Analytic Functions

3
Book: A First Course in Complex Analysis with Applications by Dennis G. Zill and
Patrick D. Shanahan.
Contents
• Derivative of Complex Function
• Analytic Functions
• Cauchy-Riemann Equations
• Harmonic Functions
• Applications
Cauchy-Riemann Equations
Criterion for Analyticity: Suppose the real functions 𝑢(𝑥, 𝑦) and 𝑣(𝑥, 𝑦) are continuous and
have continuous first-order partial derivatives in a domain 𝐷. If 𝑢 and 𝑣 satisfy the Cauchy-
Riemann equations (CR-equations):
𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑣
= and =− , (1)
𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥
at all points of 𝐷, then the complex function 𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦) is analytic in 𝐷.
Remark: The CR-equations are not only necessary condition for a function to be analytic but
are also sufficient condition.
Polar Coordinates: For complicated complex functions it is better to use Euler’s
formula which require that the CR-equations are obtained in (𝑟, 𝜃). The CR-equations
in polar coordinates are given by:
𝜕𝑢 1 𝜕𝑣 𝜕𝑣 1 𝜕𝑢
= and = − , (2)
𝜕𝑟 𝑟 𝜕𝜃 𝜕𝑟 𝑟 𝜕𝜃
 Harmonic Function
• A real-valued function 𝜑 𝑥, 𝑦 , that has continuous first and second-order partial
derivatives in a domain 𝐷 and satisfies Laplace’s equation 𝛻 2 𝜑 = 𝜑𝑥𝑥 + 𝜑𝑦𝑦 = 0, is
said to be harmonic function in 𝐷.
• If a function 𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦) is analytic in a domain 𝐷, then its real and
imaginary parts are necessarily harmonic in 𝐷, i.e., both 𝑢(𝑥, 𝑦) and 𝑣(𝑥, 𝑦) have
continuous second partial derivatives in 𝐷 and satisfy the Laplace’s equation
𝛻 2 𝑢 = 𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0,
and 𝛻 2 𝑣 = 𝑣𝑥𝑥 + 𝑣𝑦𝑦 = 0,
in 𝐷.
• Harmonic functions are encountered in the study of temperatures and potentials.
Example: The function 𝑓 𝑧 = 𝑧 2 = (𝑥 2 − 𝑦 2 ) + 𝑖(2𝑥𝑦) is analytic everywhere. The
functions 𝑢(𝑥, 𝑦) = 𝑥 2 − 𝑦 2 and 𝑣(𝑥, 𝑦) = 2𝑥𝑦 are necessarily harmonic in any domain
𝐷 of the complex plane.
Harmonic Conjugate Functions
• Now suppose 𝑢(𝑥, 𝑦) is a given real function that is known to be harmonic in
D.

• If it is possible to find another real harmonic function 𝑣(𝑥, 𝑦) so that 𝑢 and 𝑣

satisfy the Cauchy-Riemann equations throughout the domain 𝐷, then the
function 𝑣(𝑥, 𝑦) is called a harmonic conjugate of 𝑢(𝑥, 𝑦). (Of course, this has
absolutely nothing to do with the use of “conjugate” for 𝑧.)

• By combining the functions as 𝑢 𝑥, 𝑦 + 𝑖𝑣(𝑥, 𝑦) we obtain a function that is

analytic in 𝐷.
Example: Verify that the function 𝑢 𝑥, 𝑦 = 𝑥 3 − 3𝑥𝑦 2 − 5𝑦 is harmonic in the entire
complex plane. Moreover, determine the harmonic conjugate function of 𝑢.
Remark:
• If 𝑣(𝑥, 𝑦) is a harmonic conjugate of 𝑢(𝑥, 𝑦) in some domain 𝐷, then 𝑢(𝑥, 𝑦) is, in
general, not a harmonic conjugate of 𝑣(𝑥, 𝑦).
Example:
The function 𝑓 𝑧 = 𝑧 2 = (𝑥 2 − 𝑦 2 ) + 𝑖(2𝑥𝑦) is analytic everywhere. The functions
𝑢(𝑥, 𝑦) = 𝑥 2 − 𝑦 2 and 𝑣(𝑥, 𝑦) = 2𝑥𝑦 are harmonic in any domain 𝐷 of the complex
plane. In this case 𝑣(𝑥, 𝑦) is a harmonic conjugate of 𝑢(𝑥, 𝑦) throughout the plane but
𝑢(𝑥, 𝑦) cannot be a harmonic conjugate of 𝑣 𝑥, 𝑦 since the function:
𝑓 𝑧 = 2𝑥𝑦 + 𝑖(𝑥 2 − 𝑦 2 )
is not analytic anywhere. (Verify!!!!)
Examples:
1. Find the harmonic conjugates of the following:
• 𝑢 𝑥, 𝑦 = 𝑒 2𝑥 sin 2𝑦.
• 𝑢 𝑥, 𝑦 = 𝑥 2 − 𝑦 2 .

2. Find the value of 𝑘 for which 𝑢 𝑥, 𝑦 = 𝑒 𝑘𝑥 sin 4𝑦 satisfy Laplace equation.

3. For an analytic function 𝑓(𝑧) = 𝑢(𝑟, 𝜃) + 𝑖𝑣(𝑟, 𝜃) in the domain 𝐷 not containing
the origin, the Laplace’s equation in polar coordinates is written as:
𝑟 2 𝑢𝑟𝑟 + 𝑟𝑢𝑟 + 𝑢𝜃𝜃 = 0.
Using above equation verify that the function 𝑢 𝑟, 𝜃 = 𝑒 −𝜃 cos(ln 𝑟) is harmonic in
a domain 𝐷 not containing the origin. Moreover, determine the harmonic conjugate
of 𝑢(𝑟, 𝜃) and the corresponding analytic function.
 Applications:
• If the function 𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦) is analytic in a domain 𝐷, then the real and
imaginary parts of 𝑓 are harmonic, i.e., that is, both 𝑢 and 𝑣 have continuous second-partial
derivatives and satisfy Laplace’s equation in 𝐷:
𝑢𝑥𝑥 + 𝑢𝑦𝑦 = 0, and 𝑣𝑥𝑥 + 𝑣𝑦𝑦 = 0. (1)
Conversely, if we know that a function 𝑢(𝑥, 𝑦) is harmonic in 𝐷, we can find a unique (up to
an additive constant) harmonic conjugate 𝑣(𝑥, 𝑦) and construct a function 𝑓(𝑧) that is
analytic in 𝐷.
• In physical sciences and engineering, Laplace’s equation is often encountered as a
mathematical model of some time-independent phenomenon, and in that context the
problem is to solve the equation subject to certain physical side conditions called boundary
conditions.
• Because of the link displayed in (1), analytic functions are the source of an unlimited
number of solutions of Laplace’s equation, and we may be able to find one that fits the
problem at hand.
• This is just one reason why the theory of complex variables is so essential in the serious
study of applied mathematics.
 Orthogonal Families
• Suppose the function 𝑓(𝑧) = 𝑢(𝑥, 𝑦) + 𝑖𝑣(𝑥, 𝑦) is analytic in some domain 𝐷. Then the
real and imaginary parts of 𝑓 can be used to define two families of curves in 𝐷. The
equations
𝑢 𝑥, 𝑦 = 𝑐1 and 𝑣 𝑥, 𝑦 = 𝑐2 (1)
where 𝑐1 and 𝑐2 are arbitrary real constants, are called level curves of 𝑢(𝑥, 𝑦) and
𝑣 𝑥, 𝑦 , respectively.
• The level curves (1) are orthogonal families. Roughly, this means that each curve in one
family is orthogonal to each curve in the other family.
• More precisely, at a point of intersection 𝑧0 = 𝑥0 + 𝑖𝑦0 , where we shall assume that
𝑓 ′ (𝑧0 ) ≠ 0, the tangent line 𝐿1 to the level curve 𝑢(𝑥, 𝑦) = 𝑢0 and the tangent line 𝐿2 to
the level curve 𝑣(𝑥, 𝑦) = 𝑣0 are perpendicular.
• The numbers 𝑢0 and 𝑣0 are defined by evaluating 𝑢 and 𝑣 at 𝑧0 , i.e., 𝑐1 = 𝑢 𝑥0 , 𝑦0 = 𝑢0
and 𝑐2 = 𝑣 𝑥0 , 𝑦0 = 𝑣0 .
Orthogonal Families
The tangents 𝐿1 and 𝐿2 are perpendicular at 𝑧0 ,
means that the product of the two slopes is −1, i.e.,
𝑚1 𝑚2 = −1,
where:
𝑢𝑥 𝑣𝑥
𝑚1 = − and 𝑚2 = − .
𝑢𝑦 𝑣𝑦
Example:
For 𝑓(𝑧) = 𝑧 2 we identify 𝑢 𝑥, 𝑦 = 𝑥 2 − 𝑦 2 and 𝑣(𝑥, 𝑦) = 2𝑥𝑦. For this function, the
families of level curves 𝑥 2 − 𝑦 2 = 𝑐1 and 2𝑥𝑦 = 𝑐2 are two families of hyperbolas. Since 𝑓
is analytic for all 𝑧, these families are orthogonal. At a specific point, say, 𝑧0 = 2 + 𝑖 we
find 𝑐1 = 3 and 𝑐2 = 4 and two corresponding orthogonal curves are 𝑥 2 − 𝑦 2 = 3 and
𝑥𝑦 = 2. Figure shows 𝑥 2 − 𝑦 2 = 3 in color and 𝑥𝑦 = 2 in black; the curves are orthogonal
at 𝑧0 = 2 + 𝑖 (and at −2 − 𝑖, by symmetry of the curves).

Since,
2𝑥 𝑥 2𝑦 𝑦
𝑚1 = − = and 𝑚2 = − =− ,
(−2𝑦) 𝑦 2𝑥 𝑥

therefore, 𝑚1 𝑚2 = −1. This shows that the tangents 𝐿1

and 𝐿2 are perpendicular at 𝑧0 .