Chapter 2B Determinants

When we look at a particular square matrix, the question of whether it is

nonsingular is one of the first things that we ask. This chapter develops a
formula to determine this.

 2B.1 The Determinant and Evaluation of a Matrix

 2B.2 Properties of Determinants
 2B.3 Eigenvalues and Application of Determinants
 2B.4 Geometry of Determinants: Determinants as Size
Functions

3-1
2B.1 The Determinant of a Matrix
 The determinant of a 2 ×2 matrix:
 a11 a12 
A  
 21 22 
a a

 det( A)  | A |  a11a22  a21a12

 Note:  a11 a12  a11 a12
a  
 21 a22  a21 a22

2 3
 2(2)  1( 3)  4  3  7
1 2

3-2
 Minor of the entry aij :
a11 a12 L a1( j 1) a1( j 1) L a1 n
M M M
a( i 1)1 L a( i 1)( j 1) a( i 1)( j 1) L a( i 1) n
M ij  a ij
a( i 1)1 L a( i 1)( j 1) a( i 1)( j 1) L a( i 1) n
M M M M
a n1 L an( j 1) an( j 1) L ann

 Cofactor of a ij :

Cij  (1)i  j Mij

3-3
 Ex:
 a11 a12 a13   a11 a12 a13 
A  a21 a22 a23  A  a21 a22 a23 
   
a31 a32 a33  a31 a32 a33 
a12 a13 a11 a13
 M 21  M 22 
a32 a33 a31 a33

 C21  ( 1) 21 M 21   M 21 C22  ( 1) 22 M 22  M 22

     L 
     L 
 Notes: Sign pattern for cofactors  
     L 
 
     L 
     L 
 
M M M M M 
3-4
 Thm 3B.1: (Expansion by cofactors)
Let A is a square matrix of order n, then the determinant of A is
given by
n
(a ) det( A) | A |  aij Cij  ai 1Ci 1  ai 2Ci 2  L  ainCin
j 1

(Cofactor expansion along the i-th row, i=1, 2,…, n )

or
n
(b) det( A) | A |  aij C ij  a1 j C1 j  a2 j C 2 j  L  anj C nj
i 1

(Cofactor expansion along the j-th column, j=1, 2,…, n )

3-5
 Ex: The determinant of a matrix of order 3
 a11 a12 a13 
A  a21 a22 a23 
 
a31 a32 a33 

 det( A)  a11C11  a12C12  a13C13

 a21C21  a22C22  a23C23
 a31C31  a32C32  a33C33
 a11C11  a21C21  a31C31
 a12C12  a22C22  a32C32
 a13C13  a23C23  a33C33

3-6
 Ex 5: (The determinant of a matrix of order 3)
0 2 1 
A  3  1 2  det( A)  ?
4  4 1
Sol:
1 1 1 2 3 2
C11  ( 1) 7 C12  ( 1)1 2  (3  8)  5
4 1 4 1

1 3 3 1
C13  ( 1)  12  4  8
4 4

 det( A)  a11C11  a12C12  a13C13

 (0)(7)  (2)(5)  (1)( 8)  2

3-7
 The determinant of a matrix of order 3:
 a11 a12 a13 
A   a21 a22 a23 
 a31 a32 a33 
 det( A) | A | a11a22 a33  a12 a23a31  a13a21a32  a31a22 a13
 a32 a 23 a11  a33a21a12

–4 0 6
 Ex : 0 2 1 0 2
A   3  1 2 3  1
 
4  4 1 4  4
0 16 –12
 det( A) | A | 0  16  12  (4)  0  6  2

3-8
 Upper triangular matrix:
All the entries below the main diagonal are zeros.
 Lower triangular matrix:
All the entries above the main diagonal are zeros.
 Diagonal matrix:
All the entries above and below the main diagonal are zeros.

a11 a12 a13   a11 0 0  a11 0 0 

 0 a22 a23  a21 a22 0   0 a22 0 
0 0 a  a a a  0 0 a 
 33   31 32 33   33 

upper triangular lower triangular diagonal

3-9
 Theorem 2B.2: Determinant of a Triangular Matrix
If A is an nxn triangular matrix (upper triangular, lower
triangular, or diagonal), then its determinant is the
product of the entries on the main diagonal. That is

det( A) | A | a11a22a33 L ann

At this moment, our primary way to decide whether a matrix is singular is to do

Gaussian reduction and then check whether the diagonal of resulting echelon form
matrix has any zeroes.

We will look for a family of functions with the property of being unaffected by row
operations and with the property that a determinant of an echelon form matrix is the
product of its diagonal entries.

3 - 10
 Ex: Find the determinants of the following triangular matrices.

 1 0 0 0 0
2 0 0 0  0 3 0 0 0
 4 2 0 0 
(a) A    (b) B   0 0 2 0 0
 5 6 1 0  
1 3  0 0 0 4 0
 5 3  0 0 0 0  2
Sol:

(a) |A| = (2)(–2)(1)(3) = –12

(b) |B| = (–1)(3)(2)(4)(–2) = 48

3 - 11
Keywords in This Section:

 determinant : 行列式
 minor : 子行列式
 cofactor : 餘因子
 expansion by cofactors : 餘因子展開
 upper triangular matrix: 上三角矩陣
 lower triangular matrix: 下三角矩陣
 diagonal matrix: 對角矩陣

3 - 12
 2B.2 Evaluation of a determinant using elementary operations

 Theorem 2B.3: Elementary row operations and determinants

Let A and B be square matrices,
(a ) B  Pij ( A)  det( B)   det( A)
(b) B  M i (k )( A)  det( B)  k det( A)
(c ) B  C ij ( k )( A)  det( B)  det( A)

3 - 13

Ex:
1 2 3
A  0 1 4  det( A)  2
 
1 2 1
4 8 12 0 1 4  1 2 3
A1  0 1 4  A2  1 2 3 A3   2  3  2
1 2 1  1 2 1  1 2 1 

A1  M1 (4)( A)  det( A1)  4det( A)  (4)( 2)  8

A2  P12 ( A)  det( A2)   det( A)  (2)  2

A3  C12 (2)( A)  det( A3)  det( A)  2

3 - 14
Note:
A row-echelon form of a square matrix is always upper triangular.

 Ex: Evaluation a determinant using elementary row operations

2  3 10 
A  1 2  2  det( A)  ?
0 1  3

Sol:
2 3 10 1 2 2
12
det( A)  1 2 2    2 3 10
0 1 3 0 1 3

3 - 15
 1 2 2 1
1 2 2
 2 1   2  2 1
   0 7 14 
7
 (  1)( 1 ) 0 1 2
0 1 3 7
0 1 3
1 2 2
 2  3
 7 0 1 2  (7)(1)(1)( 1)  7
0 0 1

3 - 16
  Theorem 2B.4: Conditions that yield a zero determinant
If A is a square matrix and any of the following conditions is true,
then det (A) = 0.
(a) An entire row (or an entire column) consists of zeros.
(b) Two rows (or two columns) are equal.
(c) One row (or column) is a multiple of another row (or column).
The theorem states that : a matrix with two identical rows or two linear dependent
rows has a determinant of zero. A matrix with a zero row has a determinant of zero.
Note that a matrix is nonsingular if and only if its determinant is nonzero and the
determinant of an echelon form matrix is the product down its diagonal.

This theorem provides a way to compute the value of a determinant function on a

matrix: Do Gaussian reduction, keeping track of any changes of sign caused by row
swaps and any scalars that are factored out, and then finish by multiplying down the
diagonal of the echelon form result.

3 - 17
 Note:

Cofactor Expansion Row Reduction

Order n Additions Multiplications Additions Multiplications

3 5 9 5 10

5 119 205 30 45

10 3,628,799 6,235,300 285 339

3 - 18
 Ex: (Evaluating a determinant)
 2 0 1 3 2 
 2 1 3 2 1

A   1 0 1 2 3
 
 3 1 2 4 3 
 1 1 3 2 0 
Sol:
2 0 1 3 2 2 0 1 3 2
2 1 3 2 1 2 1 3 2 1
det( A)  1 0 1 2 3  1 0 1 2 3
 2  4
3 1 2 4 3   2   5 1 0 5 6 4
1 1 3 2 0 3 0 0 0 1
2 1 3 2
1 1 2 3
 (1)(  1)2 2
1 5 6 4
3 0 0 1
3 - 19
 8 1 3 2
8 1 3 0 0 5
 3  4  1 8 1 2 3
  (1)(  1)4 4 8 1 2  8 1 2
13 5 6 4  2  1
13 5 6 13 5 6
0 0 0 1
8 1
 5(  1)1 3
13 5
 (5)( 40  13)  (5)( 27)
 135

3 - 20
 2B.2 Properties of Determinants

 Theorem 2B.5: Determinant of a matrix product

det (AB) = det (A) det (B)

 Notes:

(1) det (EA) = det (E) det (A)

(2) det( A  B )  det( A)  det( B ), but

(3) a11 a12 a13 a11 a12 a13 a11 a12 a13
a21  b21 a22  b22 a23  b23  a21 a22 a23  b21 b22 b23
a31 a32 a33 a31 a32 a33 a31 a32 a33

3 - 21
 Ex: (The determinant of a matrix product)

1  2 2 2 0 1
A  0 3 2 B  0  1  2
1 0 1 3 1  2

Find |A|, |B|, and |AB|

Sol:

1 2 2 2 0 1
| A | 0 3 2  7 | B | 0  1  2  11
1 0 1 3 1 2

3 - 22
 1  2 2 2 0 1  8 4 1 
AB  0 3 2 0  1  2  6  1  10
1 0 1 3 1  2 5 1  1 

8 4 1
| AB | 6  1  10  77 | A || B |
5 1 1

3 - 23
 Theorem 2B.6: Determinant of a scalar multiple of a matrix
If A is an n × n matrix and c is a scalar, then
det (cA) = cn det (A)
 Ex:
 10  20 40 1 2 4
A   30 0 50, 3 0 5 5
 
 20  30 10  2 3 1
Find |A|.
Sol:
 1  2 4 1 2 4

A  10 3 0 5   A  10 3
3 0 5  (1000)(5)  5000
 
  2  3 1 2 3 1
3 - 24
 Thm 2B.7: Determinant of an invertible matrix
A square matrix A is invertible (nonsingular) if and only if
det (A)  0
 Ex: (Classifying square matrices as singular or nonsingular)
0 2  1 0 2  1
A  3  2 1  B  3  2 1 
3 2  1 3 2 1 
Sol:
A 0  A has no inverse (it is singular).

B  12  0  B has inverse (it is nonsingular).

3 - 25
 Thm 2B.8: Determinant of an inverse matrix
1
If A is invertible, then det(A 1 )  .
det(A)
 Thm 2B.9: Determinant of a transpose

If A is a square matrix, then det( AT )  det( A).

Ex:
1 0 3

1
(a) A  ? AT  ?
A  0  1 2
(b)

2 1 0
Sol:
1 0 3 1 1
 A1  
Q | A | 0 1 2  4 A 4
2 1 0 AT  A  4
3 - 26
 Equivalent conditions for a nonsingular matrix:

If A is an n × n matrix, then the following statements are

equivalent.
(1) A is invertible.

(2) Ax = b has a unique solution for every n × 1 matrix b.

(3) Ax = 0 has only the trivial solution of zero column vector.

(4) A is row-equivalent to In

(5) A can be written as the product of elementary matrices.

(6) det (A)  0

3 - 27
 Ex: Which of the following system has a unique solution?

(a) 2 x2  x3  1
3x1  2 x2  x3  4
3x1  2 x2  x3  4

Sol: 0 2 1
Ax  b Q A  3 2 1  0
3 2 1

 This system does not have a unique solution.

3 - 28
(b) 2 x2  x3  1
3x1  2 x2  x3  4
3x1  2 x2  x3  4
Sol:
Bx  b

Q B  12  0

 This system has a unique solution.

3 - 29
 2B.3 Introduction to Eigenvalues
 Eigenvalue problem:
If A is an nn matrix, do there exist nonzero n1 matrices x
such that Ax is a scalar multiple of x？

 Eigenvalue and eigenvector:

A：an nn matrix
：a scalar
x： a n1 nonzero column matrix

Eigenvalue
(The fundamental equation for
Ax  x
the eigenvalue problem)
Eigenvector
3 - 30
 Ex 1: (Verifying eigenvalues and eigenvectors)
1 4  1 2
A  x1    x2   
 2 3 1  1
Eigenvalue

1 4 1 5 1

Ax1          5   5 x1
2 3 1 5 1
Eigenvector
Eigenvalue
1 4   2    2  2
Ax2          1   (1) x2
2 3  1  1   1
Eigenvector
3 - 31
 Question:
Given an nn matrix A, how can you find the eigenvalues and
corresponding eigenvectors?

 Note:
Ax   x   Ix  ( I  A) x  0 (homogeneous system)
Note: If (I  A) x  0 has nonzero solutions iff det(I  A)  0.

 Characteristic equation of AMnn:

det( I  A)  ( I  A)   n  cn1 n1  L  c1  c0  0

3 - 32
 Ex: (Finding eigenvalues and eigenvectors)
1 4 
A 
 2 3 

Sol: Characteristic equation:

 1 4
(I  A) 
2  3
 2  4  5  (  5)(  1)  0
   5,  1
Eigenvalue: 1  5, 2  1

3 - 33
  4 4   x1   0 
(1) 1  5  (1I  A) x    x    
 2 2   2   0 
 x1   t  1
x      t  , t  0
 2  t  1

 2 4   x1   0 
(2) 2  1  (  2 I  A) x    x    
 2 4   2   0 
 x1   2t  2
x      t  , t  0
 2  t    1

3 - 34
Application Example of Eigenvalue-Eigenvector Problem
The equations of motion for identical mass m1  m2
and spring constant k1  k2  k can be described by
d 2 x1
m 2  kx1  k3 ( x1  x2 ) and
dt
d 2 x2
m 2  kx2  k3 ( x2  x1 )
dt

 x1 (t )   Aeit 
Try the solution x (t )      it  and plug this into the differential equations:
 x2 (t )   Be 

 m 2 A   kA  k3 ( A  B )
We can obtain
 m 2 B   kB  k3 ( B  A)

Rearrange these to put them into a neater form

(  m 2  k  k3 ) A  k3 B  0
 k3 A  (  m 2  k  k3 ) B  0

3 - 35
A nontrivial solution occurs when the determinant is zero
( k  k3  m 2 )(k  k3  m 2 )  (k3 )2  0 , which yields the following solutions

(eigenvalues): 1  k m , 2  (k  2k3 ) m

With the given eigenvalues, we can find the corresponding eigenvectors

(normal modes) to be
r  x1 ( t )   Ae  i1t 
1  k m  B  A therefore x( t )       i1t 
 2   Ae
x ( t ) 
r  x1 ( t )   Ae  i2t 
 2  ( k  2k3 ) m  B   A therefore x( t )     i 2t 
 2  
x ( t )  Ae 

3 - 36
 2.3 Applications of Determinants
 Matrix of cofactors of A:
 C11 C12 L C1n 
C C 22 L C 2 n 
Cij    M M
 21

M
Note Cij  (1)i  j Mij
 
 C n1 C n 2 L C nn 

 Adjoint matrix of A:
 C11 C 21 L C n1 
C C 22 L C n 2 
adj( A)   Cij  
T
 12

 M M M
 
C1n C 2 n L C nn 

3 - 37
 Thm 2B.10: The inverse of a matrix given by its adjoint
If A is an n × n invertible matrix, then
1
A 1  adj ( A)
det( A)
 Ex:
a b   d  c  d  b
A  CA    adj ( A)   
 c d    b a    c a 
 det( A)  ad  bc
1
1 1  d  b
A  adj ( A) 
det  A ad  bc  c a 

3 - 38
 Ex:
 1 3 2 (a) Find the adjoint of A.
A   0  2 1 
 1 0  2 (b) Use the adjoint of A to find A1

Sol: Q Cij  (1)i  j Mij

2 1 0 1 0 2
 C11    4, C12    1, C13   2
0 2 1 2 1 0
3 2 1 2 1 3
C21    6, C22    0, C23   3
0 2 1 2 1 0

3 2 1 2 1 3
C31    7, C 32    1, C33   2
2 1 0 1 0 2

3 - 39
  cofactor matrix of A  adjoint matrix of A
 4 1 2 4 6 7
 C ij    6 0 3  adj ( A)  C ij    1 0 1 
T

 
 7 1 2   2 3 2 

 inverse matrix of A
1
A 
1
adj( A) Q det  A  3
det  A
4 6 7  43 2 7
3 
 13 1 0 1   13 0 1
   3

2 3 2  23 1 3
2


 Check: AA1  I
3 - 40
 Thm 2B.11: Cramer’s Rule
a11 x1  a12 x2  L  a1n xn  b1

a21 x1  a22 x2  L  a2 n xn  b2
M
an1 x1  an 2 x2  L  ann xn  bn
 x1   b1 
x  b 
Ax  b A   aij    A(1) , A(2) ,L , A( n )  x   2 b   2
n n  M  M
   
a11 a12 L a1n  xn   bn 

a21 a22 L a2 n
det( A)  0
M M M (this system has a unique solution)
a n1 a n 2 L ann
3 - 41
Aj   A(1) , A(2) , L , A( j 1) , b, A( j 1) , L , A( n ) 

 a11 L a1( j 1) b1 a1( j 1) L a1n 

 
 a21 L a2( j 1) b2 a2( j 1) L a2 n 

M O M
 
 an1 L an( j 1) bn an ( j 1) L ann 

( i.e., det( Aj )  b1C1 j  b2C2 j  L  bnCnj )

1
x  A 1 b  adj ( A)  b
det  A

det( A j )
 xj  , j  1, 2 , L , n
det( A)

3 - 42
 Pf:
A x = b, det( A)  0
1
1
x A b  adj ( A)b
det( A)
 C11 C 21 L C n1   b1 
 C n 2   b2 
1  C12 C 22 L

det( A)  M M M  M
  
 C1 n C 2 n L C nn   bn 

 b1C11  b2C 21  L  bnC n1 

  b2C 22  L  bnC n 2 
1  b1C12

det( A)  M M 
 
 b1C1n  b2C 2 n  L  bnC nn 

3 - 43
 1
 xj  (b1C1 j  b2C 2 j  L  bnC nj )
det( A)
det( A j )
 j  1, 2 , L , n
det( A)

3 - 44
 Ex: Use Cramer’s rule to solve the system of linear equations.
 x  2 y  3z  1
2x  z  0
3x  4 y  4 z  2
Sol: 1 2  3 1 2 3
det( A)  2 0 1  10 det( A1 )  0 0 1 8
3 4 4 2 4 4
1 1  3 1 2 1
det( A2 )  2 0 1  15, det( A3 )  2 0 0  16
3 2 4 3 4 2
det( A1 ) 4 det( A2 )  3 det( A3 )  8
x  y  z 
det( A) 5 det( A) 2 det( A) 5
3 - 45
Keywords in This Section:
 matrix of cofactors : 餘因子矩陣
 adjoint matrix : 伴隨矩陣
 Cramer’s rule : Cramer 法則

3 - 46
2B.4 Geometry of Determinants: Determinants as Size Functions

 We have so far only considered whether or not a determinant is

zero, here we shall give a meaning to the value of that
determinant.

O
One way to compute the area that it encloses is to draw this rectangle and subtract
the area of each subregion.

3 - 47
The properties in the definition of determinants make reasonable postulates for
a function that measures the size of the region enclosed by the vectors in the
matrix.
See this case:

r r
The region formed by kv and w is bigger, by a factor of k, than the
r r r r
shaded region enclosed by v and w . That is, size ( kv , w) = k ·
r r
size( v , w ).

3 - 48
 Another property of determinants is that they are unaffected by pivoting.
Here are before-pivoting and after-pivoting boxes (the scalar used is
 k = 0.35). r
kv r
v

r r r
Although the region on the right, the box formed by v and kv  w , is more
slanted than the shaded region, the two have the same base and the same
height and hence the same area.

This illustrates that

3 - 49
 r r
 That is, we’ve got an intuitive justification to interpret det ( v1 , . . . , vn )
as the size of the box formed by the vectors.

Example The volume of this parallelepiped, which can be found by the

usual formula from high school geometry, is 12.

3 - 50
 The only difference between
r them is in the order
r in which the vectors
are taken. If we take u first and then go to v , follow the
counterclockwise are shown, then the sign is positive. Following a
clockwise are gives a negative sign. The sign returned by the size
function reflects the ‘orientation’ or ‘sense’ of the box.

3 - 51
 Volume, because it is an absolute value, does not depend on the order in
which the vectors are given. The volume of the parallelepiped in the
following example, can also be computed as the absolute value of this
determinant.

The definition of volume gives a geometric interpretation to something in

the space, boxes made from vectors.

3 - 52
Application of the map t represented with respect to the standard bases by

will double sizes of boxes, e.g., from this

r r
r r
w v   t ( w ) t (v ) 
   
     
 
 2 1 to  3 3 
   
 1 2   4 2 
   
   

3 - 53