1 Introduction/Preliminaries

Many mathematicians have studied the q-Bernoulli, q-Euler polynomials and related topics (see [111]). It is worth that Açikgöz et al. [1] give a new approach to the q-Bernoulli polynomials and the q-Bernstein polynomials and show some properties. That is, Açikgöz et al. introduced a new generating function related the q-Bernoulli polynomials and gave a new construction of these polynomials related to the second kind Stirling numbers and the q-Bernstein polynomials in [1]. The purpose of this paper is to redefine a generating function related the q-Bernoulli polynomials and numbers and correct its wrong properties and rebuild its theorems.

In this paper, we assume that q ( ) is indeterminate with |q| < 1. The q-number is defined by [ x ] q = q x - 1 q - 1 (see [49]).

It is known that the Bernoulli polynomials are defined as

t e t - 1 e x t = n = 0 B n ( x ) t n n ! for | t | < 2 π
(1.1)

and that B n (0) = B n are called the Bernoulli numbers.

The recurrence formula for the classical Bernoulli numbers B n is as follows:

B 0 = 1 and ( B + 1 ) n - B n = 0 if n > 0 .
(1.2)

The q-extension of the following recurrence formula for the Bernoulli numbers is given by

B 0 , q = 1 and q ( q B + 1 ) n - B n , q = 1 if n = 1 0 if n > 1
(1.3)

with the usual convention of replacing B q n by B n,q (see [2, 4]).

2 On the q-Bernoulli polynomials and numbers

In this section, we first recall the q-Bernoulli polynomials and numbers, then indicate the ambiguities on the Açikgöz et al. [1]'s definition for the q-Bernoulli polynomials and redefine it. Counter-examples show that some properties are incorrect. Specially, these examples show that the concept on the generating function of the q-Bernoulli polynomials is unreasonable.

Definition 2.1 (Açikgöz et al. [1]) For q with |q| < 1, let us define the q-Bernoulli polynomials as follows,

D q ( t , x ) = - t y = 0 q y e [ x + y ] q t = n = 0 B n , q ( x ) t n n ! .
(2.1)

Note that

lim q 1 D q ( t , x ) = t e t - 1 e x t = n = 0 B n ( x ) t n n ! for | t | < 2 π ,
(2.2)

where B n (x) are the classical Bernoulli polynomials.

In the special case x = 0, B n,q (0) = B n,q are called the q-Bernoulli number.

That is,

D q ( t ) = D q ( t , 0 ) = - t y = 0 q y e [ y ] q t = n = 0 B n , q t n n ! .
(2.3)

Remark 2.2 Definition 2.1 (Açikgöz et al. [1]) is unreasonable, since it is not the generating functions of the q-Bernoulli polynomials and numbers. This can be seen the following counter-examples.

Counter-example 2.3 If we take t = 0 in (2.2) of Definition 2.1 (Açikgöz et al. [1]), then we have lim q→1 D q (0, x) = 0. But lim t 0 t e t - 1 e x t =1 does not hold in the sense of Definition 2.1 (Açikgöz et al. [1]).

Counter-example 2.4 From (2.1) of Definition 2.1 (Açikgöz et al. [1]),

D q ( t , x ) = n = 0 B n , q ( x ) t n n ! (1) = B 0 , q ( x ) + n = 1 B n , q ( x ) t n n ! , (2) (3)
(2.4)

and

D q ( t , x ) = - t y = 0 q y e [ x + y ] q t (1) = - t y = 0 q y n = 0 [ x + y ] q n t n n ! (2) = n = 0 - 1 ( 1 - q ) n l = 0 n ( n l ) ( - 1 ) l q l x y = 0 q ( l + 1 ) y t n + 1 n ! (3) = n = 0 - n ( 1 - q ) n - 1 l = 0 n - 1 ( n - 1 l ) ( - 1 ) l q l x l 1 - q l + 1 t n n ! . (4) (5)
(2.5)

Comparing these identities (2.4) and (2.5), we obtain

B 0 , q ( x ) = 0 and B n , q ( x ) = - n ( 1 - q ) n - 1 l = 0 n - 1 ( n - 1 l ) ( - 1 ) l q l x l 1 - q l + 1 .
(2.6)

This cannot satisfy some well-known results related the Bernoulli polynomials and numbers. For example, B 0 = 1.

Counter-example 2.5 From Definition 2.1 (Açikgöz et al. [1]), we note that

q D q ( t , 1 ) - D q ( t ) = - t y = 0 q y + 1 e [ 1 + y ] q t - t y = 0 q y e [ y ] q t (1) = t , (2) (3)
(2.7)

and

q D q ( t , 1 ) - D q ( t ) = q n = 0 B n , q ( 1 ) t n n ! - n = 0 B n , q t n n ! (1) = n = 0 ( q B n , q ( 1 ) - B n , q ) t n n ! . (2) (3)
(2.8)

From (2.7) and (2.8), we can easily derive that

B n , q = 0 and q B n , q ( 1 ) - B n , q = 1 if n = 1 0 if n > 1 .
(2.9)

From (2.1) of Definition 2.1 (Açikgöz et al. [1]),

n = 0 B n , q ( x ) t n n ! = D q ( t , x ) (1) = - t y = 0 q y e [ x + y ] q t (2) = e [ x ] q t 1 q x D q ( t q x ) (3) = l = 0 [ x ] q l t l l ! × m = 0 B m , q q ( m - 1 ) x t m m ! (4) = n = 0 m = 0 n ( n m ) B m , q q ( m - 1 ) x [ x ] q n - m t n n ! . (5) (6)
(2.10)

If we compare the coefficients on the both sides in (2.10),

B n , q ( x ) = m = 0 n ( n m ) B m , q q ( m - 1 ) x [ x ] q n - m .
(2.11)

From (2.9) and (2.11),

B 0 , q ( x ) = 1 q x B 0 , q = 0 .
(2.12)

However, these are also incorrect.

Next, we redefine the q-Bernoulli polynomials and numbers.

Definition 2.6 For q with |q| < 1, let us define the q-Bernoulli polynomials B n,q (x) as follows,

F q ( t , x ) = q - 1 l o g q e 1 1 - q t - t m = 0 q x + m e [ x + m ] q t = n = 0 B n , q ( x ) t n n ! .
(2.13)

Note that

lim q 1 F q ( t , x ) = t e t - 1 e x t = n = 0 B n ( x ) t n n ! for | t | < 2 π ,
(2.14)

where B n (x) are the classical Bernoulli polynomials.

In the special case x = 0, B n,q (0) = B n,q are called the q-Bernoulli numbers. That is,

F q ( t ) = F q ( t , 0 ) = n = 0 B n , q t n n ! .
(2.15)

By simple calculations, we get

n = 0 B n , q ( x ) t n n ! = F q ( t , x ) (1) = e [ x ] q t F q ( q x t ) (2) = m = 0 [ x ] q m t m m ! × l = 0 B l , q q l x t l l ! (3) = n = 0 l = 0 n ( n l ) B l , q q l x [ x ] q n - l t n n ! . (4) (5)
(2.16)

Comparing the coefficients on the both sides in (2.16), we obtain

B n , q ( x ) = l = 0 n ( n l ) B l , q q l x [ x ] q n - l .
(2.17)

From (2.13) and (2.15), we derive the following equation.

B 0 , q = q - 1 l o g q and B n , q ( 1 ) - B n , q = 1 if n = 1 0 if n > 1 .
(2.18)

By (2.17) and (2.18), we can see that

B 0 , q = q - 1 l o g q and l = 0 n ( n l ) B l , q q l - B n , q = 1 if n = 1 0 if n > 1 .
(2.19)

Theorem 2.7* For n ∈ ℕ*, we have

B 0 , q = q - 1 l o g q and ( q B q + 1 ) n - B n , q = 1 if n = 1 0 if n > 1 .
(2.20)

with the usual convention of replacing B q n by B n,q .

Remark 2.8 Theorem 2.7* is a revised theorem of Theorem 2.1 in [1].

From (2. 13), we have

n = 0 B n , q ( x ) t n n ! = F q ( t , x ) (1) = q - 1 l o g q e 1 1 - q t - t m = 0 q x + m e [ x + m ] q t (2) = q - 1 l o g q n = 0 1 ( 1 - q ) n t n n ! - m = 0 q x + m n = 0 n [ x + m ] q n - 1 t n n ! (3) = n = 0 q - 1 l o g q 1 ( 1 - q ) n - n m = 0 q x + m [ x + m ] q n - 1 t n n ! (4) = n = 0 - ( 1 - q ) n l o g q - n ( 1 - q ) n - 1 m = 0 q x + m l = 0 n - 1 ( n - 1 l ) ( - 1 ) l q ( x + m ) l t n n ! (5) = n = 0 ( q - 1 ) 1 - n l o g q + n ( 1 - q ) n - 1 l = 0 n - 1 ( n - 1 l ) ( - 1 ) l + 1 q ( l + 1 ) x 1 1 - q ( l + 1 ) t n n ! (6) = n = 0 1 ( 1 - q ) n l = 0 n ( n l ) ( - 1 ) l q l x l [ l ] q t n n ! . (7) (8) 
(2.21)

By (2.21), we obtain the following theorem.

Theorem 2.9* For n ∈ ℕ*, we have

B 0 , q = q - 1 l o g q and B n , q ( x ) = 1 ( 1 - q ) n l = 0 n ( n l ) ( - 1 ) l q l x l [ l ] q .
(2.22)

Remark 2.10 Theorem 2.9* is a revised theorem of Theorem 2.3 in [1].