LECTURE NO TWO

BANACH ALGEBRA, NOTATIONS AND BASIC DEFINITIONS

SAHAR MOHAMED ALI ABOU BAKR

we shall study a general class of algebras, namely, Banach algebras. These are
of interest in their own right. We start with the following:

Definition 1. A complex algebra is a nonempty set A admitting three operations,

vector addition +, scalar multiplication ., and vector multiplication o
+ : A × A → A, +(x, y) = x + y ∀ x, y ∈ A,

. : C × A → A, .(λ, x) = λ.x := λ x ∀ x ∈ A and λ ∈ C,

and
o : A × A → A, o(x, y) = xoy := xy ∀ x, y ∈ A
respectively, such that
(1) A is a linear space with respect to the operations of vector addition and
scalar multiplication;
(2) A is a ring with respect to vector addition and vector multiplication; that
is, the multiplication operation is associative,
x(yz) = (xy)z,
and the multiplication is distributed on addition
x(y + z) = xy + xz and (x + y)z = xz + yz, ∀ x, y, z ∈ A.
(3) A satisfies the scalar multiplication property, for a scalar λ ∈ C and any
two vectors x and y ∈ A,
λ(xoy) = (λx)oy = xo(λy).
Simply we write
λ(xy) = (λ x)y = x(λ y) ∀ x, y ∈ A and λ ∈ C.
Definition 2. A complex algebra A is said to be a commutative algebra if and
only if
xy = yx ∀x, y ∈ A,
and it said to be unital algebra if and only if it contains a multiplicative identity
(unit), i.e, there is an element 1 ∈ A such that
xo1 = 1ox ∀x ∈ A, simply we write x1 = 1x ∀x ∈ A.
Remark. If an algebra A has identity, then it is unique. In fact, if 1 and 1′ are
two identity elements in A, then the property of identity implies that
1 = 1 1′ = 1 ′ .
Such a unique identity (unit) will be denoted by 1A .
1
2 SAHAR MOHAMED ALI ABOU BAKR

Definition 3. Let A be a complex unital algebra and x ∈ A. Then

(1) x is said to have a right inverse if and only if there is y ∈ A such that
xy = 1A .
(2) x is said to have a left inverse if and only if there is z ∈ A such that zx = 1A .
(3) x is said to be invertible if and only if it has both left and right inverses.
Remark. If A is a unital algebra and x ∈ A is an element that has both right and
left inverses y and z respectively, xy = 1A and zx = 1A , then
y = 1A y = (zx)y = z(xy) = z1A = z,
hence y = z. That is, if the element has right and left inverses, then they are
necessarily equal.
According to the last remark, we have the following equivalent definition:
Definition 4. An element x of a complex unital algebra A is said to be invertible
if and only there exists an element y ∈ A such that
xy = yx = 1A .
Remark. If x is invertible element in a unital complex algebra A, then it has a
unique inverse. Indeed; if y and y ′ are two inverses of x, then y ′ x = x y ′ = 1A and
xy = y x = 1A , consequently
y = y1A = y(x y ′ ) = (yx) y ′ = 1A y ′ = y ′ .
Such a unique inverse if exists will be denoted by x−1 .
Lemma 1. If A is a unital algebra, x and y are two commutative elements, xy = yx
and xy is invertible, then both of x and y are invertible.
Proof. If x and y are commutative, then xy = yx and if xy is invertible, then the
associative property of multiplication operation gives
1A = (xy)(xy)−1 = x[y(xy)−1 ]
and
1A = (xy)−1 (xy) = (xy)−1 (yx) = [(xy)−1 y]x,
hence x has the right inverse [y(xy)−1 ] and it has the left inverse [(xy)−1 y]. Since
x has both left and right inverses, the element x should be invertible. Using the
commutativity of x and y with the associative property of multiplication operation,
we confirmed that the left and the right inverses of x are equal as follows:
1A = [(xy)−1 y]x = [[(xy)−1 y]1A ]x
= [[(xy)−1 y][(xy)(xy)−1 ]]x = [[(xy)−1 y(xy)][(xy)−1 ]]x
= [[(xy)−1 (yx))][y(xy)−1 ]]x = [[(xy)−1 (xy))][y(xy)−1 ]]x because xy = yx
= [[1A ][y(xy)−1 ]]x = [y(xy)−1 ]x.
Hence x is invertible and x−1 = [(xy)−1 y] = [y(xy)−1 ]. Similar conclusion for y,
y −1 = x(xy)−1 = (xy)−1 x.

 BANACH ALGEBRA 3

One way to show that an element is not invertible is to show that it’s a zero
divisor. An element x is a zero divisor if it divides 0, the identity element with
respect to addition, i.e. if there is some element y ̸= 0 such that xy = 0 or
yx = 0. A zero divisor can’t be invertible because if xy = 0 and x−1 existed, then
y = x−1 xy = x−1 0 = 0. However, in many algebra (rings in particular) there are
elements that are not invertible without being zero divisors.
Definition 5. If (A, ∥.∥) is normed space and A is complex normed algebra, then
(A, ∥.∥) is said to be a normed algebra if and only if the norm ∥.∥ satisfies the norm
sub-multiplicative property, that is; for any two vectors x and y ∈ A,
∥x oy∥ ≤ ∥x∥ ∥y∥.
Simply, we write
∥x y∥ ≤ ∥x∥ ∥y∥.
Definition 6. If (A, ∥.∥) is complex normed algabra, then A is said to be a Banach
algebra if and only if (A, ∥.∥) is complete normed space. That is; if and only if
every Cauchy sequence in A converges to some limit point in A.
Lemma 2. If (A, ∥.∥) is a unital normed algebra with unit 1A , then ∥1A ∥ ≥ 1.
Proof. Since 1A = 1A 1A and the norm on A is norm sub-multiplicative, we have
∥1A ∥ = ∥1A 1A ∥ ≤ ∥1A ∥∥1A ∥,
this implies 1 ≤ ∥1A ∥. 

Lemma 3. If (A, ∥.∥) is a normed algebra, then the multiplicative operation o :

A × A → A is jointly continuous.
Proof. If {xn }n∈N and {yn }n∈N are two convergent sequences, hence both are
bounded, we set M = supn∈N ∥xn ∥, xn →n→∞ x and yn →n→∞ y for some two
elements x, y ∈ A, then using the norm sub-multiplicative property, we have
∥xn yn − xy∥ = ∥xn yn − xn y + xn y − xy∥ = ∥xn (yn − y) + (xn − x)y∥
≤ ∥xn (yn − y)∥ + ∥(xn − x)y∥
≤norm sub-multiplicative property ∥xn ∥ ∥yn − y∥ + ∥xn − x∥ ∥y∥
≤ [sup ∥xn ∥] ∥yn − y∥ + ∥xn − x∥ ∥y∥ = M ∥yn − y∥ + ∥xn − x∥ ∥y∥
n∈N
→n→∞ M × 0 + 0 × ∥y∥ = 0.
This shows that limn→∞ xn yn = x y and multiplicative operation is jointly contin-
uous. 

Example 1. The set of all complex numbers C with the usual algebraic operations
and with the absolute value as a norm on C is a commutative Banach algebra
with identity 1. This is the only Banach algebra up to isomorphism in which
every nonzero element has an inverse. Indeed;
√ if λ ∈ C is an arbitrarily nonzero
complex number λ = a + bι, where ι = −1, then it can be easily verified that
λ−1 = λ1 = √a2a+b2 + √a−b
2 +b2
ι. (Such algebras are known as division algebras, every
non-zero element in A is invertible).
4 SAHAR MOHAMED ALI ABOU BAKR

Example 2. The set of all square matrices of complex entries A = Mn×n =

{{aij }n,n
i,j=1 : aij ∈ C} with the usual operations of addition, scalar multiplica-
tion, and multiplication of matrices is unital algebra, the identity element 1Mn×n
of Mn×n is the matrix its main diagonal’s entries all are the number one while all
the other entries are zeros.
 
1 0 ... 0 0
 0 1 0 ... 0 
 
 
1Mn×n =  ... ... . . . . . . 0 
 
 0 0 ... 1 0 
0 0 ... 0 1
The matrix A ∈ Mn×n has an inverse if and only if it has a non-zero determinant,
det(A) ̸= 0, starting from Gauss-Jordan method, there are many method that can
be used to determined this inverse. For the case of 2 × 2 matrices A = M2×2 , we
have ( ) ( )
0 0 1 0
0M2×2 = , 1M2×2 =
0 0 0 1
and
( ) ( )
1 a22 −a12 a11 a12
A−1 = , where A = , |A| = a11 a22 −a12 a21 ̸= 0.
|A| −a21 a11 a21 a22
It is not commutative, for example
( ) ( ) ( ) ( ) ( ) ( )
1 0 0 1 0 1 0 1 1 0 0 0
× = , while × =
0 0 0 0 0 0 0 0 0 0 0 0
Since for any two square matrices, A, B ∈ M2×2 we have det(A B) = det(A)det(B),
then the product of two matrices is invertible if and only if both of the two matrices
are invertible. Any of the following norms on Mn×n is a sub-multiplicative norm
(1) An important example of matrix norms is given by the norm induced by a
vector norm, that is; the norm of any element A = {aij }n,n
i,j=1 ∈ Mn×n is
the norm of its representative linear bounded operator, A : Cn → Cn ,
∥A∥ = sup{∥A(x)∥ : x ∈ Cn , ∥x∥ = 1}.
If (Cn , ∥.∥∞ ), then

n ∑
n
∥A∥∞ = ∥{aij }n,n
i,j=1 ∥∞ = max |aij |.
i=1
j=1

If (Cn , ∥.∥1 ), then

n ∑
n
∥A∥1 = ∥{aij }n,n
i,j=1 ∥1 = max |aij |.
j=1
i=1

(2) The Frobenius norm defined by

√
∥A∥F := tr(A∗ A),
where tr(.) is the trace of the matrix and A∗ is its adjoint. Indeed, if A,
B, and C are arbitrarily elements in Mn×n and λ1 ≥ · · · ≥ λn ≥ 0 denotes
 BANACH ALGEBRA 5

the eigenvalues of the Hermitian matrix C ∗ C (or A∗ A or B ∗ B), then

v
u n
√ u∑ √
(0.1) ∥C C∥F = tr(C C) = t
∗ ∗ λ2i ≤ tr(C ∗ C) = ∥C∥F .
i=1

√ √
∥A B∥F = tr((A B)∗ (A B)) = tr((B ∗ A∗ ) (A B))
√ √ √
(0.2) = tr(B ∗ B A∗ A) ≤ tr(A∗ A) tr(B ∗ B)
≤ ∥A∗ A∥F ∥B ∗ B∥F .

Substituting from (0.1) in (0.2), we get

∥A B∥F ≤ ∥A∥F ∥B∥F .
This proves the sub-multiplicativity property of the Frobenius norm.
For such defined sub-multiplicative norms, the space Mn×n is Banach space, hence,
it is unital Banach algebra.
Example 3. Let (X, ∥.∥) be complex Banach space, and B(X) be the Banach space
of bounded (not necessarily a linear) self operators on X equipped with the usual
linear structures (usual algebra operations of addition and scalar multiplication)
and operator norm,
+ : B(X) × B(X) → B(X), (+(A, B))(x) := (A + B)(x) = A(x) + B(x).

. : C × B(X) → B(X), (.(λ, A))(x) := (λ + A)(x) = λ(A(x)).

∥A∥ = sup{∥A(x)∥ : x ∈ X, ∥x∥ = 1} ∀A ∈ B(X).

In addition, the composition operation of operators is taken to be the multiplication
operation on B(X) defined as:
o : B(X) × B(X) → B(X), (o(A, B))(x) := (AoB)(x) = A(B(x)).
Then B(X) with these three operations and operator norm is a unital Banach alge-
bra. Indeed, the operator norm is satisfying the norm sub-multiplicative property
with respect to composition because
∥AoB∥ = sup{∥(AoB)(x)∥ : x ∈ X, ∥x∥ = 1}
= sup{∥(A(B((x))∥ : x ∈ X, ∥x∥ = 1}
= sup{∥A∥∥B(x)∥ : x ∈ X, ∥x∥ = 1}
≤ ∥A∥ sup{∥B(x)∥ : x ∈ X, ∥x∥ = 1}
= ∥A∥ ∥B∥ ∀A, B ∈ B(X)

The identity element 1B(X) ∈ B(X) is the identity operator on X defined by

1B(X) : X → X, 1B(X) (x) = x ∀x ∈ X.

The operator A ∈ B(X) has an inverse A−1 if it is bijective and the closed graph
theory ensures that every linear one-to-one operator (equivalently; ker(A) = {0})
is invertible, [the inverse A−1 is restricted on range A].
6 SAHAR MOHAMED ALI ABOU BAKR

Example 4. Let (X, ∥.∥) be a unital Banach algebra with identity 1X and B(X)
be the Banach space of all bounded self operators on X equipped with the usual
linear structures (usual algebra operations of addition and scalar multiplication)
and operator norm,
+ : B(X) × B(X) → B(X), (+(A, B))(x) := (A + B)(x) = A(x) + B(x).
. : C × B(X) → B(X), (.(λ, A))(x) := (λ + A)(x) = λ(A(x)).
∥A∥ = sup{∥A(x)∥ : x ∈ X, ∥x∥ = 1} ∀A ∈ B(X).
The third operation on B(X) that we add is the point-wise multiplication operation
of operators on B(X) as:
⋄ : B(X) × B(X) → B(X), (⋄(A, B))(x) := (A ⋄ B)(x) = A(x)B(x).
Then B(X) with these three operations and operator norm is a unital Banach alge-
bra. Indeed, this point-wise multiplication operation has the following properties:
(1) Associative:
[(A ⋄ B) ⋄ C](x) = [(A ⋄ B)(x)]C(x) = [A(x)B(x)]C(x)
= A(x)[B(x)C(x)] because multiplication on X is associative
= A(x)[(B ⋄ C)(x)]
= [A ⋄ (B ⋄ C)](x) ∀x ∈ X.
This proves that, (A ⋄ B) ⋄ C = A ⋄ (B ⋄ C).
(2) Distributive:
[(A + B) ⋄ C](x) = (A + B)(x)C(x) = [A(x) + B(x)]C(x)
= A(x)C(x) + B(x)C(x) because multiplication on X is distributive
= (A ⋄ C)(x) + (B ⋄ C)(x)
= [(A ⋄ C) + (B ⋄ C)](x) ∀x ∈ X.
This proves that, (A + B) ⋄ C = (A ⋄ C) + (B ⋄ C). Similarly; A ⋄ (B + C) =
(A ⋄ B) + (A ⋄ C) for every A, B, and C in B(X).
(3) Scalar multiplication property, for a scalar λ ∈ C and arbitrarily two vectors
A and B in B(X), we have
[λ(A ⋄ B)](x) = λ[(A ⋄ B)(x)] = λ[(A(x)B(x)].
= (λA(x))C(x) because of the scalar multiplication property of X
= (λA)(x) C(x) = [(λA) ⋄ C)(x) ∀x ∈ X.
This proves that, λ(A ⋄ B) = (λA) ⋄ B. Similarly; λ(A ⋄ B) = A ⋄ (λB) for
every λ ∈ C and A, B ∈ B(X).
The operator norm is satisfying the norm sub-multiplicative property with respect
to the operation ⋄. Indeed;
∥A ⋄ B∥ = sup{∥(A ⋄ B)(x)∥ : x ∈ X, ∥x∥ = 1} = sup{∥A(x)B(x)∥ : x ∈ X, ∥x∥ = 1}
≤ sup{∥A(x)∥∥B(x)∥ : x ∈ X, ∥x∥ = 1}[norm sub-multiplicative on X]
≤ sup{∥A(x)∥ : x ∈ X, ∥x∥ = 1} sup{∥B(x)∥ : x ∈ X, ∥x∥ = 1}
= ∥A∥ ∥B∥ ∀ A, B ∈ B(X)
The identity element 1B(X) with respect to point-wise multiplication is given by
1B(X) : X → X, 1B(X) (x) = 1X ∀ x ∈ X.
 BANACH ALGEBRA 7

Indeed; 1B(X) ⋄ A = A ⋄ 1B(X) = A for every A ∈ B(X) because for every x ∈ X,

we have
(1B(X) ⋄ A)(x) = 1B(X) (x) A(x) = 1X A(x) = A(x)
and
[A ⋄ 1B(X) ](x) = A(x) 1X = A(x) 1B(X) (x) = A(x).
−1
If [A(x)] exists for every x ∈ X, then the operator B : X → X defined by
B(x) := [A(x)]−1 is satisfying the following:
(B ⋄ A)(x) = B(x)A(x) = [A(x)]−1 A(x) = 1X = 1B(X) (x)
= A(x)[A(x)]−1 = A(x)B(x) = (A ⋄ B)(x) ∀ x ∈ X.
Hence B is the point-wise multiplicative inverse of the operator A, B = A−1 : X →
X. The operator A has no inverse if and only if there exists x ∈ X such that A(x)
has no inverse. Note that for linear operators A(x) = 0 has no inverse for every
x ∈ Ker(A).

In particular, for B([0; 1]), the Banach space of continuous complex-valued func-
tions defined on the interval [0, 1] equipped with the sup-norm, namely, ∥f ∥ =
supx∈[0,1] |f (x)|, and with multiplication defined point-wise (f g)(x) = f (x)g(x) for
x ∈ [0, 1] is a commutative unital Banach algebra; the constant function 1 is its
identity element.
The following are examples of algebras without identity:
Example 5. Suppose that X is a locally compact Hausdorff space. Let A = C0 (X)
denote the space of all complex valued continuous functions that vanish at ∞,
C0 (X) := {f : f : X → C, f continuous and vanishes at ∞ }.
We say f ∈ C(Ω) vanishes at ∞ if and only if for every ϵ > 0, there exists a compact
subset Kϵ of Ω such that |f (t)| < ϵ for every t ∈ Kϵc the complement of Kϵ . Then it
can be verified easily that C0 (X) with point-wise multiplication and the supremum
norm is a commutative Banach subalgebra of B(X) and C0 (X) is unital if and only
if X is compact in which case C0 (X) = B(X). In fact; the identity element 1B(X)
is vanishing at ∞ if and only if X is compact. In particular; the Banach algebra
C0 (C) does not contain the identity, 1B(C) ∈/ C0 (C).
Example 6. Let (A, +, ., ×) denote the set consists of all 2 × 2 complex matrices
each of which at least of its rows consists of zero entries, with the usual algebraic
structure and any of the norms given in example (2),
( ) ( )
a b 0 0
A={A:A= or A = where a, b ∈ C }
0 0 a b
Clearly A is a Banach algebra without identity
( )
1 0
∈/ A.
0 1
Example 7. Let (A, +, ., ×) denote the set consists of all even integers, A := {2x :
x ∈ Z}, where +, ., × are the usual operations of addition, scalar multiplication, and
multiplication of real numbers, (the usual algebraic structure). Then (A, +, ., ×) is
commutative algebra that contains no identity element.
8 SAHAR MOHAMED ALI ABOU BAKR

Example 8. Let (A, +, .) denote the linear space of all real valued integrable func-
tions, ∫ ∞
A := {f : |f (x)| dx < ∞ },
−∞
where +, . are the usual algebraic structure (usual operations of addition and scalar
multiplications of functions, respectively. Endow A with the integral norm,
∫ ∞
∥f ∥ = |f (x)| dx.
−∞
Then (A, ∥.∥) is a normed space (not Banach). Let ∗ be the convolution operation
on A, ∫ ∞
(f ∗ g)(t) =: f (x) g(t − x) dx ∀f, g ∈ A.
−∞
Thus (A, +, ., ∗, ∥.∥) is a non commutative normed algebra, such normed algebra
contains no identity, otherwise; if there is element I ∈ A we would have f ∗ I = f =
I ∗ f ∀f ∈ A,
∫ ∞
(f ∗ I)(t) = f (x)I(t − x)dx = f (t) = (I ∗ f )(t)
−∞
∫ ∞
= f (t − x)I(x)dx ∀ t ∈ R, ∀f ∈ A.
−∞
This is a contradiction.
0.1. Problems. Give two example of two unital Banach algebras and two examples
of Banach algebra without identity elements.
(Sahar Mohamed Ali Abou Bakr) Department of Mathematics, Faculty of Science, Ain
Shams University, Cairo, Egypt.
E-mail address: Sahar Mohamed Ali Abou Bakr: saharm ali@yahoo.com
E-mail address: Sahar Mohamed Ali Abou Bakr: saharm ali@sci.asu.edu.eg