One of the Eight Numbers ζ(5),ζ(7),...,ζ(17),ζ(19) Is Irrational

Similar documents
Irrationality of the Sums of Zeta Values

The Analytic and Arithmetic Mystery of Riemann s Zeta Function at Positive Integers

Irrationality exponent and rational approximations with prescribed growth

Yet another failed attempt to prove the Irrationality of ζ(5)

involving odd zeta values

Mathematische Annalen

Algebraic relations for reciprocal sums of Fibonacci numbers

On Some Mean Value Results for the Zeta-Function and a Divisor Problem

Beukers integrals and Apéry s recurrences

Irrationality proofs, moduli spaces, and dinner parties

arxiv: v1 [math.mg] 15 Jul 2013

Journal of Number Theory

Mahler measure as special values of L-functions

On the Number of Divisors of n 2 1

A REMARK ON ZAGIER S OBSERVATION OF THE MANDELBROT SET

Algebra 2 (2006) Correlation of the ALEKS Course Algebra 2 to the California Content Standards for Algebra 2

Generalized Fibonacci numbers of the form 2 a + 3 b + 5 c

PRIME NUMBER THEOREM

Mathematics 324 Riemann Zeta Function August 5, 2005

GAPS IN BINARY EXPANSIONS OF SOME ARITHMETIC FUNCTIONS, AND THE IRRATIONALITY OF THE EULER CONSTANT

The Riemann and Hurwitz zeta functions, Apery s constant and new rational series representations involving ζ(2k)

THE NUMBER OF DIOPHANTINE QUINTUPLES. Yasutsugu Fujita College of Industrial Technology, Nihon University, Japan

Transcendental Number Theory: Recent Results and Conjectures

On a diophantine inequality involving prime numbers

arxiv: v2 [math.nt] 19 Apr 2017

Number Theory Solutions Packet

arxiv: v2 [math.nt] 28 Feb 2010

arxiv: v3 [math.nt] 1 Jun 2017

On repdigits as product of consecutive Lucas numbers

The Greatest Common Divisor of k Positive Integers

Math 118: Advanced Number Theory. Samit Dasgupta and Gary Kirby

NOTES ON IRRATIONALITY AND TRANSCENDENCE

arxiv:math/ v2 [math.nt] 27 Mar 2008

Gauss map on the theta divisor and Green s functions

Zero estimates for polynomials inspired by Hilbert s seventh problem

Analytic Number Theory Solutions

#A5 INTEGERS 18A (2018) EXPLICIT EXAMPLES OF p-adic NUMBERS WITH PRESCRIBED IRRATIONALITY EXPONENT

Properties of Rational and Irrational Numbers

arxiv: v1 [math.nt] 22 Jan 2019

NOTES ON INTEGERS. 1. Integers

Approximation exponents for algebraic functions in positive characteristic

ARITHMETIC WITH HYPERGEOMETRIC SERIES

Rational approximations to π and some other numbers

Math 229: Introduction to Analytic Number Theory The product formula for ξ(s) and ζ(s); vertical distribution of zeros

Introduction to Diophantine methods: irrationality and transcendence

Harmonic sets and the harmonic prime number theorem

ON THE LINEAR INDEPENDENCE OF THE SET OF DIRICHLET EXPONENTS

Some Arithmetic Functions Involving Exponential Divisors

PERIODICITY OF SOME RECURRENCE SEQUENCES MODULO M

MATHEMATICS OF COMPUTATION, VOLUME 26, NUMBER 118, APRIL An Algorithm for Computing Logarithms. and Arctangents. By B. C.

arxiv: v6 [math.nt] 12 Sep 2017

UNIVERSALITY OF THE RIEMANN ZETA FUNCTION: TWO REMARKS

On the growth of the identities of algebras

x 9 or x > 10 Name: Class: Date: 1 How many natural numbers are between 1.5 and 4.5 on the number line?

P -adic root separation for quadratic and cubic polynomials

The universality of quadratic L-series for prime discriminants

Florian Luca Instituto de Matemáticas, Universidad Nacional Autonoma de México, C.P , Morelia, Michoacán, México

PUTNAM PROBLEMS SEQUENCES, SERIES AND RECURRENCES. Notes

One-sided power sum and cosine inequalities

CHAPTER 2 Review of Algebra

arxiv: v1 [math.nt] 29 Dec 2017

Verbatim copying and redistribution of this document. are permitted in any medium provided this notice and. the copyright notice are preserved.

A PROBABILISTIC PROOF OF WALLIS S FORMULA FOR π. ( 1) n 2n + 1. The proof uses the fact that the derivative of arctan x is 1/(1 + x 2 ), so π/4 =

Riemann Zeta Function and Prime Number Distribution

G. NADIBAIDZE. a n. n=1. n = 1 k=1

On the modification of the universality of the Hurwitz zeta-function

ALGEBRA I TRANSITION MATH GLEs

Revista Colombiana de Matemáticas Volumen 40 (2006), páginas 81 86

On the discrepancy estimate of normal numbers

Even perfect numbers among generalized Fibonacci sequences

32 Divisibility Theory in Integral Domains

On distribution functions of ξ(3/2) n mod 1

On the digital representation of smooth numbers

New Multiple Harmonic Sum Identities

The zeros of the derivative of the Riemann zeta function near the critical line

The Fibonacci Zeta Function. M. Ram Murty 1

not to be republished NCERT REAL NUMBERS CHAPTER 1 (A) Main Concepts and Results

MATH10040: Numbers and Functions Homework 1: Solutions

A Pellian equation with primes and applications to D( 1)-quadruples

Math 259: Introduction to Analytic Number Theory More about the Gamma function

On the Voronin s universality theorem for the Riemann zeta-function

LINEAR FORMS IN LOGARITHMS

The Convolution Square Root of 1

arxiv: v3 [math.nt] 9 May 2011

Paper: 03 Class-X-Math: Summative Assessment - I

Math Review. for the Quantitative Reasoning measure of the GRE General Test

Math 201C Homework. Edward Burkard. g 1 (u) v + f 2(u) g 2 (u) v2 + + f n(u) a 2,k u k v a 1,k u k v + k=0. k=0 d

On the irreducibility of certain polynomials with coefficients as products of terms in an arithmetic progression

On irrationality and transcendency of infinite sums of rational numbers

arxiv:math/ v2 [math.nt] 3 Dec 2003

Absolute zeta functions and the absolute automorphic forms

Standard forms for writing numbers

Elliptic Curves Spring 2013 Lecture #8 03/05/2013

MATH 51H Section 4. October 16, Recall what it means for a function between metric spaces to be continuous:

Explicit Bounds Concerning Non-trivial Zeros of the Riemann Zeta Function

METRIC HEIGHTS ON AN ABELIAN GROUP

Analytic Aspects of the Riemann Zeta and Multiple Zeta Values

Content. CIMPA-ICTP Research School, Nesin Mathematics Village The Riemann zeta function Michel Waldschmidt

1 Euler s idea: revisiting the infinitude of primes

Places of Number Fields and Function Fields MATH 681, Spring 2018

Transcription:

Mathematical Notes, vol. 70, no. 3, 2001, pp. 426 431. Translated from Matematicheskie Zametki, vol. 70, no. 3, 2001, pp. 472 476. Original Russian Text Copyright c 2001 by V. V. Zudilin. One of the Eight Numbers ζ(5,ζ(7,...,ζ(17,ζ(19 Is Irrational V. V. Zudilin Received May 3, 2001 Key words: Riemann zeta-function, irrationality of values of the zeta-function, Nikishin Rivoal arithmetical scheme. 1. INTRODUCTION In 1978 R. Apéry [1] produced a sequence of rational approximations proving the irrationality of the number ζ(3. Despite the apparent simplicity, attempts to generalize the Apéry result and prove the irrationality of the values ζ(5,ζ(7,... of the Riemann zeta-function were futile. In 2000 T. Rivoal proposed a construction [2] which allowed us to obtain nice linear forms with rational coefficients in the values of the zeta-function ζ(s at odd positive points. The generalization of this construction [3 6] leads to linear forms in 1 and in the numbers ζ(s, ζ(s +2,..., ζ(s +2m, where s 3isodd, (1 and for a properly chosen integer m 1 one can prove the irrationality of at least one element of the set (1. As a consequence, the assertion that one of the numbers ζ(5, ζ(7,..., ζ(19, ζ(21 is irrational was obtained. Two different proofs of this assertion were proposed in [3] and [4]. In the present paper, we prove the following result. Theorem 1. At least one of the eight numbers ζ(5, ζ(7, ζ(9, ζ(11, ζ(13, ζ(15, ζ(17, ζ(19 (2 is irrational. The main difference of the construction given below from those used in [3, 4] is in the arithmetical part; the analytical part the asymptotics of linear forms is fully described in [4, Sec. 2]; therefore, we restrict ourselves, where necessary, to references to the relevant assertions in [4]. 2. ANALYTIC CONSTRUCTION Choose positive integer parameters a, b, c, d,whereb 3 is odd, a bd/c is even, and d 2c. For each positive integer n, consider the odd rational function ( b (t ± (cn +1 (t ± (cn ± R n (t =t ( a (2cn!a t(t ± 1 (t ± cn (! 2b = t Γ(t +(c + +1b Γ(t cn a+b (2cn!a Γ(t (c + b Γ(t + cn +1 a+b (! 2b =( 1 (c+ sinb πt π b t(±t +(c + b (2cn a (t + cn a+b ( 2b Γ(±t +(c + b Γ(t cn a+b Γ(2cn a Γ(t + cn a+b Γ( 2b (3 426 0001-4346/2001/7034-0426$25.00 c 2001 Plenum Publishing Corporation

ONE OF THE EIGHT NUMBERS ζ(5,ζ(7,...,ζ(17,ζ(19 IS IRRATIONAL 427 and assign to it the number I n = 1 (b 1! t=cn+1 d b 1 R(t dt b 1 = 1 2πi M+i M i π b cot b πt R n (tdt, where M is an arbitrary constant fromthe interval cn < M < (c + and cot b z = ( 1b 1 (b 1! d b 1 cot z dz b 1, b =1, 2,... (see [4, Lemma 2.4]. As is readily verified, the number I n is a linear formwith rational coefficients in the numbers 1, ζ(b +2, ζ(b +4,..., ζ(a + b 2 (4 (see [4, Lemma 1.1]. Setting r = d/(2c, replacing (! 2b by (2cn! bd/c and cn by n in the definition of R n (t, we obtain the linear forms I n (see [4] in the numbers (4 (for even a and odd b. The only difference from[4] is the appearance of the additional factor t in (3; this makes it possible to consider the numbers a, b of different parity (this device was used in [3], which does not affect the asymptotics of I n as n. Therefore, we can use Proposition 2.3 from[4]. Lemma 1. Suppose that for a real root µ 1 (c + d, + of the polynomial the following inequality is valid: (τ + c + d b (τ c a+b (τ c d b (τ + c a+b (5 { µ 1 c + d +(2c + dd min b 8c(a + b, 1 12(c + d and τ 0 is a complex root of (5 in the domain Re τ>0, Im τ>0 with is the maximum possible real part Re τ 0. Further, suppose that the function f 0 (τ =b(c + dlog(τ + c + d+b(c + dlog( τ + c + d (a + bc log(τ c (a + bc log(τ + c+2ac log(2c 2bd log d }, satisfies the condition Then Im f 0 (τ 0 0(modπZ. log I n lim =Ref 0 (τ 0. 3. DENOMINATORS OF THE LINEAR FORMS To study of the arithmetical properties of the linear forms I n, let us express the function (3 as R n (t =th n (t a G n (t b,where H n (t = (2cn! t(t ± 1 (t ± cn, G n(t = (t ± (cn +1 (t ± (cn + (! 2. The following assertion is a generalization of the arithmetical Nikishin Rivoal scheme.

428 V. V. ZUDILIN Lemma 2 [4, Lemma 4.1]. Suppose that for a polynomial P (t, deg P (t <m(n +1, the rational (not necessarily irreducible function R(t = P (t ( (t + s(t + s +1 (t + s + n m satisfies, for each j =0, 1,...,m 1, the inclusions Dn j d j ( R(t(t + k m j! dt j Z, k = s, s +1,...,s+ n, (6 where D n is the least common multiple of the numbers 1, 2,...,n. Then the inclusions (6 are satisfied for all nonnegative integers j. It follows from Lemma 2 that for all nonnegative integers j we have D j 2cn j! d j ( Hn dt j (t(t + k Z, k =0, ±1,...,±cn. (7 In addition, from the properties of integer-valued polynomials (see, for example, [7] (t +(cn +1 (t +(cn +, (! (t (cn +1 (t (cn + (! and fromthe Leibniz rule for differentiating a product, we obtain the inclusions D j j! d j dt j G n(t Z, k =0, ±1, ±2,.... (8 Lemma 3 (cf. [4, Lemma 4.5]. For each prime p, let us define the exponents { } ((c + + k! ((c + k! ν p = m in ord p k=0,±1,...,±cn (cn + k! (cn k! (! 2 (9 and set Π n = p ν p. (10 p : (2c+<p Then for all nonnegative integers j we have Π 1 n D j d j dt j G n(t Z, k =0, ±1,...,±cn. (11 Proof. Since G n (t ((c + + k! ((c + k! =( 1 (cn + k! (cn k! (! 2, k =0, ±1,...,±cn, the inclusions (11 for j = 0 follow from(9, (10. For any prime p, let us prove by induction on j that ( ord p Π 1 n D j d j dt j G n(t 0, k =0, ±1,...,±cn. (12

ONE OF THE EIGHT NUMBERS ζ(5,ζ(7,...,ζ(17,ζ(19 IS IRRATIONAL 429 We shall prove the estimates (12 for j + 1, assuming them to be proved for all the previous values of j. If the prime p does not divide Π n,thenord p Π 1 n = 0 and relations (12 follow from(8. Therefore, we next assume that p is a divisor of Π n, whence, in particular, p> (2c + and p. Setting g n (t = G n (t cn+ G n (t = 1 t ± l, we obtain For m j,wehave d j+1 G n (t dt j+1 = dj dt j ( gn (tg n (t = l=cn+1 j ( m j m=0 dj m g n (t dt j m dm G n (t dt m. (13 ord p ( d j m g n (t dt j m ( cn+ =ord p l=cn+1 k =0, ±1,...,±cn, 1 (l ± k j m+1 (j m +1, (14 since p> (2c + and l ± k (2c + for all the denominators in (14; we also have since p ; and, finally, ord p ( Π 1 n D m dm G n (t dt m ord p D j m+1 j m +1, (15 0, k =0, ±1,...,±cn, (16 by the induction assumption. Substituting t = k into (13 and using the estimates (14 (16, we find that the estimates (12 are satisfied for j + 1. Thus the induction step is justified. The lemma is proved. Set a 0 = bd/c bd/c. From Lemma 3, the inclusions (7, and the Leibniz rule, in view of d 2c, for all nonnegative integers j we obtain Π b n D j d j ( Gn dt j (t b H n (t a 0 (t + k a 0 Z, k =0, ±1,...,±cn. Hence, in particular, for j =0, 1,...,a 0 1wehave ( a0 1 a 0!Π b D n Dj 2cn d j ( Gn D 2cn j! dt j (t b H n (t a 0 (t + k a 0 Z, k =0, ±1,...,±cn. (17 By Lemma 2, the inclusions (17 are valid for all nonnegative integers j,since deg G n (t b =2b < a 0 (2cn +1=degH n (t a 0. Therefore, by standard arguments (see, for example, [4, Lemma 1.4] or [3, Lemma 2], we obtain the following assertion.

430 V. V. ZUDILIN Lemma 4. The numbers Ĩ n = a 0!Π b n are linear forms in the numbers (4 with integer coefficients. Da 0 1 D a+b a 0 2cn I n (18 The asymptotics of D and D 2cn in (18 as n is determined by the asymptotic law of distribution of the primes: log D n lim =1. (19 To calculate the asymptotic behavior of Π n as n, we should note that, by (9, for the primes p> (2b + cn we have ( ( n n n ν p ϕ = ϕ p p, (20 p where { } ϕ(x =min (c + dx + y + (c + dx y cx + y cx y 2 dx. y R (21 The function under the sign of the minimum in (21 is periodic (with period 1 in each argument; therefore, the minimum can be taken only for y [0, 1. Using the arguments of Chuovskii and Hata, by (10, (20, (19, we finally obtain log Π n ϖ = lim 1 0 ϕ(xdψ(x 1/d 0 ϕ(x dx x 2 (22 (see [4, Lemma 4.4 and the proof of Lemma 4.5], where ψ(x is the logarithmic derivative of the gamma-function, while the subtraction on the right-hand side of (22 eliminates the primes p>. 4. PROOF OF THEOREM 1 To obtain the result announced in the title, we assume a =18, b =3, c =3,and d =7. Then a 0 = bd/c =7 and µ 1 10.305445, τ 0 9.856603 + 0.197639i, f 0 (τ 0 123.071169 30.779083i. In this case, for x [0, 1 the function (21 is of the form 1 if x [ 1 10, [ 7 1 1 4, [ 2 7 2 5, [ 3 7 11 20, [ 4 7 13 ϕ(x = [ 7 10, [ 7 5 4 5, [ 5 6 17 20, [ 6 7 19 20, 1, 0 for other values of x [0, 1 ; therefore, by (22, we have ϖ 1.150969. Finally, for the linear forms (18 with integer coefficients in 1 and in the numbers (2, by Lemmas 1 and 4, we have 20, 2 3 log Ĩn lim = bϖ +(a 0 1d +2(a + b a 0 c +Ref 0 (τ 0 < 0.524077, and hence there is an irrational number among the numbers (2.

ONE OF THE EIGHT NUMBERS ζ(5,ζ(7,...,ζ(17,ζ(19 IS IRRATIONAL 431 5. OTHER RESULTS ON IRRATIONALITY The construction described in this paper also allows us to prove the irrationality of one of the numbers in the set (1 taken for s =7 and s = 9 with a lesser value of m than that in Theorem1 from[4 6]. Indeed, setting a =32, b =5, c =5, d =12 and a =46, b =7, c =1, d =2, we obtain the following results. Theorem 2. At least one of the fifteen numbers is irrational. ζ(7, ζ(9, ζ(11,..., ζ(33, ζ(35 Theorem 3. At least one of the (twenty two numbers is irrational. ζ(9, ζ(11, ζ(13,..., ζ(49, ζ(51 ACKNOWLEDGMENTS The author wishes to express his thanks to Professor Yu. V. Nesterenko for consultations concerning the subject matter of this paper. This research was supported in part by INTAS and the Russian Foundation for Basic Research under grant no. IR-97-1904. REFERENCES 1. R. Apéry, Astérisque, 61 (1979, 11 13. 2. T. Rivoal, C. R. Acad. Sci. Paris Sér. I Math., 331 (2000, no. 4, 267 270. 3. T. Rivoal, Irrationalité d au moins un des neuf nombres ζ(5,ζ(7,...,ζ(21, Acta Arith. (2001 (to appear, in: E-print math.nt/0104221. 4. V. V. Zudilin, On the irrationality of values of the Riemann zeta-function, Izv. Ross. Akad. Nauk Ser. Mat. [Russian Acad. Sci. Izv. Math.] (2002 (to appear. 5. V. V. Zudilin, Uspekhi Mat. Nauk [Russian Math. Surveys], 56 (2001, no. 2, 215 216. 6. V. V. Zudilin, On the irrationality of values of the zeta-function, in: Reports of the 23d Conference of Young Scientists (Moscow State University, 9 14 April, 2001 [in Russian], Moscow, 2001 (to appear, in: E-print math.nt/0104249. 7. E. M. Matveev, Mat. Zametki [Math. Notes], 54 (1993, no. 4, 76 81. M. V. Lomonosov Moscow State University E-mail: wadim@ips.ras.ru