COMPLEX MATRICES

If at least one element of a matrix is a complex number  + , where a, b are real andi = √−1,
then the matrix is called a complex matrix.
The matrix obtained by replacing he elements of a complex matrix A by the
corresponding conjugate complex number is called the conjugate of the matrix A and is denoted
by ̅.
2 + 3 −7 2 − 3
7
Thus, if = , then ̅ =  
5 1− 51+
 ) and the transposed
It is easy to see that the conjugate of the transpose of . . , (′
 ) are equal, Each of them is denoted by A*.
conjugate of A . . , (′
Thus, ( ) ( ) = ∗ .
A square matrix  = [ ]is said to be Hermitian if  ∗=  or  =  .
A square matrix  = [ ] is said to be skew. Hermitian if ∗ = − or  = −.
In a Hermitian matrix, the diagonal elements are all real, while every other element is the
conjugate complex of the element in the transposed position. For example.
5 2+ −3
 = 2 −  −3 1 −   is a Hermitian matrix.
3 1+ 0

In a skew-Hermitian matrix, the diagonal elements are zero or purely imaginary of the
form , where  is real. Every other element is the negative of the conjugate complex of the
element in the transposed position.
3 1+ 7
For example,  = −1 +  0 −2 −   is a skew-Hermitian matrix
−7 2− −
A square matrix A is said to be unitary if ∗ =  = ∗ 
The determinant of a unitary matrix is of unit modulus. For a matrix to be unitary, it must
be non-singular.
A square matrix A is said to be orthogonal if AA'=I= A'A or A' =A-1
Note. The following results hold:
(i) (̅) =  (ii)  +  − ̅ + 
 (iii) 
 = 

(iv)  
 =  (v) (A*)*=A (vi) (A+B)* = A*+B*
(vii) ()∗ = ̅∗ (viii) ()∗ =  ∗ ∗

51
 ILLUSTRATIVE EXAMPLES

2 +  3 −1 + 3
Example1.   =  , verify that A*A is a Hermitian matrix where
−5  4 − 2
A* is the conjugate transpose of A.
2+ −5

Sol.  = 3  
−1 + 3 4 − 2
2− −5
 = ( ) = 
∗
3 − 
−1 − 3 4 + 2
2− −5
2+ 3 −1 + 3
Now, ∗  =  3 −   
−5  4 − 2
−1 − 3 4 + 2
30 6 − 8 −19 + 17
=  6 + 8 10 −5 + 5  = ()
−19 − 17 −5 − 5 30
30 6 + 8 −19 − 17
∴   =  6 − 8 10 −5 − 5 
−19 + 17 −5 + 5 30
30 6 − 8 −19 + 17
Now,  ) =  6 + 8
 ∗ = ( 10 −5 + 5  = 
−19 − 17 −5 − 5 30
⇒  = ∗  is a Hermitian matrix.

Example 2. Show that the matrix

 +  − + 
 is unitary if   +   +   +   = 1.
 +   − 
 +  − + 
Sol. Let = 
 +   − 
 −  − − 
∴ ̅ =  
 −   + 
 −   − 
⇒ ∗ = (̅) =  
− −   + 
For a square matrix A to be unitary,
∗ =  = ∗  …(1)
  +  +   +   0
Now, ∗ =  
0  +  +  + 
 

52
   +  +   +   0
Also, ∗  =  
0  +  +  + 
 

Eqn. (1) is satisfied only when

  +  +   +   = 1
Example 3. If A and B are Hermitian, show that AB-BA is skew-Hermitian.
Sol. A and B are Hermitian ⇒ A* = A and B* = B
Now, ( − )*−()*
 ∗ ∗ − ∗  ∗ =  −  = −( − )
⇒ AB - BA is skew-Hermitian.
Example 4. If A is  skew-Hermitian matrix ⇒ ∗ = −
Now, ()∗ =  ∗ = (−)(−) = 
⇒ iA is a Hermitian matrix.
Example 5. Show that every square motrix is expressible as the sum of of a Hermitian
matrix and a skew-0 Hermitian matrix.
Sol. Let A be any square matrix.
Since ( + ∗ )∗ = ∗ + (∗ )∗ = ∗ +  =  + ∗
and ( − ∗ )∗ = ∗ − (∗ )∗ = ∗ −  = −( − ∗ )
∴  + ∗ is Hermitian and  − ∗ is skew-Hermitian.
 
Now,  = ( + ∗ ) +  ( − ∗ ) =  +  (say)


where P is Hermitian and Q is skew- Hermitian. Thus, every square matrix can be
expressed as the sum of a Hermitian matrix and a skew-Hermitian matrix.
0 1 + 2
Example 6. If  =  , obtain the matrix ( − )( + ), and shw
−1 + 2 0
that it is unitary
1 0 0 1 + 2 1 −1 − 2
Sol. 1− = − = 
0 1 −1 + 2 0 1 − 2 1
1 0 0 1 + 2 1 1 + 2
1+ = + = 
0 1 −1 + 2 0 −1 + 2 1
1 1 + 2
| + | =   = 1 − (4  − 1) = 6
−1 + 2 1

( + )

( + ) =
 1 −1 − 2
  
||  1 − 2 1

53
 1 −1 − 4  1 −1 − 2
∴ ( − )( + ) =   
1 − 2 1 1 − 2 1
 −4 −2 − 4
=   = ()
2 − 4 −4
 −4 2 + 4
 =   
−2 − 4 −1
  −4 2 + 4
( ) =  =  
 −2 + 4 −4
 −4 2 + 4  −4 −2 − 4
      
−2 + 4 −4  2 − 4 −4
 36 0 1 0
 = =
 0 36 0 1
⇒  = ( − )( + ) is unitary

TEST YOUR KNOWLEDGE

1. Show that the matrix A is Hermitian and iA is Skew-Hermitian where A is.

3 5 + 2 −3
2 3 − 4
(i)   (ii) 5 − 2 7 4 
3 + 4 2
−3 −4 5
−1 2+ 5 − 3 2 3 + 2 −4
(iii)  2 −  7 5  (iv) 3 − 2 5 6 
5 + 3 −5 2 −4 −6 3
 2 − 3 4 + 5
2. (i) Express the matrix  = 6 +  0 4 − 5 . as a sum of Hermitian and Skew-
− 2− 2+
Hermitian matrix.
1 − 1+
(ii) Express the Hermitian matrix  = 1   0 2 − 3  as  +  where P is a
1 −  2 + 3 2
real symmetric and Q is a real skew-symmetric matrix.
 0 0
3. Show that  = 0 0   is skew- Hermitian and also unitary
0  0
4. (i) If A is any square matrix, prove that  + ∗ , ∗ , ∗  are all Hermitian and  − ∗ is
Skew-Hermitian.
(ii) If A, B are hermitian or Skew-Hermitian, then so is A + B.
(iii) Show that the matrix  ∗  is Hermitian or Skew-Hermitian as A is Hermitian or
Skew-Hermitian.

54
 (iv) If A is a Hrmitian matrix, then show that  is a Skew-Hermitian matrix.
5. (i) Define unitary matrix. Show that the following matrix is unitary.
 1 1+
 
√ 1− −1
 1+ −1 + 
(ii) Prove that    is a unitary matrix.
1+ 1−
1 1 1

(iii) Show that  =
√
1    is a unitary matrix, where  is complex cube root
1  
of unity.
 
 
6. Verify that the matrix  =  
 have eigen values with unit modulus.
 

1.36. CHARACTERISTIC EQUATION:

If A is square matrix of order , we can form the matrix  − , where  is a scalar and I
is the unit matrix of order , The determinant of this matrix equated to zero, . .,
 −   … 
| − | =    −  …   = 0 is called the characteristic equation of
  …  − 
A.
On expanding the determinant, the characteristic equation can be written as a
polynomial equation of degree  in  of the form (−1)  +   +   + ⋯ +
 = 0.
The roots of this equation are called the characteristic roots or latent roots or eigen
values of A.
The set of eigen values of a square matrix A is called the spectrum of A.
Note. The sum of the eigen values of a matrix A is equal to trace of A.
[The trace of square matrix I the sum of the diagonal elements]
1.37. EIGEN VECTORS
Consider the linear transformation  =  …(1)
Which transforms the column vector X into the column vector Y. In practice, we are
often required to find those vectors X which transform into scalar multiples of
themselves.
Let X be such a vector which transforms into  ( being a non -zero scalar) by the
transformation (1).

55
 Then  =  …(2)
From (1) and (2),  =  ⇒  −  =  ⇒ ( − ) =  …(3)
This matrix equation gives  homogeneous linear equations
( − ) +  + ⋯ +   = 0
  +( − ) +. . +  = 0  …(4)
  +  +. . +( − ) = 0
These equations will have a not-trivial solution only if the co-efficient matrix  −  is
singular
. . if | − | = 0 …(5)
This is the characteristic equation of the matrix A and has  roots which are the eigen
values of A. Corresponding to each root of (5), the homogeneous system (3) has a non-
zero solution.


 =    which is called an eigen vector or latent vector.
⋮

Note. If X is a solution of (3), then so is , where l is an arbitrary constant. Thus, the
eigen vector corresponding to an eigen value is not unique.
1.38. THE CHARACTERISTIC ROOTS OF A UNITARY MATRIX ARE OF UNIT
MODULUS
Let A be a unitary matrix so that
∗  =  = ∗ …(1)
If  is a characteristic root of the matrix A and X is its latent vector, then we have
 =  …(2)
Taking transpose conjugate of (2), we obtain
()∗ = ()∗ …(3) |∴ ∗ = 
 ∗ ∗ = ̅ ∗
On multiplying (2) and(3), we get
( ∗ ∗ )() = ̅ ∗ ()

⇒  ∗ (∗ ) =  ( ∗ ) | (1)

⇒  ∗  = ( ∗ ) …(4)

⇒ 1 −  ∗  = 0

56
 Since X is a characteristic vector,  ≠ 
Consequently, ∗ ≠ 
Hence equation (4) gives

1 −  = 0

⇒  = 1
⇒ || = 1 ⇒ || = 1
Hence the characteristic roots of a unitary matrix are of unit modulus
1.39. THE LATENT ROOTS OF A HERAMITIAN MATRIX ARE ALL REAL
Let  be the characteristic or latent root of a Hermitian matrix A. Then ∃ a non-zero
latent vector X such that
 =  …(1)
Pre-multiplying both sides of (1) by  ∗ , we get
 ∗  =  ∗  …(2)
Transpose conjugate of (2) gives
( ∗ )∗ = ( ∗ )∗
⇒  ∗ ∗ ( ∗ )∗ =  ∗ ( ∗ )∗ ∗ |  

⇒  ∗ ∗  =  ∗  ∵ ∗ = 
But A is a Hermitian matrix so that ∗ 
Hence above equation becomes

 ∗  =  ∗  …(3)
From (2) and (3), we have

 ∗ ( ∗ )

⇒  −  ∗  = 0 …(4)
Since X is a non-zero latent vector
∴ ∗ ≠ 
Hence from (4), we have

− =0 ⇒ = 
Which is possible only when  is real.
Hence the latent roots of a Hermitian matrix are all real.

57
1.40. THE CHARACTERISTIC ROOTS OF A SKEW-HERMITIAN MATRIX IS
EITHER ZERO OR PURELY AN IMAGINARY NUMBER
Since A is a Skew- Hermitian matrix
∴  is Hermitian matrix.
Let  be a characteristic root of A.
Then,  =  ⇒ () = ()
⇒  is a characteristic root of matrix .
But  is a Hermitian matrix.
Therefore,  should be real.
Hence,  is either zero or purely imaginary.
1.41. THE CHARACTERISDTIC ROOTS OF AN IDEMPOTERNT MATRIX ARE
EITHER ZERO OR UNITY
Since A is an idempotent matrix. ∴  = .
Let X be a latent vector of the matrix A corresponding to the latent root  so that
 =  …(1)
⇒ ( − ) =  such that  ≠ 
Per - multiplying (1) by A
() = () = ()
⇒ () = () | (1)
⇒   =   ⇒  =   |∵  = 
⇒  =   | (1)
⇒ ( − ) =  ⇒  −  = 0 (  ≠ )
⇒ ( − 1) = 0
⇒  = 0, 1

58
 ILLUSTRATIVE EXAMPLES

Example 1. Show that if  , …  are the latent roots of the matrix A, then  has the
latent roots  ,  , …  .
Sol. Let  be a latent root of the matrix A. Then there exists a non- zero vector X such
that
 =  …(1)
⇒  () =  () ⇒   = ( )
But   = () = () | (1)|
= () = () =  
∴   = ( ) =  
⇒  is a latent root of  .
∴ If  ,  , …  are the latent roots of A, then  ,  , …  are then latent roots of
 .
Example2. If  ,  …  are eigen values of A, then find eigen values of the matrix
( − )
Sol. ( − ) =  − 2 +   
=  − 2 +  
Eigen values of  are  ,  , … 
Eigen values of 2 are 2 , 2 , …  .
Eigen values of   are 
∴ Eigen values of ( − ) are
 − 2 +  ,  − 2 +  , … ,  − 2 + 
or ( − ) , ( − ) , … , ( − ) .
Example3. If  is an eigen value of a non-singular matrix A, show that
(i)  is an eigen value of  .
||
(ii) 
is an eigen value of adj. A.

Sol. (i)  is an eigen value of A

⇒ There exists a non-zero matrix X such that AX =X
⇒  =  ()
⇒  = ( )
59
 
⇒

 =  

⇒   =  
⇒  is an eigen value of .
(ii)  is an eigen value of A
⇒ There exists a non-zero matrix X such that AX = X
⇒ (. )() = (. )()
⇒ (. ) = (. )
⇒ || = (. ) [∵ (. ) = || ]
⇒ || = (. )
||
⇒  = (. )

||
⇒ (. ) =


||
⇒ 
is an eigen value of adj. A.

Example4. Find the eigen values of the following matrices:

1 0 4 2 5 7
(i)  = 0 2 0 (ii)  = 5 3 1
3 1 −3 7 0 2
Sol. (i) Characteristic equation is
| − | = 0
1− 9 4
⇒  0 2− 0 =0
3 1 −3 − 
⇒  − 19 + 30 = 0
⇒  = 3, −5, 2
(ii) Characteristic equation is
| − | = 0
2− 5 7
⇒  5 3− 1 =0
7 0 2−
⇒  − 7 − 58 + 150 = 0
⇒  = 11.026, −6.215, 2.1888

60
Example5. Show that for any square matrix A,
(i)         .
(ii) The product of all eigen values of A is equal to determinant (A).
Sol. Let A be a square matrix.
(i) The characteristic equation of A is | − | = 0 …(1)
Let A' be the transpose of A.
Then the characteristic equation of A' will be
| − |=0 …(2)
Since the interchange of rows and columns does not alter the value of the
determinant we have,
| − | = | − |
|∵ | − |= | − |′ = |′ − ′|= | − | as I' = I
Hence the eigen values of matrix A and its transpose A' are identical.
(ii) Let  = [ ]× be a given square matrix and  ,  , ,…  be its eigen values. If
(λ) be the characteristic polynomial then,
() = | − |
 −   … . 
  − … . 
=   
⋮ ⋮ ⋮ ⋮
  … .  − 
= (−1)  +   +   + ⋯ +  
= (−1) ( −  )( −  )( −  ) … ( −  )
Putting  = 0, we get
(0) = (−1) (−1)   … 
|| =    … 
Hence the product of all eigen values of A is equal to determinant (A).
Example6.Show that for a square matrix,
(i) There are infinitely many eigem vectors corresponding to a single eigen value
(ii) Every eigen vector corresponds to a unique eigen value.
Sol. (i) Let X be a characteristic vector of a square matrix A correspoinding to a single
eigen value . Then we have,

61
  = 
Let k be any non -zero scalar. Than,
() = () ⇒ () = ()
Therefore, kX is also a characteristic vector of A corresponding to the same characteristic root 
Since k is any non-zero scalar, ∃ infinitely many eigen vectors corresponding to a single
eigen value
(ii) Let there exist two distinct eigen values  and  corresponding to an eigen vector X of a
square matrix A. Then, we have
 =   | ≠ 
 =  
∴  =   =  
⇒ ( −  )X = O
⇒ = |∵  −  ≠ 0
Which is impossible since X is a non-zero vector. Hence every eigen vector corresponding to a
unique eighen value
1 −2
Example 8. Find the eigen values and eigen vectors of matrix  =  
−5 4
Sol. The characteristic equation of the given matrix is
| = | = 0
1− −2
Or  =0
−5 4−
⇒  − 5 − 6 = 0
⇒  = 6, −1.
Thus, the eigen values of A are 6, −1,
Corresponding to  = 6, the eigen vectors are given by
( − 6) = 
1 − 6 −2 
or    = 
−5 4 − 6 
−5 −2 
or    = 
−5 −2 
We get only one independent equation −5 − 2 = 0
 
⇒ = =  ()
 

62
  = 2 ,  = −5
1
∴ The eigen vectors are  =   .
1
Example 8. Find the eigen values and eigen vectors of the matrix is.
Sol. The characteristic equation of the given matrix is
| − | = 0
−2 −  2 −3
or  2 1− −6 = 0
−1 −2 −
or (−2 − )[−(1 − ) − 12] − 2[−2 − 6] − 3[−4 + 1(1 − )] = 0
or  +  − 21 − 45 = 0
By trail,  = −3 satisfies it.
∴ ( + 3)( − 2 − 15) = 0 ⇒( + 3)( + 3)( − 5) = 0 ⇒  = −3, −3, 5
Thus, the eigen values of A are −3, −3, 5.
Corresponding to  = −3, the eigen vectors are given by
( + 3) = 
−1 2 −3 
or 2 4 −6   = 
−1 −2 3 
We get only one independent equation  + 2 − 3 = 0
Let  =  ,  =  then  = 3 − 2
∴ The eigen vectors are given by
3 − 2 3 −2
 =    =  0 +   1 
 1 0
Corresponding to  = 5, the eigen vectors are given by ( − 5) = 

−7 2 −3  0
⇒  2 −4 −6  = 0
−1 −2 −5  0

⇒ −7 + 2 − 3 = 0

 − 2 − 3 = 0
 − 2 − 5 = 0
From first two equation,

63
   
= =
10 − 6 3 + 5 −2 − 2
  
⇒ = = =  (say)
  

∴  =  ,  + 2 ,  = −
Hence the eigen vectors are given by.
1
 =   2 
−1

Example. 9. Prove that

() 0 is a characteristic root of a matrix if and only if the matrix is singular.
() The characteristic root of a real symmetric matrix are all real.
() If A and B are square matrices of same type and if P be invertible then A and  AP have
same eigen values.
() The sum of the eigen values of a square matrix is equal to the sum of the elements of its
principal diagonal.
Sol. () The characteristic roof of a matrix A is given by | − | = 0
If  = 0, then it gives || = 0
⇒ A is singular.
Again if matrix A is singular, then
| − | = 0
⇒ || − || = 0
⇒ 0 − .  = 0
⇒ =0
() Let A be a real symmetric matrix
∵ A is real

∴ =

⇒ () = ′
⇒ ∗ =  |∵   
⇒ A is Hermitian.

64
Hence the characteristic roots of A are all real
() Let  =  
Then  −  =   − 
=   −   =  ( − )
∴ | − | = | ( − )| = | | | − |||
= | − || | || = | − || − |
= | − | || = | − | |∵ || = 1
Hence matrices A and   have the same characteristic roots.
(iv) Le = [ ]× be a square matrix of order 3. The characteristic equation is
| − | = 0
 −   
⇒    −    = 0
   − 
⇒ − +  ( +  +  ) − ⋯ = 0 … (1)
But we know that,
| − | = (−1) ( −  )( −  )( −  )
= − +  ( +  +  ) …. …(2)
Comparing equations (1)and (2), we get
 +  +  =  +  + 

TEST YOUR KNOWLEDGE

1. Fine the eigen values and corresponding eigen vectors of the following matrices:
1 2 2  ℎ 
−5 2 5 4
(i)  (ii)   (iii)  0 2 1 (iv) 0  0
2 −2 1 2
−1 2 2 0 0 
1 2 2
2. (i) Find the characteristic equation of the matrix  =  0 2 1
−1 2 2

65
 2 −3 1
(ii) Find the eigen values of the matrix  =  3 1 3
−5 2 −4
1 −1
(iii) Find the eigen vectors for the matrix  =  
2 −1
1 2 −3
 
3. Find the eigen values of 3 + 5 − 6 + 2 where  = 0 3 2
0 0 −2
4 2 101
4. Find the eigen value of matrix   corresponding to the eigen vector  .
2 4 101
5. (i) Show that if  ios a characteristic root of the matrix  then  +  is a characteristic
root of the matrix  + .
(ii) Show that if  (1 ≤  ≤ ) are the eigen values of a square matrix  then  gas the
eigen values  (1 ≤) ≤ , m beigen a positive integer

(iii) Prove that the characteristic roots of a diagonal matrix are the diagonal elements of
the matrix
0 ℎ  0  ℎ 0  
6. Verify that the matrices  =  ℎ 0   ,  =  0   ,  +  0 ℎ have same
  0 ℎ  0  ℎ 0
characteristic equation.
  
7. If  +  +  = 0, Find the characteristic roots of the matrix  =    
  
2 1 1
8. Prove that for matrix  =  2 3 4 , all its eigen values are distinct and real.
−1 −1 −2
Hence find the corresponding eigen vectors.
3 10 5
9. Show that the matrix −2 −3 −4 has leas than three linearly independent eigen
3 5 7
vectors. Also find them.
0 1 1 + − −

10. If  = 1 0 1 and  =  −   +   − , Show that the transform of  byM

1 1 0 − − +
. .,  is a diagonal matrix and hence find the eigen values of .
3 2 2 −4
11. Find characteristic equation and eigen values of the matrix  =  2 3 2 −1 
1 1 2 −1
2 2 2 −1

66
 4 −1 −1
 
12. If  −1 4 −1 and  = 1 − , then show that  = 1 −  , Where  and  are
 
−1 −1 4
the eigen values of  and  respectively.

ANSWERS

1 2 1 4
1. (i) −1, −6;    ,    (ii) 1, 6;    ,   
2 −1 −1 1
1 2 1 ℎ 
(iii) 1, 2, 2 ;   1  ,  1 (iv) , ,  ;  0 ,   −  ,   0 
−1 0 0 0 −
1 1
2. (i)  − 5 + 8 − 4 = 4 (ii) 0, 1, −2 (iii)    ,   
1− 1+
3. 4, 110, 10
4. 6
 /
7.  = 0, =  ( +   +   )

2 0 1
8. 3, −1, 1;   3  ,  −1 ,  −1
−1 1 0
5 1
9. 2, 2, 3;   2  ,   1 
−5 −2
10. , , 
11.  − 7 + 17 − 17 + 6 = 0;  = 1, 1, 2, 3

67