3 Analytic

Functions

2
3.1 Differentiability and Analyticity

1 3.2 Cauchy-Riemann Equations

3.3 Harmonic Functions
0
3.4 Applications
–1
Chapter 3 Review Quiz
–2

Introduction In the preceding chapter we intro-

–3 duced the notion of a complex function. Anal-
–3 –2 –1 0 1 2 3
ogous to the calculus of real functions we can
Level curves for f ( z ) = 1/ z . develop the notions of derivatives and integrals
See page 170. of complex functions based on the fundamental
concept of a limit. In this chapter our principal
focus will be on the definition and the properties
of the derivative of a complex function.

141
3.1 Differentiability and Analyticity
The calculus of3.1
complex functions deals with the usual concepts of derivatives and integrals
of these functions. In this section we shall give the limit definition of the derivative of a
complex function f (z). Although many of the concepts in this section will seem familiar to
you, such as the product, quotient, and chain rules of differentiation, there are important
differences between this material and the calculus of real functions f (x). As the subsequent
chapters of this text unfold, you will see that except for familiarity of names and definitions,
there is little similarity between the interpretations of quantities such as f
(x) and f
(z).

The Derivative Suppose z = x+iy and z0 = x0 +iy0 ; then the change

in z0 is the difference ∆z = z − z0 or ∆z = x − x0 + i(y − y0 ) = ∆x + i∆y. If
a complex function w = f (z) is defined at z and z0 , then the corresponding
change in the function is the difference ∆w = f (z0 + ∆z) − f (z0 ). The
derivative of the function f is defined in terms of a limit of the difference
quotient ∆w/∆z as ∆z → 0.

Definition 3.1 Derivative of Complex Function

Suppose the complex function f is defined in a neighborhood of a point

z0 . The derivative of f at z0 , denoted by f
(z0 ), is

f (z0 + ∆z) − f (z0 )

f
(z0 ) = lim (1)
∆z→0 ∆z
provided this limit exists.

If the limit in (1) exists, then the function f is said to be differentiable

at z0 . Two other symbols denoting the derivative of w = f (z) are w
and
dw/dz. If the latter notation is  used, then the value of a derivative at a
dw 
specified point z0 is written .
dz z=z0

EXAMPLE 1 Using Definition 3.1

Use Definition 3.1 to find the derivative of f (z) = z 2 − 5z.

Solution Because we are going to compute the derivative of f at any point,

we replace z0 in (1) by the symbol z. First,

f (z + ∆z) = (z + ∆z)2 − 5(z + ∆z) = z 2 + 2z∆z + (∆z)2 − 5z − 5∆z.

Second,

f (z + ∆z) − f (z) = z 2 + 2z∆z + (∆z)2 − 5z − 5∆z − (z 2 − 5z)

= 2z∆z + (∆z)2 − 5∆z.
Then, finally, (1) gives

2z∆z + (∆z)2 − 5∆z

f
(z) = lim
∆z→0 ∆z

∆z(2z + ∆z − 5)
= lim
∆z→0 ∆z

= lim (2z + ∆z − 5).

∆z→0

The limit is f
(z) = 2z − 5.

Rules of Differentiation The familiar rules of differentiation in the

calculus of real variables carry over to the calculus of complex variables. If
f and g are differentiable at a point z, and c is a complex constant, then (1)
can be used to show:

Differentiation Rules
d d
Constant Rules: c = 0 and cf (z) = cf
(z) (2)
dz dz
d
Sum Rule: [f (z) ± g(z)] = f
(z) ± g
(z) (3)
dz
d
Product Rule: [f (z)g(z)] = f (z)g
(z) + f
(z)g(z) (4)
dz
 
d f (z) g(z)f
(z) − f (z)g
(z)
Quotient Rule: = (5)
dz g(z) [g(z)]2

d
Chain Rule: f (g(z)) = f
(g(z)) g
(z). (6)
dz

The power rule for differentiation of powers of z is also valid:

d n
z = nz n−1 , n an integer. (7)
dz

Combining (7) with (6) gives the power rule for functions:

d n n−1

[g(z)] = n[g(z)] g (z), n an integer. (8)
dz
 EXAMPLE 2 Using the Rules of Differentiation
Differentiate:
z2  5
(a) f (z) = 3z 4 − 5z 3 + 2z (b) f (z) = (c) f (z) = iz 2 + 3z
4z + 1
Solution
(a) Using the power rule (7), the sum rule (3), along with (2), we obtain

f
(z) = 3 · 4z 3 − 5 · 3z 2 + 2 · 1 = 12z 3 − 15z 2 + 2.

(b) From the quotient rule (5),

(4z + 1) · 2z − z 2 · 4 4z 2 + 2z
f
(z) = = .
(4z + 1)2 (4z + 1)2

(c) In the power rule for functions (8) we identify n = 5, g(z) = iz 2 + 3z, and
g
(z) = 2iz + 3, so that

f
(z) = 5(iz 2 + 3z)4 (2iz + 3).

For a complex function f to be differentiable at a point z0 , we know from

f (z0 + ∆z) − f (z0 )
the preceding chapter that the limit lim must exist and
∆z→0 ∆z
equal the same complex number from any direction; that is, the limit must
exist regardless how ∆z approaches 0. This means that in complex analysis,
y
the requirement of differentiability of a function f (z) at a point z0 is a far
greater demand than in real calculus of functions f (x) where we can approach
a real number x0 on the number line from only two directions. If a complex
z
function is made up by specifying its real and imaginary parts u and v, such
∆z = ∆x
as f (z) = x + 4iy, there is a good chance that it is not differentiable.
x

(a) ∆z → 0 along a line parallel EXAMPLE 3 A Function That Is Nowhere Differentiable

to x-axis
Show that the function f (z) = x + 4iy is not differentiable at any point z.
y
Solution Let z be any point in the complex plane. With ∆z = ∆x + i∆y,
∆z = i∆y
z
f (z + ∆z) − f (z) = (x + ∆x) + 4i(y + ∆y) − x − 4iy = ∆x + 4i∆y

f (z + ∆z) − f (z) ∆x + 4i∆y

x and so lim = lim . (9)
∆z→0 ∆z ∆z→0 ∆x + i∆y

(b) ∆z → 0 along a line parallel

to y-axis
Now, as shown in Figure 3.1(a), if we let ∆z → 0 along a line parallel to
the x-axis, then ∆y = 0 and ∆z = ∆x and
Figure 3.1 Approaching z along a
f (z + ∆z) − f (z) ∆x
horizontal line and then along a lim = lim = 1. (10)
vertical line
∆z→0 ∆z ∆z→0 ∆x
 On the other hand, if we let ∆z → 0 along a line parallel to the y-axis as
shown in Figure 3.1(b), then ∆x = 0 and ∆z = i∆y so that

f (z + ∆z) − f (z) 4i∆y

lim = lim = 4. (11)
∆z→0 ∆z ∆z→0 i∆y

In view of the obvious fact that the values in (10) and (11) are different,
we conclude that f (z) = x + 4iy is nowhere differentiable; that is, f is not
differentiable at any point z.

The basic power rule (7) does not apply to powers of the conjugate of
z because, like the function in Example 3, the function f (z) = z̄ is nowhere
differentiable. See Problem 21 in Exercises 3.1.

Analytic Functions Even though the requirement of differentiability

is a stringent demand, there is a class of functions that is of great importance
whose members satisfy even more severe requirements. These functions are
called analytic functions.

Definition 3.2 Analyticity at a Point

A complex function w = f (z) is said to be analytic at a point z0 if f

is differentiable at z0 and at every point in some neighborhood of z0 .

A function f is analytic in a domain D if it is analytic at every point

in D. The phrase “analytic on a domain D” is also used. Although we shall
not use these terms in this text, a function f that is analytic throughout a
domain D is called holomorphic or regular.
Very Important ☞ You should reread Definition 3.2 carefully. Analyticity at a point is not
the same as differentiability at a point. Analyticity at a point is a neighbor-
hood property; in other words, analyticity is a property that is defined over
2
an open set. It is left as an exercise to show that the function f (z) = |z| is
differentiable at z = 0 but is not differentiable anywhere else. Even though
2
f (z) = |z| is differentiable at z = 0, it is not analytic at that point because
there exists no neighborhood of z = 0 throughout which f is differentiable;
2
hence the function f (z) = |z| is nowhere analytic. See Problem 19 in Exer-
cises 3.1.
In contrast, the simple polynomial f (z) = z 2 is differentiable at every
point z in the complex plane. Hence, f (z) = z 2 is analytic everywhere.

Entire Functions A function that is analytic at every point z in the

complex plane is said to be an entire function. In view of differentiation
rules (2), (3), (7), and (5), we can conclude that polynomial functions are
differentiable at every point z in the complex plane and rational functions
are analytic throughout any domain D that contains no points at which the
denominator is zero. The following theorem summarizes these results.
 Theorem 3.1 Polynomial and Rational Functions

(i ) A polynomial function p(z) = an z n + an−1 z n−1 + · · · + a1 z + a0 ,

where n is a nonnegative integer, is an entire function.
p(z)
(ii ) A rational function f (z) = , where p and q are polynomial
q(z)
functions, is analytic in any domain D that contains no point z0 for
which q(z0 ) = 0.

 
Singular Points Since the rational function f (z) = 4z/ z 2 − 2z + 2
1 ± i are zeros of the denominator ☞ is discontinuous at 1 + i and 1 − i, f fails to be analytic at these points. Thus
of f . by (ii ) of Theorem 3.1, f is not analytic in any domain containing one or both
of these points. In general, a point z at which a complex function w = f (z)
fails to be analytic is called a singular point of f . We will discuss singular
points in greater depth in Chapter 6.
If the functions f and g are analytic in a domain D, it can be proved that:

Analyticity of Sum, Product, and Quotient

The sum f (z) + g(z), difference f (z) − g(z), and product f (z)g(z) are
analytic. The quotient f (z)/g(z) is analytic provided g(z) = 0 in D.

An Alternative Definition of f
(z) Sometimes it is convenient

to define the derivative of a function f using an alternative form of the differ-
ence quotient ∆w/∆z. Since ∆z = z − z0 , then z = z0 + ∆z, and so (1) can
be written as
f (z) − f (z0 )
f
(z0 ) = lim . (12)
z→z0 z − z0
In contrast to what we did in Example 1, if we wish to compute f
at a
general point z using (12), then we replace z0 by the symbol z after the limit
is computed. See Problems 7–10 in Exercises 3.1.
As in real analysis, if a function f is differentiable at a point, the function
is necessarily continuous at the point. We use the form of the derivative given
in (12) to prove the last statement.

Theorem 3.2 Differentiability Implies Continuity

If f is differentiable at a point z0 in a domain D, then f is continuous

at z0 .

f (z) − f (z0 )
Proof The limits lim and lim (z − z0 ) exist and equal f
(z0 )
z→z0 z − z0 z→z0
and 0, respectively. Hence by Theorem 2.2(iii ) of Section 2.6, we can write
the following limit of a product as the product of the limits:

f (z) − f (z0 )
lim (f (z) − f (z0 )) = lim · (z − z0 )
z→z0 z→z0 z − z0
f (z) − f (z0 )
= lim · lim (z − z0 ) = f
(z0 ) · 0 = 0.
z→z0 z − z0 z→z0

From lim (f (z) − f (z0 )) = 0 we conclude that lim f (z) = f (z0 ). In view of
z→z0 z→z0
Definition 2.9, f is continuous at z0 . ✎
Of course the converse of Theorem 3.2 is not true; continuity of a func-
tion f at a point does not guarantee that f is differentiable at the point. It
follows from Theorem 2.3 that the simple function f (z) = x + 4iy is contin-
uous everywhere because the real and imaginary parts of f , u(x, y) = x and
v(x, y) = 4y are continuous at any point (x, y). Yet we saw in Example 3
that f (z) = x + 4iy is not differentiable at any point z.
As another consequence of differentiability, L’Hôpital’s rule for computing
limits of the indeterminate form 0/0, carries over to complex analysis.

Theorem 3.3 L’Hôpital’s Rule

Suppose f and g are functions that are analytic at a point z0 and

f (z0 ) = 0, g(z0 ) = 0, but g
(z0 ) = 0. Then

f (z) f
(z0 )
lim =
. (13)
z→z0 g(z) g (z0 )

The task of establishing (13) is neither long nor difficult. You are guided
through the steps of a proof in Problem 33 in Exercises 3.1.

EXAMPLE 4 Using L’Hôpital’s Rule

z 2 − 4z + 5
Compute lim .
z→2+i z 3 − z − 10i
Solution If we identify f (z) = z 2 − 4z + 5 and g(z) = z 3 − z − 10i, you
should verify that f (2 + i) = 0 and g(2 + i) = 0. The given limit has the
indeterminate form 0/0. Now since f and g are polynomial functions, both
functions are necessarily analytic at z0 = 2 + i. Using

f
(z) = 2z − 4, g
(z) = 3z 2 − 1, f
(2 + i) = 2i, g
(2 + i) = 8 + 12i,

we see that (13) gives

z 2 − 4z + 5 f
(2 + i) 2i 3 1
lim =

= = + i.
z→2+i z − z − 10i
3 g (2 + i) 8 + 12i 26 13
 In part (b) of Example 4 in Section 2.6 we resorted to the lengthy proce-
dure of factoring and cancellation to compute the limit
z 2 − 2z + 4
lim√ √ . (14)
z→1+ 3i z − 1 − 3i
A rereading of that example shows that the limit (14) √ has the indeterminate
form 0/0. With f (z) = z 2 − 2z + 4, g(z) = z − 1 − 3i, f
(z) = 2z − 2, and
g
(z) = 1, L’Hôpital’s rule (13) gives immediately
√  
z 2 − 2z + 4 f
(1 + 3i) √ √
lim√ √ = = 2 1 + 3i − 1 = 2 3i.
z→1+ 3i z − 1 − 3i 1

Remarks Comparison with Real Analysis

(i ) In real calculus the derivative of a function y = f (x) at a point

x has many interpretations. For example, f
(x) is the slope of the
tangent line to the graph of f at the point (x, f (x)). When the slope
is positive, negative, or zero, the function, in turn, is increasing,
decreasing, and possibly has a maximum or minimum. Also, f
(x)
is the instantaneous rate of change of f at x. In a physical setting,
this rate of change can be interpreted as velocity of a moving object.
None of these interpretations carry over to complex calculus. Thus
it is fair to ask: What does the derivative of a complex function
w = f (z) represent? Here is the answer: In complex analysis the
primary concern is not what a derivative of function is or represents,
but rather, it is whether a function f actually has a derivative. The
fact that a complex function f possesses a derivative tells us a lot
about the function. As we have just seen, when f is differentiable at
z and at every point in some neighborhood of z, then f is analytic at
the point z. You will see the importance of analytic functions in the
remaining chapters of this book. For example, the derivative plays
an important role in the theory of mappings by complex functions.
Roughly, under a mapping defined by an analytic function f , the
magnitude and sense of an angle between two curves that intersect
a point z0 in the z-plane is preserved in the w-plane at all points at
which f
(z) = 0. See Chapter 7.
2
(ii ) We pointed out in the foregoing discussion that f (z) = |z| was
differentiable only at the single point z = 0. In contrast, the real
2
function f (x) = |x| is differentiable everywhere. The real func-
tion f (x) = x is differentiable everywhere, but the complex function
f (z) = x = Re(z) is nowhere differentiable.
(iii ) The differentiation formulas (2)–(8) are important, but not nearly
as important as in real analysis. In complex analysis we deal with
functions such as f (z) = 4x2 − iy and g(z) = xy + i(x + y), which,
even if they possess derivatives, cannot be differentiated by formulas
(2)–(8).
 (iv ) In this section we have not mentioned the concept of higher-order
derivatives of complex functions. We will pursue this topic in depth
in Section 5.5. There is nothing surprising about the definitions of
higher derivatives; they are defined in exactly the same manner as in
real analysis. For example, the second derivative is the derivative of
the first derivative. In the case f (z) = 4z 3 we see that f
(z) = 12z 2
and so the second derivative is f

(z) = 24z. But there is a major
difference between real and complex variables concerning the exis-
tence of higher-order derivatives. In real analysis, if a function f
possesses, say, a first derivative, there is no guarantee that f pos-
sesses any other higher derivatives. For example, on the interval
(−1, 1), f (x) = x3/2 is differentiable at x = 0, but f
(x) = 32 x1/2 is
not differentiable at x = 0. In complex analysis, if a function f is
analytic in a domain D then, by assumption, f possesses a derivative
at each point in D. We will see in Section 5.5 that this fact alone
guarantees that f possesses higher-order derivatives at all points in
D. Indeed, an analytic function f is infinitely differentiable in D.
(v ) The definition of “analytic at a point a” in real analysis differs from
the usual definition of that concept in complex analysis (Definition
3.2). In real analysis, analyticity of a function is defined in terms of
power series: A function y = f (x) is analytic at a point a if f has a
Taylor series at a that represents f in some neighborhood of a. In
view of Remark (iv ), why are these two definitions not really that
different?
(vi ) As in real calculus, it may be necessary to apply L’Hôpital’s rule
several times in succession to calculate a limit. In other words, if
f (z0 ), g(z0 ), f
(z0 ), and g
(z0 ) are all zero, the limit lim f (z)/g (z)
z→z0
may still exist. In general, if f , g, and their first n − 1 derivatives
are zero at z0 and g (n) (z0 ) = 0, then

f (z) f (n) (z0 )

lim = (n) .
z→z0 g(z) g (z0 )

EXERCISES 3.1 Answers to selected odd-numbered problems begin on page ANS-12.

In Problems 1–6, use (1) of Definition 3.1 to find f
(z) for the given function.
1. f (z) = 9iz + 2 − 3i 2. f (z) = 15z 2 − 4z + 1 − 3i
1
3. f (z) = iz 3 − 7z 2 4. f (z) =
z
1
5. f (z) = z − 6. f (z) = −z −2
z
In Problems 7–10, use the alternative definition (12) to find f
(z) for the given
function.
7. f (z) = 5z 2 − 10z + 8 8. f (z) = z 3
1
9. f (z) = z 4 − z 2 10. f (z) =
2iz