HIGHER ALGEBRA

60

2.12. Exponential function.

infinite series
When r is real the
1+r+5+t
denoted bv c. e
for all r and the sum isThis is called is a
converges
real variable z defined for all z. the funcion d,
exponential
of a real variable
z.
For a complex variable z = T+ iy., the exponential
junchun
is
written as exp 2. is defined by exp(* + iy) = e(cos yti sin y).
This definition agrees with the real exponential function whhen
purely real. z
When z is purely real, y = 0 and exp z =e (cos 0+i sin
exp r = e.
0), ie
When z is purely imaginary, z=( 0 and exp Z = (COs y +i
i.e., exp(iy) = cos y+i sin y.
Since e >0 for all real . e(cos y +i sin y) represents acomnl
number in polar form, e being the mÍdulus and y being an ampitv
of exp 2.
Since e 0for any real number T, exp Z 1s a
non-zero comple
mumber for any complex number z.
Let u + iv be a non-zero complex number and let its
polar represen.
tation be r(cos +isin 0). Since r is positive, log r is real and r can
be expressed as r =elog r.
Therefore u + iv eog (cos +i sin 0)
exp (log r+i0).
Thus when u + iv is a given non-zero complex number,
a complex number z = log r+i8 such that there exiSt
exp z =u
that the range of the exponential function of z is the entire+iv. This mean
complex plaut
excluding the origin.
Properties.
V. exp Z1.exp'22 =exp(z + z), where z1,Z are complex
numbers:
Proof. Let z = I1 +iy1, z2=
+ iy2
Then zË t z) = (1 t+ 2) +i(y+
exp zË =e (cOs UI +i y2)
sin V1),exp z) =ea(cos y2 +i sin y2).
"(cos U1 +i sin y1)ea(cos V2 +i sin y2)
e1t(eos(u+ y2) +i sin(y1 + a)l
exp|( + 2) +i(y1 + y2)]
exp(z1 +22).
 COMPLEX NUMBERS 67

vh. exp
exp 32
= exp(21 - 22).
exp Z1 js defned.
Proof. Since exp z is a non-zero complex number, exp o
Let z1 =T1 + iy1, Z2 =T2 + iy2.
Then ~1 - 72 = (Ti - T2) + i(yn -V2).
exp z1 = e (cos y1 +isin yi),exp z =ea(cos y2 +isin y2).
exp 31 = eT1-F2 COS V1+i sin y1
exp cOs y2+i sin v2
= ei-2 (cos(y1 -y2) 4i sin(y1 -y2))
exp[(*1 - T2) + i(y1 -y2)l
exp(z1 - z2).
Corollary. pE exp(-z).
This follows from the property since exp(0) = 1
x3. Ifn be an integer, (exp z) = exp(nz).
This follows from the property 1 and the relation (exp z) =
exp(-z).
Note. If e be real and n be an integer, it follows that (exp it)" = exp int,
i.e., (cos +isin O)= cos no +isin n&. This is De Moivre's theorem.
4. If n be an fraction say p/q, (exp z)" has g distinct values but exp(nz)
is unique. In this case, exp(nz) is one of the values of (exp z)".
Note. If 0 be real and n be a fraction, it follows that (exp ino) is one
of the values of (exp i0)", i.e., cos no + i sin no one of the values of
(cos 8 + i sin 0)".
5. If nbe an integer, exp(z+ 2nmi) =exp z.
This follows from the property 1 and the relation exp(2nmi) = 1.
This states that exponential function is periodic with period 2ri.
Note. A complex function f is said to be a periodic function on its
domain DcCif there exists a non-zero constant w such that for all
integers ., f(z + nw) = f(z) (*) hÍlds for all z e D. If no submultiple
of wsatisfies the relation (*), then/w is said to be the period of f.

Let w be a non-zero complex number. Then there always exists a

Complex numberz such that exp z=W.
Dy the property 5, exp z W’ exp(z + 2nmi) = w, where n is an
Integer.
Ihus for a non-zero complex number w there exist infinitely many
COmplex numbers z 8uch that exp 2 =W.
 HIGHER ALGEBRA

68

Worked Examples.
sum
real, find the
1. If a, g be
cos(a + 0) + cos(a +20)+ . to n terms;
() cos a + to n terms
sin a + sin(a+)+ sin(a + 2) +
(ii)
[Here the angles are in arithmetic progression with the frst term a
common
difference 3)
and the
cos(a + ) +*+cos(a + n - IB).
Let ch = cOs a +
sin(a+n- 1O).
Sn =sin a +sin(a+B) +*+
i(a +n-Ta
Then ch + is, = exp ia + exp i(a + ) +**+ exp
i3)n-1]
= (expia)[1 + (expi8) + (expi8)+**+ (exp
= (expia)[]

+isin a)[-cos nß-i sinng1

= (cos a 1-cos B-i sin
a)[l-cos n/3-i sin n8)(1-cos +i sin g)1
= (cos a +isin (1-cos )2+sin B

i2 sin cos ng12sin"+12sin cos

= (cos a +i sin a)[2 sin 4sin?

= (cos a+ isin a)(-2isin )(cos +isin )(2isin )

4sin?

sin (cos(a+ )+isin(a + ).

sin n sin
Therefore h =[cos(a
sin
+ )] and sn = sin
sin(a + 2

2. Find all complex numbers z such that exp z =-1.

Let z = t+ iy.
Then exp 2 =-1 implies e (cos y +i sin y) =-1.
Therefore e cos y=-1,et sin y =0.
We have e = 1 and cos y=-1,sin y = 0.
e 1 T= 0, integer.

Cn y -1 and sin y =0’y= (2n +1)T, where n is an

Therefore z = (2n+ 1)mi.
3. Find sll complex numbers z 8uch that exp(2z + 1) =t:
Let z =r+ iy. Then 2z +l (2c +1) +i2y.
exp(22 + 1) =iinplis ee (cos 2y + i sin 2y) =i.
 COMPLEX NUMBERS 69

Therefore e2z+ cos 2y = 0, e2r+ sin 2y = 1.

We have er+1 = l and cos 2y = 0,
e2r+1 =l’=-, sin 2y = 1.
cos 2y =0and sin 2y = 1 ’y= (4n +1) ,
where nis an integer.
Thereforez=- +(4n +1)i.
Solve exp z =1+ 3.
Let z =I+iy.
Then exp z=1+V3i implies e (cos y+isin y)
Therefore e cos y=1,e* sin y = =1+W3
V3.
We have e = 2 and cos y
=,sin y = .
These determine z = log 2 and y = 2nT
+5, where n is an integer.
Therefore z = log 2+ (2nm +)i.
2.13. Logarithmic function.
Let z be a non-zero complex number. Then
complex number wsuch that exp w=z. there always exists a
w is said to be a logarithm of z.
Again exp w = exp(w + 2nri), where n is an integer.
that if w is a logarithm of z, then w +2nTi This shows
is also a logarithm of z.
This means that "logarithm of z" is a
This is denoted by Log z= w+2nmi. many-valued function of z.
Of the many values of logarithm of z, a
particular one is called the
principal logarithm of z and is denoted by log z.
Since z is a non-zero complex number, z has a polar
Let z = r(cos +isin ), -T <8 <T (a polar representation.
form with amp z).
Let w =u+iv be a logarithm of z. Then
exp w=z. This gives
e(cos t +isin v) =r(cos +i sin 0).
Therefore e cos V=T Cos , e sin v =r sin 0.
We have e = r and cos v= Cos , sin v= sin .
Thse deternine u= log r and v = +2n7, where
n ís an integer.
Therelore w= log r+i(0+ 2r), -T <8ST
ie., Loy z = log r+i(0+2nr)
log | z |+i(arg z +2nm).
The principal logarithm of z, denoted by log z,
Therefore log z log r 4 i),-<0<
corresponds to n = 0.
log | | iarg z.
70 HIGHER ALGEBRA

Note 1. Log z has been expressed as Log z log r + i0

8 is the principal argument of z and n is an integer.
+2nTi, where
An equivalent expresion is Log z =log r+- ia + 2pri, where
argument of : andip is an integer, since a = 2mr +0 fofor some a is any
The difference in these two expressions is to be noted integer m
The frst expression gives the p.v. of Log z(i.e., log z) when
but the latter does not give log z when p = 0. n=0,
Therefore in order to find log z (the p.v. of Log z) we shall ner
stick to the polar representation of z with principal argument.
Note 2. From the definition of Log z it follows that
(i) exp(Log 2) =z for all z #0 and
(iü) one of the values of Log(exp ) is z, the other values he
z+2nmi, nbeing a non-zero integer.
The principal logarithmic function is the inverse function of th
exponential function and
(i) exp (log 2)=z for all z 0; and (ii) log (exp z)=z for all z ¬C
Worked Examples.
1. Find Log z and log z, where
(i) z=1, (i) z=-1, (ii) z =i, (iv) z=-i.
(i) 1= 1(cos 0+isin 0). (polar form with principal argument)
Log 1 = log 1 + i(0 + 2n), where n is an integer.
Therefore Log 1 = 2niand log 1 = 0. (corresponding to n =0)
(ü) -1=l(cos T+i sin T).
Log(-1) = log 1+ i(T + 2nr), where n is an integer.
Therefore Log (-1) = (2r + 1)i and log (-1) = #i. (correspondiy
to n= 0)
(iii) i= 1(cos + i sin ).
Logi= log 1+i(;+2nm). where nis an integer.
Therefore Log i= (4n + 1)%i and log i=i.
(iv) -í = l cos ti sin ).
n i ) og 1 +i( +2nT), where nis an integer.
Therefore Loyl-) - (4n - 1)3i and log(-i) = i:
Noe that Log(-í) cnn also be expressed as ( + 2nm)i.
 COMPLEX NUMBERS 71

+2. Express Log(z + iy), (æ, v) # (0,0), in the form A+ iB where A and
Bare real and find log( + iy).
Since z+iy is a non-zero complex number, it has a polar representa
tion. Let z +iy = r(cos 0 +i sin 0), -T <0<T.
Then r =mod(r + iy) and = arg(r + iy) (principal argument).
Let Log( + iy) = u + iv.
Then exp(u +iv) =+iy =r(cos 0 +i sin )
or, e"(cos v +i sin v) = r(cos +i sin ).
This gives e cos v= cos 6, e sin v=r sin .
We have =2e"=ru= log r=}log(r +y).
Since e" =T, we have cos U= COS , sin v = sin 6.
Therefore v = 0+ 2n, where n is an integer.
Hence Log(r +iy) =u+iv= log(r² +y') + i(8 +2nm).
Therefore A= log( +y) and B = + 2nm =Arg(z + iy).
The principal logarithm of c+iy corresponds to n =0 (since is the
principal argument).
Hence log(r +iy) = }log(r² + y') +i arg(r + iy).
3. Find Log zand log 2, where z=1+itan 0,5 <8<T.
Let z = r(cos +i sin ) Then cos = 1,r sin = tan .
We have r = sec and this gives r = - sec 8 since sec 8<0.
Therefore cos =-cos 8 and sin =-sin 0.
These determine =T+0.
As << 27. is not the principal argument of z.
arg z = -2r = 0 -7.
Hence Log z = log(-sec 0) + i(0- T +2nm), where n is an integer
snd log z= log(- sec 0) +i(0 -).
Properties.
f z,. be two distinct complex numbers such that z122 0, then

Lt z =rË(cos 0, +i sin 01), z2 = Ta(co8 O) +i sin 2).

Then z22=riralcom(, + 02) +i sin(e, + 0,)]
Log zË log r1 +i(01 + 2ru), where n is n integer ;
Log z2 log ry +i(02 + 2nn), where n la an integer ;
Log (z122) = log(r1ra) + i(0, 4+ 0, + 2pn), where p 0s an integer.
 HIGHER ALGEBRA
72

Log z1+Log z2 log ri + log r2 + i(81+ 2+ 2n +

=log(r1ra))+ i(01 +82 + 2qr), where 2mm)
q
=ntm,
Since p and qare arbitrary integers, Log z1+Log zz
Note 1. If z1 = Z2, Log Z1+ Log Z2 = 2log rË +i(2 8,+ 4nn).
=Log(z1z).
+ i(2 6 +
n is an integer; and Log z12 = 2log rË 2pr), where whe
integer. pis
The set of the general values of Log z1+ LOg z is a proper subseto
the set of the general values of Log Z1Z2.
Hence Log z1+ Log z2 Log z1z2 if zË = Z2.
Note 2. log zË + log z2 is not necessarily equal to log(z1za)
For example, let Z1 =i, z2 =-1. Then z1z =-i.
Jzl=1, z2=1, |zz2|=1, arg z1 = arg z2 = T, arg(z1z) s
log z1 = i, log Z2 = Ti and log(21z2) =-i.
Hence log 1 +log z = i log(z1z2).
2. If z1 and z2 be two distinct complex numbers such that z122 0, the
Log z1-Log z Log
Proof. z1z2 0 ’ z1 0,z20.
Let zË =ri(cos 81 +i sin 1),z=T2(cos 82 +i sin 82).
Then =(cos(0, - 02) +i sin(, - 02)].
Log zË = log r1+i(bË +2nm), where n is an integer ;
Log z log r2 + i(®2 + 2rna), where m is an integer :
Log = log() +i(01 - b2 +2pm), where p is an integer.
Log z1- Log z2 = log rË - log r2 +i(012+ 2nm -2rnr)
= log()+i(01 -2+2gn), where q =n- m.
Since p and q are arbitrary integers, Log z1-Log z2 =Lo8
Note 1. If = 2, Log zË- Log z = 0 and Log = Log 12
where n is an integer. Hence Log z1 Log z2 # Log:
Note 2. log z1 - log Z2 is not necessarily equal to log
For exanmple, let z1 =-1,z=-i. Then = -i.
|z=1, |z2 = 1, |=1, arg z1 = T, arg z22=-, arg()=
log zË =mi log z2 = -i and log() =-i
Hence log Z- log z, =ilog().
3. If z 0 and n.be a positivo integor, Log zn mLog }.
Proof. Iet z = r(oOs 0 +i sin 0).
Then 2" "(co8 m0+ i sin m ).
 COMPLEX NUM1BERS 73

Log 2 = log r + i(8 + 2nm), where n is an integer;

Log 2" = log rn + i(m8+ 2pr), where p is an integer.
m Log z m log r+i(m0 + 2nn)
= log rn +i(mb + 2p1 ), where p1 = mT.
Since p is arbitrary and p1 is a multiple of m, each value of m Log z
is a value of Log 2n but not conversely.
So the set of values of m Log z is a proper subset of the set of values
of Log 2"m, Therefore Log zn # mLog z.
For example, let z = i,m = 2.
2 Log z =2 Log i= (4n + 1)ri, where n is an integer.
Log 2 =Log(-1) = (2k + 1) mi, where k is an integer.
Each value of 2 Log i is a value of Log i but not conversely.
Hence Log i #2 Log i.
4. If z #0 and m be a positive integer, Log z/m= Log z.
Proof. Let z = r(cos +i sin ).
Then zl/m = r(cos 2kTt +i sin 2kT+O), where k = 0, 1,...,m -1.
Log z = log r +i(0+ 2nm), where pis an ipteger;
Log zl/m = log r +i(2T+ + 2pT), wherep is an integer
m

log r +i + 2(k+mp),r).
Since 0<k< n - 1 and pis an arbitrary integer, k+ mp is also an
arbitrary integer. Let k + mp = g.
Then Log z/m= log r +i( + 4), where q is an integer and
Logz = log r +i(+n), where n is an integer.
Therefore Log z1/n=Log z.
Worked Examples.
Verify that Log(-i)/2 = }Log(-i).
-i= cos(-)+i sin(-).
Log(-i) =(2n7- )1= (n7 -)i, where n is an integer.
Two values of (-1)1/2 are cos i+isin ,k= 0,1
ie, coe(-) +i sin(-), cos +i sin .
Now Log cos(-) + i sin()] = (2rnT - )i, and
Log (com +i sin ]= (2pr + )i= (2p+ 1)r - i where m,p
areintegers.

Cla-6
 ALGEBRA
74 HÊGHER,

where n is
The values of Log (-i)/2 can be exhibited as (nT)i,
an integer. Hence Log (-i)?2 =}Log(-i).
 COMPLEX NUMBERS 75

4. Express Log [Log (cos +isin 0)](0<0 < ) in the form A+iB,
where A and B are real.

Log (cos 0+isin ) = (0+ 2nm)i, where n is an integer.

Case I. Let n >0. Then + 2rm > 0.
mod [(®+ 2n)i] =+2nm and arg [(0 +2nn)i = .
Therefore Log Log(cos 8 + i sin 0)] =Log [(e + 2nm)i
= log(0 t+ 2n) + (2kT +), where k is an integer.
Case II. Let n <0. Then 0 +2nT < 0.
mod[( + 2nm)i] = -(8 + 2rnm) and arg ((8 + 2nn)i] = -5
Therefore Log [Log(cos + i sin 0)]=Log[(O + 2nm)i]
= log-(0 +2nm)] + (2kr )i, where k is.an integer.
2,14. Complex exponents.
If a be a non-zero complex number and z be any complex number, a?
is defined by
a² = exp(z Log a).
Since Log a is many-valued, a* is a many-valued function. The prin
cipal value of a² corresponds to the principal logarithm of a.
Let a = r(cos +isin b), -T <<T: and 2 =+ iy.
Then Log a = (r + iy)[log r+i(2nn + 0)), where n is an integer
= (r log r- y(2nn + 0)) + ia(2n7 + 0) ty logr].
Therefore a = e logr-y(2nn+o) cos{z(2nm + ) + y logr} +
isin{z(2rm + 0) +y logr}}, where n is an integer.
The principal value of a* corresponds to n =0.
The p.v. of a = e logr-yo cos(z® + y logr) + i sin(z8 +y logr)
= exp[( + iy) (logr + i0)
= exp(z log a).
Particular cases.
L. Let a be a positive real number and z be a complex number +y.
Then r = 4,0 = 0.
a =azthy =e loga-2ny eos(2nms+y loga) +i sin(2nrety log a)l
where n is an integer.
The p.v. of a* is e lognfcos(y loga) +i sin(y log a)}:
In particular, if a = e, then
76 HIGHER ALGEBRA

eztiy = e-2nmy |cos (2rnr + v) +i sin(2nn + v)), where n is an

integer. Therefore e+y has many values.
The p.v. of etiy = e (cos y+i sin y) = exp( + iy).
Thus e is a many-valued function and the principal value of e* is exp

2. Let a be a negative real number and z be a complex number z +iy.

Then a=rcos 8+i sin , where r =b= -a and 0 =T.
artiy = e log b-(2n+1)y [cos (2n+1)nz+y log b} +isin{(2n+1)mz+
y log b}l, where n is an integer.
The p.v. of atiy= elog b-my[cos(T+ ylogb) + i sin(rz +y log b)]
=et log b[cos(ylogb) + i sin(y log b)].e-TV (cos mr +i sin me).
Now the p.v. of (-1)+iy exp[(* + iy) log(-1)]
= exp[( + iy)ri)
e(cos TI +i sin mt).
of (-1)z+iy.
Hence the p.v. of a+iy=(the p.v. of Utiy).(the p.v.
z = 1, real number.
3. Let a be a positive real number and
Then r=a,8 =0,y = 0.
where n is an integer and th
a = e log a (cos(2nr) + isin(2nn)),
p.v. of a* = e
log a
having the same modulus eloga, In th
a hasinfinitely many values points on a circle whose centre
complex plane they are represented by
elog a
the origin and radius is
Subcase (i). Let z be an integer.
is the unique real nunb
Then cos(2n) +isin(2nmr) = l and a
e log a,
where p andg are integ
Subcase (ii). Let r be a rational number ,q
prime to each other (g > 1).
only q distinct values and t
Then cos(2) +i sin(n2) has
correspond to n =0,1,...,g-1.
Therefore a has only fnite number of values.
n =0:
The principal value of ai is eoga (corresponding to
Subcase (iii), Let s be an irrational real number.
Therefore a h
Then sin(2nm) #0 for all integers n #0.principal value are
number of values, all of wlhich excepting the
complex numbers. The principal value of a
is e log4,
 COMPLEX NUMBERS 77

4. Let abe a negative real number and z=, a real number.

Let a = -b, where b> 0. Then r = b,= T, y=0.
a =elog [cos(2n + 1)r +i sin(2n + 1)m, where n is an integer.
The p.v. of a is e log bcos mr +i sin r].
Now the p.v. of (-1)" = exple(ri))
(cos TE +i sin mr).
Hence the p.v. of a =(the p.v. of U).(the p.v. of (-1)).
5. Let a = Cos +i sin , where is real and z =+ iy.
Then a* = (cos +isin 0)+iy
exp[( + iy) Log (cos 8+isin 9)]
exp[(z + iy)(2nm +0)il
= exp[-v(2nn + 0) + (2nm + 0)i]
e v2nrte) [cos u(2nn+ 8) + i sin z(2nm + 0)]
e-y Arg [cos(z Arg a) + isin(z Arg a)].
Hence thep.v. of (cos +sin )tiy
=ey arg (cos(z arg a) +i sin(z arg a)]
5a. Let a = cos 8+i sin 8, where is real and z=T. a real number.
Then a = (cos @+i sin 0) = [cos z(2nm +) + i sin z(2nr + 0)],
where n is an integer
and the p.v. of (cos +i sin ) = cos ( arg a)+isin(z arg a), where
a = cos +i sin .

Subcase (i). Let z be an integer.

Then (cos +i sin ) = cos To +i sin z.. In this case (cos 8 +
isin ) has only one value.
Subcase (ii). Let z be a rational number , where p and q are integers
prine to each other (g > 1).
Then cos(2nn + 0) + i sin(2n7 + 0) has only qdistinct values.
So in this case (cos 0+i sin 0)² has only a finite number of distinct
valus and cos zh + i sin z (corresponding to n = 0) is one of these.
The principal value of (cos 0+i sin 0) is cos( arg a) +i sin(arg a),
Wee a O 0 +isin 0.
Subease (ii). Let z be an irrational real nunber, then (cos 8+i sin 0) =
o z(2 + 0) + ibin z(2nn +0), whero n is an integer and cos ad t
Isiu z (corrempodiny, to n 0) is one of these values.
The prinripal valu of (cos 0 4isin 0) IN cos arg a+is0n arg a,
where a= COs 0 +i sin ,
 HIGHER ALGEBRA

0 + i sin 0, where -7 <8<T,

if a = cos
(i) the
Thusprincipal value of (cos +i sin e) is cos 28 +i sin Bo,
prime to each other:
positive integers
p.q are + ii sin 0) is cos To +i
principal value
of (cos sin z,
(ü) the
irrationa.
z,
Examples.

r.y be real,
1. if1rtiy =exp[( +iy)Log 1]=exp[( +iy)(2nmi), where nis an integ
=e-2nrV(cos 2nmT + isin 2ne].

In particular, if y = 0, then is an integer.

1 = (cos 2nr +isin 2nTT), where n
(i) If z be an integer, (cos 2nmT + i sin 2nnz] = 1 for all n. Therein
1 =1.
(ii) If z be a rational number , where p, q are integers prime to ed
other (q > 1), (cos 2nm. +isin 2nr.] has q distinct values. Therefe
this case.
there are q distinct values of 1 in
(iii) If z be an irrational number, 1 = (cos 2rm +i sin 2nnzl, where a
is an integer.
Since z irrational, sin 2nTT #0for all integers n #0. For n=l2
sin 2nI = 0and cos 2nTT = 1. Therefore there are infinite number dp
values of 1 of which one real and all others are non-real. F

2. If z,y be real,
(-1)*+iy = expl( + iy)Log (-1)]
= exp[( + iy) (2n + 1)ri)], where n is an integer
=e-(2n+1)my|cos (2n+ 1)Tu + i sin (2n + 1)nr
In particular, if y = 0, then integer.
(-1) =[cos (2n +1)mz +isin (2n +1)mz, where n is an aCCordi
(i) If r be an integer, [cos (2n +1)mz +ii sin (2n +1)wr] =ti
orQd even
as r is even or odd. Therefore t is
e(-1)* = t1 according as
to ead
(ü) If z be a rational number , where P,q are integers
prime
ialir
otler lq> 1), fcos (2n +1)m. +i sin (2n 1)m.4+ has g distinct

Trelone there are qdistinct values of (-1)" in this case.

(2
(i) 1f be ns + i sin
l)x, where n irrational
1)r A
nunber, (-1)" = (cos (2r +
is an integer. n. Therko

Sinee zis irrational, sin(2n+ 1)na 0 for all integers

there
n o n - r e a l

ae inuite nunbet of are

valu8 of (-l)" all of which
 COMPLEX NUMBERS 79

Properties. p r i r c p value

1. If z1,22 and a are complex numbers where a #0.

but (the p.v. of a ).(the p.v. of aa) = the p.v. of aitza,

Proof. a* = exp(21 Log a)
= expz1 (log a + 2rni)], where n is an integer;
a» = exp22(log a + 2mni)], where m is an integer;
aita =exp[(z1 + 2) (log a + 2pmi)]
= exp[(zi + z2) log a + 2p(z1 + 22)mi], where p is an integer.
a'.ata = exp[zi (log a + 2nzi)].
exp(2(log a + 2mmi)]
exp[z1(log a + 2nai) +z2(log a +2mni)]
exp[(z1 + z2) log a +2(nz1 + mzz)mi).
When m, n,p are arbitrary integers, the set of complex numbers
p(21 + 22) is a subset of the set of complex numbers nz1 + mz2, but
not conversely.
Therefore a.a²? at,
But the p.v. of ai = exp(z1 log a), the p.v. of a²? = exp(z2 log a)
and the p.v. of a*itia = exp[(21 + 22) log aj.
Therefore (the p.v. of q ).(the p.v. of aa) = the p.v. of ait»,
2. If z, aand b are complex numbers and ab 0, (ab)* = a*.b², but the
p.v. of (ab)= # (the p.v. of a).(the p.v. of 6).
Proof. a = exp(z Log a), b² = exp(z Log b),
(ab)² = exp[z Log(ab)].
explz Log (ab)] exp(z Log a + z Log b)
= exp(z Log a). exp(z Log b)
Therefore (ab)² a.b*.
The p.v. of a² = exp(z log a), the p.v. of b = exp(z log b) and the
p.v. of (ab)* = exp(z log ab).
But log ab # log a+ log a, in general,
Therelore the p.v. of (ab)* (the p.v. of a²).(the p.v. of b).
. Ifa and z are complex numbers and a #0,
Log a = z Log a + 2nmi, where n is an integer.
Pruf. et = + iy, a = r(cos +i sin 0), where -T<OST.
Then z Loy a = (r+ iv)[log r+i(0 + 2m)), where m is an integer
rlog r - v(2nm +0) + ir(2mm + 0) +y log rli
and a a gr y2mn40)cos(r(2mnr + 0) +y log r) +i sin{r(2mr +
6) +y log r}).
80 HIGHER ALGEBRA

Hence Log a² = (z log r-v(2mn +0)] +i[z(2mn+0)+y log r+2nn],

where is an integer
=z Log a + 2nmi.

Worked Examples.
Find the general values of i'.
exp(i Log i)
expi(2nT + )], where n is an
integer
exp[-(4n +1)] = e-(an+1) }.
Note. The values of i are all real.

2. Find the principal value of (1 +

i)
1+i=2(cos +i sin ). log(1+i)=
=}log 2+i(2n7+), where n is an integer and
Log (1+i)
log 2 + i. exp[i log(1 + i)]
Hence the p.v. of (1+i)' = exp[-+(log 2)1]
=
= etlcos( log 2) + isin(3log 2)].
principal values of ( + iy)+10 and (
3. Show that the ratio of the
iy)-10, where , y, a, b are real, is
2(blog r t ab) + isin 2(b log rt ab), where r = r + iyl, 0 =
(i) cos
arg( + iy) ;
arg(r + iy) = 1.
(ii) e-2bmlcos(2b log r) + isin(26 log r)], where r = ||,
Then r = lz+ iy
Let a + iy = r(cos + isin ), -T < 8< T. + isin(-0)].
e= arg(r+iy). - iy = r(cos 8- isin 8) =T,r(cos(-9) = 1.
arg(r - iy) = -0 provided #T. If0 = then arg(r - iy)
Case 1. 8 = arg( +iy) T.
The p.v. of (z +iy)ati0 = exp[(a + ib) log(z+ iy)]
= exp[(a + ib) (log r + i0)]
= exp[(a log r b0) + i(blog r + a)].
The p.v. of (z - iy)o-ib = exp[(a - ib) log(z - iy))
= exp[(a - ib) (log r i0)]
= exp[(alog r - b0) -i(b log r + a)].
The ratio = exp[2i(b log r + a0)]
= cos 2(b1log r + a) +isin 2(blog r + a0).
 COMPLEX NUMBERS 81

Case 2. = arg(r + iy) = m.

The p.v.of (z+iy)a+i = exp[(alog r-bm) + i(b log r + an)].
The p.v. of (r - iy)-t = exp[(a - ib) (log r+ in)
= exp[(a log r + bn) - i(b log r- an)].
The ratio = exp[-2bT + 2ib log r]
=e-2b* [cos(2blog r) + isin(2b log r)]l.
4. Discuss the reality of z, where a,y are both irrational.
Case 1. T>0.

z= exp[y Log z] = expv(logz + 2rni)], where n is an integer

= eV logz(cos 2ny + sin 2ny].
For n = 0, sin 2ny =0 and cos 2nyr = l. sin 2nyT 0 for all n 0.
Therefore a' has infinite number of values of which only one is real
and all others are non-real.
Case 2. I <0. Let a = -t, where t>0.
=(-t)' = exply Log (-t)] = exply{logt + (2rn + 1)i}]
= eylogt(cos(27 + 1)yr +isin(2n + 1)yr].
Since y is irrational, sin(2n + 1)yr is non-zero for all integers .
Therefore z' has infinite number of values all of which are non-real.

.2.14.1. Definition of Logaz, where a and z are non-zero complex

numbers.
We define Logaz= w such that z is any value of aw
We have a" = exp (w Log a). Therefore z =exp (w Log a).
This gives Log z =w Log a
Log z
or, w = Tog a
Thus w is a doubly infinitely many-valued function of z.
For example,
(2n+1)Ti
) Logi(-1)= B where m and n are integers;
2nni
(ü) Loge(1)= iWhere n and n are integers;
(2n+1)ni
(i1) Log:(-1) = u Where m and n are integers.
 Exercises 2B
1. (i) If exp z is positive real. pruve that
Im:= 2n, where nis an integ
(ii) If exp z is negative real, prove that
Im : =(2n +1), here nsan
integr
(ii) If exp z is purely imaginay, pruve tat
Im:=(2n+1). where n is an intgr
v2. Find all values of z such that
(i) exp z =-2, (iü)exp 2= 41. (i) expi-1+1
)exp(22 + ) =3+41.
(-3. If aexp(i6) + bexp(3+) =c where a,b.c are all re
a+b= tc. or b(b -a) =
4. Prove that

(i) cos +cos +cos+coe+ +cos =

J5. Find (i) Log 4. log 4. (u) Log (-4).kg (-4).
(i) Log 4i, kog 4i; (v) Log (-4:), kg (-4:).
J6. Show that (i) log + kog(l +kg (1).
4(ü) log i+ log(-1 + )kg -1+.
(ii) Log(1+)31Log (1 +)but kg(l+)3ogl +
V7. Express in the fotm A+B hete A at B ate real
(i) log(sin 0+i cs ), 0<t
(üi) log(sin 0-i cos e). 0<0
(ii) log(1+icot e). <<
(iv) log(1l+ cos 24+isin 2e). b<
8. If a> 0.b>0 and z = h a w that
log z =i tan ifa >b
i(T +t a n )ifa<b
=ij ifa =b.
Hint. Let =r(s 8+i sin ).-
If a > b. e = <e<r. Theaar=lce #
tan-.Ifa<i,e=r+ta-
,9. If r be real, prove that ilog=-r+2tan -r. ifr 20
ifr <0.
=7+2tan- r .
 COMPLEX NUMBERS 83

10. If r be real, prove that ,1+iz

ilogi-it = -2tana.
11. Show that

(i) if zbe a non-zero complex number, Log =-Log z but log may not
be equal to - log z;
(ii) if aand z are complex numbers and u 0, a = 1/u' and also the
D.v. of a-= the p.v. of 1/a.
A2. Find the general values and the principal value of
ppuing
()2, () 24, (i) 2, kiv) (-1)a) im generl.d

() 3, (vi) (-1)", (vi) (1+ i), (vii) los(1+i),

.13. Show that
4n+1 where m, nare integers.
) Log:i=i
(2n+1)mi
(ii) Log2(-1) =g92;, where m, n are integers.
14. Show that

() the p.v. of (2)' is equal to the ratio of the principal values of and
(-)';
) the p.v. of( ) is not equal to the ratio of the principal values of (-i)
and i.

.15. Show that (i) yi(-1+) #ViV-1+i; (1) Vi(1 -i) = v/1-i.
Show that the ratio of the principal values of (1 +i)l-i and (1 -i)+* is
sin(log 2) +i cos(log 2).
17 If a, b, z are real and <a + ib| = 1, prove that (a + ib)i is purely real.
18. If z be a non-zero real number, prove that
z=e2rn" [cos (log ) +i sin(log )], when > 0
=e n1) (cos log(-z) +i sin log(-)), when z < 0.
19. If a, bare positive2=Og
real numbers
b+2mi
and a² = b, show that the general values
of z are given by log a+2nmi ,m, nbeing integers.
20. If i =ishow that z is real and the general values of z are given by
4m +l m.n being integers.
41+1

3. Find the yeneral values and the principal value of itiv where #, yare real.
Show that the principal vnlue is purely real or purely imaginary according
as z is an even or an odll integer.