Next Article in Journal
Failed Skew Zero Forcing Numbers of Path Powers and Circulant Graphs
Previous Article in Journal
Exploring the Influence of Oblateness on Asymptotic Orbits in the Hill Three-Body Problem
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Evaluation of Harmonic Number Series Involving the Binomial Coefficient C(3n,n) in the Denominator by Integration

1
Department of Physics and Engineering Physics, Obafemi Awolowo University, Ile-Ife 220005, Nigeria
2
Independent Researcher, 72762 Reutlingen, Germany
*
Author to whom correspondence should be addressed.
AppliedMath 2025, 5(1), 31; https://doi.org/10.3390/appliedmath5010031
Submission received: 29 October 2024 / Revised: 4 February 2025 / Accepted: 25 February 2025 / Published: 19 March 2025

Abstract

:
Two classes of series involving differences of harmonic numbers and the binomial coefficients C ( 3 n , n ) are evaluated in closed form. The classes under consideration are k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k m z k and k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k m z k , where z is a complex number and m (a non-negative integer) is an additional parameter. The tool that will be applied is integration in combination with complex analysis and partial fraction decomposition. Several remarkable integral values and difficult series identities are stated as special cases of the main results.
MSC:
11G55; 33B30; 65B10

1. Introduction, Motivation and Preliminaries

Integration is a classical tool in the evaluation of infinite series. The approach was revived recently in the papers by Sofo and Nimbran [1], Stewart [2], and Li and Chu [3,4,5,6,7]. For instance, Li and Chu review in [6] a few (known) series involving harmonic numbers H n and odd harmonic numbers O n such as
n = 1 H n n 2 , n = 1 ( 1 ) n 1 H n n 2 , n = 1 O n n 2 , n = 1 ( 1 ) n 1 O n n 2 ,
obtained by using definite integrals. The authors evaluate difficult addition series involving alternating harmonic and odd harmonic numbers in closed form by employing calculus and complex analysis. Here, as usual, harmonic numbers H n and odd harmonic numbers O n are defined by H 0 = 0 , O 0 = 0 , and
H n = k = 1 n 1 k , O n = k = 1 n 1 2 k 1 .
Obvious relations between harmonic numbers H n and odd harmonic numbers O n are the following:
H 2 n = 1 2 H n + O n and H 2 n 1 = 1 2 H n 1 + O n .
For more information about H n and O n we refer to the papers [8,9,10,11,12,13].
In their paper [7] Li and Chu use integration techniques to prove (among other things) the series expression
k = 0 4 k 2 k x ( 1 x ) 4 k 2 H 4 k 3 H 2 k + H k = 1 x 2 x 1 ln ( 1 x ) ,
which converges for | 4 x ( 1 x ) | < 1 , i.e., for ( 1 2 ) / 2 < x < 1 / 2 , and which was conjectured by Sun in his paper [14]. Another conjecture from Sun’s paper that is strongly related to our study is concerned with the two evaluations
k = 1 H 3 k H k k 2 k 3 k k = 2 5 G + ln 2 2 π 2 24
and
k = 1 H 3 k H k k 2 2 k 3 k k = 11 4 ζ ( 3 ) π 2 24 ln 2 π G ,
where G is Catalan’s constant. This is Conjecture 2.4 in [14] (Equations (2.6) and (2.7), respectively). In addition, as reported by Sun in [14], the similar conjectured identities
k = 1 H 2 k H k k 2 k 3 k k = 3 10 ln 2 2 + π 20 ln 2 π 2 60
and
k = 1 H 2 k H k k 2 2 k 3 k k = 33 32 ζ ( 3 ) + π 2 24 ln 2 π G 2
were proved by Au in the recent paper [15]. Finally, we mention another recent paper by Sun [16] dealing with similar series.
In this paper, we proceed in the same direction and study series involving differences of harmonic numbers in the numerator and the binomial coefficients 3 k k in the denominator. To be more precise, using integrals in combination with complex analysis and partial fraction decompositions we will evaluate in closed form the Euler-type series
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k m z k and k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k m z k ,
for all m 0 and all z C with | z | < 1 . For instance, we will prove that
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k 2 k + 1 = π 2 48 ln 2 2 10 + 2 5 G ,
where G = j = 0 ( 1 ) j / ( 1 + 2 j ) 2 is Catalan’s constant. Another difficult evaluation that will be derived is
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k k 4 k + 1 = 3 448 ln 2 9 512 ln 2 2 3 128 arctan 2 7 5 1 224 89 6272 ln 2 7 arctan 7 5 .
We proceed wit two special functions that will be needed. Let Li 2 ( z ) be the dilogarithm defined by (see Lewin [17])
Li 2 ( z ) = k = 1 z k k 2 , | z | 1 ,
having the special values
Li 2 ( 1 ) = π 2 6 , and Li 2 1 2 = π 2 12 ln 2 ( 2 ) 2 .
Let also Cl 2 ( z ) be the Clausen’s function defined by [17,18]
Cl 2 ( z ) = n = 1 sin ( n z ) n 2 = 0 z ln | 2 sin ( θ / 2 ) | d θ .
This function has the functional relations
Cl 2 ( π + θ ) = Cl 2 ( π θ ) ,
Cl 2 ( θ ) = Cl 2 ( 2 π θ ) ,
1 2 Cl 2 ( 2 θ ) = Cl 2 ( θ ) Cl 2 ( π θ ) ;
and the special values
Cl 2 ( n π ) = 0 , n Z + ,
and
Cl 2 ( π / 2 ) = G = Cl 2 ( 3 π / 2 ) ,
where G is Catalan’s constant.
We conclude this section with a motivation of our approach. We start with the Beta integral [19]:
0 1 x a 1 ( 1 x ) b 1 d x = B ( a , b ) = Γ ( a ) Γ ( b ) Γ ( a + b ) , a , b > 0 .
Differentiating the above definition with respect to a and using the fact that
d d a x a 1 = x a 1 ln ( x )
we get
0 1 x a 1 ( 1 x ) b 1 ln ( x ) d x = Γ ( a ) Γ ( b ) Γ ( a + b ) ψ ( a ) ψ ( a + b ) ,
where ψ ( x ) = Γ ( x ) / Γ ( x ) is the psi or digamma function, Γ ( x ) being the Gamma function. This function is related to harmonic numbers via ψ ( n + 1 ) = H n γ , where γ is the Euler-Mascheroni constant. From here we can make the transformations a k a + 1 and b 2 k b + 1 to obtain
0 1 x k a ( 1 x ) 2 k b ln ( x ) d x = Γ ( k a + 1 ) Γ ( 2 k b + 1 ) Γ ( k ( a + 2 b ) + 2 ) ψ ( k a + 1 ) ψ ( k ( a + 2 b ) + 2 ) = H k a H k ( a + 2 b ) + 1 ( k ( a + 2 b ) + 1 ) k ( a + 2 b ) k a .
Also, by symmetry or by applying the transformations a 2 k a + 1 and b k b + 1 we obtain
0 1 x 2 k a ( 1 x ) k b ln ( x ) d x = H 2 k a H k ( 2 a + b ) + 1 ( k ( 2 a + b ) + 1 ) k ( 2 a + b ) k b .
Now, let a = b = 1 . Then
0 1 x k ( 1 x ) 2 k ln ( x ) d x = H k H 3 k + 1 ( 3 k + 1 ) 3 k k ,
and we can consider the series (for all z C with | z | 1 )
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k z k = 0 1 ln ( x ) 1 z x ( 1 x ) 2 d x ,
which can also be written as
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k 1 z k + 1 = 0 1 ln x x ( 1 x ) 2 z d x , | z | 1 .
In particular,
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k 1 2 k + 1 = 0 1 ln ( x ) ( x 2 ) ( x 2 + 1 ) d x
and
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k ( 1 ) k 4 k + 1 = 0 1 ln ( x ) ( x + 1 ) ( x 2 3 x + 4 ) d x .
Similarly, we obtain
0 1 x 2 k a ( 1 x ) k a ln ( x ) d x = 1 0 ( 1 x ) 2 k a x k a ln ( 1 x ) d x = H 2 k a H k ( 2 a + b ) + 1 ( k ( 2 a + b ) + 1 ) k ( 2 a + b ) k b .
This gives for all z C with | z | 1
k = 0 H k a H 2 k a ( 3 k a + 1 ) 3 k a k a z k = 0 1 ln x 1 x 1 z x a ( 1 x ) 2 a d x
or the particular relation
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k z k = 0 1 ln x 1 x 1 z x ( 1 x ) 2 d x .
The evaluation of the integrals is not trivial but can be done by applying some additional theory.

2. Main Results, Part 1

In this section, we explicitly deal with the series on the left hand side of (12).
Lemma 1.
For λ [ 0 , 1 ) the following identities hold:
0 1 ln x x λ d x = Li 2 1 λ ,
0 1 ln ( x / ( 1 x ) ) x λ d x = 1 2 ln 2 λ 1 λ .
Proof. 
Identity (17) follows immediately from the fact that
ln x x λ d x = Li 2 x λ + ln x ln λ x λ + c o n s t .
Now,
0 1 ln ( x / ( 1 x ) ) x λ d x = 0 1 ln x x λ d x 0 1 ln ( 1 x ) x λ d x .
Let
I = 0 1 ln ( 1 x ) x λ d x .
A change of variable u = 1 x gives
I = 0 1 ln u u ( 1 λ ) d u = Li 2 1 1 λ
on account of (17). Thus
0 1 ln ( x / ( 1 x ) ) x λ d x = Li 2 1 λ + Li 2 1 1 λ = 1 2 ln 2 λ 1 λ
since ([17], p. 283)
Li 2 ( x ) + Li 2 x x 1 = 1 2 ln 2 ( 1 x ) , x < 1 .
Lemma 2
([17], p. 291). For θ [ 2 π , 2 π ] , we have the following relation
Li 2 e i θ = π 2 6 + θ 2 2 π | θ | 4 + i Cl 2 ( θ ) ,
where Cl 2 ( z ) is Clausen’s function.
Proposition 1.
We have
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k 2 k + 1 = π 2 48 ln 2 ( 2 ) 10 + 2 5 G .
Proof. 
Using the decomposition
1 ( x 2 ) ( x 2 + 1 ) = 1 ( 4 i 2 ) 1 x + i 1 ( 4 i + 2 ) 1 x i + 1 5 1 x 2
we have
0 1 ln x ( x 2 ) ( x 2 + 1 ) d x = 1 4 i 2 0 1 ln x x + i d x 1 4 i + 2 0 1 ln x x i d x + 1 5 0 1 ln x x 2 d x = 1 4 i 2 Li 2 e i π / 2 1 4 i + 2 Li 2 e i π / 2 + 1 5 Li 2 1 2 ,
in view of (17). Now, (20) gives
Li 2 ( e i π / 2 ) = π 2 48 + i G , Li 2 ( e i π / 2 ) = π 2 48 + i Cl 2 ( π / 2 ) .
Using (9) with θ = π / 2 gives
Cl 2 ( π / 2 ) = Cl 2 ( 3 π / 2 ) = G , by ( 11 ) ;
so that
Li 2 ( e i π / 2 ) = π 2 48 i G .
Thus, using (22) and (24), and the evaluation of Li 2 ( 1 / 2 ) from (6), we have
0 1 ln x ( x 2 ) ( x 2 + 1 ) d x = 1 4 i 2 π 2 48 + i G 1 4 i + 2 π 2 48 i G + 1 5 π 2 12 1 2 ln 2 2
and hence (21), in view of (14). □
Alternative Proof of Proposition 1. 
We begin with the partial fraction decomposition
1 ( x 2 ) ( x 2 + 1 ) = 1 5 1 x 2 x + 2 x 2 + 1 .
Hence,
0 1 ln x ( x 2 ) ( x 2 + 1 ) d x = 1 5 0 1 ln x x 2 d x 0 1 ( x + 2 ) ln x x 2 + 1 d x .
As
ln x x 2 d x = ln ( 2 ) ln ( | x 2 | ) Li 2 2 x 2 + c o n s t ,
we get
0 1 ln x x 2 d x = π 2 12 ln 2 ( 2 ) 2 .
Next,
x ln x x 2 + 1 d x = 1 2 Li 2 ( i x ) + Li 2 ( i x ) + ln ( x ) ln ( 1 + x 2 ) + c o n s t . ,
and this gives
x ln x x 2 + 1 d x = 1 2 Li 2 ( i ) + Li 2 ( i ) = π 2 48 ,
where the relation (see also [17])
Li 2 ( x ) + Li 2 ( x ) = 1 2 Li 2 ( x 2 ) ,
was used. Finally, from the indefinite integral
2 ln x x 2 + 1 d x = i Li 2 ( i x ) Li 2 ( i x ) + ln ( x ) ln i + x i x + c o n s t ,
we get
2 0 1 ln x x 2 + 1 d x = i ( Li 2 ( i ) Li 2 ( i ) ) = 2 j = 0 ( 1 ) j ( 1 + 2 j ) 2 = 2 G .
Putting everything together we obtain the claimed result. □
Proposition 2.
We have
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k ( 1 ) k 4 k + 1 = π 2 96 + 1 8 Li 2 3 8 + i 7 8 + 5 7 56 Li 2 3 8 + i 7 8 .
Proof. 
Consider the partial fraction decomposition
1 x ( 1 x ) 2 + 4 = 1 ( x + 1 ) ( x 2 3 x + 4 ) = 1 8 1 x + 1 1 112 7 + 5 i 7 x ( 3 + i 7 ) / 2 1 112 7 5 i 7 x ( 3 i 7 ) / 2 ;
which allows the integral on the rhs of (15) to be written as
0 1 ln x d x ( x + 1 ) ( x 2 3 x + 4 ) = 1 8 0 1 ln x d x x + 1 7 + 5 i 7 112 0 1 ln x d x x ( 3 + i 7 ) / 2 7 5 i 7 112 0 1 ln x d x x ( 3 i 7 ) / 2 ,
which upon using (17) gives
0 1 ln x d x ( x + 1 ) ( x 2 3 x + 4 ) = 1 8 Li 2 ( 1 ) 7 + 5 i 7 112 Li 2 2 3 + i 7 7 5 i 7 112 Li 2 2 3 i 7 = 1 8 Li 2 ( 1 ) 1 56 ( 7 + 5 i 7 ) Li 2 2 3 + i 7 .
Use of (26) in (15) gives (25), after simplification. In taking the real part in (26), it is convenient to use the fact that ( f g ) = f g f g for arbitrary f and g. □
Differentiating (12) m times with respect to z and thereafter replacing z with 1 / z gives
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k m z k + 1 = ( 1 ) m 0 1 x m ( 1 x ) 2 m ln x x ( 1 x ) 2 z m + 1 d x , | z | 1 .
Theorem 1.
Let m be a non-negative integer and let z 1 , z 2 and z 3 be the distinct roots of x ( 1 x ) 2 z = 0 where z is a real number such that | z | 1 . Then the following identity holds:
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k m z k + 1 = ( 1 ) m k = 1 3 j = 0 m a j ( z k ) C j ( z k ) ;
where, for 0 r m ,
a r ( λ ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x λ ) m + 1 x ( 1 x ) 2 z m + 1 x = λ ,
C 0 ( λ ) = Li 2 1 λ ,
and for r a positive integer and λ [ 0 , 1 ] ,
C r ( λ ) = ( 1 ) r r p = 1 r 1 1 p λ r p 1 ( λ 1 ) p 1 λ p ( 1 ) r r λ r ln λ 1 λ .
Proof. 
We start with the partial fraction decomposition
x m ( 1 x ) 2 m x ( 1 x ) 2 z m + 1 = j = 0 m a j ( z 1 ) ( x z 1 ) j + 1 + a j ( z 2 ) ( x z 2 ) j + 1 + a j ( z 3 ) ( x z 3 ) j + 1 = k = 1 3 j = 0 m a j ( z k ) ( x z k ) j + 1 ,
where, for 0 r m , the coefficients a r ( z 1 ) , a r ( z 2 ) , a r ( z 3 ) are found from (29).
We therefore have
0 1 x m ( 1 x ) 2 m ln x x ( 1 x ) 2 z m + 1 d x = k = 1 3 j = 0 m a j ( z k ) 0 1 ln x ( x z k ) j + 1 d x .
Differentiating (17) j times with respect to λ gives
0 1 ln x ( x λ ) j + 1 d x = ( 1 ) j j p = 1 j 1 1 p λ j p 1 ( λ 1 ) p 1 λ p ( 1 ) j j λ j ln λ 1 λ = C j ( λ ) , j 0 ,
which, when utilized in (33) yields an evaluation of the integral on the right hand side of (27) and hence (28). □
Corollary 1.
Let m be a non-negative integer and let ( λ 1 , λ 2 , λ 3 ) = ( 2 , i , i ) where i is the imaginary unit. Then
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k m 2 k + 1 = ( 1 ) m k = 1 3 j = 0 m a j ( λ k ) C j ( λ k ) ;
where, for 0 r m , the coefficients are given by
a r ( 2 ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x 2 + 1 ) m + 1 x = 2 ,
a r ( i ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x 2 ) ( x i ) m + 1 x = i ,
a r ( i ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x 2 ) ( x + i ) m + 1 x = i ;
C 0 ( 2 ) = Li 2 1 2 = π 2 12 1 2 ln 2 2 , C 0 ( i ) = π 2 48 + i G , C 0 ( i ) = π 2 48 i G , by ( 30 ) ;
and for 1 r m , C r ( 2 ) , C r ( i ) and C r ( i ) are found from (31).
Proof. 
Set z = 2 in Theorem 1. □
We now list some examples from (35):
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k 2 k + 1 = 3 π 100 3 π 2 400 + 3 25 ln 2 + 9 250 ln 2 2 13 125 G ,
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k ( k 1 ) 2 k + 2 = 3 100 29 π 1250 + 149 π 2 30000 58 625 ln 2 149 6250 ln 2 2 + 243 3125 G ,
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k 3 2 k + 1 = 13 375 + 1529 π 75000 577 π 2 150000 + 752 9375 ln 2 + 577 31250 ln 2 2 1903 31250 G .
Corollary 2.
Let m be a non-negative integer and let ( γ 1 , γ 2 , γ 3 ) = ( 1 , ( 3 + i 7 ) / 2 , ( 3 i 7 ) / 2 ) where i is the imaginary unit. Then
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k ( 1 ) k k m 4 k + 1 = ( 1 ) m + 1 k = 1 3 j = 0 m a j ( γ k ) C j ( γ k ) ;
where, for 0 r m , the coefficients are given by
a r ( γ 1 ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x 2 3 x + 4 ) m + 1 x = 1 ,
a r ( γ 2 ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x + 1 ) ( x ( 3 i 7 ) / 2 ) m + 1 x = ( 3 + i 7 ) / 2 ,
a r ( γ 3 ) = 1 ( m r ) ! d m r d x m r x m ( 1 x ) 2 m ( x + 1 ) ( x ( 3 + i 7 ) / 2 ) m + 1 x = ( 3 i 7 ) / 2 ;
C 0 ( γ 1 ) = Li 2 ( 1 ) = π 2 12 , C 0 ( γ 2 ) = Li 2 2 3 + i 7 , C 0 ( γ 3 ) = Li 2 2 3 i 7 , by ( 30 ) ;
and for 1 r m , C r ( γ 1 ) , C r ( γ 2 ) and C r ( γ 3 ) are found from (31).
Proof. 
Set z = 4 in Theorem 1. Note that γ 1 = 1 , γ 2 = ( 3 + i 7 ) / 2 and γ 3 = ( 3 i 7 ) / 2 are the roots of x ( 1 x ) 2 + 4 = 0 . □
We give one example from (43). At m = 0 we have
k = 0 H 3 k + 1 H k 3 k + 1 3 k k ( 1 ) k 4 k + 1 = π 2 96 4 7 Li 2 2 3 + i 7 5 + i 7 ,
since (43) at m = 0 gives
k = 0 H 3 k + 1 H k 3 k + 1 3 k k ( 1 ) k 4 k + 1 = π 2 96 + 2 i 7 Li 2 2 / 3 + i 7 5 + i 7 2 i 7 Li 2 2 / 3 i 7 5 i 7 ;
the right hand side of which can be simplified using
f g + f * g * = 2 f g 2 f g ,
for arbitrary functions f and g.
Differentiating (13) m times with respect to z gives
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k + m k z k + m + 1 = ( 1 ) m 0 1 ln x x ( 1 x ) 2 z m + 1 d x , | z | 1 .
Note that (49) holds for every real number m that is not a negative integer.
Setting z = 2 in (49) gives
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k + m k 2 k + m + 1 = ( 1 ) m 0 1 ln x ( ( x 2 ) ( x 2 + 1 ) ) m + 1 d x .
Theorem 2.
Let m be a non-negative integer and let z 1 , z 2 and z 3 be the distinct roots of x ( 1 x ) 2 z = 0 where z is a real number with | z | 1 . Then
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k + m k z k + m + 1 = ( 1 ) m k = 1 3 j = 0 m b j ( z k ) C j ( z k ) ;
where, for 0 r m ,
b r ( λ ) = 1 ( m r ) ! d m r d x m r ( x λ ) m + 1 x ( 1 x ) 2 z m + 1 x = λ ,
and C r ( z 1 ) , C r ( z 2 ) and C r ( z 3 ) are calculated using (30) and (31).
Proof. 
Consider the partial fraction decomposition
1 x ( 1 x ) 2 z m + 1 = j = 0 m b j ( z 1 ) ( x z 1 ) j + 1 + b j ( z 2 ) ( x z 2 ) j + 1 + b j ( z 3 ) ( x z 3 ) j + 1 = k = 1 3 j = 0 m b j ( z k ) ( x z k ) j + 1 ,
where, for 0 r m , the coefficients b r ( z 1 ) , b r ( z 2 ) , b r ( z 3 ) are found from (52). We therefore have
0 1 ln x x ( 1 x ) 2 z m + 1 d x = k = 1 3 j = 0 m b j ( z k ) 0 1 ln x ( x z k ) j + 1 d x = k = 1 3 j = 0 m b j ( z k ) C j ( z k ) , by ( 34 ) ;
and hence (51). □
Corollary 3.
Let m be a non-negative integer and let ( λ 1 , λ 2 , λ 3 ) = ( 2 , i , i ) where i is the imaginary unit. Then
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k + m k 2 k + m + 1 = ( 1 ) m k = 1 3 j = 0 m b j ( λ k ) C j ( λ k ) ;
where, for 0 r m , the coefficients are given by
b r ( 2 ) = 1 ( m r ) ! d m r d x m r 1 ( x 2 + 1 ) m + 1 x = 2 ,
b r ( i ) = 1 ( m r ) ! d m r d x m r 1 ( x 2 ) ( x i ) m + 1 x = i ,
b r ( i ) = 1 ( m r ) ! d m r d x m r 1 ( x 2 ) ( x + i ) m + 1 x = i ;
and C r ( 2 ) , C r ( i ) and C r ( i ) can be readily obtained from (30) and (31).
Proof. 
Set z = 2 in Theorem 2. □
Here are a couple of evaluations using (55):
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k + 1 2 k + 2 = 3 π 200 + π 2 150 + 3 50 ln 2 4 125 ln 2 2 + 37 G 250 ,
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k ( k + 1 ) ( k + 2 ) 2 k + 4 = 3 400 + 23 π 2500 + 27 π 2 10 , 000 + 23 625 ln 2 81 6250 ln 2 2 + 843 G 12 , 500 ,
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k k + 3 k 2 k + 4 = 83 12 , 000 + 3059 π 600 , 000 + 11 π 2 9375 + 1517 75 , 000 ln 2 88 15 , 625 ln 2 2 + 8137 G 250 , 000 .
Corollary 4.
Let m be a non-negative integer and let ( γ 1 , γ 2 , γ 3 ) = ( 1 , ( 3 + i 7 ) / 2 , ( 3 i 7 ) / 2 ) where i is the imaginary unit. Then
k = 0 H 3 k + 1 H k ( 3 k + 1 ) 3 k k ( 1 ) k k + m k 4 k + m + 1 = k = 1 3 j = 0 m b j ( γ k ) C j ( γ k ) ;
where, for 0 r m , the coefficients are given by
b r ( γ 1 ) = 1 ( m r ) ! d m r d x m r 1 ( x 2 3 x + 4 ) m + 1 x = 1 ,
b r ( γ 2 ) = 1 ( m r ) ! d m r d x m r 1 ( x + 1 ) ( x ( 3 i 7 ) / 2 ) m + 1 x = ( 3 + i 7 ) / 2 ,
b r ( γ 3 ) = 1 ( m r ) ! d m r d x m r 1 ( x + 1 ) ( x ( 3 + i 7 ) / 2 ) m + 1 x = ( 3 i 7 ) / 2 ;
C 0 ( γ 1 ) = Li 2 ( 1 ) = π 2 12 , C 0 ( γ 2 ) = Li 2 2 3 + i 7 , C 0 ( γ 3 ) = Li 2 2 3 i 7 , by ( 30 ) ;
and for 1 r m , C r ( γ 1 ) , C r ( γ 2 ) and C r ( γ 3 ) are found from (31).
Proof. 
Theorem 2 with z = 4 . □

3. Main Results, Part 2

This section deals with the second category of series, i.e., series of the form
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k z k = 0 1 ln x 1 x 1 z x ( 1 x ) 2 d x .
Differentiating (16) m times with respect to z and thereafter replacing z with 1 / z gives
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k m z k + 1 = ( 1 ) m 0 1 x m ( 1 x ) 2 m ln ( x / ( 1 x ) ) x ( 1 x ) 2 z m + 1 d x .
Theorem 3.
Let m be a non-negative integer and let z 1 , z 2 and z 3 be the distinct roots of x ( 1 x ) 2 z = 0 where z is a real number such that | z | 1 . Then
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k m z k + 1 = ( 1 ) m k = 1 3 j = 0 m a j ( z k ) C j ( z k ) + ( 1 ) j C j ( 1 z k ) ,
where a r ( z 1 ) , a r ( z 2 ) , a r ( z 3 ) are as defined in Theorem 1 and C 0 and C r can be found from (30) and (31) in Theorem 1.
Proof. 
The proof is similar to that of Theorem 1. Note that
0 1 ln ( x / ( 1 x ) ) ( x λ ) r + 1 d x = 0 1 ln x ( x λ ) r + 1 d x + ( 1 ) r 0 1 ln x ( x ( 1 λ ) ) r + 1 d x = C r ( λ ) + ( 1 ) r C r ( 1 λ ) .
Corollary 5.
Let m be a non-negative integer and let ( λ 1 , λ 2 , λ 3 ) = ( 2 , i , i ) where i is the imaginary unit. Then
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k m 2 k + 1 = ( 1 ) m k = 1 3 j = 0 m a j ( λ k ) C j ( λ k ) + ( 1 ) j C j ( 1 λ k ) ,
where a r ( 2 ) , a r ( i ) , a r ( i ) are as defined in (36)–(38) in Corollary 1 and C r are found from (30) and (31).
Proof. 
Theorem 3 with z = 2 . □
Examples from Corollary 5 include
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k 2 k + 1 = π 200 + 9 π 2 4000 + 3 100 ln 2 + 27 1000 ln 2 2 13 1000 π ln 2 ,
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k ( k 1 ) 2 k + 2 = 2 π 625 149 π 2 100 , 000 51 5000 ln 2 447 25 , 000 ln 2 2 + 243 25 , 000 π ln 2 .
Corollary 6.
Let m be a non-negative integer and let ( γ 1 , γ 2 , γ 3 ) = ( 1 , ( 3 + i 7 ) / 2 , ( 3 i 7 ) / 2 ) where i is the imaginary unit. Then
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k k m 4 k + 1 = ( 1 ) m + 1 k = 1 3 j = 0 m a j ( γ k ) C j ( γ k ) + ( 1 ) j C j ( 1 γ k ) ,
where a r ( γ 1 ) , a r ( γ 2 ) , a r ( γ 3 ) are as defined in (44)–(46) in Corollary 2 and C r are found from (30) and (31).
Proof. 
Theorem 3 with z = 4 . □
Here we present a couple of examples from (73).
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k k 4 k + 1 = 3 448 ln 2 9 512 ln 2 2 3 128 arctan 2 7 5 1 224 89 6272 ln 2 7 arctan 7 5 ,
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k k ( k 1 ) 4 k + 1 = 219 25 , 088 ln 2 + 93 4096 ln 2 2 + 31 1024 arctan 2 7 5 + 1 256 6651 351 , 232 ln 2 7 arctan 7 5 .
As counterpart of (49), we have
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k + m k z k + m + 1 = ( 1 ) m 0 1 ln x 1 x x ( 1 x ) 2 z m + 1 d x , | z | 1 .
Note that (76) holds for every real number m that is not a negative integer.
Setting z = 2 in (76) gives
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k + m k 2 k + m + 1 = ( 1 ) m 0 1 ln x 1 x ( ( x 2 ) ( x 2 + 1 ) ) m + 1 d x ,
while setting z = 4 gives
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k k + m k 4 k + m + 1 = 0 1 ln x 1 x ( ( x + 1 ) ( x 2 3 x + 4 ) ) m + 1 d x .
Proposition 3.
We have
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k 2 k + 1 = π ln ( 2 ) 20 3 40 ln 2 ( 2 ) π 2 160 .
Proof. 
Using (2) and (18), we have
0 1 ln x 1 x ( x 2 ) ( x 2 + 1 ) d x = 1 4 i 2 0 1 ln x 1 x x + i d x 1 4 i + 2 0 1 ln x 1 x x i d x + 1 5 0 1 ln x 1 x x 2 d x = 1 4 i 2 1 2 ln 2 i 1 i 1 4 i + 2 1 2 ln 2 i 1 i + 1 5 1 2 ln 2 2 1 2 = 1 4 i 2 1 2 ln 2 ( 1 i ) 1 4 i + 2 1 2 ln 2 ( 1 + i ) + 1 5 1 2 ln 2 2 ,
which simplifies to (79). □
Setting m = 0 in (78) gives
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k 4 k + 1 = 0 1 ln ( x / ( 1 x ) ) d x ( x + 1 ) ( x 2 3 x + 4 ) .
The integral occurring on the RHS can be evaluated. The result is stated in Proposition 4.
Proposition 4.
We have
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k 4 k + 1 = 3 64 ln 2 2 + 1 16 arctan 2 7 5 5 7 112 ln 2 arctan 7 5 .
Proof. 
We wish to evaluate the integral in (80). Proceeding as in Proposition 2, we have
0 1 ln ( x / ( 1 x ) ) d x ( x + 1 ) ( x 2 3 x + 4 ) = 1 8 0 1 ln ( x / ( 1 x ) ) d x x + 1 7 + 5 i 7 112 0 1 ln ( x / ( 1 x ) ) d x x ( 3 + i 7 ) / 2 7 5 i 7 112 0 1 ln ( x / ( 1 x ) ) d x x ( 3 i 7 ) / 2 ,
so that upon using (18) we obtain
0 1 ln ( x / ( 1 x ) ) d x ( x + 1 ) ( x 2 3 x + 4 ) = 1 16 ln 2 2 + 7 + 5 i 7 224 ln 2 1 + i 7 3 + i 7 + 7 5 i 7 224 ln 2 1 i 7 3 i 7 = 1 16 ln 2 2 + 1 112 ( 7 + 5 i 7 ) ln 2 1 + i 7 3 + i 7 ;
which simplifies to the RHS of (81) upon using ( f g ) = f g f g . □
Theorem 4.
Let m be a non-negative integer and let z 1 , z 2 and z 3 be the distinct roots of x ( 1 x ) 2 z = 0 where z is a real number such that | z | 1 . Then
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k + m k z k + m + 1 = ( 1 ) m k = 1 3 j = 0 m b j ( z k ) C j ( z k ) + ( 1 ) j C j ( 1 z k ) ,
where b r ( z 1 ) , b r ( z 2 ) , b r ( z 3 ) are as defined in Theorem 2 and C 0 and C r can be found from (30) and (31) in Theorem 1.
Proof. 
The proof is similar to that of Theorem 2. We evaluate the integral on the RHS of (76). □
Corollary 7.
Let m be a non-negative integer and let ( λ 1 , λ 2 , λ 3 ) = ( 2 , i , i ) where i is the imaginary unit. Then
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k + m k 2 k + m + 1 = ( 1 ) m k = 1 3 j = 0 m b j ( λ k ) C j ( λ k ) + ( 1 ) j C j ( 1 λ k ) ,
where b r ( 2 ) , b r ( i ) , b r ( i ) are as defined in (56)–(58) in Corollary 3 and C r are found from (30) and (31).
Proof. 
Theorem 4 with z = 2 . □
At m = 1 and m = 2 in (83) we obtain
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k + 1 2 k + 2 = π 400 + 37 2000 π ln 2 π 2 500 + 3 200 ln 2 3 125 ln 2 2
and
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k k + 2 k 2 k + 3 = 17 π 10 , 000 + 843 100 , 000 π ln 2 81 π 2 100 , 000 + 249 20 , 000 ln 2 243 25 , 000 ln 2 2 .
Corollary 8.
Let m be a non-negative integer and let ( γ 1 , γ 2 , γ 3 ) = ( 1 , ( 3 + i 7 ) / 2 , ( 3 i 7 ) / 2 ) where i is the imaginary unit. Then
k = 0 H 2 k H k ( 3 k + 1 ) 3 k k ( 1 ) k k + m k 4 k + m + 1 = k = 1 3 j = 0 m b j ( γ k ) C j ( γ k ) + ( 1 ) j C j ( 1 γ k ) ,
where b r ( γ 1 ) , b r ( γ 2 ) , b r ( γ 3 ) are as defined in (63)–(65) in Corollary 4 and C r are found from (30) and (31).
Proof. 
Use Theorem 4 with z = 4 . □

4. Concluding Remarks

In this paper we have applied integration techniques combined with complex analysis and partial fraction decompositions to derive closed forms for two general classes of series involving differences of harmonic numbers in the numerator. Such series may have applications in combinatorics, signal processing, analysis of algorithms and more.
Theorems similar to those in the previous section can be stated for alternating sums. There does not appear to be values of z, however, for which the integrals can be evaluated in terms of elementary functions. For a slightly different direction of future research we mention that replacing z by i z and comparing the real and imaginary parts gives integral relations of the form
k = 0 ( 1 ) k H 6 k + 1 H 2 k ( 6 k + 1 ) 6 k 2 k z k = 0 1 ln ( x ) 1 + z 2 x 2 ( 1 x ) 4 d x ,
k = 0 ( 1 ) k H 6 k + 4 H 2 k + 1 ( 6 k + 4 ) 6 k + 3 2 k + 1 z k = z 0 1 x ( 1 x ) 2 ln ( x ) 1 + z 2 x 2 ( 1 x ) 4 d x ,
k = 0 ( 1 ) k H 4 k H 2 k ( 6 k + 1 ) 6 k 2 k z k = 0 1 ln x 1 x 1 + z 2 x 2 ( 1 x ) 4 d x ,
and
k = 0 ( 1 ) k H 4 k + 2 H 2 k + 1 ( 6 k + 4 ) 6 k + 3 2 k + 1 z k = z 0 1 x ( 1 x ) 2 ln x 1 x 1 + z 2 x 2 ( 1 x ) 4 d x .
In each case one can attempt to evaluate the integrals appearing on the right hand side.

Author Contributions

Conceptualization, K.A. and R.F.; methodology, K.A. and R.F.; writing-original draft preparation, K.A.; writing-review and editing, K.A. and R.F.; All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

No new data were created or analyzed in this study. Data sharing is not applicable to this article.

Acknowledgments

We wish to thank the four referees for their time and the helpful comments that resulted in an improvement of the exposition.

Conflicts of Interest

The authors declare that there are no conflicts of interest.

References

  1. Sofo, A.; Nimbran, A.S. Euler sums and integral connections. Mathematics 2019, 7, 833. [Google Scholar] [CrossRef]
  2. Stewart, S.M. Explicit evaluation of some quadratic Euler-type sums containing double-index harmonic numbers. Tatra Mt. Math. Publ. 2020, 77, 73–98. [Google Scholar] [CrossRef]
  3. Li, C.; Chu, W. Evaluation of infinite series by integrals. Mathematics 2022, 10, 2444. [Google Scholar] [CrossRef]
  4. Li, C.; Chu, W. Improper integrals involving powers of inverse trigonometric and hyperbolic functions. Mathematics 2022, 10, 2980. [Google Scholar] [CrossRef]
  5. Li, C.; Chu, W. Infinite triple series by integrals. Publ. Inst. Math. Nouv. Ser. 2022, 112, 131–142. [Google Scholar] [CrossRef]
  6. Li, C.; Chu, W. Evaluating infinite series involving harmonic numbers by integration. Mathematics 2024, 12, 589. [Google Scholar] [CrossRef]
  7. Li, C.; Chu, W. Parametric integrals for binomial series with harmonic polynomials. Axioms 2024, 13, 885. [Google Scholar] [CrossRef]
  8. Adegoke, K.; Frontczak, R. Combinatorial identities with multiple harmonic-like numbers. AppliedMath 2024, 4, 986–998. [Google Scholar] [CrossRef]
  9. Adegoke, K.; Frontczak, R. Some notes on an identity of Frisch. Open J. Math. Sci. 2024, 8, 216–226. [Google Scholar] [CrossRef]
  10. Choi, J. Finite summation formulas involving binomial coefficients, harmonic numbers and generalized harmonic numbers. J. Inequalities Appl. 2013, 2013, 49. [Google Scholar] [CrossRef]
  11. Choi, J.; Srivastava, H.M. Some summation formulas involving harmonic numbers and generalized harmonic numbers. Math. Comput. Model. 2011, 54, 2220–2234. [Google Scholar] [CrossRef]
  12. Chu, W. Summation formulae involving harmonic numbers. Filomat 2012, 26, 143–152. [Google Scholar] [CrossRef]
  13. Wang, J.; Wei, C. Derivative operator and summation formulae involving generalized harmonic numbers. J. Math. Anal. Appl. 2016, 434, 315–341. [Google Scholar] [CrossRef]
  14. Sun, Z.-W. Series with summands involving harmonic numbers. arXiv 2023, arXiv:2210.07238v8. [Google Scholar]
  15. Au, K.C. Iterated integrals and multiple polylogarithm at algebraic arguments. arXiv 2022, arXiv:2201.01676. [Google Scholar]
  16. Sun, Z.-W. Infinite series involving binomial coefficients and harmonic numbers. SCIENTIA SINICA Math. 2024, 54, 765–774. [Google Scholar]
  17. Lewin, L. Polylogarithms and Associated Functions; Elsevier/North-Holland: Amsterdam, The Netherlands, 1981. [Google Scholar]
  18. Trićković, S.B.; Stanković, M.S. On the closed form of Clausen functions. Integral Transform. Spec. Funct. 2023, 34, 469–477. [Google Scholar] [CrossRef]
  19. Srivastava, H.M.; Choi, J. Series Associated with the Zeta and Related Functions; Springer Science+Media: Berlin/Heidelberg, Germany, 2001. [Google Scholar]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Adegoke, K.; Frontczak, R. Evaluation of Harmonic Number Series Involving the Binomial Coefficient C(3n,n) in the Denominator by Integration. AppliedMath 2025, 5, 31. https://doi.org/10.3390/appliedmath5010031

AMA Style

Adegoke K, Frontczak R. Evaluation of Harmonic Number Series Involving the Binomial Coefficient C(3n,n) in the Denominator by Integration. AppliedMath. 2025; 5(1):31. https://doi.org/10.3390/appliedmath5010031

Chicago/Turabian Style

Adegoke, Kunle, and Robert Frontczak. 2025. "Evaluation of Harmonic Number Series Involving the Binomial Coefficient C(3n,n) in the Denominator by Integration" AppliedMath 5, no. 1: 31. https://doi.org/10.3390/appliedmath5010031

APA Style

Adegoke, K., & Frontczak, R. (2025). Evaluation of Harmonic Number Series Involving the Binomial Coefficient C(3n,n) in the Denominator by Integration. AppliedMath, 5(1), 31. https://doi.org/10.3390/appliedmath5010031

Article Metrics

Back to TopTop