R E S E A R C H
Open Access
Certain combinatoric Bernoulli polynomials
and convolution sums of divisor functions
Daeyeoul Kim
1and Nazli Yildiz Ikikardes
2**Correspondence: nyildizikikardes@gmail.com 2Department of Elementary Mathematics Education, Necatibey Faculty of Education, Balikesir University, Balikesir, 10100, Turkey Full list of author information is available at the end of the article
Abstract
It is known that certain convolution sums can be expressed as a combination of divisor functions and Bernoulli formula. One of the main goals in this paper is to establish combinatoric convolution sums for the divisor sums
σ
ˆs(n) =
d|n(–1)
n d–1ds.
Finally, we find a formula of certain combinatoric convolution sums and Bernoulli polynomials.
MSC: 11A05; 33E99
Keywords: Bernoulli numbers; convolution sums
1 Introduction
The symbolsN and Z denote the set of natural numbers and the ring of integers, respec-tively. The Bernoulli polynomials Bk(x), which are usually defined by the exponential
gen-erating function text et– = ∞ k= Bk(x) tk k!,
play an important and quite mysterious role in mathematics and various fields like analysis, number theory and differential topology. The Bernoulli polynomials satisfy the following well-known identities: N j= jk=Bk+(N + ) – Bk+() k+ (k≥ ) = k+ k j= (–)j k+ j BjNk+–j. (.)
The Bernoulli numbers Bk are defined to be Bk:= Bk(). For n∈ N, k ∈ Z, we define
some divisor functions
σk(n) := d|n dk, σk∗(n) := d|n n dodd dk, σk(n) := d|n (–)d–dk, ˆσk(n) := d|n (–)nd–dk, σk,l(n; ) := d|n d≡l(mod ) dk.
©2013 Kim and Yildiz Ikikardes; licensee Springer. This is an Open Access article distributed under the terms of the Creative Com-mons Attribution License (http://creativecomCom-mons.org/licenses/by/2.0), which permits unrestricted use, distribution, and repro-duction in any medium, provided the original work is properly cited.
It is well known that σk∗(n) = σk(n) – σk(n) and ˆσk(n) = σk(n) – σk(n) [, (.)]. The identity n– k= σ(k)σ (n – k) = σ(n) + – n σ(n)
for the basic convolution sum first appeared in a letter from Besge to Liouville in []. Hahn [, (.)] considered m<n ˆσ (m) ˆσ(n – m) = ⎧ ⎨ ⎩ –ˆσ(n) + σ(n) if n is odd, –ˆσ(n) – σ(n) + σ(n) if n is even. (.)
For some of the history of the subject, and for a selection of these articles, we mention [, ] and [], and especially [, ] and []. The study of convolution sums and their ap-plications is classical, and they play an important role in number theory. In this paper, we investigate the combinatorial Bernoulli numbers and convolution sums. For k and n being positive integers, we show that the sum
k j= k + j Bjˆσk+–j(n)
can be evaluated explicitly in terms of divisor functions and a combinatorial convolution sum. We prove the following.
Theorem Let k, n be positive integers. Then
k j= k + j Bjˆσk+–j(n) = (k + )σk+∗ (n) – k + ˆσk+(n) – k(k + ) ˆσk–(n) – (k + ) k– s= k s + n– m= ˆσk–s–(m)ˆσs+(n – m).
Remark Let n be positive integers. In Theorem , replace k by , we find easily that
n– m= ˆσ(m) ˆσ(n – m) = σ ∗ (n) – ˆσ(n) – ˆσ(n), (.) and in particular, if q∈ N, p = q + , an odd prime integer, then
q m= ˆσ(m) ˆσ(p – m) = q(q + )(q + ) = q m= k= B(q + ). (.) Equations (.) and (.) are in (.) and [, Corollary .]. Using these combinatoric convolution sums, we obtain the following.
Theorem If k is a positive integer, then u+v+w=k+ v–k– k + u, v, w Bv· (l + )w= Bk+(l + ), whereuk+,v,w=(k+)!u!v!w! and l= , , , , , , , , , , , .
Thus, we can pose a general question regarding Bernoulli polynomials.
Question For all k, l∈ N, does the identity u+v+w=k+ v–k– k + u, v, w Bv· (l + )w= Bk+(l + ) hold?
The problem of convolution sums of the divisor function σ(n) and the theory of
Eisenstein series has recently attracted considerable interest with the emergence of quasi-modular tools. In connection with the classical Jacobi theta and Euler functions, other aspects of the function σ(n) are explored by Simsek in []. Finally, we prove the
follow-ing.
Theorem If a(≥ ) and k are positive integers, then (i) k– s= k s + a– m= σk–s–, m ; σs+ a– m = (k + ) Bk+ a+ – k (k + ) a– i= Bk+ i+ (k+)a– k+– – a++ , (ii) k– s= k s + a– m= σk–s–, m ; σs+ a– m = + k (k + ) a– i= Bk+ i+ (k+)a– k+– +
k(a+)– k(a+)–+ ka–– k– + a– – (k–)(a+)– (k–)a+ k–– + a–– , (iii) k– s= k s + a– m= σk–s– m σs+ a– m = (k + ) Bk+ a+ (k + ) a– i= Bk+ i
+ (k+)a– k+– +
k(a+)– k(a+)–+ ka–– k– + a– – (k–)(a+)– (k–)a+ k–– .
2 Properties of convolution sums derived from divisor functions Proposition ([]) Let k, n be positive integers. Then
k j= k + j Bjσk+–j(n) = – k + σk+(n) – (k + ) k – n σk–(n) + (k + ) k– s= k s + n– m= σk–s–(m)σs+(n – m).
Proposition ([, ]) Let k, n be positive integers. Then
(i) k– s= k s + n– m= σk–s–∗ (m)σs+∗ (n – m) = σk+∗ (n) – nσk–∗ (n), (ii) k– s= k s + n m= σk–s–(m – )σs+(n – m + ) = σ ∗ k+(n).
Proof of Theorem Let k, n∈ N. By Proposition and Proposition , we obtain
T := k– s= k s + n– m= ˆσk–s–(m)ˆσs+(n – m) = k– s= k s + n– m= σk–s–(m) – σk–s– m × σs+(n – m) – σs+ n– m . It is easily checked that
σk–s– m σs+(n – m) + σk–s–(m)σs+ n– m = σk–s– m σs+ n– m + σk–s–(m)σs+(n – m) + σk–s–(m) – σk–s– m σs+(n – m) – σs+ n– m . Thus, T = k– s= k s + n– m= σk–s–(m)σs+(n – m) + σk–s– m σs+ n– m – k– s= k s + n– m= σk–s– m σs+ n– m + σk–s+(m)σs+(n – m)
– k– s= k s + n– m= σk–s–(m) – σk–s– m × σs+(n – m) – σs+ n– m = k– s= k s + n– m= σk–s–(m) – σk–s– m σs+(n – m) – σs+ n– m – k– s= k s + n– m= σk–s–(m)σs+(n – m) – σk–s– m σs+ n– m = k– s= k s + n– m= σk–s–∗ (m)σs+∗ (n – m) – k + k + σk+(n) + k – n σk–(n) + k + k j= k + j Bjσk+–j(n) + k + k + σk+ n + k – n σk– n + k + k j= k + j Bjσk+–j n =σk+∗ (n) – nσk–∗ (n)– k + k + σk+(n) – σk+ n –k σk–(n) – σk– n + n σk–(n) – σk– n – k + k j= k + j Bj σk+–j(n) – σk+–j n = σk+∗ (n) – k + k + ˆσk+(n) – k ˆσk–(n) – k + k j= k + j Bjˆσk+–j(n).
This proves the theorem.
Example Let n be a positive integer. In Theorem , put k = , we get
n– m= ˆσ(m) ˆσ(n – m) = σ ∗ (n) – ˆσ(n) – ˆσ(n) + ˆσ(n).
Corollary Let k, n be positive integers. Then, we obtain (i) k j= k + j Bjˆσk+–j(n) = k + σk+∗ (n) – k + ˆσk+(n) – k(k + ) ˆσk–(n) – (k + ) k– s= k s + n– m= ˆσk–s–(m)ˆσs+(n – m),
(ii) k j= k + j Bj σk+–j(n) + σk+–j n = –σk+(n) – (k + )σk+ n –(k + )(k – n) σk–(n) –(k + )(k – n) σk– n + (k + ) k– s= k s + n– m= σk–s– m σs+(n – m), (iii) k j= k + j Bjσk+–j∗ (n) = kσk+∗ (n) –k(k + ) σ ∗ k–(n) – n(k + ) ˆσk–(n) – (k + ) k– s= k s + n– m= σk–s–∗ (m)ˆσs+(n – m) = –(k + )σk+∗ (n) –(k – n)(k + ) σ ∗ k–(n) + n(k + ) σk– n + (k + ) k– s= k s + n– m= σk–s–∗ (m)σs+(n – m).
Proof (i) We note that
k– s= k s + n– m= σk–s–∗ (m)σs+∗ (n – m) = k– s= k s + n– m= σk–s–∗ (m)σs+∗ (n – m) – k– s= k s + n– m= σk–s–∗ (m – )σs+∗ n– (m – ).
(ii) and (iii) are applied in a similar way.
3 Bernoulli polynomials and convolution sums Proposition ([]) Let k, n be positive integers. Then
k– s= k s + n– m= k–s–σk–s–(m/)σs+(n – m) = σk+(n/) – σk(n) – k+σk(n/) – kσk(n/)
–n σk–(n) + kσk–(n/) + σk,(n; ) + k– k + k j= k + j Bjσk+–j(n/) + (k + ) k j= k + j Bjk+–jσk+–j(n/) + (k + ) k j= k + j Bjσk+–j,(n; ) – (k + ) u+v+w=k+ v– k + u, v, w Bvσw,(n; ).
It is well known that σk–s–,(m; ) = k–s–σk–s–(m). Using Proposition , we get
this lemma.
Lemma Let k, n be positive integers. Then (i) k– s= k s + n– m= σk–s–, m ; σs+(n – m) = σk+ n + (k + )σk+,(n; ) + (k + )σk+, n ; + (k + )σk,(n; ) – σk(n) + σk,(n; ) – σk,(n; ) + σk,(n; ) –n σk–(n) + k σk–,(n; ) + k – n σk–, n ; + k σk–,(n; ) + k (k + ) k j= k + j Bjσk+–j n + (k + ) k j= k + j Bjσk+–j, n ; + (k + ) k j= k + j Bjσk+–j,(n; ) – (k + ) u+v+w=k+ v– k + u, v, w Bvσw,(n; ), (ii) k– s= k s + n– m= σk–s–, m ; σs+(n – m) = (k + ) σk+(n) + k+ k + σk+ n + k– n σk–(n) + k– n σk– n – σk+ n
– (k + ) σk+,(n; ) – (k + ) σk+, n ; – (k + ) σk,(n; ) + σk(n) – σk,(n; ) + σk,(n; ) – σk,(n; ) +n σk–(n) – k σk–,(n; ) – k – n σk–, n ; – k σk–,(n; ) – k (k + ) k j= k + j Bjσk+–j n – (k + ) k j= k + j Bjσk+–j, n ; – (k + ) k j= k + j Bjσk+–j,(n; ) + (k + ) u+v+w=k+ v– k + u, v, w Bvσw,(n; ).
Remark (i) Using Lemma , we obtain u+v+w=k+ v– k + u, v, w Bvσw,(n; ) = k + σk+ n – σk(n) + σk,(n; ) + σk, n ; – n(k + ) σk–(n) + σk–, n ; + (k + )σk,(n; ) + k k j= k + j Bjσk+–j n + k j= k + j Bjσk+–j, n ; + k j= k + j Bjσk+–j,(n; ) – (k + ) k– s= k s + n– m= σk–s–, m ; σs+(n – m) = – k + σk+ n – nσk–, n ; + σk, n ; – k j= k + j Bjσk+–j(n) + k– k j= k + j Bjσk+–j n + k j= k + j Bjσk+–j, n ; + k j= k + j Bjσk+–j,(n; ) + (k + ) k– s= k s + n– m= σk–s–, m ; σs+(n – m).
(ii) If n is an odd integer, then u+v+w=k+ v– k + u, v, w Bvσw,(n; ) = (k + ) k– s= k s + n– m= σk–s–, m ; σs+(n – m). (.)
(iii) In (.), put k = , we get u+v+w= v– u, v, w Bvσw,(n; ) = n– m= σ, m ; σ(n – m), and thus, n– m= σ, m ; σ(n – m) = σ(n) – σ(n) . In (.), replace k by , we find that u+v+w= v– u, v, w Bvσw,(n; ) = n– m= σ, m ; σ(n – m) + n– m= σ, m ; σ(n – m) , and thus, n– m= σ, m ; σ(n – m) + n– m= σ, m ; σ(n – m) = σ(n) – σ(n) + σ(n) .
Proof of Theorem If n = , compare both sides of (.), we obtain u+v+w=k+ v– k + u, v, w Bv= . (.) If we put n = in (.), we obtain u+v+w=k+ v– k + u, v, w Bv· + w = u+v+w=k+ v– k + u, v, w Bv· w = (k + ) k– s= k s + . (.)
From (.) and (.), we get u+v+w=k+ v– k + u, v, w Bv· w= (k + )k. (.)
By combining (.) and (.), we obtain u+v+w=k+ v– k + u, v, w Bv· w= kBk+().
Others cases follow in a similar way. This completes the proof.
Proof of Theorem (i) If a is a positive integer, then u+v+w=k+ v– k + u, v, w Bv· σw, a; = u+v+w=k+ v– k + u, v, w Bv=
by (.). According to Remark (i), we deduce that
= (k + ) k– s= k s + a– m= σk–s–, m ; σs+ a– m – k j= k + j Bjσk+–j a+k– k j= k + j Bjσk+–j a + k j= k + j Bjσk+–j, a ; – k + σk+ a – a+ .
If a = , it is clearly evident. We suppose that a > . We check that
k j= k + j Bj· ak+–j = k j= k + j (–)jBj· ak+–j+ k + B· ak = (k + ) a j= jk+ (k + ) – ak = Bk+ a (.) by (.).
(ii) and (iii) are applied in a similar way.
Remark If p is a prime integer, then u+v+w=k+ v– k + u, v, w Bv· pw= (k + ) k– s= k s + p– m= σk–s–, m ; σs+(p – m) by (.) and (.).
Competing interests
The authors declare that they have no competing interests. Authors’ contributions
All authors contributed equally and significantly in writing this article. All authors read and approved the final manuscript. Author details
1National Institute for Mathematical Sciences, Yuseong-daero 1689-gil, Yuseong-gu, Daejeon, 305-811, South Korea. 2Department of Elementary Mathematics Education, Necatibey Faculty of Education, Balikesir University, Balikesir, 10100, Turkey.
Received: 26 July 2013 Accepted: 23 September 2013 Published:08 Nov 2013 References
1. Hahn, H: Convolution sums of some functions on divisors. Rocky Mt. J. Math. 37(5), 1593-1622 (2007) 2. Besge, M: Extrait d’une lettre de M. Besge à M. Liouville. J. Math. Pures Appl. 7, 375-376 (1862)
3. Cheng, N, Williams, KS: Evaluation of some convolution sums involving the sum of divisors functions. Yokohama Math. J. 52, 39-57 (2005)
4. Lahiri, DB: On Ramanujan’s functionτ(n) and the divisor functionσ(n), I. Bull. Calcutta Math. Soc. 38, 193-206 (1946) 5. Melfi, G: On some modular identities. In: Number Theory (Eger, 1996), pp. 371-382. de Gruyter, Berlin (1998) 6. Huard, JG, Ou, ZM, Spearman, BK, Williams, KS: Elementary evaluation of certain convolution sums involving divisor
functions. In: Number Theory for the Millennium, vol. II, pp. 229-274 (2002)
7. Kim, D, Kim, A, Sankaranarayanan, A: Bernoulli numbers, convolution sums and congruences of coefficients for certain generating functions. J. Inequal. Appl. 2013, 225 (2013)
8. Williams, KS: Number Theory in the Spirit of Liouville. London Mathematical Society Student Texts, vol. 76. Cambridge University Press, Cambridge (2011)
9. Kim, D, Kim, A, Kim, M: A remark on algebraic curves derived from convolution sums. J. Inequal. Appl. 2013, 58 (2013) 10. Simsek, Y: Elliptic analogue of the Hardy sums related to elliptic Bernoulli functions. Gen. Math. 15(2-3), 3-23 (2007) 11. Kim, D, Bayad, A: Convolution identities for twisted Eisenstein series and twisted divisor functions. Fixed Point Theory
Appl. 2013, 81 (2013)
12. Cho, B, Kim, D, Park, H: Combinatorial convolution sums derived from divisor functions and Faulhaber sums (2013, submitted)
10.1186/1687-1847-2013-310
Cite this article as: Kim and Yildiz Ikikardes: Certain combinatoric Bernoulli polynomials and convolution sums of