MODULE 5.

EIGENVALUES AND EIGENVECTORS

Time Frame: 4 weeks

Objectives: At the end of this chapter, the student is expected to

1. Find the eigenvalues and eigenvectors of a matrix
2. Find the characteristic polynomial of a matrix
3. Diagonalize a matrix

Definition. Let A be an n x n matrix. The real number λ is called an eigenvalue of A if

there exists a nonzero vector X in ℝ𝑛 such that AX = λX. Every nonzero vector X
satisfying AX = λX is called an eigenvector of A associated with the eigenvalue 𝛌.

Note: eigen in German means proper

Eigenvalues are called proper values, characteristic values, and latent values.
Eigenvectors are called proper vectors, characteristic vectors, and latent vectors.

Example 1. If A is the identity matrix 𝐼𝑛 , then the only eigenvalue is λ = 1. Every nonzero

vector in ℝ𝑛 is an eigenvector of A associated with the eigenvalue λ = 1: 𝐼𝑛 𝑋 = 1𝑋.

0 1
Example 2. Let A = [ 1 2 ]. Then
2 0

1
1 0 1 1 1 1 1
A [ ] = [ 1 2 ] [ ] = [21] = 2 [ ] so that 𝑋1 = [ ] is an
1 2 0 1 1 1
2

1
eigenvector of A associated with eigenvalue λ1 = . Also,
2

−1
1 0 1 1 1 1 1
A [ ] = [ 1 2 ] [ ] = [ 21 ] = − 2 [ ] so that 𝑋2 = [ ] is an eigenvector of A
−1 2 0 −1 −1 −1
2

1
associated with the eigenvalue λ2 = − .
2

0 0 1 0 0 1 0 1 1
Example 3. Let A = [ ]. Then A [ ] = [ ] [ ] = [ ] = 0 [ ] so that 𝑋1 = [ ] is an
0 1 0 0 1 0 0 0 0

eigenvector of A associated with eigenvalue λ1 = 0. Also,

0 0 0 0 0 0 0
A[ ] = [ ] [ ] = [ ] = 1 [ ] so that 𝑋2 = [ ] is an eigenvector of A associated with the
1 0 1 1 1 1 1

MAT121 Module 5 I Linear Algebra I 1

 MODULE 5. EIGENVALUES AND EIGENVECTORS
eigenvalue λ2 = 1.

1 1
Example 4. Let A = [ ]. Find the eigenvalues of A and their associated eigenvectors.
−2 4

Solution:
𝑥1 1 1 𝑥1 𝑥1
Let X = [𝑥 ]. Then [ ] [𝑥 ] = λ [𝑥 ] . Thus
2 −2 4 2 2

𝑥1 + 𝑥2 = λ𝑥1 or (λ − 1) 𝑥1 − 𝑥2 = 0
−2 𝑥1 + 4𝑥2 = λ𝑥2 2 𝑥1 + (λ − 4)𝑥2 = 0

Note that if A is an nxn matrix, then the homogeneous system AX = 0 has a nontrivial
solution if and only if |𝐴| = 0. We have

λ−1 −1
| | = 0. This means that (λ − 1)( λ − 4) + 2 = 0 or
2 λ−4

λ2 − 5λ + 6 = 0 = (λ − 3)(λ − 2). Hence λ=3 and λ = 2 are the eigenvalues of A.

1 1 𝑥1 𝑥1
If λ = 2 . Then AX = 2X or [ ] [𝑥 ] = 2 [𝑥 ]. Thus
−2 4 2 2

𝑥1 + 𝑥2 = 2𝑥1 or 𝑥1 − 𝑥2 = 0
−2𝑥1 + 4𝑥2 = 2𝑥2 2𝑥1 − 2𝑥2 = 0.

Hence, 𝑥1 = 𝑥2 where 𝑥2 = any real number r. Therefore, all eigenvectors

𝑟
associated with the eigenvalue λ = 2 are [ ], where r is any nonzero real number.
𝑟

1 1 𝑥1 𝑥1
If λ = 3 . Then AX = 3X or [ ] [𝑥 ] = 3 [𝑥 ]. Thus
−2 4 2 2

𝑥1 + 𝑥2 = 3𝑥1 or 2𝑥1 − 𝑥2 = 0
−2𝑥1 + 4𝑥2 = 3𝑥2 2𝑥1 − 𝑥2 = 0.

1
Hence, 𝑥1 = 2 𝑥2 where 𝑥2 = any real number r. Therefore, all eigenvectors

𝑟
associated with the eigenvalue λ = 3 are [ 2 ], where r is any nonzero real number.
𝑟

Definition. Let A = [𝑎𝑖𝑗 ] be an nxn matrix. The determinant

λ − 𝑎11 −𝑎12 … −𝑎1𝑛

−𝑎 λ − 𝑎22 … −𝑎2𝑛
𝑓(λ) = |λ𝐼𝑛 − 𝐴| = | 21 |
⋮ ⋮ ⋮
−𝑎𝑛1 −𝑎𝑛2 … λ − 𝑎𝑛𝑛

MAT121 Module 5 I Linear Algebra I 2

 MODULE 5. EIGENVALUES AND EIGENVECTORS

is called the characteristic polynomial of A. The equation

𝑓(λ) = |λ𝐼𝑛 − 𝐴| = 0

is called the characteristic equation of A.

Theorem. The eigenvalues of A are the real roots of the characteristic polynomial of A.

1 2 −1
Example. Let A = [1 0 1], find
4 −4 5

a. characteristic polynomial of A
b. eigenvalues of A
c. eigenvectors of A

Solution:

1 0 0 λ 0 0
a. Note that λ𝐼𝑛 = λ | 0 1 0 | = | 0 λ 0 |
0 0 1 0 0 λ

λ−1 −2 1
𝑓(λ) = |λ𝐼𝑛 − 𝐴| = | −1 λ −1 |
−4 4 λ−5

= [(λ − 1)(λ)( λ − 5) + (−2)(−1)(−4) + 1(−1)(4)] −

[(-4)( λ)(1) + 4(−1)(λ − 1) + (λ − 5)(−1)(−2)]

= λ3 − 6λ2 + 11λ − 6.

Hence, the characteristic polynomial of A is 𝑓(λ) = λ3 − 6λ2 + 11λ − 6.

b. The possible integer roots of 𝑓(λ) = λ3 − 6λ2 + 11λ − 6 are ±1, ±2, ±3, and ±6.

If λ = 1. Then 𝑓(1) = 1 − 6 + 11 − 6 = 0. Thus λ = 1 is a root of 𝑓(λ). Hence (λ − 1)

Is a factor of 𝑓(λ). Dividing 𝑓(λ) by (λ − 1) we have

𝑓(λ) = (λ − 1)(λ2 − 5λ + 6) = (λ − 1)(λ − 2)(λ − 3) .

MAT121 Module 5 I Linear Algebra I 3

 MODULE 5. EIGENVALUES AND EIGENVECTORS

Thus the eigenvalues of A are λ1 = 1, λ2 = 2, and λ3 = 3. The same

eigenvalues are obtained if λ is 2 or 3.

c. If λ1 = 1. Then we form a system (1𝐼3 − 𝐴)𝑋 = 0,

1−1 −2 1 𝑥1 0
[ −1 1 − 1 ] [𝑥2 ] = [0]
−4 4 1 − 5 𝑥3 0

or

0 −2 1 𝑥1 0
[−1 1 ]
−1 2[ 𝑥 ] = [ 0].
−4 4 −4 3 𝑥 0

We have
−2𝑥2 + 𝑥3 = 0
−𝑥1 + 𝑥2 −𝑥3 = 0
−4𝑥1 + 4𝑥2 − 4𝑥3 = 0

Thus 𝑥3 = 2𝑥2 where 𝑥3 = any real number r.

Using Gauss-Jordan Elimination, we have

1 −𝑟
0 −2 1 ⋮ 0 1 0 2 ⋮ 0 2
[−1 1 −1⋮ 0] → [0 1 − 1⋮ 0] → [ 𝑟 ]
−4 4 −4⋮ 0 0 0 2⋮ 0 2
0 𝑟

−𝑟
2
Hence, all eigenvectors associated with the eigenvalue λ1 = 1 are [ 𝑟 ], where r is any
2
𝑟
nonzero real number.
−1
If r = 2 then 𝑋1 = [ 1 ] is an eigenvector of A associated with λ1 = 1.
2

If λ2 = 2. Then we form a system (2𝐼3 − 𝐴)𝑋 = 0,

2−1 −2 1 𝑥1 0
[ −1 2 − 1 ] [𝑥2 ] = [0]
−4 4 2 − 5 𝑥3 0

or

MAT121 Module 5 I Linear Algebra I 4

 MODULE 5. EIGENVALUES AND EIGENVECTORS

1 −2 1 𝑥1 0
[ −1 2 − 1 ] [𝑥2 ] = [0].
−4 4 − 3 𝑥3 0

We have
𝑥1 − 2𝑥2 + 𝑥3 = 0
−𝑥1 + 2𝑥2 −𝑥3 = 0
−4𝑥1 + 4𝑥2 − 3𝑥3 = 0.

Using elimination on equations 2 and 3, we get −2𝑥1 − 𝑥3 = 0. Thus 𝑥3 = −2𝑥1 where

𝑥3 = any real number r. Using Gauss-Jordan Elimination, we have

1 𝑟
0 −2 1 ⋮ 0 1 0 2 ⋮ 0 2
[−1 2 −1⋮ 0] → [0 1 − 1⋮ 0] → [ 𝑟 ]
−4 4 −3⋮ 0 0 0 4⋮ 0 4
0 𝑟

−𝑟
2
Hence, all eigenvectors associated with the eigenvalue λ2 = 2 are [ 𝑟 ], where r is any
4
𝑟
nonzero real number.
−2
If r = 4 then 𝑋2 = [ 1 ] is an eigenvector of A associated with λ2 = 2 .
4

If λ3 = 3. Then we form a system (3𝐼3 − 𝐴)𝑋 = 0,

3−1 −2 1 𝑥1 0
[ −1 3 𝑥
− 1 ] [ 2 ] = [0]
−4 4 3 − 5 𝑥3 0

or

2 −2 1 𝑥1 0
[ −1 3 𝑥
− 1 ] [ 2 ] = [0].
−4 4 − 2 𝑥3 0

We have
2𝑥1 − 2𝑥2 + 𝑥3 = 0
−𝑥1 + 3𝑥2 −𝑥3 = 0
−4𝑥1 + 4𝑥2 − 2𝑥3 = 0.

MAT121 Module 5 I Linear Algebra I 5

 MODULE 5. EIGENVALUES AND EIGENVECTORS
Using elimination on equations 2 and 3, we get −8𝑥1 + 2𝑥3 = 0. Thus 𝑥3 = 4𝑥2 where

𝑥3 = any real number r. Using Gauss-Jordan Elimination, we have

1 −𝑟
0 −2 1 ⋮ 0 1 0 2 ⋮ 0 2
[−1 1 −1⋮ 0] → [0 1 − 1⋮ 0] → [ 𝑟 ]
−4 4 −4⋮ 0 0 0 2⋮ 0 2
0 𝑟
−𝑟
4
Hence, all eigenvectors associated with the eigenvalue λ3 = 3 are [ 𝑟 ], where r is any
4
𝑟
nonzero real number.
−1
If r = 4 then 𝑋3 = [ 1 ] is an eigenvector of A associated with λ3 = 3 .
4

Note: The characteristic polynomial of a given matrix may have imaginary roots and it may
even have no real roots.

0 1
Example 2. Let A = [ ]. Then the characteristic polynomial of A is
−1 0

𝑓(λ) = λ2 + 1 which has no real roots. Thus A has no eigenvalues.

MAT121 Module 5 I Linear Algebra I 6

 MODULE 5. EIGENVALUES AND EIGENVECTORS

Exercises
A. Find the characteristic polynomial of the following

2 1
1. [ ]
−1 3

1 2 1
2. [ 0 1 2]
−1 3 2

4 −1 3
3. [0 2 1]
0 0 3

B. Find a. characteristic polynomial

b. eigenvalues
c. eigenvectors

2 2 3
1 −1
1. [ ] 4. [1 2 1]
2 4
2 −2 1

1 0 0 2 0 0
2. [−1 3 0] 5. [3 −1 0]
3 2 −2 0 4 3

1 2 3 4
2 −2 3
0 −1 3 2
3. [0 3 −2] 6. [ ]
0 0 3 3
0 −1 2
0 0 0 2

MAT121 Module 5 I Linear Algebra I 7

 MODULE 5. EIGENVALUES AND EIGENVECTORS

Definition. A matrix B is said to be similar to matrix A if there is a nonsingular matrix P

such that 𝐁 = 𝑷−𝟏 𝑨𝑷.

1 1 1 1 2 −1
Example. Let A = [ ] and P = [ ]. Then 𝑃−1 = [ ] and
−2 4 1 2 −1 1
2 −1 1 1 1 1 2 0
B = 𝑃 −1 𝐴𝑃 = [ ][ ] [ ]= [ ]. Thus B is similar to A.
−1 1 −2 4 1 2 0 3

Properties
1. A is similar to A.
2. If B is similar to A, then A is similar to B.
3. If A is similar to B and B is similar to C, then A is similar to C.

Definition. Matrix A is diagonalizable if it is similar to a diagonal matrix or we can say that

A can be diagonalized.

Theorem. An n x n matrix A is diagonalizable if and only if has n linearly independent

eigenvectors. In this case A is similar to a diagonal matrix D, with 𝑃−1 𝐴𝑃 = 𝐷, whose
diagonal elements are the eigenvalues of A, while P is a matrix whose columns are n,
linearly independent eigenvectors of A.

1 1
Example 1. Let A = [ ] . From page 2 example 4, the eigenvalues are λ = 2, 3 and
𝑟
−2 4
𝑟
eigenvectors: [ ] , [ 2 ]. Consider λ1 = 2 λ2 = 3. If we let r = 1 and r = 2, then we have the
𝑟 𝑟

following eigenvectors
1 1
𝑋1 = [ ] and 𝑋2 = [ ]
1 2

which are linearly independent . Hence A is diagonalizable. Here

1 1 2 −1
P=[ ] and 𝑃−1 = [ ]. Thus
1 2 −1 1

2 −1 1 1 1 1 2 0
𝑃−1 𝐴𝑃 = [ ][ ][ ]=[ ] = 𝐷 . On the other hand, if we let λ1 = 3
−1 1 −2 4 1 2 0 3
1 1
λ2 = 2 then 𝑋1 = [ ] and 𝑋2 = [ ] . Thus
2 1
1 1 −1 1
P=[ ] and 𝑃−1 = [ ] . Hence
2 1 2 −1
−1 1 1 1 1 1 3 0
𝑃−1 𝐴𝑃 = [ ][ ][ ]=[ ] = 𝐷.
2 −1 −2 4 2 1 0 2

MAT121 Module 5 I Linear Algebra I 8

 MODULE 5. EIGENVALUES AND EIGENVECTORS

1 1
Example 2. Let A=[ ] . Then the eigenvalues are λ1 = 1 λ2 = 1. Eigenvectors
0 1
𝑟
associated with λ1 and λ2 are vectors of the form [ ], where r is any nonzero real
0

number. Since A does not have two linearly independent eigenvectors, thus A is not
diagonalizable.

Theorem A matrix is diagonalizable if all the roots of its characteristic polynomial are real
and distinct.

0 0 1
Example 3. Let A = [0 1 2]. Is A diagonalizable?
0 0 1

Solution:

λ 0 − 1
𝑓(λ) = |λ𝐼𝑛 − 𝐴| = | 0 λ − 1 − 2 |
0 0 λ−1

= λ ( λ − 1 ) (λ − 1)
= λ( λ − 1 ) 2

Thus, the characteristic polynomial of A is 𝑓(λ) = λ( λ − 1 ) 2 . Hence the eigenvalues of A

are λ1 = 0, λ2 = 1 and λ3 = 1. Thus λ2 = 1 is an eigenvalues of multiplicity 2.

If λ2 = 1. Then we form a system (1𝐼3 − 𝐴)𝑋 = 0,

1 0 − 1 𝑥1 0
𝑥
[ 0 0 − 2 ] [ 2 ] = [0].
0 0 0 𝑥3 0

Using Gauss-Jordan Elimination, we have

1 0 −1⋮ 0 1 0 0⋮ 0 1 0 0⋮ 0 0
[0 0 −2⋮ 0] → [0 0 1⋮ 0] → [0 0 0⋮ 0] → [𝑟 ]
0 0 0⋮ 0 0 0 0⋮ 0 0 0 1⋮ 0 0

0
Thus all eigenvectors associated with the eigenvalue λ2 = 1 are [𝑟 ], where r is any
0

nonzero real number. The dimension of the solution space of (1𝐼3 − 𝐴)𝑋 = 0 is 1.

There do not exist two linearly independent eigenvectors associated with λ2 = 1 . Hence,

A cannot be diagonalized.

MAT121 Module 5 I Linear Algebra I 9

 MODULE 5. EIGENVALUES AND EIGENVECTORS

0 0 0
Example 4. Let A = [0 1 0]. Is A diagonalizable?
1 0 1

Solution:

The characteristic polynomial of A is 𝑓(λ) = λ( λ − 1 ) 2 . Thus the eigenvalues of A are

λ1 = 0, λ2 = 1 and λ3 = 1. Hence λ2 = 1 is an eigenvalues of multiplicity 2.

If λ1 = 0. Then we form a system (0𝐼3 − 𝐴)𝑋 = 0,

0 0 0 𝑥1 0
[ 0 − 1 0 ] [𝑥2 ] = [0].
−1 0 − 1 𝑥3 0

Using Gauss-Jordan Elimination, we have

0 0 0⋮ 0 0 0 0⋮ 0 𝑟
[ 0 1 0 ⋮ 0] → [0 1 0⋮ 0] → [0]
−1 0 −1⋮ 0 1 0 1⋮ 0 −𝑟

𝑟
Thus all eigenvectors associated with the eigenvalue λ1 = 0 are [ 0 ], where r is any
−𝑟
1
nonzero real number. Thus 𝑋1 = [ 0 ] is an eigenvector associated with λ1 = 0.
−1

If λ2 = 1. Then we form a system (1𝐼3 − 𝐴)𝑋 = 0,

1 0 0 𝑥1 0
[ 0 0 0 ] [𝑥2 ] = [0].
−1 0 0 𝑥3 0

Using Gauss-Jordan Elimination, we have

1 0 0⋮ 0 1 0 0⋮ 0 0
[ 0 0 0⋮ 0] → [0 0 0⋮ 0] → [𝑟 ]
−1 0 0⋮ 0 0 0 0⋮ 0 𝑠

0
Thus all eigenvectors associated with the eigenvalue λ2 = 1 are [𝑟 ], where r and s are
𝑠
0 0
any nonzero real number. Hence we take the eigenvectors 𝑋2 = [1] and 𝑋3 = [0].
0 1

MAT121 Module 5 I Linear Algebra I 10

 MODULE 5. EIGENVALUES AND EIGENVECTORS
𝑋1 𝑋2 𝑋3
1 0 0⋮ 0 1 0 0⋮ 0
Since [ 0 1 0⋮ 0] → [0 1 0⋮ 0] which means 𝑐1 = 𝑐2 = 𝑐3 = 0, then
−1 0 1⋮ 0 0 0 1⋮ 0

1 0 0
𝑋1 = [ 0 ] , 𝑋2 = [1] and 𝑋3 = [0] are linearly independent. Hence A is diagonalizable.
−1 0 1

2 3 0
Example 5. Let A = [0 1 0].
0 0 2

a. Determine the eigenvalues of A and their multiplicities.

b. For each eigenvalues λ in (a), find a basis for the eigenspace of A corresponding to
λ.
c. Is A diagonalizable? If it is not, explain why? If it is, specify a nonsingular P and a
diagonal D such that 𝑃−1 𝐴𝑃 = 𝐷

Solution:

a. The characteristic polynomial of A is 𝑓(λ) = (λ − 1)( λ − 2 ) 2 . Thus the

eigenvalues of A are λ1 = 1, λ2 = 2 and λ3 = 2. Hence λ2 = 2 is an
eigenvalues of multiplicity 2.

b. If λ1 = 1. Then we form a system (𝐼3 − 𝐴)𝑋 = 0,

−1 − 3 0 𝑥1 0
[ 0 0 0 ] [𝑥2 ] = [0].
0 0 − 1 𝑥3 0

Using Gauss-Jordan Elimination, we have

−1 −3 0 ⋮ 0 1 3 0⋮ 0 −3𝑠
[0 0 0 ⋮ 0] → [0 0 0⋮ 0] → [ 𝑠 ]
0 0 −1⋮ 0 0 0 1⋮ 0 0

−3𝑠
Thus the eigenspace associated with the eigenvalue λ1 = 1 is [ 𝑠 ], where s is any
0
−3
nonzero real number. Tthus 𝑋1 = [ 1 ] is the basis for the eigenspace associated with
0
λ1 = 1.

If λ2 = 2. Then we form a system (2𝐼3 − 𝐴)𝑋 = 0,

MAT121 Module 5 I Linear Algebra I 11

 MODULE 5. EIGENVALUES AND EIGENVECTORS

0 − 3 0 𝑥1 0
[ 0 1 0 ] [𝑥2 ] = [0].
0 0 0 𝑥3 0

Using Gauss-Jordan Elimination, we have

0 −3 0⋮ 0 0 0 0⋮ 0 𝑟
[0 1 0⋮ 0] → [0 1 0⋮ 0] → [0]
0 0 0⋮ 0 0 0 0⋮ 0 𝑠

𝑟
Thus the eigenspace associated with the eigenvalue λ2 = 2 is [0], where r and s are any
𝑠
1 0
nonzero real number. Thus 𝑋2 = [0] and 𝑋2 = [0] are the basis for the eigenspace
0 1
associated with λ2 = 2.

𝑋1 𝑋2 𝑋3
−3 1 0⋮ 0 1 0 0⋮ 0
c. Since [ 1 0 0⋮ 0] → [0 1 0⋮ 0] which means 𝑐1 = 𝑐2 = 𝑐3 = 0, then
0 0 1⋮ 0 0 0 1⋮ 0

−3 1 0
𝑋1 = [ 1 ] , 𝑋2 = [0] and 𝑋3 = [0] are linearly independent. Hence A is
0 0 1

𝑋1 𝑋2 𝑋3
−3 1 0 0 1 0
diagonalizable. Then P=[ 1 0 0] and 𝑃−1 = [1 3 0]
0 0 1 0 0 1

(Note in finding the inverse [𝑃 ⋮ 𝐼3 ] → [ 𝐼3 ⋮ 𝑃−1 ] . Note that

λ1 0 0 1 0 0
D = [ 0 λ2 0 ] = [0 2 0] . Hence
0 0 λ3 0 0 2

0 1 0 2 3 0 −3 1 0 1 0 0
𝑃−1 𝐴𝑃 = [1 3 0] [0 1 0] [ 1 0 0 ] = [0 2 0] = 𝐷
0 0 1 0 0 2 0 0 1 0 0 2

MAT121 Module 5 I Linear Algebra I 12

 MODULE 5. EIGENVALUES AND EIGENVECTORS

Exercises
A. Tell whether the following matrices are diagonalizable or not.

1 2 3
1 4
1. [ ] 4. [0 −1 2]
1 −2
0 0 2

3 1 0
1 0
2. [ ] 5. [0 3 1]
−2 1
0 0 3

1 1 −2
3. [4 0 4]
1 −1 4

B. Find for each matrix A, if possible, a nonsingular matrix P such that

𝑃−1 𝐴𝑃 = 𝐷.

3 −2 1
0 −1
1. [ ] 4. [0 2 0]
2 3
0 0 0

4 2 3 1 1 2
2. [ 2 1 2] 5. [0 1 0]
−1 −2 0 0 1 3

1 2 3
3. [0 1 0]
2 1 2

C. Find for each matrix A.

a. Determine the eigenvalues of A and their multiplicities.

b. For each eigenvalues λ in (a), find a basis for the eigenspace of A corresponding
to λ.
c. Is A diagonalizable? If it is not, explain why? If it is, specify a nonsingular P and a
diagonal D such that 𝑃−1 𝐴𝑃 = 𝐷

2 1 1 1 2 2 3 4
1 2 1 1 0 2 3 2
1. [ ] 2. [ ]
1 1 2 1 0 0 1 1
0 0 0 1 0 0 0 1

MAT121 Module 5 I Linear Algebra I 13

 MODULE 5. EIGENVALUES AND EIGENVECTORS

Diagonalization of Symmetric Matrices

Remarks
1. All roots of the characteristic polynomial of symmetric matrix are real numbers.
2. If A is a symmetric matrix all of whose eigenvalues are distinct, then A is
diagonalizable.
3. If A is a symmetric matrix, the eigenvectors that belong to distinct eigenvalues of A
are orthogonal.

Definition. A nonsingular matrix A is called orthogonal if 𝐴−1 = 𝐴𝑇 .

Remark: A nonsingular matrix A is orthogonal if 𝐴𝑇 𝐴 = 𝐼𝑛 .

2 −2 1
3 3 3
2 1 −2
Example. Show that A =
3 3 3
is an orthogonal matrix.
1 2 2
[3 3 3]

Solution:

2 2 1
3 3 3
𝑇 −2 1 2
𝐴 =
3 3 3
1 −2 2
[ 3 3 3]

2 2 1 2 −2 1
3 3 3 3 3 3 1 0 0
−2 1 2 2 1 −2
𝐴𝑇 𝐴 = = [0 1 0] = 𝐼3 .
3 3 3 3 3 3
1 −2 2 1 2 2 0 0 1
[3 3 3] [3 3 3]

Hence A is an orthogonal matrix.

Theorem. The nxn matrix A is orthogonal if and only if the columns (and rows) of A form
an orthogonal set of vectors in 𝑅 𝑛 .

Theorem. If A is a symmetric nxn matrix., then there exists an orthogonal matrix P such
that
𝑃−1 𝐴𝑃 = D, a diagonal matrix. The eigenvalues of A lie on the main diagonal of D.

0 0 −2
Example 1. Let A = [ 0 −2 0 ] . Then the characteristic polynomial of A is
−2 0 3

MAT121 Module 5 I Linear Algebra I 14

 MODULE 5. EIGENVALUES AND EIGENVECTORS

𝑓(λ) = (λ + 2)(λ − 4)(λ + 1). Thus the eigenvalues are λ1 = −2, λ2 = 4, λ3 = −1. Hence

0 −1 2
the eigenvectors are 𝑋1 = [1], 𝑋2 = [ 0 ], and 𝑋3 = [0]. Since
0 2 1
𝑋1 ∙ 𝑋2 ∙ 𝑋3 = 0(−1)(2) + 1(0)(0) + 0(2)(1) = 0, { 𝑋1 , 𝑋2 , 𝑋3 } is orthogonal. If we normalize
1 2
0 − √5 √5
{ 𝑋1 , 𝑋2 , 𝑋3 }, we find that T = {[1] , [ 0 ] , [ 0 ]} is an orthonormal set of vectors. The
0 2 1
√5 √5
matrix P such that 𝑃−1 𝐴𝑃 is diagonal is the matrix whose columns are the vectors in T.
1 2 0 0
0 − √5 √5 1 1 2
Thus P = [1 0 0 ]. Since P is orthogonal, 𝑃 −1 = 𝑃𝑇 = [−√5 0 √5]. Hence,
0 2 1 2
0 1
√5 √5 √5 √5

0 0 1 2
1 1 2 0 0 −2 0 − √5 √5 −2 0 0
𝑃−1 𝐴𝑃 = 𝑃𝑇 𝐴𝑃 = [−√5 0 √5] [ 0 −2 0 ] [1 0 0 ] = [ 0 4 0 ] = 𝐷 .
2 1
0 −2 0 3 0 2 1 0 0 −1
√5 √5 √5 √5

0 2 2
.Example 2. Diagonalize matrix A = [2 0 2] by an orthogonal matrix.
2 2 0

Solution:

The characteristic polynomial of A is 𝑓(λ) = (λ + 2)2 (λ − 4), so the eigenvalues are

λ1 = −2, λ2 = −2, λ3 = 4. That is, -2 is an eigenvalue whose multiplicity is 2.

To find the eigenvectors associated with λ1 and λ2 , we solve the homogeneous linear
system (−2𝐼3 − 𝐴)𝑋 = 0:
−2 − 2 − 2 𝑥1 0
solution space: [−2 − 2 − 2] [𝑥2 ] = [0]
−2 − 2 − 2 𝑥3 0

A basis for the solution space is consist of the eigenvectors

−1 −1
𝑋1 = [ 1 ] and 𝑋2 = [ 0 ]
0 1
Since 𝑋1 ∙ 𝑋2 = (−1)(−1) + 1(0) + (0)1 ≠ 0, 𝑋1 𝑎𝑛𝑑 𝑋2 are not orthogonal.

We can use the Gram-Schmidt process to obtain an orthonormal basis for the solution
space above .

MAT121 Module 5 I Linear Algebra I 15

 MODULE 5. EIGENVALUES AND EIGENVECTORS

−1
Let 𝑌1 = 𝑋1 = [ 1 ]
0
and
1
−1 −1 −2
𝑋2 ∙𝑌1 −1(−1)+1(0)+0(1)
𝑌2 = 𝑋2 − ( 𝑌 ∙𝑌 ) 𝑌1 = [ 0 ] − ( (−1)2+12 +02 ) [ 1 ] = [− 1].
1 1 2
1 0
1
−1
Let 𝑌2 = 2𝑌2 = [−1] . The set {𝑌1 , 𝑌2 ∗ } is an orthogonal set of vectors.
∗

2
Normalizing these eigenvectors, we have
1
−
1 −
−1 −1 √6
√2
𝑌 1 𝑌∗ 1 1
𝑍1 = |𝑌1 | = [ 1 ]=[ 1 ] and 𝑍2 = |𝑌2 ∗ | = [−1] = − 6.
1 √2 2 √6 √
0 √2 2 2
0 [ √6 ]
The set {𝑍1 , 𝑍2 } is an orthonormal basis of eigenvectors of A for the solution space above.

The basis for the solution space of (4𝐼3 − 𝐴)𝑋 = 0:

4 − 2 − 2 𝑥1 0
[ −2 4 − 2 ] [𝑥2 ] = [0],
−2 − 2 4 𝑥3 0
1
1 1 √3
1 1
is consist of 𝑋3 = [1] and normalizing this vector, we have 𝑍3 = [1] = as an
√3 √3
1 1 1
[√3]
orthonormal basis for (4𝐼3 − 𝐴)𝑋 = 0.

Since eigenvectors associated with distinct eigenvalues are orthogonal, 𝑍3 is orthogonal to

both 𝑍1 and 𝑍2 . Thus the set {𝑍1 , 𝑍2, , 𝑍3 } is an orthonormal basis of 𝑅 3 consisting of
eigenvectors of A. Hence
−1 1
−1
√6 √3
√2 −1 1
P= 1 . (Note that 𝑃−1 = 𝑃𝑇 )
√6 √3
√2 2 1
[0 √6 √3]

−1 1 −1 1
√2 √2
0 −1
√6 √3
−1 −1
2 0 2 2 √2 −1 1
−2 0 0
Therefore, 𝑃−1 𝐴𝑃 = √6 [2 0 2] 1 = [ 0 −2 0] = 𝐷.
√6 √6 1 √6 √3
1 1 2 2 0 √2 2 1 0 0 4
[√3 √3
√3] [0 √6 √3]

MAT121 Module 5 I Linear Algebra I 16

 MODULE 5. EIGENVALUES AND EIGENVECTORS

1 2 0 0
2 1 0 0
Example 3. Diagonalize A=[ ] and find an orthogonal matrix P such that
0 0 1 2
0 0 2 1
𝑃−1 𝐴𝑃 is diagonal.

Solution:

The characteristic polynomial of A is 𝑓(λ) = (λ + 1)2 (λ − 3)2 . so the eigenvalues are

λ1 = −1, λ2 = −1, λ3 = 3 and λ4 = 3. That is, -1 and 3 are an eigenvalue whose
multiplicity is 2.

The basis for the solution space of (−1𝐼3 − 𝐴)𝑋 = 0 consist of the eigenvectors

1 0
−1 0
𝑋1 = [ ] and 𝑋2 = [ ] which are orthogonal.
0 1
0 −1
Normalizing 𝑋1 and 𝑋2 , we have
1 0
√2 0
1 1
𝑍1 = − and 𝑍2 = as an orthonormal basis of eigenvectors for the solution
√2 √2
0 1
[ 0 ] [− √2]

space (−1𝐼3 − 𝐴)𝑋 = 0.

The basis for the solution space of (3𝐼3 − 𝐴)𝑋 = 0 consist of the eigenvectors

1 0
1 0
𝑋3 = [ ] and 𝑋4 = [ ] which are orthogonal.
0 1
0 1

Normalizing 𝑋3 and 𝑋4 , we have

1 0
√2 0
1 1
𝑍3 = √2 and 𝑍4 = as an orthonormal basis of eigenvectors for the solution
√2
0 1
[0] [√2]

space (3𝐼3 − 𝐴)𝑋 = 0.

Since eigenvectors associated with distinct eigenvalues are orthogonal, thus {𝑍1 , 𝑍2 , 𝑍3 , 𝑍4 }
is an orthonormal basis for 𝑅 4 consisting of eigenvectors of A. Hence, matrix

MAT121 Module 5 I Linear Algebra I 17

 MODULE 5. EIGENVALUES AND EIGENVECTORS
1 0 1 0
√2 0 √2 0
1 1 1 1
P= − 2 .
√ √2 √2 √2
0 1 0 1
[ 0 − 2 ]
√ 0 √2

Remark. Suppose that A is an nxn matrix for which we can find an orthogonal matrix P
such that 𝑃−1 𝐴𝑃 is a diagonal matrix D. Thus 𝑃 −1 𝐴𝑃 = 𝐷 or A = PD𝑃−1. Since
𝑃−1 = 𝑃𝑇 , we can write A = PD𝑃𝑇 . Then A𝑇 = (PD𝑃𝑇 )𝑇 = (𝑃𝑇 )𝑇 𝐷 𝑇 𝑃𝑇 = 𝑃𝐷𝑃𝑇 = 𝐴
(𝐷 = 𝐷 𝑇 , since D is a diagonal matrix). Thus A is symmetric.

Exercises

Diagonalize the following matrix and find an orthogonal matrix P such that 𝑃−1 AP is
diagonal.

0 0 0 0
−1 2 2
0 0 0 0
1. [ 2 −1 2 ] 7. [ ]
0 0 0 0
2 2 −1
0 0 1 0

2 2 0 0
0 −1 −1
2 2 0 0
2. [−1 0 −1] 8. [ ]
0 0 2 2
−1 −1 0
0 0 2 2

0 0 0 1
1 1 0
0 0 0 0
3. [1 1 0] 9. [ ]
0 0 0 0
0 0 1
1 0 0 0

1 0 0
4. [0 1 1]
0 1 1

2 1 1
5. [1 2 1]
1 1 2

−3 0 −1
6. [ 0 −2 0 ]
−1 0 −3

MAT121 Module 5 I Linear Algebra I 18