Introduction to the square root of a 2 by 2 matrix

Yue Kwok Choy

The square root of a 2 by 2 matrix A is another 2 by 2 matrix B such that  = !

, where !

stands for the matrix product of B with itself. We write ( = . In general, there can be no, two, four
'

or even an infinite number of square root matrices.

1 2 1 2 7 10
For example: - 1- 1=- 1
3 4 3 4 15 22

7 10 7 10 6 1 2
5

If we take =- 1 , then ( = -

'
1 =- 1
15 22 15 22 3 4
1 2 7 10
Note that =- 1 is one of the square roots of  = - 1 .
3 4 15 22
There may be another root(s) for .

Questions (See answers at the end of this file)

1 1 1 1 2 2 1 1
1. Evaluate : - 1- 1. Hence find two square roots of (a) - 1 (b) - 1 .
1 1 1 1 2 2 1 1

4 0 6 a 0 6
5 5

2. Find four values of - 1 . Hence find four values of - 1 , where a, b ≥ 0 .

0 9 0 b
a b
3. List all possible matrices  = - 1 , where a, b, c and d can be either 0 or 1, such that (
'

c d
does not exist.

(A) Solving equations method

1 2
Suppose we like to find the square root of  = - 1. So we write:
2 5
a b a b 1 2 a! + bc b(a + d) 1 2
- 1- 1=- 1⟹B ! D = -2 51
c d c d 2 4 c(a + d) cb + d
We get four equations:
a! + bc = 1 … (1)
b(a + d) = 2 … (2)
c(a + d) = 2 … (3)
cb + d! = 5 …(4)
(2) − (3), (b − c)(a + d) = 0
Since a + d ≠ 0 (otherwise (2) or (3) give absurdity), b − c = 0, b = c.
Hence we get: a! + b! = 1 … (5)
b(a + d) = 2 … (6)

1
 b! + d! = 5 … (7)

(6)! , b! (a + d)! = 4 , b! = (JKL)6 … (8)

I

(8) ↓ (5), a! + = 1, … (9)

I
(JKL)6

(7) − (5), d! − a! = 4 … (10)

d−a= … (11) (a + d ≠ 0)
I
JKL

(11) ↓ (9), a! + = 1, aP + a! d + d − a = a + d, aP + a! d − 2a = 0
LOJ
JKL

a(a! + ad − 2) = 0

∴ a = 0 or d = … (12)
!OJ6
J

(12) ↓ (10), a = 0, d = ±2 or a = ± ,d = ±
S P
√! √!

The rest can be solved easily, we therefore have:

± ±
S S
1 2 6 0 ±1
5

- 1 =- 1 , U √! PV
√!
2 5 ±1 ±2 ±
S
±
√! √!

0 1 6
5

3. Prove that - 1 does not exist.

0 0
a b
List, without prove, all possible matrices  = - 1 , where a, b, c and d can be either 0 or 1,
c d

such that ( has no real solution(s).

'

1 2 6
5

4. Use algebraic method to find - 1.

0 1
5. Check that the square root of the identity matrix is given by:

1 0 6
5
±d c
1 = Y±d
SOL6
- \ , YSOL6 \ ,
0 1 ∓d
Z
c ∓d Z

±1 0 ±1 0 ±1 c
where ± - 1,- 1,- 1 are limiting cases.
0 ∓1 c ∓1 0 ∓1
1 2
6. (a) Let  = - 1, find || and ( . Hence find ( .
'

2 4
a b || = 0.
(b)  = - 1, where Given that tr() = a + d > 0.
c d

Show that ( = tr() . Hence find ( .

'

2
(B) Diagonalization of Matrrix

Solving equation method in finding the square root of a matrix may not be easy. It involves
41 12 6
5

solving four non-linear equations with four unknowns. You may try this: - 1 , and soon
12 34
may give up.

We note that the square root of a diagonal matrix can be found easily:
a 0 6 a 0 − a 0 a 0 − a 0
5

- 1 = B√ D,B √ D , B√ D,B √ D.
0 b 0 √b 0 √b 0 −√b 0 −√b
If a matrix is NOT a diagonal matrix, we devise a method called diagonalization to help us.

We proceed with the finding of the eigenvalue(s) and eigenvector(s) of A .

A real number λ is said to be an eigenvalue of a matrix A if there exists a non-zero

column vector v such that Av = λv or (A − λI)v = 0 .

33 24 6
5

We like to find - 1
48 57
(1) Eigenvalues
33 24 x 33 − λ 24 x
A=- 1, v = -y1, (A − λI)v = - 1 -y1 = 0
48 57 48 57 − λ
Now, (A − λI)v = 0 has non-zero solution, |A − λI| = 0
33 − λ 24
f f=0
48 57 − λ
(33 − λ)(57 − λ) − 48 × 24 = 0
λ! − 90λ + 729 = 0
(λ − 81) (λ − 9) = 0
∴ λ = 9 or λ = 81 , and these are the eigenvalues.
eigenvalues
(2) Eigenvectors
We usually would like to find the eigenvector corresponding to each eigenvalue.
The process is called normalization.
For λS = 9,
33 24 xS xS 33xS + 24yS = 9xS i
- 1 -y 1 = 9 -y 1 h
48 57 S S 48xS + 57yS = 9yS
Choose for convenience 24xS + 24yS = 0 ⇔ xS + yS = 0
xS 1
∴ -y 1 = - 1 , which is a eigenvector.
eigenvector
S −1
For λ! = 81,
33 24 x! x! 33x! + 24y! = 81x! i
- 1 -y 1 = 81 -y 1 ⇔ h
48 57 ! ! 48xS + 57yS = 81yS
3
 Choose −48xS + 24yS = 0 ⇔2xS − yS = 0
x! 1
∴ -y 1 = - 1 , which is another eigenvector.
! 2

(3) Diagonalization of matrix

xS x! 1 1
We place two eigenvectors together. Let j = -y y! 1 = -−1 21
S

Consider another matrix:

1 1 OS 33 24 1 1
= j OS j = - 1 - 1- 1
−1 2 48 57 −1 2
S 2 −1 33 24 1 1
= - 1- 1- 1
P 1 1 48 57 −1 2
9 0
=- 1
0 81
At last we get a matrix B which is diagonal with eigenvalues as entries in the main
diagonal.
9 0 6 3 0 −3 0 3 0 −3 0
5
5
=- 1 =- 1,- 1,- 1,- 1
0 81 0 9 0 −9 0 −9 0 9
6

Note:
The diagonalization of a matrix may not be a simple subject since |A − λI| = 0
may have equal roots or even complex roots. Although most matrices are not
diagonal, they can be diagonalized. Not all square matrices can be diagonalised.A
thorough study of diagonalization of a matrix is not discussed here.

(4) Finding
Finding the square root of the original matrix A

Since = j OS j, we have  = j j O'

-j (j O' j)
'
j O' 1 -j
'
j O' 1 = j
' '
j O' = j
' '
( ( ( ( ( ( j O' = j j O' = 

!
So -j
'
( j O' 1 = 

1 1 3 0 1 1 OS 1 1 3 0 S 2 −1
(a) 6 = j
5 '
j O' = - 1- 1- 1 =- 1- 1l - 1m
−1 2 0 9 −1 2 −1 2 0 9 P 1 1
(

3 9 S 2 −1 1 3 2 −1 5 2
=- 1 - 1=- 1- 1=- 1
−3 18 P 1 1 −1 6 1 1 4 7
1 1 −3 0 1 1 OS 1 1 −3 0 S 2 −1
(b) A6 = PB 6 P OS = -
5 5
1- 1- 1 =- 1- 1l - 1m
−1 2 0 9 −1 2 −1 2 0 9 P 1 1
−3 9 S 2 −1 −1 3 2 −1 1 4
=- 1 - 1=- 1- 1=- 1
3 18 P 1 1 1 6 1 1 8 5

33 24 6 n ( −n −( ' o −' −o
5

∴- 1 =- 1,- 1,- 1,- 1

48 57 o p −o −p q n −q −n

4
Questions
7. Real matrix may have irrational square roots. Check by multipication:
multipication
1 2 6 1 √3 − 1D , B1 −√3 − 1D , B−1 √3 + 1D , B−1 −√3 + 1D
5

- 1 =B
0 3 0 √3 0 −√3 0 √3 0 −√3

8. A simple real matrix may have complex square roots. Check by multipication:
multipication
0 1 6 S 1+i 1−i S 1−i 1 + i S −1 − i −1 + i S −1 + i −1 − i
5

- 1 = !- 1, - 1, - 1, - 1
1 0 1−i 1+i ! 1+i 1 − i ! −1 + i −1 − i ! −1 − i −1 + i

9. A matrix can have both integral and fractional square roots. Use Diagonalization Method to show:
! I
0 4 6 2 −4
5

- 1 = ±- 1 , ± U PS sV
P
−1 5 1 −3 − P P

Answers

1 1 1 1 2 2
1. - 1- 1=- 1 , Hence:
1 1 1 1 2 2

2 2 6 1 1 −1 −1
5

(a) - 1 =- 1, - 1
2 2 1 1 −1 −1
1 1 1 1 2 2 1 1 1 1 S 1 1 S 1 1
(b) - 1- 1=- 1 = 2- 1⟹- 1 = l± - 1m l± - 1m
1 1 1 1 2 2 1 1 1 1 √! 1 1 √! 1 1

− −
S S S S
1 1 6 S 1 1
5

- 1 =± - 1 = U√! S V,
√!
U √! SV
√!
1 1 √! 1 1 S
−
S
−
√! √! √! √!

4 0 6 2 0 −2 0 2 0 −2 0
5

2. - 1 =- 1,- 1,- 1,- 1

0 9 0 3 0 3 0 −3 0 −3
a 0 6 a 0 − a 0 a 0 − a 0
5

- 1 = B√ D,B √ D , B√ D,B √ D
0 b 0 √b 0 √b 0 −√b 0 −√b

a b a b 0 1 a! + bc b(a + d) 0 1
3. - 1- 1=- 1⟹B ! D = -0 01
c d c d 0 0 c(a + d) cb + d
a! + bc = 0 = cb + d! ⟹ a = ±d
(a) Since b(a + d) = 1, a + d ≠ 0, and so a = d ≠ 0.
(b) Lastly, since c(a + d) = 0 ⟹ c = 0 and so a! + bc = a! = 0, contradicts with (a).
0 1 6 0 0 6
5 5

- 1 ,- 1 does not exist

0 0 1 0
0 1 6 1 1 6
5 5

- 1 - 1 has no real solution(s).

1 1 1 0
a b a b 1 2 a! + bc b(a + d) 1 2
4. - 1- 1=- 1⟹B D=- 1
c d c d 0 1 c(a + d) cb + d! 0 1
5
 We get four equations:
a! + bc = 1 … (1)
b(a + d) = 2 … (2)
c(a + d) = 0 … (3)
cb + d! = 1 …(4)
From (3), c = 0 … (5) (a + d = 0 gives contradiction in (2))
From (1), a! = 1, a = ±1.
From (3), d! = 1, d = ±1
1 2 6 1 1 −1 −1
5

Hence, - 1 =- 1,- 1
0 1 0 1 0 −1
1 2 || = 0 , ( = -1 2 1 2 5 10 1 2
6. (a)  = - 1, 1- 1=- 1 = 5- 1 = 5.
2 4 2 4 2 4 10 20 2 4
S !
!
1 2
Hence  = t  = - 1 .  = = -
'
1 = U√t IV .
S ( S S S √t
√t 2 4
(
√t √t !
√t √t

a b || = ad − bc = 0 ⟹ ad = bc … (1)
(b)  = - 1,
c d
a b a b a! + bc b(a + d) a! + ad b(a + d)
( = - 1- 1=B D = B D , by (1).
c d c d c(a + d) cb + d! c(a + d) ad + d!
a(a + d) b(a + d) a b
=B D = (a + d) - 1 = tr()
c(a + d) d(a + d) c d
!
∴= ( = B D , where tr() > 0.
' '
uv() wuv()

Hence, ( =
'
 .
'
wuv()

9. Eigenvalues : 4, 1
1 4
Eigenvectors: - 1 , - 1
1 1
xS x! 1 4 S −1 4 4 0
j = -y y! 1 = -1 1 , j O' = P - 1 , = - 1
S 1 1 −1 0 1

4 0 6 2 0 −2 0 2 0 −2 0
5
5
=- 1 =- 1,- 1,- 1,- 1
0 1 0 1 0 −1 0 −1 0 1
6

! I
1 4 2 0 1 4 OS 1 4 2 0 S −1 4
(a)  = j
5 '
j O'
=- 1- 1- 1 =- 1- 1l - 1m = U P S sV
P
1 1 0 1 1 1 1 1 0 1 P 1 −1
6 (
−
P P

1 4 −2 0 1 4 OS 1 4 −2 0 S −1 4 2 −4
(b) 6 = j
5 '
j O' = - 1- 1- 1 =- 1- 1 lP - 1m = - 1
1 1 0 1 1 1 1 1 0 1 1 −1 1 −3
(

! I
0 4 6 2 −4
5

∴- 1 = ±- 1 , ± U PS sV
P
−1 5 1 −3 −
P P
6