Let us start with the complex number
c = a + bi
where a and b are real (b = 0) and attempt to find an explicit representation for its square
root. Of course, every complex number (other than 0) will have two square roots. If w is
one square root, then the other one will be −w. We will find the one whose real part is
non-negative.
Let us assume that a square root of c is p + qi where p and q are real. Then we have
(p + qi)
2 = a + bi.
Equating the real and imaginary parts gives us the two equations
p2 − q2 = a (1)
2pq = b. (2)
We must have p = 0 since b = 0. Solving equation (2) for q gives
q = b
2p (3)
and we can substitute this value for q into equation (1) to get
p2 −
b
2p
2
= a
Reprinted from Mathematics and Informatics Quarterly, 3(1993)54–56
2
or
4p4 − 4ap2 − b2 = 0.
This is a quadratic in p2, so we can solve for p2 using the quadratic formula. We get
(taking just the positive solution):
p2 = a + √a2 + b2
2
so that
p = 1
√2
a + a2 + b2 .
From equation (3), we find
q = b
2p = b
√
2
2
√
a2 + b2 + a
= b
√
2
2
√
a2 + b2 + a
·
√a2 + b2 − a
√
a2 + b2 − a
= b
√2
√
a2 + b2 − a
(a2 + b2) − a2
= b
√2
√a2 + b2 − a
√
b2 = b
√2
√a2 + b2 − a
|b|
= sgn b
√2
a2 + b2 − a .
Note that √
b2 = |b|, so that b/|b| = sgn(b), the sign of b (defined to be +1 if b > 0 and -1
if b < 0).
Thus we have our answer:
Theorem 1. If a and b are real (b = 0), then
√
a + bi = p + qi
where p and q are real and are given by
p = 1
√2
a2 + b2 + a
and
q = sgn b
√2
a2 + b2 − a .
3
In practice, square roots of complex numbers are more easily found by first converting
to polar form and then using DeMoivre’s Theorem. Any complex number a + bi can be
written as
r(cos θ + isin θ)
where
r = a2 + b2, cos θ = a
r
, and sin θ = b
r (4)
DeMoivre’s Theorem states that if n is any positive real number, then
(a + bi)
n = rn(cos nθ + isin nθ).
In particular, if n = 1/2, we have
√
a + bi = √r
cos
θ
2 + isin θ
2
. (5)
This gives us a straightforward way to calculate √a + bi.
This method also gives us an alternate proof of Theorem 1. If we apply the half-angle
formulae
cos
θ
2 = ±
1 + cos θ
2
and
sin θ
2 = ±
1 − cos θ
2
to equation (5), we get
√
a + bi = √r
1 + cos θ
2
± i
1 − cos θ
2
where we have arbitrarily chosen the “+” sign for the first radical. Using the value for
cos θ from equation (4), we get
√
a + bi = √r
1 + a/r
2
± i
1 − a/r
2
=
r + a
2
± i
r − a
2
=
√
a2 + b2 + a
2
± i
√
a2 + b2 − a
2
which is equivalent to Theorem 1. As before, the “±” sign should be chosen to be the
same as the sign of b.
4
We sometimes need to find the square root of an expression of the form s+√−d where
s and d are real numbers and d > 0. We can use Theorem 1 to get an explicit formula for
this square root which is of the form p+qi where p and q are real. Since s+√−d = s+i
√
d,
we can let a = s and b = √
d in Theorem 1, to get the result:
Theorem 2. If s and d are real with d > 0, then
s + √
−d = 1
√2
s2 + d + s + i
1
√2
s2 + d − s .
Reference
[1] A. Mostowski and M. Stark, Introduction to Higher Algebra. Pergamon Press. New
York: 1964.