This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: CHAPTER 2 FUNCTIONS OF A COMPLEX VARIABLE 2.1 COMPLEX NUMBERS In the course of study of roots of algebraic equations and in particular the cubic
equation, it has been found convenient to introduce the concept of a number
whose square is equal to —1. By a wellestablished tradition, this number is
denoted by i, and we write i2 = —— l and i = V — 1. If we allow ito be multiplied
by real numbers, we obtain the socalled imaginary numbers* of the form bi (where
b is real). If the usual rules of multiplication are extended to imaginary numbers,
then we must conclude that the products of imaginary numbers are real numbers;
moreover, their squares are negative real numbers. For instance, (3i)(4i) = (3)(4)i2 = (—12)(1) =
(51‘)2 = (—5)2i2 = —25. If imaginary numbers are adjoined to real numbers, we have a system within
which we can perform multiplication and division (except by zero, of course).
We say that such a system is closed under multiplication and division. However,
our system is not closed under addition and subtraction.1‘, To eliminate this de
ﬁciency, socalled complex numbers are introduced. These are numbers which are
most often written in the form a + bi (a, b = real numbers) and are assumed to obey appropriate algebraic rules. As will be shown below,
the system of complex numbers is closed under addition, subtraction, multiplica
tion, and division plus the "extraction of roots” operation. In short, it has all the
desirable algebraic characteristics and represents an extension of the real number
system. The study of complex numbers is invaluable for every physicist because
the description of physical laws is much more complicated without them. * Imaginary numbers are also called pure imaginary numbers to stress the distinction from
the more general case of complex numbers. The name originated from the belief that
imaginary numbers, as well as complex numbers, do not represent directly observable
quantities in nature. While this point of view is now mostly abandoned, the original
nomenclature still exists. T The system is not closed under the operation of extraction of the square root either;
for example, x/i is neither real nor (pure) imaginary 44 2.2 BASIC ALGEBRA AND GEOMETRY or COMPLEX NUMBERS 45 2.2 BASIC ALGEBRA AND GEOMETRY OF COMPLEX NUMBERS Ifcomplex numbers are written in the usual form a + ib (or a + bi) then the usual
algebraic operations with them are deﬁned as follows. 1. Addition; (01+ib1)+(a2 +1172): (41+ 02)+i(b1+ b2)
2. Multiplication: (01 + ibi) ' ((12 +1132): (aiaz  [91132) + [(ale + 02171) The second rule is easy to follow if we recognize that the expressions a + ib are
multiplied in the same manner as binomials, using the distributive and associative
laws, and i2 is replaced by — 1. Complex numbers of the form a + [O are tacitly identiﬁed with real numbers
since they obey the same algebraic rules and are generally indistinguishable from
each other.* Complex numbers of the form 0 + ib are then (pure) imaginary
numbers. It is customary to write simply a + i0 = a and 0 + ib = ib. Sub
traction of complex numbers can be deﬁned as inverse addition so that if (‘11 + [131) — (02 + ibz) = x + i)”
then a1+ib1= (x + iy) + (az + ibz)
from which it follows thatT
x = a1 —— a2 and y =tb1— b2.
An alternative is to form the negative of a complex number,
—(a + ib) = (—l)(a + ib) = (—l + iO)(a + ib) = —a — ib, 'and reduce the subtraction to addition. The rule for division can be similarly deduced by inverting the multiplication.
A shortcut method is given by the following technique: a+ib a+ibc—id_(ac+bd)+i(bc—ad) c+id=c+idc—id_ c2+d2
_ac+bd .bC—ad 2 2
—__c2+d2+lc———_2+d2 (c +d #0). It is readily seen that the divisor can be any complex number except zero (namely
the number 0 + [0, which is unique and is written simply 0). * In a more rigorous language, “the subset of complex numbers of the form a + [0 is
isomorphic to the set of real numbers under the correspondence 0 + i0 <—> a.” T It is tacitly postulated that x1 + iyl = x2 + iy2 if and only if x1 = x2 and y1 = y2. 46 FUNCTIONS OF A COMPLEX VARIABLE 2.2 Remarks 1. The addition of complex numbers obeys the same rule as the addition of vectors in
plane, provided a and b are identiﬁed with components of a vector. Note, however, that
the multiplication of complex numbers differs from the formation of dot and cross
products of vectors. 2. The use of the symbol i and the related binomial a + ib is conventional, but not
indispensable. It is possible to deﬁne a complex number as a pair of real numbers,
(a, b), obeying certain peculiar rules, e.g., the multiplication can be deﬁned by (a19b1)(a2) b2) = (ala2 — blb23alb2 + 02b1)a and so on. It should be clear that the form a + ib is just a representation of a complex
number. It is customary to represent complex numbers by points in the socalled com
plex plane, or Argand diagram (Fig. 2.1). If we denote the complex number
x + iy by a single symbol 2 and write 2 = x + iy, then to each 2 there cor
responds a point in the complex plane with the abscissa x and the ordinate y.
This idea also leads us to the trigonometric representation of a complex number: 2 = r(cosO+ isin6), where r = x/x2 + y2 and tan 0 = y/x. In this representation r is unique (positive square
root) but 6 is not. A common convention is to
demand that'l' lm Imaginary axis x Re ‘T < 9 S 7", Real axis along with the standard rule of quadrants, ' Figure 21
namely, 0 < O ify < 0.
The following nomenclature and notation will be widely used: If 2: x+iy= r(coso+isin6)
then x = Re 2 is the real part of z,
y = 1m 2 is the imaginary part of z, r = z is the modulus of 2, also known as the magnitude
or absolute value of z, ' 0 is the argument of 2, also called the polar angle or phase: The number x — iy is called the complex conjugate Of the number 2 = x + iy
and vice versa. We shall denote it by 2*. We can say that z and 2* represent (on
the complex plane) the reﬂections of each other with respect to the real axis. T Another commonly used convention is 0 S 0 < Zr.
1 A more precise name for 0 would be the “principal value of the argument of 2” (see p. 57). 2.2 BASIC ALGEBRA AND GEOMETRY or COMPLEX NUMBERS 47 Remarks 1. The quantity 22*
are the same). 2. The quantity 2 +
are the same). 3. The rules (21 + 22)* = 21‘ + 2'5 and (2122)“ =
remembered. is always a nonnegative real number equal to lzl 2 or to z"‘2 (which 2* is always a real number, equal to 2 Re 2 or to 2 Re 2* (which 2"sz are evident and should be Figure 2.2 Because complex numbers obey the same addition rule that applies to vectors in a plane, they can be added graphically by the parallelogram rule (Fig. 2.2a).
Conversely, vectors in a plane can be represented by complex numbers. The scalar product of two such vectors can be obtained by the rule
(2122) = Re (2122) = Re (2.2;). e vectors corresponding to complex num where it is understood that 21 and 22 ar
duct can be obtained in a similar bers 21 and 22 respectively. The vector pro fashion:
[21 x 221= 1m (2122) = —1m (2123‘). Exercise. Verify the validity of the above rules for scalar and vector products. In the theory of complex variables, the expression lzl  221 is often used. According to Fig. 2.2(b) this quantity (modulus of the complex number 21 — 22)
is equal to the distance between the points 21 and 22 in the complex plane. It
follows that the statement l2  201 < R (which often occurs in proofs of various
theorems) means geometrically that point z is within the circle of radius R drawn
around the point 20 as a center (i.e., z is in the Rneighborhood of 20; see p. 16). The following two inequalities are easily proved from geometrical considerations: 1 121 + Z2 5 [211+ 122~ (A side of a triangle is less than or equal to the sum of the other two sides.) 7 izi — Zle “21'" lzzli (The diﬁerence of two sides of a triangle is less than or equal to the third side.) 48 FUNCTIONS or A COMPLEX VARIABLE 2.3 Remark. It should be emphasized that inequalities can exist only among the moduli of
complex numbers, not among the complex numbers themselves. A complex number
cannot be greater or smaller than another complex number. Also, there are no positive or negative complex numbers. 2.3 DE MOIVRE FORMULA AND THE CALCULATION OF ROOTS While addition and subtraction of complex numbers are most easily performed
in their cartesian form 2 = x + iy, multiplication and division are easier in
trigonometric form. Ile = r1(cos 01 + isin 01) and 22 = r2(cos 92 + [sin 02),
then elementary calculation shows that 2122 = "1’2lC05 (91 + 02)+isin(01+ 92)] with the provision that if 01 + 02 happens to be greater than 1r, or less than or
equal to —1r, then the amount 21r should be added or subtracted to fulﬁll the condition —1r < (01 + 02) 5 1r. Remark. It should be emphasized that even though cos (0 :1: 21r) = cos 0 and
sin (0 5; Zn) = sin 6, the value of 0 is supposed to be uniquely speciﬁed. This will be
come evident when 0 is subjected to certain operations, e. g., in the course of evaluation
of roots. The convention 1r < 0 < 1r is not the only one possible, but some convention must be adopted and ours is just as good as any other Using the same trigonometric identities as in the above multiplication rule,
we can also obtain the socalled De Moivre formula: (cos 0 + isin 0)" = cos n6 + isin n6 (n = integer). Thus we now have the general rule for calculating the nth power ofa complex num
ber z. Ifz = r(cos 6 + [sin 0), then 2” = R(cos ¢ + i sin ¢), where R = r" and
¢ = n6 i 27rk with the integer k chosen in such a way that —1r < (1) 3 1r. The rule for calculating the nth root of a complex number can now be derived
without much diﬁ’iculty. If 2 = r(cos 6 + isin 9), then the complex number We = {7; (cos; + isin Q) n is deﬁnitely the nth root of 2 because
w}; = z. However,this is not the only nth
root of z; the numbers W], = {l/r<cos—~—— 0+ 27rk + i sin 0+ —2—7rk) n n where k =1, 2, 3, . . . , (n — l), are also
nth roots of 2 because w = 2. It is Figure 2.3 2.4 COMPLEX FUNCTIONS. EULER’S FORMULA 49 customary to call the number wo the principal root of z. The nth roots of a complex
number 2 are always located _at the vertices of a regular polygon of n sides inscribed
in a circle of radius R = {7 r about the origin (Fig. 2.3). Exercise. Verify that all possible roots of a complex number 2 are given by the above
formulas. Show that all complex numbers except one have exactly n (different) nthorder roots. Which complex number is the exception? 2.4 COMPLEX FUNCTIONS. EULER’S FORMULA Complex numbers 2 = x + iy may be considered as variables if x or y (or both)
vary. If this is so, then complex functions may be formed. For instance, consider
the equation w = 22. If we write 2 = x + iy and w = u + iv, it follows that u=x—y, v=2xy. From this we conclude that if w is a function of 2, then u and v are, in general,
functions of both x and y. Thus we are dealing with two (independent) real
functions of two (independent) real variables. lm .\ o:=l+i Invariant point Re
:=1 Figure 2.4 Figure 2.5 Graphical representation of complex functions poses a problem since we must
deal with four real variables simultaneously. The idea of mapping is most com
monly used. Two separate complex planes, the zplane and the wplane, are
considered side by side, and a point 20 is said to be mapped ont the point
We = f(zo). For instance, formula w = 22 maps 21 ’= i onto wl = i2 = ——1;
it also maps 22 = l + ionto W2 = 2i, 23 = 1 onto w3 = l, and so on. This is
illustrated in Fig. 2.4, where it is also indicated that the horizontal line y = l
in the zplane is mapped onto the parabola u = 2\/u + l in the wplane. Some
times it is convenient to superimpose the two planes. Then the images of various
points are located on the same plane and the function w = f(z) is said to trans
form the complex plane into itself (or a part of itself), as in Fig. 2.5, for the same function w = 22. 50 FUNCTIONS or A COMPLEX VARIABLE 24 Exercise. Show that the function w = iz represents counterclockwise rotation of the
complex plane by 90°. How would you describe a rotation by 180°? How would you
describe a clockwise rotation by 90° ? Algebraic functions of a complex variable are deﬁned by algebraic operations
which are directly applicable to complex numbers. Transcendental functions,
however, may require special deﬁnitions. Consider, for instance, the exponential function e‘ (real x). Its basic properties are
1. em”2 = exle”, 2. (e2)“ = e” It is desired to deﬁne a com lex ex onential function e2 with the same ro er
P P P P ties. Write z = x + I'y; then
e2 = ez+iy = e’ei”. The quantity e is a well deﬁned real number, but how shall we deﬁne e1”?
One possible method is as follows. Assume that e‘” can be represented by the usual power series 2 3
=1+(iy)+(—2——ly,) +(——'y,) + Then, rearranging the terms, we have 2 y4 3 5
ew‘<1‘_+l—!“”)+i< _§_! %_...)
=cosy+isiny. The validity of this procedure can be established after the development of the
theory of convergence for complex series. However, at this stage we may simply
deﬁne the function e‘” by means of e” = cosy + isin y.
This is Euler’s formula. The desired properties,
ei‘ylﬂ’z) = emei“, (eiy)" = emy , (n = integer),
follow from the identities (cosy1+isiny1)(cosy2 + isinyz) = 0050/1 + M) + isin (y; + M) d . . . .
an (cosy + Ism y)" = cos ny + zsm ny. The deﬁnition of a complex exponential function is then given by the formula
e2 = e‘(cosy + isin y) which has the desired properties and reduces to the real exponential function if
Im z = O. 2.5 APPLICATIONS OF EULER’S FORMULA 51 2.5 APPLICATIONS OF EULER’S FORMULA
Euler’s formula leads to the compact polar representation of complex numbers, z= x+iy= r(cosB+isin0)= ref”. Suppose that a complex number 2 is multiplied by ei", where a is a real con
stant. Then The new number can be obtained by rotating the point 2 about the origin by an angle a. This fact has many important applications.
Euler’s formula also permits the description of sinusoidally varying real quantities by means of complex exponentials. A general form of such quantity is
f(t) = acos (wt — 0), where a (amplitude), to (angular frequency), and 0 (phase) are constants, and t is
a real variable (usually time). Consider the complex function of the real' variable g(t) where B is a complex constant. Set B = aew; then II W m
l
2 g(t) = ae‘ee’iw’ = acos (0 — wt) + ia sin (0 — wt)
= acos (wt — 0) — ia sin (wt — 6). In other words, f (t) = Re {g(t)}.
Complex functions of a real variable can of real variables. For instance, if g0) = “(0 + 1'00) (11,0 be treated by the methods of calculus real functions), then '
[email protected] .__’
E’d:+’dz and so on. Differentiation of Be’i‘”l is very simple: ﬂ (Be—M) = —z'wBe dt —1"w t The use of complex exponentials is illustrated in the following example.
Consider a (damped) harmonic oscillator subject to a harmonically varying
external force. The differential equation to be solved reads 5: + 2ax + 02%): = Fcos (wt — ¢) (x = (dx/dl) etc.), Where the constants a, coo, F, to, and ¢ are real, and both variables x and t are real. 34 rUNCHUNs or A COMPLEX VARIABLE 2.6 2.6 MULTIVALUED FUNCTIONS AND RIEMANN SURFACES Certain complex functions are multivalued and they are usually considered as
consisting of branches, each branch being a singlevalued function of 2. For
instance, f (z) = \/E can be split into two branches according to the usual formula
for the roots (2 = re”): 1. Principal branch, f1(2) = v; em“),
2. Second branch, f2(z) = \ﬂ eil<9+21r)/21_ Strictly speaking, f1(z) and f 2(2) are two separate functions but they are intimately
connected and for this reason they are treated together as two branches of a
(doublevalued) function f (z) = x/E. Note that the principal branch does not map the zplane onto the entire
wplane, but rather onto the right halfplane (Re w > 0) to which the positive
imaginary semiaxis is added. The negative imaginary semiaxis is not included.
The second branch, which has no special name, maps the zplane onto the left
halfplane (Re w < 0) plus the negative imaginary semiaxis. Except for z = 0,
no other point on the wplane (image plane) is duplicated by both mappings. Also observe another important feature of the two branches. Each branch
taken separately is discontinuous on the negative real semiaxis. The meaning of
this is as follows: The points i(1r—5) 21 = e t(—1r+6), and 22 = e
where 6 is a small positive number, are very close to each other. However, their
images under the principal branch mapping, namely f1(21) = awn—M2) and f1(22) = e—“ﬂz—é/m, are very far from each other. On the other hand, note that the image of 22 under
the mapping/2(2), namely,
f2(22) ___ ei(1r/2+6/2)’ is very close to the pointf1(zl). It appears that the continuity of mapping can be
preserved if we switch branches as we cross the negative real semiaxis. To give this idea a more precise meaning we must deﬁne the concept of con
tinuous function of a complex variable. Let w = f(z) be deﬁned in some neigh
borhood (see pp. 47 and 16) of point 20 and let f(zo) = W5. We say that f(z)
is continuous at 20 if* f(z) —> wo whenever z —> 20 in the sense that given 6 > 0
(arbitrarily small), the inequality lf(z) — wol < 6 holds whenever [z — 20] < 6
holds, for sufﬁciently small 6. It is readily shown'l' that if w = u(x, y) + iv(x, y),
then the continuity of w implies the continuity of u(x, y) and v(x, y) and vice versa. * Also written as limb“0 f (z) = f(zo).
T For example, see Kaplan, p. 495. joir Two edges
joined here ; 2.6 MULTIVALUED FUNCTIONS AND RIEMANN SURFACES 55 Riemann proposed an ingenious device to represent both branches by means
of a single continuous mapping: Imagine two separate zplanes cut along the
negative real semiaxis from “minus inﬁnity” to zero. Imagine that the planes are
superimposed on each other but retain their separate identity in the manner of
two sheets of paper laid on top of each other. Now suppose that the second
quadrant of the upper sheet is joined along the cut to the fourth quadrant of the
lower sheet to form a continuous surface (Fig. 2.6). It is now possible to start
a curve C in the third quadrant of the upper sheet, go around the origin, and cross
the negative real semiaxis into the third quadrant of the lower sheet in a con
tinuous motion (remaining on the surface). The curve can be continued on the
lower sheet around the origin into the second quadrant of the lower sheet. Lower sheet Figure 2.6 Figure 2.7 Now imagine the second quadrant of the lower sheet joined to the third
quadrant of the upper sheet along the same cut (independently of the ﬁrst joint and
actually disregarding its existence). The curve C can then be continued onto the
upper sheet and may return to the starting point. This process of cutting and cross
joining two planes leads to the formation of a Riemann surface which is thought of
as a single continuous surface formed of two Riemann sheets (Fig. 2.7). An important remark is now in order: The line between the second quadrant
of the upper sheet and the third quadrant of the lower sheet is to be considered as
distinct from the line between the second quadrant of the lower sheet and the third
quadrant of the upper one. This is where the paper model fails us. According to
this model the negative real semiaxis appears as the line where all four edges of
our cuts meet. However, the Riemann surface has no such property; there are
two real negative semiaxes on the Riemann surface just as there are two real posi
tive semiaxes. The mapping [(2) = \/3 may help to visualize this: The principal
branch maps the upper Riemann sheet (negative real semiaxis excluded) onto
the region Re w > O of the wplane. The line joining the second upper with the
third lower quadrants is also mapped by the principal branch onto the positive
imaginary semiaxis. The lower Ri...
View
Full Document
 Fall '09
 brewer

Click to edit the document details