DETERMINANT

Determinant of a 2x2 matrix

a b 
For a 2  2 matrix A =   , the determinant of A, denoted as det(A) or A is given by
c d 

a b
det ( A ) = = ad − bc .
c d

Example

1 2
= (1)(4) − (2)(3) = −2
3 4

Determinant of 3 by 3 or higher order matrices

 a11 a12 a13 

For a 3  3 matrix A =  a21 a22 a23  , the determinant of A can be obtained by expanding it
 a31 a32 a33 
across the first row (say) as follow: The ( + or − ) signs are assigned according to
the following “place sign” pattern:
a11 a12 a13
+ − +
det ( A ) = a21 a22 a23
− + −
a31 a32 a33
+ − +
= a11C11 + a12C12 + a13C13
= a11 ( + M 11 ) + a12 ( − M 12 ) + a13 ( M 13 )
a a23   a21 a23   a21 a22 
= a11  22  + a12  −  + a13  
 a32 a33   a31 a33   a31 a32 

where M ij is the minor of aij which is the determinant of the submatrix of A obtained from
A by deleting the ith row and the jth column, as shown at below:

a11 a12 a13 a11 a12 a13

a a23 a a23
M11 = a21 a22 a23 = 22 , M12 = a21 a22 a23 = 21 ,
a32 a33 a31 a33
a31 a32 a33 a31 a32 a33

a11 a12 a13

a a22
M13 = a21 a22 a23 = 21 .
a31 a32
a31 a32 a33
and Cij is called the cofactor of aij , which is given by:

Cij = (place sign) M ij

= ( - 1)i + j M ij

Remark

When evaluating a determinant, you can expand it across any row or down any column you
choose.

Example

1 −1 4 
For A = 5 −2 −3 , compute det(A) by expanding it across the
1 3 −6 

(a) first row,

(b) third row,

(c) second column.

Solution

1 −1 4
−2 −3 5 −3 5 −2
(a) det ( A ) = 5 −2 −3 = (1) − ( −1) + ( 4)
3 −6 1 −6 1 3
1 3 −6
= (12 + 9 ) + ( −30 + 3) + 4 (15 + 2 )
= 62

1 −1 4
(b) det ( A ) = 5 −2 −3
1 3 −6
−1 4 1 4 1 −1
= (1) − ( 3) + ( −6 )
−2 −3 5 −3 5 −2
= 62

(c) Do It Yourself.
Cramer’s Rule

Let Ax = b be an n  n linear system, with det ( A )  0 . Let A i be the matrix obtained from
A by replacing the ith column of A by the vector b, then

det( Ai )
xi = , i = 1, 2, ,n
det( A)

is the unique solution of the system.

Example

Solve the following system by Cramer’s Rule.

2 x1 − x2 + x3 = 8
−2 x1 + 3x2 + x3 = −4
4 x1 + 2 x2 − 3x3 = −3

Solution

2 −1 1 8 −1 1
det( A) = −2 3 1 det( A1 ) = −4 3 1 = −72
4 2 −3 −3 2 −3

3 1 −2 1 −2 3
= ( 2) − (−1) + (1)
2 −3 4 −3 4 2
= −36

2 8 1 2 −1 8
det( A 2 ) = −2 −4 1 = 36 det( A3 ) = −2 3 −4 = −108
4 −3 −3 4 2 −3

By Cramer’s rue,

det( A1 ) −72
x1 = = = 2,
det( A) −36

det( A 2 ) 36
x2 = = = −1 ,
det( A) −36

det( A3 ) −108
x3 = = = 3.
det( A) −36
Determinant of Triangular Matrices

Theorem

Let A be any triangular matrix with the main diagonal elements a11 , a22 , , ann , then the
determinant of A is given by,

det(A) = a11a22 ann

Example

2 22 222
(a) 0 4 44 = 2  4  6 = 48
0 0 6

9 99 999 9999
0 8 88 888
(b) = 98 0 6 = 0
0 0 0 77
0 0 0 6

Finding Determinant by Using Elementary Row Operations

Suppose a matrix A is reduced to triangular form through a sequence of ERO, as illustrate in

the figure below:

 a11 a12 a13  Elementary Row Operations e11 e12 e13 

A =  a21 a22 a23  0 e
 22 e23  = E
 a31 a32 a33   0 0 e33 

then determinant of A can be obtained as

det ( A ) = k det ( E) = k e11e22e33

where k is a correction factor which can be easily figured out by tracking the operations used
in reducing A to E.

Effects of ERO on Determinants

In order to figure out the correction factor k, we need to know the effect of ERO on
determinants.
(1) If matrix B is obtained from matrix A by interchanging two rows (or two columns) of
A, i.e.,

det ( B ) = − det ( A )
R R
A ⎯⎯⎯→
i j
B then,

(2) If matrix B is obtained from matrix A by multiplying a constant m to a row (or

column) of A, i.e.,

det ( B ) = m det ( A )
R →mR
A ⎯⎯⎯⎯
i i
→B then,

(3) If matrix B is obtained from matrix A by adding to a row (or column) of A a multiple
of any other row (or column, respectively) of A, i.e.,

det ( B ) = det ( A )
R →R + mR
A ⎯⎯⎯⎯⎯
i i j
→B then,

Example

Find the determinant of A by using ERO, where

3 5 −2 6  3 2 0 0 2 0 −4 6 
1 2 −1 1 6 8 0 0  4 5 1 0 
(a) A =  (b) A =  (c) A = 
2 4 1 5 0 0 4 7 0 2 6 −1
     
3 7 5 3 0 0 2 5  −3 8 9 1

Solution

3 5 −2 6  1 2 −1 1 R2 → R2 −3 R1  1 2 −1 1
1 2 −1 1 R1  R2 3 5 −2 6  R3 → R3 −2 R1 0 −1 1 3
(a) A= ⎯⎯⎯→  ⎯⎯⎯⎯→
2 4 1 5 2 4 1 5 R2 → R2 −3 R1 0 0 3 3
     
3 7 5 3 3 7 5 3 0 1 8 0

 1 2 −1 1  1 2 −1 1
0 −1 1 3 R4 → R4 − R3 0 −1 1 3

R3 →3 R3
⎯⎯⎯⎯ → ⎯⎯⎯⎯→ =B
R4 → R4 + R2
0 0 9 9 0 0 9 9
   
0 0 9 3 0 0 0 −6

det ( B ) = (1)(1)( 3)(1)(1)(1)( −1) det ( A )

−1
 det ( A ) = det ( B )
3
−1
= (1 (−1)  9  (−6) )
3
= −18
 3 2 0 0 3 2 0 0 3 2 0 0
6 8 0 0 R2 → R2 − 2 R1 0 4 0 
0 R4 → R4 − R3 0 4 0 0 
(b) A= ⎯⎯⎯⎯⎯ → ⎯⎯⎯⎯→  =B
0 0 4 7  R3 →(1/2) R3 0 0 2 3.5 0 0 2 3.5
     
0 0 2 5 0 0 2 5 0 0 0 1.5

det ( B ) = (1)(1/ 2 )(1) det ( A )

 det ( A ) = 2 det ( B )
= 2 ( 3  4  2 1.5 )
= 72

(c) We may use ERO to reduce a determinant to lower order, instead of carrying out the
ERO until a triangular form is achieved, as illustrate below:

2 0 −4 6  2 0 −4 6 
4 5 1 0  R2 → R2 − 2 R1 0 5 9 −12 
A= ⎯⎯⎯⎯⎯ → =B
0 2 6 −1 R4 → R4 +1.5 R1  0 2 6 −1 
   
 −3 8 9 1 0 8 3 10 

det ( B ) = (1)(1) det ( A )

det ( A ) = det ( B ) Note:
Expanding across the 1st column
5 9 −12
= 2 2 6 −1 − ( 0 ) M 21 + ( 0 ) M 31 − ( 0 ) M 41
8 3 10
 6 −1 2 −1 2 6 
= 2  (5) − (9) + (−12) 
 3 10 8 10 8 3 
= 2  (5)(63) − (9)(28) + ( −12)(−42) 
= 1134

Exercise

 a b c
1. Let A =  p q r  . If det(A) = 18, find the det(B) where
 x y z 
 1 1 1 
 −5 p − q
5
− r 
5
 
B =  2a 2b 2c  .
 1 1 1 
x + a y + b z + c
 3 3 3 
 3 1 0 5p 4p −2 p
2. Given −2 −4 3 = −1 . Find 2 4 −3 .
5 4 −2 3 − 5q 1 − 4q 2q

1 1 1 1
r 1 1 1
3. Show that = (1 − r )3
r r 1 1
r r r 1

a+b c c
4. Show that a b+c a = 4abc .
b b c+a

Answers

1. 36/5

2. −p

Further Properties of Determinant

Suppose A and B are n  n matrices, then

• det(AT) = det(A).

• det(A) = 0, if any zero row or zero column.

• det(A) = 0, if some rows or columns are proportional.

• det(AB) = det(A) det(B) = det(B) det(A) = det(BA).

• det(cA) = cn det(A).

• For a matrix A-1, such that A-1A = I, then

det(A-1) = 1/det(A).

Example

Find det(A) by inspection, where

 1 −2 5 4 
2 3 6 −6 
A=  .
 3 −4 7 8
 
4 1 8 −2 
Solution

det ( A ) = 0 because the 2nd column and the 4th column are proportional.

Example

Suppose A and B are 3  3 matrices where det ( A ) = 3 and det ( B ) = 9 . Find det ( 2A2B−1 ) .

Solution

det ( 2 A 2 B −1 ) = det ( 2 A 2 ) det ( B −1 )

1
= det ( 2 A ) det ( A )
det ( B )
1
= 23 det ( A ) det ( A )
det ( B )
1
= 8 3 3
9
=8

Exercise

1 1 1
1. Show that a b c = 0 by applying one step of ERO.
b+c a+c a+b

2. Show that det ( A ) = 0 by applying two steps of ERO, where

1 2 3 4
5 6 7 8
A= .
 9 10 11 12 
 
13 14 15 16 

Rank and Determinant

Theorem:

For an n  n matrix A,

• det ( A )  0 implies rank(A) = n

 • det ( A ) = 0 implies rank(A) < n

Homogeneous Linear Systems

For homogeneous linear systems, consistency is never an issue, because homogeneous linear
system always consistent. There exists at least a solution, x1 = 0, x2 = 0, , xn = 0 which is
called the trivial solution for every homogeneous linear system.

Now, the issue is, in what conditions will the system has not only the trivial solution but also
the non-trivial solutions.

We give the answer as a theorem below:

Theorem

Suppose Ax = 0 is a homogeneous linear system of m equations in n unknowns.

• If rank ( A ) = n , then the system has only the trivial solution.

• If rank ( A )  n , then the system has non-trivial solutions as well as the trivial solution.

Example

Find the value(s) of k such that the following homogeneous linear system has non-trivial
solutions.

2x + k y + z = 0
(k − 1) x − y + 2z = 0
4x + y + 4z = 0

Solution

For non-trivial solution exist, require rank(A) < n = 3, or equivalently, det(A) = 0.

2 k 1
k − 1 −1 2 = 0
4 1 4
−1 2 k −1 2 k − 1 −1
2 −k +1 =0
1 4 4 4 4 1

2 ( −4 − 2 ) − k ( 4k − 4 − 8 ) + ( k − 1 + 4 ) = 0
−4k 2 + 13k − 9 = 0
( −4k + 9 )( k − 1) = 0
k = 9 / 4 or k = 1
Exercise

Consider the following homogeneous linear system.

2x − y + 5z = 0
3x − y + 7z = 0
4x − 2 y +  z = 0

(a) Find the value(s) of  such that the system has non-trivial solutions exist.
(b) Solve the system for this value of  .
(c) If  = 5 , write down the solution of the linear system without any calculations. Give
your reason.

Answer

(a)  = 10
 x  −2 
   
(b)  y = k  1 , k 
z 1
   

 x  0
   
(c) As long as   10 , the system has only the trivial solution, i.e.,  y  =  0  .
 z  0
   