Introduction on Bernoulli's numbers Paseal Sebah and Xavier Gourdon numbers. computation. frec.fr/Constants/constants. html dune 12, 2002 Abstract ‘This essay is a general and elementary overview of some of the prop- certies of the famous numbers introduced by Bernoulli and used by Euler to express the value of the zeta function at integer even values 1 Introduction Bernoulli's numbers play an important and quite mysterious role in mathemat- ies and in various places like analysis, number theory and dilferential topology. They first appeared in Ars Conjectandi, page 97, a famous (and posthumous} treatise published in 1713, by Jakob Bernoulli (1654-1705) when he studied the sums of powers of consecutive integers se, w spl IM? where p and n are two given positive integers. Bernoulli's numbers also appear in the computation of the numbers 1 $0) = gE and in the expansion of many usual functions as tan(cr), tanh(x}, 1/ sin(2), « Perhaps one of the most important result is Buler-Maclaurin summation formula, where Bernoulli's numbers ate contained and which allows to accelerate the computation of slow converging series (see the essay on Euler's constant at i9|}. They also appear in numbers theory (Fermat's theorem) and in many other domains and have caused the ereation of a hnige literature (see the 270¢ and more entries enumerated in |6)), According to Louis Saalschiitz |17), the term Bernoulli's numbers was used for the first time by Abraham De Moivre (1667-1754) and also by Leonhard Enler (1707-1783) in 1755.2. Bernoulli’s approach During the first years of the Calculus period and in its first integral computations of the function x ++ 2?, Pierre de Fermat (1601-1665) in 1636 had to evaluate the sums sp(n) defined by (1). You ean see this by replacing the area under the curve «++ 2 by it’s rectangular approximations aud naturally comes the need to compute sp(n) Also in 1631, Johann Faulhaber (1580-1635) developed explicit formulas for these sums up to p = 17 (read the excellent |12| for the beginnings of integration and 18] for some excerpts of Bernoulli's work). ‘Thus, it was already known tc Takoh Bernonli that soln) = 0 s(n) = sa(n) = sa(n) = nln) = n= en aH n—1) ela) = want n= N= Jakob Bernoulli then, empirically, noticed that the polynomials s(n) have the form Lip y PnP 40x ne? Snr + Bae) 40% nh In this expression, the numbers (1,—1/2,1/12,0,...) are appearing and de not depend on p. More generally, the sums s,(n) can be written in the form 2 Be pete s(n) = Hato" al Bort! | By, B ptr” where the By, are numbers which are independent of » and called Bernoulli's numbers. We find by identification Bs pi By Ph Baty ayn? + + BP Bo =1,Bi "To illustrate the usefulness of his formula, Bernoulli computed the astonish ing value of s10(1000) with little effort (in less than “half a quarter of an hour”he says ... |18|) sangozazarazey p24 2424250 (vou can check it !). To achieve this he needed to find Bo up to Bi. ‘There is good evidence that the famous Japanese mathematician, Sekt Takakazt (1642-1708) also discovered Bernoulli's numbers at the same time. ‘The famous Indian mathematician Srinivasa Ramanujan (1887-1920) independently studied and rediscovered those numbers in 1904. He wrote one of his first article on this subject in 1911 [15 3 A more modern definition An equivalent definition of the Bernoulli expansion of the function =/(e* — 1) s numbers is obtained from the series lel< on Q Im other words, the generating function of the Bernoulli's numbers By is 1). The first terms of the expansion of this function are 2 a * 30240 1209000 *” Further, we observe that a+ 3 =3( = Joh 2 where coth is the hyperbolic tangent, hence G(z) + 2/2 is an even function and consequently every Bernoulli's numbers of the form Byi.41 (k > 0) is null, 3.1 Bernoulli’s polynomials With a little modification it’s possible to define Bernoulli's polynomials By(x) by:and G(2,0) = G(2) hence the value of the function B(x) at x = 0 is By and hecanse KBa(e) ‘Then, it’s easy to deduce that By(:r) are polynomials of degree k, and the first one are Bo(z) Bio) = Br(2) Bala) = Bale) 54,3,3_1 Bole) = 2° Sat Ba — 2 ‘Thanks to Bernoulli's polynomials, it’s possible to rewrite the expression of the sums s9(n) as 1 pai Brealn) — Boss) ‘There are many relations with these polynomials, for example Bu(l—z) (-1)*Bx(2), (-2)ABg(—e) = Bela) + kak! |Bae(z)| << |Bas| k= and 0 + 5B; + By ths sp.+ (+00. +1)=0 therefore By = 4.2 Bernoulli’s numbers and the zeta function In 1735, the solution of the Basel problem, expressed by Jakob Bernoulli some years before, was one of Buler’s most sensational discovery. ‘Uhe problem was to find the limit of c= he found it to be x*/6, He also discovered the values of the sums 1 60K) = oe for k up to 13 (|8) and for more details |7).It’s an extraordinary result that ¢(2k) can be expressed with Bernoulli's numbers ; the values of these sums are given by co =A Hale go, W (a proof based on two different expansions of 2 cot(2) is given in [4] p. 383) [No similar expression is known for the odd values of the Zeta function, (On the extension of this function to negative values, we also have Bax (1 - 2k) = -#= k>o, which may also be used to compute Bog (80° [5)). 4.3. Asymptotic expansion of Bernoulli’s numbers From the previous relation (4) with the Zeta function, it’s clear that 2(2k)! Bae) = FEC) 6 and because, when k becomes large, thanks to Stirling's formula ck) ~ 1, (2k)! ~ (2k)*%en Var, we have pul ~a(£)" vak In {16}, the following results describes how the numerator Nox of Bax grows with fe log | Na log(k) + OCR). 4.4 Bounds Te may be useful to estimate bounds for Box, to achieve this we use the following relation between the function ¢(s) and the alternating series a(s) coy = 7S a) = po and since 1 1-5 1, we observe that the primes p= 2,3 are sueh as (p ~ 1) divides 2k, Let's illustrate this theorem with a few examples, For k= 1, it becomes 1liyt p 2°93 for k and for kCorollary 3 («due to Rado 1934) For every prime numbers k of the form 3n+1 1 By =; (mod 1) i Proof. It's an easy consequence of the Staudt’s theorem since p~1 divides 2k = 23n +1) only if p—1 is one of 1,2,3n+1,6n +2, that is p is one of 2,3,3n-+2,6n43. But 6n-+3 is divisible by 3 and 3n+2 és divisible by 2 because 3n-+1 is prime so the only primes p candidates are 2 and 3. Example 4 The first primes of the form 3n-+1 are 7,13,19,31,37,43,61,67, hence me hawe Bus os = Bea = Bra = Bus = Bizz = Biss =f (mod 2), ‘Clausen-von Staudt’s theorem also permits to compute exactly a Bernoulli's number as soon as a sufficiently good approximation of it is known. 6 Expansion of usual functions Ina previous section we gave the definition and the consequence seoth(2) zeot(2) where cot(z) = cos( is the cotangent function. ‘Now it’s possible to find the expansion for tanh(2) and tan(z), if we observe that ‘osh(2) __cosh?() + sinh(2) ~—sinh(2) cosh(=) 2eoth(22)—coth(2) = 2 Tay so that 2k eM a8 nt = 15 315 tanh(z) = )>4*(4* —1) BaxI) Baw tan(z) = Yeota - Bernoulli's numbers also occur in the expansions of other classical functions a) sox ( =) latest) (' 2), Sin(2)’ Sinh(e) 6.1 Series Setting z = 1 or z = —1 in the expansion (2) of 2/(e” ~ 1) leads to the fast converging series and Bu 2 8 (2k) ~ (ar ™ 30.48 In one of his famous notebook, Ramanujan stated without proof the following result Theorem 5 Let (a,b) two positive real numbers such as ab =, let n > 1 an integer, then OS a COD ary 0) Proof. See (3|. Corollary 6 Let n> 1, then Bata Proof. Just apply the theorem with a == x and replace n by 2n41. It’s interesting to compare this result with the classical integral representa- tion valid form 3 1 mB [ dy = (yr ar which implies that for n > 0 a0 gn Binsa Lait7 Enler-Maclanrin formula Let f(x) be a function of elass €2°*? on an interval [a,] and let h = (b—a}/m a subdivision of this interval into m equal parts then we have the important result Theorem 7 ‘There erist A ng Le = [perder 0 + s00)¢ where Ra(f,p) is the remainder bounded by Palh) 2: Se = [eben Ben-ayv0 7 1 Ltn? n(n + 1)(2n +1) 7 2 2 f(z) = Afe, FORM x) = —(2k — 1)1/2"*, Euler-Maclaurin formula yields for a given p: ant ak (1 ar) + ald 10when n — 00, the left hand side of the equality tends to + (Euler's con- stant) and the equality gives yo Bs Ralf) finally Bx 1 + (Ral EP) — Real F0))- (check that Rn(f,p) ~ Raolf.p) = O(1/n*”**) ) 3. f(z) = log(z), with the same method (eft as exercise) Euler-Maclaurin Formula becomes YS bos(t) = most mB log(n) . Bo 1 jn SE HHoal V2) aR ay aT TO so, for example, with p= log(n!) = nlog(n) —n-+log( Van) + + ES +0 (3) and taking the exponential nhante- "Vines (55 1,1 o(=))- Tn 360n8 * T260n8 * O LaF This is the asymptotie Stirling formula. Using the series expansion of the exponential function near the origin, it’s more convenient to write it as nian (14 et aghg gilt — gallo + gett, +0(4 v 12n © 288n? — 51840n7 —-2488320n4 © 209018880n* ne 8 Bernoulli’s numbers and Fermat’s last. theo- rem ‘The famous Fermat's last theorem states that the equation oy" never has non-zero integer solutions for n > 2. Since Fermat expressed this result around 1630, the pursuit of a proof occupied senerations of mathematiciansA big step was made in 1850 by Emnst Kummer (1810-1893) when he proved Fermat's theorem for n = p, whenever p is, what is called today, a regular prime. Kummer gave the beautiful regularity criterion: p 1s a regular prime if and only itp does not divide the numerator of Ba, Ba, He showed that all primes before 37 where regular, hence Fermat's theorem ‘was proved for those primes. 37 is the first non regular prime because it divides the numerator of 7709321041217 _ _ 208360028141 x 37 eS 510 510 ‘The next irregular primes (less than 300) are 99, 67, 101,103, 131, 149, 197, 283, 207, 263, 271, 283,293, For example, 157 divides the numerators of Bye and Bi10- ‘Thanks to arithmetical properties of Bernoulli's numbers, Johann Ludwig Jensen (1859-1925) proved in 1915 that the number of irregular primes is infi- nite. Even if it’s probable that the number of regular primes is infinite, a proof remains unknown [16| 9 The first. Bernoulli’s numbers 9.1 Wirst: numbers Here is the list of the first Bernoulli's numbers. Except for By numbers of the form Box-+1 are null Bo = 1 By 1/2 B= 6 By = -1/30 Bo 1/42 By = -1/30 Bw = 5/66 By —691/2730 Bu = 1/6 By = —3617/510 Bis 43867 /798 Bao ~174611/330_ Boy = 854513/138 Bas 236364091 /2730 v,8553103 /6 =23749461020 870 8615841276005 /14322 7709321041217 /510 2577687858367 /6 —26815271553053477373/ 1919190 -2929993913841550/6 —261082718496449122051 /13530 ‘More numbers are given in [1] and in [17] 9.2 Some computations Bernoulli himself computed the numbers that now bear his name up to Bun Later, Euler computed these numbers up to Ban, then Martin Ohm extended the calculation up to Byz in 1840 [14]. A few years later, in 1877, Adams made the impressive computation of all Bernoulli's numbers up to Bist (or Biz according to his convention) [2]. For instance, the numerator of Bia, has 110 digits and the denominator is the number 3#. In 1996, Simon Ploutfe and Greg J. Fee computed Bay990 @ huge number of about 800000 digits, the computation took about 2 hours on a work station, In 2002, the same authors improved the record to Beoocao Which has 2727474 digits by a 12 hours computation on @ personal computer. ‘Lhe method is based on the formula (5) which allow a direct computation of the required mumber without the need to compute the previous ones, References |1| M. Abramowitz and I. Stegun, Handbook of Mathematical Functions, Dover, ‘New York, (1964) [2] J.C. Adams, On the calculation of Bernoulli’s numbers up to Bes by means of Staudt's theorem, Rep. Brit. Ass., (1877) Bl ©. Bern, Ramanjan’s Notebooks, Part II, Springer-Verlag, New York, || J.M. Borwein and P.B. Borwein, Pi and the AGM - A study in Analytic Number Theory and Computational Complerity, A Wiley-Interscience Pub- lication, New York, (1987) [5] P. Borwein, An efficient algorithm for the Riemann Zeta function, (1995) [0] K. Dilcher, A Bibliography of Bernoulli Numbers, World Wide Web site at the address: http: //www.mscs .dal.ca/“dilcher/bernoul1i .html 18[7] W. Dunham, Buler The Master of Us All, The Mathematical Association of America, (1999) [S| L. Euler, Introduction & Vanalyse infinitésimale (french traduction by Labey), Barrois, ainé, Librairie, (original 1748, traduction 1796), vol. 1 9) X. Gourdon and P,—Sebah, Numbers, Constants and Computation, World Wide Website atthe address: http: //nunbers computation. free.fr/Constants/constants.html, (1999) {10} R.L. Graham, D.E. Knuth and O. Patashnik, Concrete Mathematics Addison-Wesley, (1994) [11] G.H. Hardy and E. M. Wright, An Introduction to the Theory of Numbers Oxford Science Publications, (1979) [12] V.J. Katz, A History of Mathematics-An Introduction, Addison-Wesley, (1998) {13| C, Maclaurin, A Treatise of flurions, Edinburgh, (1742) {14] M. Ohm, Etwas iiber die Bernoulli’schen Zahlen, J. Reine Angew. Math, (1840), vol. 20, p. 11-12 {15| 8, Ramanujan, Some properties of Bernoulli's numbers, J. Indian Math, Soe, (1911), vol. 3, p. 219-234 [16] P. Ribenboim, The new Book of Prime Number Records, Springer, (1996) [17] L, Saalschiitz, Vorlesungen iiber die Bernoullischen Zahlen, Berlin, Verlag von Julius Springer, (1893) [18] D.E. Smith, A Source Book in Mathematics, Dover Publications, New York, (1959, first edition 1929)