Preprint
Article

This version is not peer-reviewed.

Using Beta Function and Factorization to Find New π-Series

Submitted:

11 November 2023

Posted:

13 November 2023

You are already at the latest version

Abstract
Using the method of the beta function, Sun has recently evaluated some series of the type $\sum_{k=0}^{\infty}(ak+b)x^k/\binom{mk}{nk}$. By factorization of the polynomial $t^{m-n}(1-t)^n-x$, we will give a general method to find new $\pi$-series.
Keywords: 
;  ;  

1. Introduction

The classical rational Ramanujan-type series for π introduced by Cooper[6] have the form
k = 0 b k + c m k a ( k ) = λ d π .
In 1905, Glaisher[7] proved that
k = 0 ( 4 k 1 ) 2 k k 4 ( 2 k 1 ) 4 256 k = 8 π 2 .
Chan, Chan and Liu[4] proved that
n = 0 5 n + 1 64 n D n = 8 3 π ,
where D n denotes the Domb number k = 0 n n k 2 2 k k 2 n 2 k n k .
Let b , c Z , for each n N , the coefficient of x n in the expansion of ( x 2 + b x + c ) n is denoted by T n ( b , c ) , Sun[11] gave many conjectural series for 1 π containing the T n ( b , c ) such as
k = 0 3990 k + 1147 ( 288 ) 3 k T k ( 62 , 95 2 ) 3 = 432 95 π ( 94 2 + 195 14 ) .
In[10], Sun derived several identities involving π by the telescoping method. For example, from Bauer’s series[2]
k = 0 ( 4 k + 1 ) 2 k k 3 ( 64 ) k = 2 π
and the telescoping sum
k = 0 n ( 16 k 3 4 k 2 2 k + 1 ) 2 k k 2 ( 2 k 1 ) 2 ( 64 ) k = 8 ( 2 n + 1 ) ( 64 ) n 2 n n 3 ,
he deduced
k = 0 k ( 4 k 1 ) 2 k k 3 ( 2 k 1 ) 2 ( 64 ) k = 1 π .
By Gosper’s algorithm[8], Hou and Li[9] give a systematic method to construct π -series of the form (1). For more results on the π -series, we refer to [3][5] .
In 1974 Gosper announced the new identity
k = 0 25 k 3 2 k 3 k k = π 2 .
Motivated by Gosper’s identity, Almkvist, Krattenthaler and Petersson[1] found some new identities of the type
k = 0 P ( k ) m k n k x k = π ,
where P ( k ) is a polynomial in k. If P ( k ) has degree d, they also get that
k = 0 P ( k ) m k n k x k = 0 1 S ( t , x ) ( t n ( 1 t ) m n x ) d + 2 d t ,
where S ( t , x ) is a polynomial in t of degree m ( d + 1 ) . They find a good way to get π is to have a r c t a n ( t ) or a r c t a n ( t 1 ) after integration of (2). Then they get t n ( 1 t ) m n x must have the factor t 2 + 1 or ( t 1 ) 2 + 1 . This restricts m and n and gives the value of x.
Using the method of the beta function, Sun[12] has recently evaluated some series of the type k = 0 ( a k + b ) x k / m k n k . For example, he completely determined the values of
k = 1 k r x k 3 k k 27 4 < x < 27 4
for r = 0 , ± 1 .
We first consider the convergence of the left of the equality (2), by Stirling’s formula
k ! 2 π k k e k , k + .
If m > n > 0 are integers, then k + , we have
1 / m k n k 2 π n ( m n ) k m n n ( m n ) m n m m .
So we get that if
| x | > n n ( m n ) m n m m ,
the left of the equality (2) is convergent.
The Gamma function is pointed out by Euler as
Γ ( x ) = 0 + t x 1 e t d t , ( x > 0 ) .
The beta function is defined as
B ( a , b ) = 0 1 x a 1 ( 1 x ) b 1 d x , a > 0 , b > 0 .
The connection between Gamma function and beta function is given by Euler as
B ( a , b ) = Γ ( a ) Γ ( b ) Γ ( a + b ) .
Now we present an auxiliary proposition.
Proposition 1. 
Let m > n > 0 be integers, 0 t 1 and | x | > n n ( m n ) m n / m m , then we have
| t n ( 1 t ) m n x | < 1 .
Proof. 
Note that for 0 t 1 , we have
t n n 1 t m n m n m n · t n + ( m n ) · 1 t m n m = 1 m ,
Hence we have
| t n ( 1 t ) m n x | n n ( m n ) m n m m | 1 x | < 1 .
By Proposition (1) and beta function we get the following lemma we will use later. □
Lemma 1. 
Let m > n > 0 be integers, and let x be real numbers with | x | > ( n n ( m n ) m n ) / m m , then we have
k = 0 1 x k m k n k = 0 1 x [ ( m 1 ) t m n ( 1 t ) n + x ) ] ( t m n ( 1 t ) n x ) 2 d t .
Proof. 
Clearly we have
k = 0 1 x k m k n k = k = 0 ( n k ) ! ( ( m n ) k ) ! x k ( m k ) ! = k = 0 ( m k + 1 ) Γ ( n k + 1 ) Γ ( ( m n ) k + 1 ) x k Γ ( m k + 2 ) = k = 0 m k + 1 x k B ( n k + 1 , ( m n ) k + 1 ) = k = 0 m k + 1 x k 0 1 t ( m n ) k ( 1 t ) n k d t = 0 1 k = 0 ( m k + 1 ) t m n ( 1 t ) n x k d t .
Let
y = t m n ( 1 t ) n x .
Then by Proposition 1 we have | y | < 1 , then we get
0 1 k = 0 ( m k + 1 ) y k d t = 0 1 ( m 1 ) y + 1 ( 1 y ) 2 d t = 0 1 x [ ( m 1 ) t m n ( 1 t ) n + x ) ] ( t m n ( 1 t ) n x ) 2 d t .
Using the same method of Lemma 1, we get more equalities we will use later.
Lemma 2. 
Let m > n > 0 be integers, and let x be real numbers with | x | > ( n n ( m n ) m n ) / m m , then we have
k = 0 1 ( m k + 1 ) x k m k n k = 0 1 x t m n ( 1 t ) n x d t .
k = 0 k x k m k n k = 0 1 t m n ( 1 t ) n [ ( m 1 ) t m n ( 1 t ) n + ( m + 1 ) x ] x ( t m n ( 1 t ) n x ) 3 d t .
k = 1 1 k x k m k n k = 0 1 n t m n ( 1 t ) n 1 [ t m n ( 1 t ) n x ] d t .
Proof. 
The proof of equality (4) and (5) is similar as the Lemma 1. To prove the equality (6), we have
k = 1 1 k x k m k n k = k = 1 1 k x k ( n k ) ! ( ( m n ) k ) ! ( m k ) ! = k = 1 n k k x k Γ ( n k ) Γ ( ( m n ) k + 1 ) Γ ( m k + 1 ) = k = 1 n x k B ( n k , ( m n ) k + 1 ) = k = 1 n 0 1 t ( m n ) k ( 1 t ) n k 1 x k d t = 0 1 k = 1 n 1 t t m n ( 1 t ) n x k d t .
Let
y = t m n ( 1 t ) n x .
Then by Proposition 1 we have | y | < 1 , then we get
0 1 k = 1 n 1 t y k d t = 0 1 n y ( 1 t ) ( 1 y ) d t = 0 1 n t m n ( 1 t ) n 1 [ t m n ( 1 t ) n x ] d t .
By equality (2) and Lemma 1 and 2, we find the denominator of the integration contains the form t m n ( 1 t ) n x , if we integrate directly, as x is a symbol we just don’t know, in most times we will fail. □
In this paper, we replace x with a suitable rational function x ( b ) , then by Mathematica we can integrate the rational function f ( t , b ) from 0 to 1 which the denominator contains the form t m n ( 1 t ) n x ( b ) . Since after integration we get equalities with b and the variable t no longer exists. Then we combine these equalities and let b be a suitable number, this gives lights to find new π -series. we will show how to get π -series containing the type 1 / 2 k k , 1 / 3 k k , 1 / 4 k k , 1 / 4 k 2 k . For example, we get new π -series as follows
Example 1. 
k = 0 ( 3 + 2 k ) 2 k 2 k k = π 2 .
k = 1 ( 9 + 35 k ) 8 k 1 k ( 1 + 3 k ) 3 k 3 k k = 3 2 π 1 .
k = 1 ( 64 + 211 k + 1805 k 2 ) 9 k k ( 1 + 4 k ) 8 k 4 k k = 45 + 16 3 π .
k = 1 3 + 14 k 2 k ( 1 + 4 k ) ( 3 ) k 4 k 2 k = 1 6 ( 6 + 3 π ) .
k = 1 ( 1 + 6 k ) 2 2 k 1 k ( 1 + 4 k ) 4 k 2 k = π 2 1 .

2. π -series containing

In this section, we will give methods to find π -series containing the type 1 / 2 k k .
Theorem 1. 
If b 0 , we have
k = 0 1 ( b 2 + 1 / 4 ) k 2 k k = 1 + 4 b 2 8 b 3 2 b + a r c t a n 1 2 b .
If | b | > 2 2 , we have
k = 0 1 ( 1 / 4 b 2 ) k 2 k k = 4 b 2 1 8 b 3 2 b a r c t a n h 1 2 b ,
Proof. 
By Lemma 1 we have
k = 0 1 x k 2 k k = 0 1 x ( t 2 t x ) ( t 2 t + x ) 2 d t .
By inequality (3) we have | x | > 1 4 . □
Since t 2 t + x = ( t 1 2 ) 2 + x 1 4 . If x > 1 4 we let x 1 4 = b 2 , then we b 0 and by Mathematica we have
k = 0 1 x k 2 k k = 0 1 x ( t 2 t x ) ( t 2 t + x ) 2 d t = 0 1 ( b 2 + 1 4 ) ( t 2 t b 2 1 4 ) ( ( t 1 2 ) 2 + b 2 ) 2 d t = 1 + 4 b 2 8 b 3 2 b + a r c t a n 1 2 b .
If x < 1 4 we let x 1 4 = b 2 , then | b | > 2 2 and by Mathematica we have
k = 0 1 x k 2 k k = 0 1 x ( t 2 t x ) ( t 2 t + x ) 2 d t = 0 1 ( b 2 1 4 ) ( t 2 t + b 2 + 1 4 ) ( ( t 1 2 ) 2 b 2 ) 2 d t = 4 b 2 1 8 b 3 2 b a r c t a n h 1 2 b .
Using the same method of Theorem 1 and by Lemma 2 we have
Theorem 2. 
If b 0 , we have
k = 0 1 ( 2 k + 1 ) ( b 2 + 1 / 4 ) k 2 k k = 1 + 4 b 2 2 b a r c t a n 1 2 b .
k = 0 k ( b 2 + 1 / 4 ) k 2 k k = ( 1 + 4 b 2 ) [ 6 b + ( 3 + 4 b 2 ) a r c t a n 1 2 b ] 64 b 5 .
k = 1 1 k ( b 2 + 1 / 4 ) k 2 k k = a r c t a n 1 2 b b .
If | b | > 2 2 , we have
k = 0 1 ( 2 k + 1 ) ( 1 / 4 b 2 ) k 2 k k = 4 b 2 1 2 b a r c t a n h 1 2 b .
k = 0 k ( 1 / 4 b 2 ) k 2 k k = ( 4 b 2 1 ) [ 6 b + ( 4 b 2 3 ) a r c t a n h 1 2 b ] 64 b 5 .
k = 1 1 k ( 1 / 4 b 2 ) k 2 k k = a r c t a n h 1 2 b b .
By Theorem 1 and 2 we can get propositions of equality.
Proposition 2. 
If b 0 , Via ( 3 ) × ( ) + 8 b 2 × ( ) we have
k = 0 24 b 2 k ( 1 / 4 + b 2 ) k 2 k k = 1 + 4 b 2 2 b a r c t a n 1 2 b
By equality (8) and (11) we let b = 1 2 , 3 6 , 3 2 , we can get series involves π , here are the examples.
Example 2. 
k = 0 2 k ( 1 + 2 k ) 2 k k = π 2 .
k = 0 3 k ( 1 + 2 k ) 2 k k = 4 π 3 3 .
k = 0 1 ( 1 + 2 k ) 2 k k = 2 π 3 3 .
k = 0 ( 3 + 2 k ) 2 k 2 k k = π 2 .
k = 0 ( 9 + 2 k ) 3 k 2 k k = 4 π 3 .
k = 0 ( 1 + 2 k ) 2 k k = 2 π 9 3 .
3. π -series containing 1 / 3 k k
In this section, we will give methods to find π -series containing the type 1 / 3 k k .
We first give an auxiliary lemma we will use.
Lemma 3. 
Let f ( t ) = t 3 t 2 + x , if | x | > 4 27 , let x = 2 b ( 1 + 2 b ) 2 , then we can factor f ( t ) as
f ( t ) = ( t 1 2 b ) ( ( t + b ) 2 + b ( 2 + 3 b ) ) ,
where b < 2 3 or b > 1 6 ( 1 + 2 ) 2 / 3 + 1 ( 1 + 2 ) 2 / 3 1 3 0.059 .
Proof. 
Consider f ( t ) = 0 , since | x | > 4 27 , it is easy to calculate that
Δ = x ( 27 x 4 ) 108 > 0
Then f ( t ) = 0 has one real roots and one pair of unequal conjugate complex roots. Suppose we factor f ( t ) as
f ( t ) = ( t + a ) ( ( t + b ) 2 + c ) ,
then we have
a + 2 b = 1 2 a b + b 2 + c = 0 a ( b 2 + c ) = x
That is to say a = ( 1 + 2 b ) and c = b ( 3 b + 2 ) and x = 2 b ( 1 + 2 b ) 2 . Since | x | > 4 27 , we have | 2 b ( 1 + 2 b ) 2 | > 4 27 , then we get b < 2 3 or b > 1 6 ( 1 + 2 ) 2 / 3 + 1 ( 1 + 2 ) 2 / 3 1 3 . □
Theorem 3. 
Let μ be defined by
μ = 1 6 ( 1 + 2 ) 2 / 3 + 1 ( 1 + 2 ) 2 / 3 1 3 .
If we define q ( b ) as
q ( b ) = a r c t a n 2 + 3 b b 1 4 b + 3 , b < 3 4 , o r b > μ π 2 , b = 3 4 a r c t a n 2 + 3 b b 1 4 b + 3 π , 3 4 < b < 3 2
If b > μ or b < 2 3 , we have
k = 0 1 3 k k ( 2 b ( 1 + 2 b ) 2 ) k = 27 b ( 1 + 2 b ) 2 ( 2 + 3 b ) ( 1 + 6 b ) 2 + 6 b ( 1 + 2 b ) ( 1 + 6 b ) 3 log 2 b 1 + 2 b + 2 ( 1 + 2 b ) ( 18 b 2 + 6 b 1 ) ( 2 + 3 b ) ( 1 + 6 b ) 3 b 2 + 3 b q ( b ) ,
Proof. 
By Lemma 1 we have
k = 0 1 x k 3 k k = 0 1 x ( 2 t 3 2 t 2 x ) ( t 3 t 2 + x ) 2 d t .
By equation (3) we have | x | > 4 27 . □
Let x = 2 b ( 1 + 2 b ) 2 , then we have t 3 t 2 + x = ( t ( 2 b + 1 ) ) ( ( t + b ) 2 + b ( 3 b + 2 ) ) . If | x | > 4 27 , we have b < 2 3 or b > μ . If b 3 4 , we have
k = 0 1 x k 3 k k = 0 1 x ( 2 t 3 2 t 2 x ) ( t 3 t 2 + x ) 2 d t = 0 1 2 b ( 1 + 2 b ) 2 ( 2 t 3 2 t 2 + 2 b ( 1 + 2 b ) 2 ) ( t ( 2 b + 1 ) ) 2 ( ( t + b ) 2 + b ( 3 b + 2 ) ) 2 d t
We will calculate the integral by Mathematica. Then we get
k = 0 1 ( 2 b ( 1 + 2 b ) 2 ) k 3 k k = b ( 2 + 3 b ) 3 / 2 1 + 2 b ( 1 + 6 b ) 3 [ 27 b 2 + 3 b ( 2 b + 1 ) ( 6 b + 1 ) + 6 b 2 + 3 b ( 2 + 3 b ) log 2 b 1 + b + 2 ( 18 b 2 + 6 b 1 ) a r c t a n 1 + b b 2 + 3 b a r c t a n b 2 + 3 b ]
Let
q ( b ) = a r c t a n 1 + b b 2 + 3 b a r c t a n b 2 + 3 b ,
It’s easy to deduce that
q ( b ) = a r c t a n 2 + 3 b b 1 4 b + 3 , b < 3 4 , o r b > 0 a r c t a n 2 + 3 b b 1 4 b + 3 π , 3 4 < b < 2 3
In the domain of b we have
b ( 2 + 3 b ) 3 / 2 b 2 + 3 b = b 2 + 3 b
and
b 2 + 3 b = b 2 + 3 b .
If b = 3 4 , the theorem also holds, so we complete the proof.
Using the same method of Theorem 3 and by Lemma 2 we have
Theorem 4. 
Let μ be defined by
μ = 1 6 ( 1 + 2 ) 2 / 3 + 1 ( 1 + 2 ) 2 / 3 1 3 .
If we define q ( b ) as
q ( b ) = a r c t a n 2 + 3 b b 1 4 b + 3 , b < 3 4 , o r b > μ π 2 , b = 3 4 a r c t a n 2 + 3 b b 1 4 b + 3 π , 3 4 < b < 3 2
If b > μ or b < 2 3 , we have
k = 0 1 ( 1 + 3 k ) 3 k k ( 2 b ( 1 + 2 b ) 2 ) k = 3 b ( 1 + 2 b ) 1 + 6 b log 2 b 1 + 2 b + 2 ( 1 + 2 b ) ( 1 + 3 b ) ( 1 + 6 b ) b 2 + 3 b q ( b ) .
k = 0 k ( 2 b ( 1 + 2 b ) 2 ) k 3 k k = 81 b ( 1 + 2 b ) 2 ( 2 + 3 b ) 2 ( 1 + 6 b ) 4 + 6 b ( 1 + 2 b ) ( 1 + 8 b + 12 b 2 ) ( 1 + 6 b ) 5 log 2 b 1 + 2 b + 2 ( 1 + 2 b ) f ( b ) ( 2 + 3 b ) 2 ( 1 + 6 b ) 5 b 2 + 3 b q ( b ) .
where f ( b ) = 1 45 b 150 b 2 + 54 b 3 + 648 b 4 + 648 b 5 .
k = 1 1 k ( 2 b ( 1 + 2 b ) 2 ) k 3 k k = 1 1 + 6 b log 2 b 1 + 2 b + 2 1 + 6 b b 2 + 3 b q ( b ) .
Remark 1. 
If b < 0 , b is defined by b = | b | i
By Theorem 3 and 4 we can get propositions of equality.
Proposition 3. 
Let μ be defined by
μ = 1 6 ( 1 + 2 ) 2 / 3 + 1 ( 1 + 2 ) 2 / 3 1 3 .
If we define q ( b ) as
q ( b ) = a r c t a n 2 + 3 b b 1 4 b + 3 , b < 3 4 , o r b > μ π 2 , b = 3 4 a r c t a n 2 + 3 b b 1 4 b + 3 π , 3 4 < b < 3 2
If b > μ or b < 2 3 , we have:
Via ( 13 ) + 3 b ( 1 + 2 b ) × ( 15 ) we get
k = 1 ( 18 b 2 + 9 b + 1 ) k + 3 b ( 1 + 2 b ) k ( 1 + 3 k ) ( 2 b ( 1 + 2 b ) 2 ) k 3 k k = 2 ( 1 + 2 b ) b 3 + 2 b q ( b ) 1
Via ( 1 + 6 b ) 2 × ( 12 ) + 2 × ( 13 ) we get
k = 0 ( 1 + 6 b ) 2 k + ( 1 + 4 b + 12 b 2 ) ( 1 + 3 k ) ( 2 b ( 1 + 2 b ) 2 ) k 3 k k = 9 b ( 1 + 2 b ) 2 2 + 3 b + 2 ( 1 + 2 b ) 2 2 + 3 b b 2 + 3 b q ( b )
By Proposition 3, by equality (16) and (17) we let b = 1 and b = 3 4 , we can get four π -series, here are the examples.
Example 3. 
k = 1 3 + 10 k k ( 1 + 3 k ) 2 k 3 k k = 1 + π 2
k = 1 ( 9 + 35 k ) 8 k 1 k ( 1 + 3 k ) 3 k 3 k k = 3 2 π 1
k = 0 9 + 25 k ( 1 + 3 k ) 2 k 3 k k = π 2 + 9
k = 0 ( 19 + 49 k ) 8 k 4 ( 1 + 3 k ) 3 k 3 k k = 27 4 + 3 π
4. π -series containing 1 / 4 k k
In this section, we will give methods to find π -series containing the type 1 / 4 k k .
We first give an auxiliary lemma we will use.
Lemma 4. 
Let f ( t ) = t 4 t 3 + x , let x = 8 b 3 ( 1 + 2 b ) 3 ( 1 + 4 b ) 2 , if x > 27 256 , then we can factor f ( t ) as
f ( t ) = t b 1 2 2 + ( 1 + 2 b ) 2 ( 1 + 4 b ) 4 ( 1 + 4 b ) ( t + b ) 2 + b 2 ( 3 + 4 b ) 1 + 4 b ,
where b > 1 4 or b < 3 4 .
If x < 27 256 , then we can factor f ( t ) as
f ( t ) = t 2 ( 1 + 2 b ) t + 2 b ( 1 + 2 b ) 2 1 + 4 b ( t + b ) 2 + b 2 ( 3 + 4 b ) 1 + 4 b ,
where 0.33 1 4 μ < b < 1 4 + μ 0.17 and b 1 4 and
μ = 2 8 3 ( 1 + 2 ) 2 / 3 + 2 ( 1 + 2 ) 1 / 3 + 3 ( 1 + 2 ) 1 / 3 0.08 .
Proof. 
Consider f ( t ) = 0 , since | x | > 27 256 , we can deduce that
Δ = 12288 x 2 ( 256 x 27 ) < 0 , x > 27 256 > 0 , x < 27 256
If x > 27 256 , then f ( t ) = 0 has two pairs of unequal conjugate complex roots. Suppose we factor f ( t ) as
f ( t ) = ( ( t + a ) 2 + c ) ( ( t + b ) 2 + d ) ,
Then we have
2 ( a + b ) = 1 a 2 + b 2 + 4 a b + c + d = 0 2 ( a 2 + c ) b + 2 ( b 2 + d ) a = 0 ( a 2 + c ) ( b 2 + d ) = x
That is to say a = 1 2 b and c = ( 1 + 2 b ) 2 ( 1 + 4 b ) 4 ( 1 + 4 b ) and d = b 2 ( 3 + 4 b ) 1 + 4 b and x = 8 b 3 ( 1 + 2 b ) 3 ( 1 + 4 b ) 2 . If x > 4 27 , we have b < 3 4 or b > 1 4 . □
If x < 27 256 , then f ( t ) = 0 has two unequal real roots and a pair of conjugate complex roots. Suppose we factor f ( t ) as
f ( t ) = ( t + a ) ( t + c ) ( ( t + b ) 2 + d ) = ( t + ( a + c ) t + a c ) ( ( t + b ) 2 + d ) ,
Then we have
a + c + 2 b = 1 a c + 2 ( a + c ) b + b 2 + d = 0 2 a c b + ( a + c ) ( b 2 + d ) = 0 a c ( b 2 + d ) = x a c
That is to say a + c = ( 1 + 2 b ) and a c = 2 b ( 1 + 2 b ) 2 1 + 4 b and d = b 2 ( 3 + 4 b ) 1 + 4 b and x = 8 b 3 ( 1 + 2 b ) 3 ( 1 + 4 b ) 2 . Since x < 27 256 and a c , we have 1 4 μ < b < 1 4 + μ and b 1 4 .
Theorem 5. 
Let
μ = 2 8 3 ( 1 + 2 ) 2 / 3 + 2 ( 1 + 2 ) 1 / 3 + 3 ( 1 + 2 ) 1 / 3 .
If b > 1 4 or b < 3 4 or 1 4 μ < b < 1 4 + μ and b 1 4 , then we have
k = 0 ( 1 + 4 b ) 2 k ( 8 b 3 ( 1 + 2 b ) 3 ) k 4 k k = 2048 b 3 ( 1 + 2 b ) 3 ( 1 + 4 b ) ( 3 + 4 b ) ( 3 + 16 b + 32 b 2 ) 2 + 6 b ( 1 + 2 b ) ( 1 + 4 b ) ( 1 + 4 b ) f 1 ( b ) ( 1 + 4 b ) 2 ( 3 + 16 b + 32 b 2 ) 3 a r c t a n ( 1 + 4 b ) ( 1 + 4 b ) 1 + 8 b 2 + 6 b ( 1 + 2 b ) ( 1 + 4 b ) ( 3 + 4 b ) f 2 ( b ) ( 3 + 4 b ) 2 ( 3 + 16 b + 32 b 2 ) 3 a r c t a n ( 1 + 4 b ) ( 3 + 4 b ) 1 + 8 b + 8 b 2 + 6 b ( 1 + 2 b ) ( 1 + 4 b ) ( 1 + 16 b + 32 b 2 ) ( 3 + 16 b + 32 b 2 ) 3 log 4 b 2 ( 1 + 2 b ) 2
where
f 1 ( b ) = 1 + 12 b + 176 b 2 + 640 b 3 + 512 b 4
and
f 2 ( b ) = 9 60 b 16 b 2 + 384 b 3 + 512 b 4 .
Proof. 
By Lemma 1 we have
k = 0 1 x k 4 k k = 0 1 x ( 3 t 4 3 t 3 x ) ( t 4 t 3 + x ) 2 d t .
By equation (3) we have | x | > 27 256 . □
Let x = 8 b 3 ( 1 + 2 b ) 3 ( 1 + 4 b ) 2 . Since | x | > 27 256 , we have b > 1 4 or b < 3 4 or 1 4 μ < b < 1 4 + μ and b 1 4 . By Mathematica we have
k = 0 1 x k 4 k k = 0 1 x ( 3 t 4 3 t 3 x ) ( t 4 t 3 + x ) 2 d t = b 3 ( 1 + 2 b ) 3 ( 1 + 4 b ) 2 ( 3 + 16 b + 32 b 2 ) 3 [ 2048 ( 1 + 4 b ) 2 ( 3 + 16 b + 32 b 2 ) 3 + 8 b + 16 b 2 + 6 ( 1 + 4 b ) 2 f 1 ( b ) b 2 ( 1 + 2 b ) 2 ( 1 + 4 b ) 1 + 4 b 1 + 4 b a r c t a n 1 1 + 4 b 1 + 4 b a r c t a n 1 + 2 b ( 1 + 2 b ) 1 + 4 b 1 + 4 b + 6 ( 1 + 4 b ) 2 f 2 ( b ) b 2 ( 1 + 2 b ) 2 ( 3 + 4 b ) 1 + 4 b 3 + 4 b a r c t a n ( 1 + b ) 1 + 4 b 3 + 4 b b a r c t a n 1 + 4 b 3 + 4 b + 6 ( 1 + 4 b ) 3 ( 1 + 16 b + 32 b 2 ) b 2 ( 1 + 2 b ) 2 log 4 b 2 ( 1 + 2 b ) 2 ] .
where
f 1 ( b ) = 1 + 12 b + 176 b 2 + 640 b 3 + 512 b 4
and
f 2 ( b ) = 9 60 b 16 b 2 + 384 b 3 + 512 b 4 .
Since in the domain of b we have
1 + 4 b 1 + 4 b a r c t a n 1 1 + 4 b 1 + 4 b a r c t a n 1 + 2 b ( 1 + 2 b ) 1 + 4 b 1 + 4 b = ( 1 + 4 b ) ( 1 + 4 b ) 1 + 4 b a r c t a n ( 1 + 4 b ) ( 1 + 4 b ) 1 + 8 b 2
and
1 + 4 b 3 + 4 b a r c t a n ( 1 + b ) 1 + 4 b 3 + 4 b b a r c t a n 1 + 4 b 3 + 4 b = ( 1 + 4 b ) ( 3 + 4 b ) 3 + 4 b a r c t a n ( 1 + 4 b ) ( 3 + 4 b ) 1 + 8 b + 8 b 2 .
Then we get the desired results.
Using the same method of Theorem 5 and by Lemma 2 we have
Theorem 6. 
Let
μ = 2 8 3 ( 1 + 2 ) 2 / 3 + 2 ( 1 + 2 ) 1 / 3 + 3 ( 1 + 2 ) 1 / 3 .
If b > 1 4 or b < 3 4 or 1 4 μ < b < 1 4 + μ and b 1 4 , then we have
k = 0 ( 1 + 4 b ) 2 k ( 8 b 3 ( 1 + 2 b ) 3 ) k ( 4 k + 1 ) 4 k k = 2 b ( 1 + 2 b ) ( 1 + 4 b + 16 b 2 ) ( 1 + 4 b ) ( 1 + 4 b ) ( 1 + 4 b ) ( 1 + 4 b ) ( 3 + 16 b + 32 b 2 ) a r c t a n ( 1 + 4 b ) ( 1 + 4 b ) 8 b 2 1 + 2 b ( 1 + 2 b ) ( 3 + 12 b + 16 b 2 ) ( 1 + 4 b ) ( 3 + 4 b ) ( 1 + 4 b ) ( 3 + 4 b ) ( 3 + 16 b + 32 b 2 ) a r c t a n ( 1 + 4 b ) ( 3 + 4 b ) 1 + 8 b + 8 b 2 2 b ( 1 + 2 b ) ( 1 + 4 b ) 3 + 16 b + 32 b 2 log 4 b 2 ( 1 + 2 b ) 2 .
k = 1 ( 1 + 4 b ) 2 k ( 8 b 3 ( 1 + 2 b ) 3 ) k k 4 k k = 3 ( 1 + 2 b ) ( 1 + 4 b ) ( 1 + 4 b ) ( 1 + 4 b ) ( 3 + 16 b + 32 b 2 ) a r c t a n ( 1 + 4 b ) ( 1 + 4 b ) 1 + 8 b 2 + 6 b ( 1 + 4 b ) ( 3 + 4 b ) ( 3 + 4 b ) ( 3 + 16 b + 32 b 2 ) a r c t a n ( 1 + 4 b ) ( 3 + 4 b ) 1 + 8 b + 8 b 2 + 3 ( 1 + 4 b ) 2 ( 3 + 16 b + 32 b 2 ) log 4 b 2 ( 1 + 2 b ) 2 .
By Theorem 5 and 6 we can get propositions of equality.
Proposition 4. 
Let
μ = 2 8 3 ( 1 + 2 ) 2 / 3 + 2 ( 1 + 2 ) 1 / 3 + 3 ( 1 + 2 ) 1 / 3 .
If b > 1 4 or b < 3 4 or 1 4 μ < b < 1 4 + μ and b 1 4 , then we have:
Via ( 3 + 4 b ) ( 3 + 16 b + 32 b 2 ) 2 × ( 22 ) + 3 ( 1 + 4 b ) ( 3 + 8 b ) × ( 23 ) 128 b 2 ( 1 + 2 b ) 3 × ( 24 ) we have
[ ( 3 + 4 b ) ( 3 + 16 b + 32 b 2 ) 2 k 2 + f 1 ( b ) k 32 b 2 ( 1 + 2 b ) 3 ] ( 1 + 4 b ) 2 k k ( 1 + 4 k ) ( 8 b 3 ( 1 + 2 b ) 3 ) k 4 k k = 3 16 b 2 ( 1 + 2 b 2 ) 16 b 2 1 a r c t a n 16 b 2 1 1 + 8 b 2 f 2 ( b ) ( 4 b 1 ) 2
where
f 1 ( b ) = 9 + 96 b + 328 b 2 + 448 b 3 + 256 b 4
and
f 2 ( b ) = ( 1 + 4 b ) ( 1 + 4 b ) ( 3 8 b + 8 b 2 ) .
Via ( 1 + 4 b ) ( 3 + 16 b + 32 b 2 ) 2 × ( 22 ) 3 ( 1 + 4 b ) ( 1 + 8 b ) × ( 23 ) 256 b 3 ( 1 + 2 b ) 2 × ( 24 ) we have
[ ( 1 + 4 b ) ( 3 + 16 b + 32 b 2 ) 2 k 2 f 3 ( b ) k 64 b 3 ( 1 + 2 b ) 2 ] ( 1 + 4 b ) 2 k k ( 1 + 4 k ) ( 8 b 3 ( 1 + 2 b ) 3 ) k 4 k k = 3 f 4 ( b ) 16 b 2 ( 1 + 2 b ) 2 ( 1 + 4 b ) ( 3 + 4 b ) a r c t a n ( 1 + 4 b ) ( 3 + 4 b ) 1 + 8 b + 8 b 2 ( 3 + 4 b ) 2
where
f 3 ( b ) = 3 + 24 b + 40 b 2 + 64 b 3 + 256 b 4
and
f 2 ( b ) = ( 1 + 4 b ) ( 3 + 4 b ) ( 3 + 16 b + 8 b 2 ) .
By Proposition 4, in equation (25) we let b = 1 2 and in (26) we let b = 1 , we can get one π -series, here is the example.
Example 4. 
k = 1 ( 64 + 211 k + 1805 k 2 ) 9 k k ( 1 + 4 k ) 8 k 4 k k = 45 + 16 3 π
5. π -series containing 1 / 4 k 2 k
In this section, we will give methods to find π -series containing the type 1 / 4 k 2 k .
We first give an auxiliary lemma we will use.
Lemma 5. 
Let f ( t ) = t 4 2 t 3 + t 2 x , if x < 1 16 , let x = b ( 1 + b ) ( 1 + 2 b ) 2 , then we can factor f ( t ) as
f ( t ) = ( t b 1 2 + b ( 1 + b ) ) ( t + b ) 2 + b ( 1 + b ) ,
where b > 2 + 2 2 4 1 2 0.0493 or b < 1 2 2 + 2 2 4 1.05 .
If x > 1 16 , let x = b 2 , then we can factor f ( t ) as
f ( t ) = ( t 1 / 2 ) 2 + b 1 / 4 ( t 1 / 2 ) 2 b 1 / 4 ,
where | b | > 1 / 4 .
Proof. 
Consider f ( t ) = 0 , since | x | > 1 16 , we can deduce that
Δ = 196608 x 2 ( 16 x 1 ) > 0 , x > 1 16 < 0 , x < 1 16
If x > 1 16 , then f ( t ) = 0 has two unequal real roots and a pair of conjugate complex roots. We then get
f ( t ) = t 4 2 t 3 + t 2 b 2 = t 2 ( t 1 ) 2 b 2 = ( t ( t 1 ) b ) ( t ( t 1 ) + b ) = ( ( t 1 / 2 ) 2 b 1 / 4 ) ( ( t 1 / 2 ) 2 + b 1 / 4 )
If x < 1 16 , then f ( t ) has two pair of conjugate complex roots. Suppose we factor f ( t ) as
f ( t ) = ( ( t + a ) 2 + c ) ( ( t + b ) 2 + d ) ,
Then we have
2 ( a + b ) = 2 a 2 + b 2 + 4 a b + c + d = 1 2 ( a 2 + c ) b + 2 ( b 2 + d ) a = 0 ( a 2 + c ) ( b 2 + d ) = x
That is to say a = ( b + 1 ) and c = b ( 1 + b ) and d = b ( 1 + b ) and x = b ( 1 + b ) ( 1 + 2 b ) 2 . Since x < 1 16 , we have b > 2 + 2 2 4 1 2 0.0493 or b < 1 2 2 + 2 2 4 1.05
We first consider the case x < 1 16 , then we have the following theorem.
Theorem 7. 
If b > 2 + 2 2 4 1 2 0.0493 or b < 1 2 2 + 2 2 4 1.05 , then we have
k = 0 1 ( b ( 1 + b ) ( 1 + 2 b ) 2 ) k 4 k 2 k = 16 b ( 1 + b ) ( 1 + 2 b ) ( 1 + 10 b + 24 b 2 + 16 b 3 ) ( 1 + 8 b + 8 b 2 ) 3 + 2 ( 1 + 2 b ) 2 ( 1 + 8 b + 8 b 2 ) b ( 1 + b ) ( 1 + 8 b + 8 b 2 ) 3 a r c t a n 1 2 b ( 1 + b ) + 2 b ( 1 + b ) ( 1 + 2 b ) ( 3 + 8 b + 8 b 2 ) ( 1 + 8 b + 8 b 2 ) 3 log b 1 + b
Proof. 
By Lemma 1 we have
k = 0 1 x k 4 k 2 k = 0 1 x ( 3 t 2 ( 1 t ) 2 + x ) ( t 2 ( 1 t ) 2 x ) 2 d t .
By inequality (3) we have | x | > 1 16 .
Let x = b ( 1 + b ) ( 1 + 2 b ) 2 , then we have
f ( t ) = ( t b 1 2 + b ( 1 + b ) ) ( t + b ) 2 + b ( 1 + b ) ,
If x < 1 16 , we have b > 2 + 2 2 4 1 2 0.0493 or b < 1 2 2 + 2 2 4 1.05 , then by Mathematica we get
k = 0 1 x k 4 k k = 0 1 x ( 3 t 2 ( 1 t ) 2 + x ) ( t 2 ( 1 t ) 2 x ) 2 d t = 2 ( 1 + 2 b ) b 1 + b ( 1 + 8 b + 8 b 2 ) 3 [ 8 b 1 + b ( 1 + 10 b + 24 b 2 + 16 b 3 ) + ( 1 + 2 b ) ( 1 + 8 b + 8 b 2 ) a r c t a n 1 + b b a r c t a n b 1 + b + ( 3 + 8 b + 8 b 2 ) b 1 + b log b 1 + b ]
Since in the domain of b we have
b 1 + b b 1 + b = b ( 1 + b )
and
b 1 + b a r c t a n 1 + b b a r c t a n b 1 + b = b ( 1 + b ) a r c t a n 1 2 b ( 1 + b ) .
Then by some simplification we get the desired results. □
Using the same method of Theorem 7 and by Lemma 2 we have
Theorem 8. 
If b > 2 + 2 2 4 1 2 0.0493 or b < 1 2 2 + 2 2 4 1.05 , then we have
k = 0 1 ( b ( 1 + b ) ( 1 + 2 b ) 2 ) k ( 4 k + 1 ) 4 k 2 k = 2 ( 1 + 2 b ) 2 b ( 1 + b ) 1 + 8 b + 8 b 2 a r c t a n 1 2 b ( 1 + b ) 2 b ( 1 + b ) ( 1 + 2 b ) 1 + 8 b + 8 b 2 log b 1 + b
k = 1 1 ( b ( 1 + b ) ( 1 + 2 b ) 2 ) k k 4 k 2 k = 4 b ( 1 + b ) 1 + 8 b + 8 b 2 a r c t a n 1 2 b ( 1 + b ) + 1 + 2 b 1 + 8 b + 8 b 2 log b 1 + b
By Theorem 7 and 8 we get some propositions of equality.
Proposition 5. 
If b > 2 + 2 2 4 1 2 or b < 1 2 2 + 2 2 4 , then we have:
Via ( 29 ) + 2 b ( 1 + 2 b ) × ( 30 ) we get
k = 1 ( 1 + 8 b + 8 b 2 ) k + 2 b ( 1 + b ) k ( 4 k + 1 ) ( b ( 1 + b ) ( 1 + 2 b ) 2 ) k 4 k 2 k = 2 b ( 1 + b ) a r c t a n 1 2 b ( 1 + b ) 1
Via ( 1 + 8 b + 8 b 2 ) 2 × ( 28 ) 2 b ( 1 + b ) ( 3 + 8 b + 8 b 2 ) × ( 30 ) we get
k = 0 ( 1 + 8 b + 8 b 2 ) 2 k 2 b ( 1 + b ) ( 3 + 8 b + 8 b 2 ) k ( b ( 1 + b ) ( 1 + 2 b ) 2 ) k 4 k 2 k = 1 + 2 b ( 1 + b ) a r c t a n 1 2 b ( 1 + b )
By Proposition 5 and in equation (31) and (32) we let b = 1 2 , 2 1 2 , 3 3 1 2 , we can get some π -series, here are the examples.
Example 5. 
k = 1 3 + 14 k 2 k ( 1 + 4 k ) ( 3 ) k 4 k 2 k = 1 6 ( 6 + 3 π )
k = 1 ( 1 + 6 k ) ( 2 ) k 2 k ( 1 + 4 k ) 4 k 2 k = 1 4 ( 4 + π )
k = 1 ( 1 + 10 k ) ( 9 ) k 6 k ( 1 + 4 k ) 4 k 2 k = 1 9 ( 9 + 3 π )
k = 1 27 + 98 k 2 k ( 3 ) k 4 k 2 k = 1 + π 2 3
k = 1 ( 5 + 18 ) ( 2 ) k 2 k 4 k 2 k = 1 + π 4
k = 1 ( 11 + 50 k ) ( 9 ) k 18 k 4 k 2 k = 1 + π 3 3
Now we consider the case x > 1 16 .
Theorem 9. 
If | b | > 1 4 , then we have
k = 0 1 b 2 k 4 k 2 k = 16 b 2 16 b 2 1 + 2 b 1 4 b ( 1 + 4 b ) a r c t a n 1 1 4 b + 2 b ( 1 + 4 b ) 1 + 4 b a r c t a n 1 1 + 4 b .
Proof. 
By Lemma 1 we have
k = 0 1 x k 4 k 2 k = 0 1 x ( 3 t 2 ( 1 t ) 2 + x ) ( t 2 ( 1 t ) 2 x ) 2 d t .
By equation (3) we have | x | > 1 16 . □
Let x = b 2 , then we have
f ( t ) = ( ( t 1 / 2 ) 2 b 1 / 4 ) ( ( t 1 / 2 ) 2 + b 1 / 4 ) .
If x > 1 16 , we have | b | > 1 4 then by Mathematica we have
k = 0 1 b 2 k 4 k k = 0 1 b 2 ( 3 t 2 ( 1 t ) 2 + b 2 ) ( ( t 1 / 2 ) 2 b 1 / 4 ) 2 ( ( t 1 / 2 ) 2 + b 1 / 4 ) 2 d t = 2 b ( 1 4 b ) 3 / 2 ( 1 + 4 b ) 3 / 2 [ 8 b 1 4 b 1 + 4 b + ( 1 + 4 b ) 3 / 2 a r c t a n 1 1 4 b ( 1 4 b ) 3 / 2 a r c t a n 1 1 + 4 b ] .
Then by some simplifications we get the desired results.
Using the same method of Theorem 9 and by Lemma 2 we have
Theorem 10. 
If | b | > 1 4 , then we have
k = 0 1 ( 4 k + 1 ) b 2 k 4 k 2 k = 2 b a r c t a n 1 1 + 4 b 1 + 4 b a r c t a n 1 1 4 b 1 4 b .
k = 0 k b 2 k 4 k 2 k = 24 b 2 ( 1 + 4 b ) 2 ( 1 + 4 b ) 2 + b ( 1 + 2 b ) ( 1 + 4 b ) 2 a r c t a n 1 1 4 b 1 4 b + b ( 1 + 2 b ) ( 1 + 4 b ) 2 a r c t a n 1 1 + 4 b 1 + 4 b .
k = 1 1 k b 2 k 4 k 2 k = 2 a r c t a n 1 1 4 b 1 4 b + a r c t a n 1 1 + 4 b 1 + 4 b .
By Theorem 10 we can get propositions of equality.
Proposition 6. 
If | b | > 1 4 , via ( 40 ) + b × ( 42 ) we have
k = 1 ( 4 b + 1 ) k + b k ( 4 k + 1 ) b 2 k 4 k 2 k = 4 b 4 b 1 a r c t a n 1 4 b 1 1 .
If | b | > 1 4 , via ( 1 + 2 b ) × ( 40 ) + 2 ( 1 + 4 b ) 2 × ( 41 ) we have
k = 0 2 ( 1 + 4 b ) 2 k ( 4 k + 1 ) + 2 b 1 ( 4 k + 1 ) b 2 k 4 k 2 k = 48 b 2 ( 1 + 4 b ) 2 + 8 b 2 ( 5 + 16 b 2 ) ( 1 + 4 b ) 2 a r c t a n 1 1 + 4 b 1 + 4 b .
By Proposition 6 and in equation (43) and (44) we let b = 1 3 , 1 2 , 1 , we can get some π -series, here are the examples.
Example 6. 
k = 1 ( 1 + 7 k ) 3 2 k 1 k ( 1 + 4 k ) 4 k 2 k = 4 3 π 9 1
k = 1 1 + 5 k k ( 1 + 4 k ) 4 k 2 k = 2 3 π 9 1
k = 1 ( 1 + 6 k ) 2 2 k 1 k ( 1 + 4 k ) 4 k 2 k = π 2 1
k = 1 ( 3 + 98 k + 392 k 2 ) 9 k 1 ( 1 + 4 k ) 4 k 2 k = 488 π 9 3 + 48
k = 1 1 + 50 k + 200 k 2 ( 1 + 4 k ) 4 k 2 k = 4 27 ( 36 + 7 3 π )
k = 1 k 4 k 4 k 2 k = π 4 + 2 3

Data Availability Statement

Data openly available in a public repository.

Acknowledgments

Acknowledgments We would like to thank the referees for their helpful comments.

Conflicts of Interest

The author declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Almkvist, G.; Krattenthaler, C.; Petersson, J. Some new formulas for π, Experiment. Math. 2003, 12, 441–456. [Google Scholar]
  2. Bauer, G. Von den Coefficienten der Reihen von Kugelfunctionen einer Variablen. J. Reine Angew. Math. 1859, 56, 101–121. [Google Scholar]
  3. Borwein, J.M.; Chamberland, M. Chamberland, Integer powers of arcsin. Int. J. Math. Math. Sci. 2007, 10, 19381. [Google Scholar]
  4. Chan, H.H.; Chan, S.H.; Liu, Z. Dombs numbers and Ramanujan-Sato type series for 1/π. Adv. Math. 2004, 186, 396–410. [Google Scholar] [CrossRef]
  5. Coffey, M.W. Integral representations of functions and Addison-type series for mathematical constants. Journal of number theory 157, 79–98. [CrossRef]
  6. S. Cooper, Ramanujans Theta Functions, Springer, Cham, (2017).
  7. Glaisher, J.W.L. On series for 1/π and 1/π2. Quart. J. Pure Appl. Math. 1905, 37, 173–198. [Google Scholar]
  8. Gosper, R.W. Decision procedure for indefinite hypergeometric summation. Proc. Natl. Acad. Sci. USA 1978, 75, 40–42. [Google Scholar] [CrossRef]
  9. Qing-Hu Hou, Guo-Jie Li, Gosper summability of rational multiples of hypergeometric terms. J. Differ. Equ. Appl. 2021, 27, 1723–1733. [CrossRef]
  10. Sun, Z. W. New series for powers of π and related congruences. Electron. Res. Arch. 2020, 28, 1273–1342. [Google Scholar] [CrossRef]
  11. Z. W. Sun, On sums related to central binomial and trinomial coefficients, in Combinato- rial and Additive Number Theory: CANT 2011 and 2012, Springer Proc. Math. Stat., 101, Springer, New York, (2014), 257–312.
  12. Z. W. Sun, Evaluations of some series of the type k = 0 ( s k + b ) x k / ( m k n k ) . arXiv:2204.08275.
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.
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
Prerpints.org logo

Preprints.org is a free preprint server supported by MDPI in Basel, Switzerland.

Subscribe

Disclaimer

Terms of Use

Privacy Policy

Privacy Settings

© 2025 MDPI (Basel, Switzerland) unless otherwise stated