbers are closed under the standard operations, we can define things which
previously made no sense: log(−1), arccos(2), (−1)π , sin(i), . . . . The complex numbers are
large enough to define every function value you might want. Note that addition, subtraction,
 Im z1 Im z1

 √ a2+b2 = r
b
r
b

 θ
a




a Re Re

Figure 1: Complex numbers can be displayed on the complex plane. A complex number
z1 = a + bi may be displayed as an ordered pair: (a, b), with the “real axis” the usual x-
axis and the “imaginary axis” the usual y-axis. Complex numbers are also often displayed
as vectors pointing from the origin to (a, b). The angle θ can be found from the usual
trigonometric functions; |z1 | = r is the length of the vector.

multiplication,
√ and division of complex numbers proceeds as usual, just using the symbol
for −1 (let’s use j):
z1 = a + bj z2 = c + dj (8)

z1 + z2 = (a + bj) + (c + dj) = (a + c) + (b + d)j

z1 − z2 = (a + bj) − (c + dj) = (a − c) + (b − d)j
z1 × z2 = (a + bj) × (c + dj) = ac + adj + bcj + bdj 2 = (ac − bd) + (ad + bc)j
1 1 1 a − bj a − bj a −b
= = × = 2 2
= 2 2
+ 2 j
z1 a + bj a + bj a − bj a +b a +b a + b2
Note in calculating 1/z1 we made use of the complex number a − bj; a − bj is called the
complex conjugate of z1 and it is denoted by z1∗ or sometimes z1 . See that zz ∗ = |z|2 .
Note that, in terms of the ordered pair representation of C, complex number addition and
subtraction looks just like component-by-component vector addition:

(a, b) + (c, d) = (a + b, c + d) (9)

Im z1 Im z1 + z2

z2
z1

Re Re

z1*

Figure 2: The complex conjugate is obtained by reflecting the vector in the real axis.
Complex number addition works just like vector addition.
Thus there is a tendency to denote complex numbers as vectors rather than points in the
complex plane.

Superposition of Oscillation

While the closure property of the complex numbers is dear to the hearts of mathematicians,
the main use of complex numbers in science is to represent sinusoidally varying quantities
in a simple way—allowing them to be combined with relative ease. (Remember that the
superposition of sinusoidal quantities is itself sinusoidal, but with a new amplitude and
phase.) For example, in a series RC circuit the voltage across the resistor might be given
by VR (t) = A cos ωt whereas the voltage across the capacitor might be given by VC (t) =
B sin ωt, and the voltage across the combination (according to Kirchhoff) is the sum:

VR (t) + VC (t) = A cos ωt + B sin ωt where: A, B ∈ R

 
p
2 2
A B
= A +B √ cos ωt + √ sin ωt
A2 + B 2 A2 + B 2
p A
= A2 + B 2 (cos δ cos ωt + sin δ sin ωt) where: cos δ = √
A2 + B 2
p
= A2 + B 2 cos(ωt − δ)

Yuck! That’s a lot of work just to add two sinusoidal functions; we seek a simpler method
(which might not seem overly simple at first glance). Note that VR can be written as
Re(Aejωt ) and VC can be written as Re(−jBejωt ) so:

VR (t) + VC (t) = Re (A − jB)ejωt

(10)

Now using the polar form of the complex number A − jB:

p
A − jB = A2 + B 2 e−jφ where: tan φ = B/A (11)

we have:

VR (t) + VC (t) = Re (A − jB)ejωt

p 
= Re A2 + B 2 e−jφ ejωt
p  
= A2 + B 2 Re ej(ωt−φ)
p
= A2 + B 2 cos(ωt − φ)

If we have more sinusoidal waves to add up, the problem is not much more difficult. Consider
the case of adding four waves, which for ease of notation we assume
√ have unit amplitude
and constant phase difference (δ). [I’m also going to switch to −1 = i; you should get
used to both notations.]

f (t) = cos(ωt) + cos(ωt + δ) + cos(ωt + 2δ) + cos(ωt + 3δ) (12)

We can write this as the real part of a complex expression:

h  i
f (t) = Re 1 + eiδ + ei2δ + ei3δ eiωt (13)
 3

1 2 3 4

-1

-2

-3

Figure 3: Consider the problem of adding four sinusoidal functions with the same frequency
and amplitude, but with different offsets (see left). The result (right) is a sinusoidal function
with the same frequency — we seek to easily determine the resulting amplitude and offset.

We’ll use some tricks below to add these four complex numbers, but for now the main point
is that they add to some complex number which can be expressed in polar form:
 
1 + eiδ + ei2δ + ei3δ = A eiφ (14)
so h i
f (t) = Re A ei(ωt+φ) = A cos(ωt + φ) (15)
Thus adding sinusoidal waves is as simple as adding the corresponding (complex) ampli-
tudes.

Now for the trick that applies to this particular problem. Recall the formula for the sum of
the geometric series:
1 − rN
1 + r + r 2 + · · · + r N −1 = (16)
1−r
(The proof of this result is easy: just multiply both sides by (1 − r), and notice that the
rhs terms telescope down to 1 − r N .) Here:
r = eiδ (17)
So the sum is:
1 − ei4δ ei2δ ei2δ − e−i2δ sin(2δ)
A eiφ = iδ
= iδ/2 iδ/2
= ei3δ/2 (18)
1−e e e −e−iδ/2 sin(δ/2)

3δ
Im

2δ

Re

Figure 4: Finding the
sum of four cosines amounts to adding four complex amplitudes:
1 + eiδ + ei2δ + ei3δ
 15

10

1 2 3 4

Figure 5: A2 is plotted as a function of δ. Note that if δ = 0, 2π, 4π, . . ., the four waves
are in phase so the amplitude is 4 (so A2 = 16). The first zero of A2 occurs when the four
amplitude vectors form a (closed) square for δ = π/2.

and hence:
sin(2δ)
A= φ = 3δ/2 (19)
sin(δ/2)
In physics we are usually most interested in the value of A2 which is plotted in Figure 5 as
function of δ.

In a more general case we might need to add N cosines with possibly different real amplitudes
ak and offsets δk :

h(t) = a1 cos(ωt + δ1 ) + a2 cos(ωt + δ2 ) + · · · + aN cos(ωt + δN ) (20)

N
X
= ak cos(ωt + δk ) (21)
k=1
h  i
= Re a1 eiδ1 + a2 eiδ1 + · · · + aN eiδN eiωt (22)
" N ! #
X
= Re ak eiδk eiωt (23)
k=1

Once again all that is required is to express the sum of the complex amplitudes in polar
fashion:

N
!
X
ak eiδk = Aeiφ (24)
k=1

then
h(t) = A cos(ωt + φ) (25)

Differential Equations and eiωt

Complex exponentials provide a fast and easy solution for many differential equations.
Consider the damped harmonic oscillator:

Fnet = −kx − bv = ma (26)

or:
d2 x b dx k
0= + + x (27)
dt2 m dt m
Seeking a more compact notation, we redefine the constants in this expression:
b k
≡ 2β ≡ ω02 (28)
m m
So:
d2 x dx
2
+ 2β + ω02 x = 0 (29)
dt dt
If we guess a solution of the form: x = Aert , we find:
 2
r + 2β r + ω02 Aert = 0

(30)
Since ert is never zero, r must be a root of the quadratic equation in square brackets.
p
−2β ± 4β 2 − 4ω02
q
r= = β ± i ω02 − β 2 (31)
2
p
where we have assumed ω0 > β. Defining the free oscillation frequency ω ′ = ω02 − β 2 , we
have a solution: h i
′
x = Re A e−βt eiω t = |A| e−βt cos(ω ′ t + φ) (32)

In the driven, damped harmonic oscillator, we have a driving force: F0 cos ωt in addition to
the other forces:
Fnet = F0 cos ωt − kx − bv = ma (33)
Defining A0 = F0 /m yields the differential equation:
d2 x dx
+ 2β + ω02 x = A0 cos(ωt) (34)
dt2 dt
If we seek a solution that oscillates at the driving frequency: x = Re Aeiωt , our differential
 

equation becomes:  2
−ω + 2βiω + ω02 Aeiωt = A0 eiωt

(35)
So:
A0
A= (36)
(ω02 − ω 2 ) + 2βiω
From which we can extract the amplitude and phase of the oscillation, for example:
A0
|A| = p 2 (37)
(ω0 − ω 2 )2 + 4β 2 ω 2
One can show that the amplitude is largest for
q
ω = ω02 − 2β 2 (38)

Finally it is customary to describe driven oscillators in terms of a dimensionless quality

factor, Q,
ω0
Q= (39)
2β
If we then let x be the dimensionless frequency ratio ω/ω0 , we can write the oscillation
amplitude in a particularly simple form:
A0 /ω02
|A| = p (40)
(1 − x2 )2 + x2 /Q2

Notice that the case of small damping (small β) corresponds to large Q.

 20

15

10

0.6 0.8 1 1.2 1.4

1
Figure 6: Resonance: the amplitude factor: √ is plotted as a function of the
(1−x2 )2 +x2 /Q2
dimensionless frequency ratio: x = ω/ω0 for the case Q = 20. Clearly the largest amplitude
occurs when x ≈ 1, i.e., ω ≈ ω0

Homework

1. Prove that when you multiply complex numbers z1 and z2 , the magnitude of the result
is the product of the magnitudes of z1 and z2 , and the phase of the product is the
sum of the phases of z1 and z2 .

Im z1

r1

z2 r2 θ2
θ1

Re
z3 = z1z2
θ3 = θ1 + θ2
z3 r3 = r1r2

2. Express the following in the r∠θ format (I bet your calculator can do this automati-
cally):  
1 3+i 2i ∗
 1 
(a) (b) (c) 25e (d) (1/(1 + i)) (e)  
1+i 1 + 3i (1 + i) 
3. Find the following in (a, b) format (I bet your calculator can do this automatically):
3i − 7 √
(a) (b) (.64 + .77i)4 (c) 3 + 4i (d) 25e2i (e) ln(−1)
i+4
4. Three cosine functions with amplitudes and offsets: a1 = 1.32, δ1 = .253 rad; a2 =
3.21, δ2 = .532 rad; and a3 = 2.13, δ3 = .325 rad are to be added together. Find the
result.

5. Find the formula for the amplitude A that results from adding 5 unit-amplitude cosine
waves with constant phase difference:

g(t) = cos(ωt) + cos(ωt + δ) + cos(ωt + 2δ) + cos(ωt + 3δ) + cos(ωt + 4δ)

= A cos(ωt + φ)

Use your favorite graphing program to make a hardcopy plot of A2 vs. δ for δ ∈ (0, 4π).
6. Consider a driven RC circuit (see below left). According to Kirchhoff’s law the voltage
drop across the resistor (IR, for current I) plus the voltage drop across the capacitor
(Q/C, for charge Q) must equal the voltage applied by the a.c. generator (V0 cos(ωt))
(where the generator is operating at an angular frequency ω). Thus:

Q
+ R I = V0 cos(ωt) (41)
C
Since the the current flowing must accumulate as charge on the capacitor we have:
dQ
I= (42)
dt
Thus we have the differential equation:
Q dQ
+R = V0 cos(ωt) (43)
C dt
Using complex variable methods and guessing a solution of the form:

Q = Q0 eiωt (44)

Determine the amplitude and phase of Q0 so you can express your final answer in real
form:
Q = A cos(ωt + φ) (45)

I I
Vo cos(ωt )
Vo cos(ωt )

7. Consider a driven RL circuit (see above right). According to Kirchhoff’s law the
voltage drop across the resistor (IR, for current I) plus the voltage drop across the
inductor (L dI/dt) must equal the voltage applied by the a.c. generator (V0 cos(ωt))
(where the generator is operating at an angular frequency ω). Thus:

dI
L + R I = V0 cos(ωt) (46)
dt
Using complex variable methods and guessing a solution of the form:

I = I0 eiωt (47)

Determine the amplitude and phase of I0 so you can express your final answer in real
form:
I = A cos(ωt + φ) (48)