Abstract
We find new representations, in terms of constant terms of powers of Laurent polynomials, for all the 15 sporadic Apéry-like sequences discovered by Zagier, Almkvist-Zudilin and Cooper. The new representations lead to binomial expressions for the sequences, which, as opposed to previous expressions, do not involve powers of 3 or 8. We use these to establish the supercongruence for all primes and integers , where Bn is a sequence discovered by Zagier, known as Sequence B. Additionally, for 14 of the 15 sequences, the Newton polytopes of the Laurent polynomials contain the origin as their only interior integral point. This property allows us to prove that these sequences satisfy a strong form of the Lucas congruences, extending work of Malik and Straub. Moreover, we obtain lower bounds on the p-adic valuation of these sequences via recent work of Delaygue.
1 Introduction
Roger Apéry’s ingenious proof of the irrationality of and [Citation1, Citation2] utilized two particular sequences:(1.1) (1.1) for . These sequences are referred to as Apéry sequences or Apéry numbers. Apéry constructed explicit rational approximations to and , with denominators being small multiples of an and bn, respectively. These rational approximations turn out to converge quickly enough (in a particular sense) to and as in order to conclude that .
Originally, Apéry defined an and bn through 3-term recurrences with given initial values:(1.2) (1.2) and showed that an and bn can be expressed as the binomial sums in (1.1). The recurrences for an and bn can be restated in terms of differential equations for the g.f.s (generating functions) namely, setting is a solution to the third-order equation(1.3) (1.3) while solves the second-order equation(1.4) (1.4)
Beukers [Citation9, Citation10] showed that enjoy modular parametrization. Namely, there is a modular function (z in the upper-half plane) with respect to the modular subgroup such that is a modular form of weight 2 with respect to . Similarly, there is a modular function with respect to such that is a modular form of weight 1 with respect to .
This is not a coincidence, and in fact related to the shape of the specific differential equations. EquationEquation (1.4)(1.4) (1.4) is a Picard-Fuchs equation, while (1.3) is a symmetric square of a Picard-Fuchs equation. We shall not discuss this geometric aspect here, and instead refer the reader to the papers [Citation45, Citation47, Citation55] and to Zagier’s survey [Citation58].
1.1 Sporadic Sequences
There is no apparent reason for solutions of the recurrences (1.2) to be integers (or, equivalently, for solutions of (1.3) or (1.4), which are regular at the origin, to have only integral coefficients in their Taylor expansion around z = 0). It is a surprise that an and bn are integers. For this reason, Zagier [Citation57] searched for triples such that the recurrence(1.5) (1.5)
has a solution in integers with . A solution un gives rise to a power series satisfying the second-order differential equation(1.6) (1.6) and vice-versa. For we already know that there is such a desired solution, namely bn. In his computer search, two well-understood families were found: Legendrian and hypergeometric. In addition, six sporadic solutions are found. The two families are easily explained, in the sense that the g.f.s have particularly simple forms, which explain the integrality. The six sporadic sequences, which include un = bn, do not fall into either of these families and are not explained as easily. They are given in the following table.
Table
Zagier conjectured that his (finite) search yielded all the integral solutions to (1.5). This is suggested by the fact that so few sporadic solutions were found in a very large parameter domain. As in the case of bn, the differential equations associated with the sequences found by Zagier are Picard–Fuchs equations, and the solutions to the differential equations have modular parametrizations [57, §5, §7].
Similar searches were conducted in relation to an. Almkvist and Zudilin [Citation5] searched for such that the recurrencehas integer solutions with . They found 6 sporadic solutions, including an, which corresponds to :
Table
In [4, Thm. 1], Almkvist, van Straten and Zudilin discovered a bijection between Zagier’s sequences and Almkvist and Zudilin’s (cf. Cooper [22, p. 400]). The sequences A- F correspond to , , in this order, via sending to , and the g.f.s for the Almkvist and Zudilin sequences are essentially the squares of the corresponding sequences of Zagier:where is the third-order sequence corresponding to the solution un of (1.5).
Later, Cooper [Citation21] introduced an additional parameter :
A solution un gives rise to a power series satisfying the third-order differential equation(1.7) (1.7) and vice-versa. Cooper found 3 additional sporadic solutions, named s7, s10 and s18. He established their modular parameterization (the subscript is related to the level of the corresponding modular form). Formulas for the terms of s7, s10 and s18 were obtained by Zudilin:
Table
In total, we have 15 sporadic solutions, which are often referred to as Apéry-like sequences.
1.2 Main results
We say that a sequence is the constant term sequence of a Laurent polynomial if un is the constant term of for every . Our main result is
Theorem 1.1.
Let be one of the 15 sporadic sequences, and denote by the order of the differential equation it satisfies. Then un is a constant term sequence of a Laurent polynomial Λ with integer coefficients in k variables . In addition,
Λ may be written as a product of Laurent polynomials, each with coefficients in the set .
Λ may be chosen in such a way that it is a (usual) polynomial divided by .
For all but possibly , Λ may be chosen in such a way that its Newton polytope contains the origin as its only interior integral point.
For all but possibly and F, Λ may be chosen to satisfy the last 3 conditions simultaneously. In the case of F, one may choose a Laurent polynomial satisfying the last two conditions, with satisfying the first condition.
Specific Laurent polynomials are given in Proposition 3.3, from which we derive new binomial sum representations for B, F and .
Proposition 1.2.
Consider the set
The sequence B can be written as
Consider the set
The sequence F can be written as(1.8) (1.8)
Consider the set
The sequence can be written as(1.9) (1.9)
1.3 Gauss congruences
Both an and bn possess a wealth of arithmetic properties. We concentrate on Gauss congruences.
Definition 1.3.
A sequence of integers is said to satisfy the Gauss congruences of order r if, for all primes and all positive integers k, n we have(1.10) (1.10)
The case r = 1 is simply called ‘Gauss congruences’, while the case r > 1 is an example of supercongruences. A well-known example of a sequence satisfying the Gauss congruences is for any choice of . Another example of such a sequence is given by [46, Ch. 5.3.3]. In fact, satisfies the Gauss congruences of order 3, as shown by Jacobsthal [Citation13] (cf. [46, Ch. 7.1.6]).
Chowla et al. [Citation16], based on numerical observations, made several conjectures on the values of an modulo primes and prime powers. Gessel [Citation29] proved generalizations of their conjectures. In particular, in [29, Thm. 3] he proved that an satisfies the Gauss congruences of order 3, but only in the special case where k = 1 in (1.10). This was established independently by Mimura [37, Prop. 2]. Finally, in his PhD thesis, Coster [Citation23] showed that the Gauss congruences of order 3 are satisfied by an (without any restriction on k), as well as by bn.
Following numerical observations, it was conjectured that all sporadic sequences satisfy the Gauss congruences of order 2 or 3 (depending on the sequence), see [Citation44, Table 2]. The fact that they all satisfy the Gauss congruences of order 1 is explained by modular parametrization, see e.g. [Citation38, Citation41, Citation56]. Establishing Gauss congruences of higher order is a harder task. In his thesis, Coster [Citation23] established such congruences for ‘generalized Apéry numbers’, that isfor and . Specifically, he proved that satisfies the Gauss congruences of order 3 as long as , or r = 1, and . The sequences include, as special cases, the sequences and D, as well as A and s10. Coster’s method for proving these congruences was extended and used by various authors:
Osburn and Sahu [Citation42] proved that C satisfies the Gauss congruences of order 3.
Later, Osburn and Sahu [Citation43] proved that satisfies the Gauss congruences of order 3 as well, extending a result of Chan, Cooper and Sica [Citation17]. They also sketched the proof of a similar result for E (using the representation (3.3)).
Osburn et al. [Citation44] proved that s18 satisfies the Gauss congruences of order 3, as well as a family that includes s7, D and . In [44, Ex. 3.1] they sketch a proof of a similar result for .
All these sequences involved only one summation variable, k. Recently, the author established that the sequence , whose definition involves summation over two variables k and l, satisfies the Gauss congruences of order 3 [30, Cor. 1.4]. Although the proof uses q-congruences, this can be avoided, and one can formulate the proof using Coster’s method.
This leaves three sequences for which little is known: B, F and . It was proved by Amdeberhan and Tauraso [Citation3] that satisfies the Gauss congruences of order 3, but only in the special case where k = 1 in (1.10). The main difficulty in dealing with B, F and has to do with the appearance of powers (of 3 and 8) in their definition. As opposed to binomial coefficients, which behave nicely modulo high powers of primes, powers’ behavior is much worse. For instance, satisfies the Gauss congruences of order 1 but not of higher order, as opposed to . In fact, it is not even known if there are infinitely many primes p such that ; such primes are known as Wieferich primes. We prove the following.
Theorem 1.4.
Sequence B satisfies the Gauss congruences of order 2.
This confirms a conjecture of Osburn, Sahu and Straub [Citation44].
Remark 1.5.
Theorem 1.1 also gives a new proof for the fact that all the sporadic sequences are integral and satisfy the Gauss congruences (of order 1), since these congruences hold for the constant terms of powers of any integral Laurent series.
2 Further results
2.1 Lucas and D3 congruences
Definition 2.1.
A sequence of integers is said to satisfy the Lucas congruences if, for any prime p, and any positive integer n with base-p expansion () we have
The name of these congruences comes from a theorem of Lucas [Citation34], stating that if and are base-p expansions of , then
Gessel [29, Thm. 1], inspired by the previously mentioned conjectures of Chowla, Cowles and Cowles, has shown that an satisfies the Lucas congruences. Deutsch and Sagan [27, Thm. 5.9] showed more generally that for any positive integers r and s, the sequencesatisfies the Lucas congruences, which implies that an and bn satisfy the Lucas congruences. They also showed that satisfies Lucas congruences [27, Thm. 4.4]. A property stronger than Lucas congruences is the D3 property, named after Dwork [Citation28].
Definition 2.2.
Let be a sequence of integers, and set . We say that satisfies the D3 congruences with respect to a prime p if(2.1) (2.1) for all . If this holds for all primes p, we simply say that satisfies the D3 congruences.
These congruences are implicit in the work of Dwork [Citation28] and have significance in p-adic analysis. We refer the reader to [Citation40, Citation52] for further information, as well as definitions of related congruences, known as D1 and D2, which together imply D3. One can show that (2.1) with s = 1 implies the Lucas congruences. Samol and van Straten [Citation52] proved that a certain class of sequences satisfies the D3 congruences. A different proof of this was found by Mellit and Vlasenko [Citation40]. Before we state their result, we recall that the Newton polytope of a Laurent polynomial is the convex hull of the support of f, that is, of . We use the notation for the constant term of f (or any Laurent series).
Theorem 2.3.
[Citation40, Citation52] Let be a Laurent polynomial. If the Newton polytope of Λ contains the origin as its only interior integral point, then the sequence satisfies the D3 congruences.
As observed, e.g., by Mellit and Vlasenko [40, §1], bn is the constant term of the nth power of(2.2) (2.2)
Straub [51, Rem. 1.4] observed that an is the constant term of the nth power of(2.3) (2.3)
A short computation shows that the Newton polytopes of (2.2) and (2.3) satisfy the conditions of Theorem 2.3, and so an and bn satisfy the D3 congruences.
2.2 D3 congruences for sporadic sequences
Malik and Straub [39, Thm. 3.1] proved that all 15 Apéry-like sequences satisfy the Lucas congruences. For all sequences except for and s18 they did this by extending and applying a general method developed by McIntosh [Citation36], while for and s18 a much more delicate analysis was needed. From Theorems 2.3 and 1.1 we immediately obtain
Corollary 2.4.
All the sporadic sequences, except possibly , satisfy the D3 congruences.
2.3 p-adic valuation of sporadic sequences
Given a sequence of integers and a prime p, let . For any non-negative integer n whose base-p expansion is , let .
Let be the p-adic valuation of an integer n. Recently, Delaygue proved that under two conditions on [26, Thm. 2]. The first condition is that for some Laurent polynomial Λ with integer coefficients, whose Newton polytope contains the origin as its only interior integral point. This holds for all the sporadic sequences except possibly . The second condition involves a restriction on the differential operator annihilating . According to (1.6), (1.7) and [39, p. 23], this condition holds for all 15 sequences. We thus obtain
Corollary 2.5.
Every sporadic sequence , except possibly , satisfies for every and every prime p.
3 Constant term and binomial representations for sporadic sequences
3.1 Review of diagonals, constant terms and binomial representations
Given a power series , we define its diagonal as the power series
Given a rational function whose denominator does not vanish at the origin, we can identify it with its Taylor expansion around . For such A we have the following result of Lipshitz.
Theorem 3.1.
[Citation33] Let (K a field of characteristic 0). If the denominator of A does not vanish at the origin, then the diagonal satisfies a homogeneous linear differential equation with coefficients in .
Given a Laurent series , we define its constant term series as the power series
The constant term series of f is also known as its fundamental period. Suppose a Laurent polynomial in d variables becomes a polynomial when multiplied by . Then it is known that A can be expressed as the diagonal of a rational function in d + 1 variables, namely(3.1) (3.1)
In general, the diagonal of a rational function cannot be expressed as the constant term series of a Laurent polynomial. However, we have the following partial converse. This converse is a special case of the MacMahon Master Theorem [35].
Lemma 3.2.
Given a square matrix , consider the rational function
We have(3.2) (3.2)
As the rational function in the right-hand side of (3.2) is homogeneous of degree 0, we may substitute xi = 1 for some i without changing the constant term series.
There are various benefits that come from representing a sequence as the diagonal of a rational function, some of which are described in [51, p. 1987]. For instance, it allows one to extract the asymptotics of the sequence directly from the rational function. A much simpler benefit, which we make use of in Proposition 1.2, is that it allows us to easily extract binomial sum representations for the sequenceFootnote1.
3.2 Known binomial sum representations
In [57, §7], Zagier explains how the modular parametrization of a sequence can be used to obtain binomial representations for it. In practice, many of the binomial representations are found using educated guesses or computer searches.
Binomial representations for A and can be found in [Citation48, Citation50], for and can be found in [Citation25] and for and can be found in [Citation24]. Additional representations for A, D, E and are found in their OEIS pages [Citation49], and we single out the following representation for E,(3.3) (3.3)
This formula for E was used in [Citation43] in establishing Gauss congruences of order 2; the original representation of E, given by Zagier, involves powers of 4, which get in the way of establishing supercongruences. A proof of (3.3) is sketched in [26, p. 257].
3.3 Known diagonal representations
Bostan et al. [11, Thm. 3.5] proved (algorithmically) that if is a binomial sum, then the g.f. of un is a diagonal of a rational function. Hence, every sporadic sequence is the diagonal of some rational function. We are aware of the following explicit diagonal representations for the 10 sequences A, B, C, D, F, , s7 and s10:
Zagier [Citation57] (cf. Verrill [Citation54]) already explained how his Sporadic sequences have constant term representations, but this has been worked out only in a few cases: A is the constant term sequence of
and so of as well, and F is the constant term sequence ofand so of as well.
Various rational functions whose diagonal is the g.f. for are given in [Citation6, Citation18, Citation19, Citation51]. Of particular interest is the one found by Straub, which involves 4 variables, while others involve between 5 to 8 variables. His rational function is [51, Thm. 1.1](3.4) (3.4)
Recently, Gheorghe Coserea added to the OEIS [Citation49] two rational functions, in 4 variables, whose diagonal is the g.f. of : and .
Straub [51, Ex. 3.4] observed that the diagonal of(3.5) (3.5)
Straub observed that(3.6) (3.6)
are both diagonals for A [51, Ex. 3.6, 3.7]. In the OEIS, Coserea provides three more rational functions, in 3 variables as well, whose diagonal is the g.f. of A:(3.7) (3.7)
and .
Straub [51, Eq. (32)] also observed that the g.f. for coincides with the diagonal of
The diagonal of is the g.f. for s10, see Straub [51, Ex. 3.6]. Coserea [Citation49] also provides the rational function .
Sequences C and are the constant term sequences of the Laurent polynomial with d = 3 and d = 4, respectively (see [15, Ex. 6], [14, Ex. 3.17], [12, Eq. (8)]). By (3.1), this also gives diagonal representations for C and .
Coserea lists the functions and as rational functions whose diagonal is the g.f. for B [Citation49].
Coserea has also found that the diagonal of is the g.f. for s7.
In [53, Ex. 4.2], Straub and Zudilin prove that the g.f. of is the diagonal of the Kauers–Zeilberger rational function , where e3 is the third elementary symmetric polynomial.
In the proof of Straub [51, Thm. 1.1], Straub observes that the rational function (3.4) may be written as whereand so by Lemma 3.2 it gives rise to a constant term representation for , the one defined by (2.3). Representations of the form are hidden in some of the other functions that we listed, as we now show. The rational function in (3.5) can be written as where and by Lemma 3.2, we find that the constant term series of (2.2) is the g.f. of D. The rational functions in (3.6) and (3.7) can be written as for
3.4 New constant term representations
The discussion in §3.3 shows that there are known constant term representations for A, B, C, F, and , and also that a previously known diagonal representation for D gives rise to constant term representation through Lemma 3.2.
We call a Laurent polynomial good if it may be factored into a product of Laurent polynomial, each of which has coefficients in the set . We have conducted a computer search of the following form: we enumerated good Laurent polynomials f (in several variables) with bounded degree and coefficients. For each such polynomial we tested whether is the g.f. of a sporadic sequence. For each of the 15 sequences, the search yielded more than one Laurent polynomial whose is its g.f. The reason for looking only at good Laurent polynomials is that they lead to binomial sum representations which do not involve powers of integers other than ±1 (as in Proposition 1.2).
We list some of the Laurent polynomials we have found in the following theorem.
Proposition 3.3.
The constant term series of the Laurent polynomials in the second column are g.f.s of the sequences in the first column:
Table
Table
As can be seen, we were always able to find good symmetric polynomials, except in the cases of , s7 and s18. For is a symmetric polynomial whose constant term series is (empirically) the g.f. of , but P is not good according to our definition. For s18, is a symmetric polynomial whose constant term series is (empirically) the g.f. of s18, and is good according to our definition, but we do not know how to prove it. Note that the first polynomial for B was given in [Citation54], although it was not factored; its factorization is useful in the proofs of Proposition 1.2 and Theorem 1.4.
Proof of Proposition 3.3.
For each of the Laurent polynomials in the proposition, we may easily compute the first three terms and see that they agree with the first three terms of the respective sequence. It now suffices to compute the differential operator annihilating the constant term series of the polynomial, and that it coincides with (1.6) or (1.7).
The proof of Theorem 3.1 is constructive and may give rise to such a differential operator (by appealing to (3.1)). However, the algorithm arising from Lipshitz’s theorem is often not efficient enough for practical purposes. However, Zeilberger’s Creative Telescoping algorithm [59], as generalized by Chyzak [Citation20], gives a much faster way to compute the operator. See [Citation8] for more details.
The algorithm is implemented in the package .Footnote2 To find an operator annihilating the diagonal of , we run the command
P_, S_, T_: = op(op(Mgfun:-creative_telescoping(F, x::diff,
[s::diff, t::diff])))on , and will be the operator. So if we want to find the operator annihilating , we shall run it on . To find an operator annihilating the diagonal of , we run the command
P_, S_, T_, X_: = op(op(Mgfun:-creative_telescoping(F, w::diff,
[s::diff, t::diff, x::diff])))on , and will be the operator. So if we want to find the operator annihilating , we shall run it on .
In the two-variate case, the algorithm returns the expected operator in under a minute on an average laptop. For the three-variate case, the algorithm takes a few minutes when we ran it on the polynomials for , s7 and s10. For and s18 we had to ‘massage’ the Laurent polynomials slightly before running the program, as we now explain.
For , we consider . We know by Lemma 3.2 that for the matrix
we have
Instead of running the algorithm on , we ran it on where , and it proved to be much quicker. The same trick works for the first polynomial of . The second polynomial for α comes from the polynomial mentioned in §3.3, in which we plugged .
For the second polynomial f for , note that for . We ran the algorithm on . For s18, the same trick works.
Finally, for we use a different implementation of Creative Telescoping, which turned out to be faster for this example, namely the Mathematica package HolonomicFunctionsFootnote3 by Christoph Koutschan. To find the recurrence satisfied by for the polynomial of , we ran
CreativeTelescoping[((z x + x y - y z - x - 1) (x y + y z - z x - y
- 1)
(y z + z x - x y - z - 1)/(x y z))^n/(x y z), Der[x], {Der[y], Der[z], S[n]}]
CreativeTelescoping[First[%], Der[y], {Der[z], S[n]}]
CreativeTelescoping[First[%], Der[z], {S[n]}]which outputs the expected recurrence. □
3.5 Proof of Proposition 1.2
We use bold Latin letters as short for sequences . If e.g. a sums to n we write for the multinomial coefficient . For the first part, note that
where . By Proposition 3.3, is the generating series for Bn, and so(3.8) (3.8) where the sum is over tuples of non-negative integers satisfying and for i = 1, 2, 3. Observe that . Since Bn is real, we may take the real part of (3.8), obtaining where we sum over tuples in S(n). The second term evaluates to by the multinomial theorem, as needed. For the second part, note that
Since is the generating series for Fn (by Proposition 3.3), one obtains (1.8). For the third part, note that
Since is the generating series for (by Proposition 3.3), one obtains (1.9). □
3.6 Proof of Theorem 1.1
For , we can check that the polynomials in Proposition 3.3 satisfy all the conditions of the theorem by drawing the corresponding Newton polygons (in case there is more than one polynomial, we check the first one only). For instance,
is the Newton polygon of the polynomial corresponding to A in Proposition 3.3. For F, the polynomial satisfies the first condition of the theorem, but not the later two. It turns out that for . Since , we see that Q(x, y) satisfies the second condition. Finally, we check the third condition, regarding the Newton polygon of Q, which is given below:Indeed, (0, 0) is its only interior point. For and , drawing a picture is harder. In , one can check whether a Laurent polynomial in 3 variables contains the origin as its only interior point using the following procedure:def checkInterior(poly):polyhedron = Polyhedron(vertices = poly.exponents())for pt in polyhedron.integral_points():if tuple(pt)! = (0,0,0) and polyhedron.interior_contains(pt):return Falsereturn polyhedron.interior_contains((0,0,0))
For instance, the following code checks that the first polynomial of (from Proposition 3.3) satisfies the conditions of the theorem:
R.<x,y,z> = LaurentPolynomialRing(QQ,3) P =
(y-z + 1)*(-x + y+z)*(x + z + 1)*(x + y-1)/(x*y*z) checkInterior(P)
≫ True
This allowed us to verify that the Newton polytopes contain the origin as their only interior point, except in the case of , where (1, 0, 0) and (1, 1, 0) (as well as their permutations) are additional interior points. □
4 Proof of Theorem 1.4
4.1 Auxiliary results
The following lemma is a classical result of Jacobsthal [Citation13], discovered independently by Kazandzidis [Citation31] (cf. [46, chap. 7.1.6]).
Lemma 4.1.
Let be a prime and be integers. Thenwhere if p = 3 and otherwise.
The identity yields the following congruence.
Lemma 4.2.
Let p be a prime and be integers. Suppose that . Then
4.2 Conclusion of proof
As satisfies the Gauss congruences of order 2 by Lemma 4.1, and so does , we deal with instead of Bn.
The vector stands for . Let and let be a prime. We write for , and if all the entries of are divisible by p. Given we write
Let p be an odd prime and let . By Proposition 1.2, we may write as(4.1) (4.1) where
Here we have used the following observation: if then .
Let with . Suppose without loss of generalization that . At least one of b1 and c1 is indivisible by p as well, because of the condition . Suppose without loss of generalization that . The integer is divisible by , which by Lemma 4.2 implies that , and so .
We now deal with S2. Let . We may write both and as a product of 6 binomial coefficients, namely
Applying Lemma 4.1 6 times, we obtain that(4.2) (4.2) where
Without loss of generalization, suppose that or . In the first case, since , Lemma 4.2 implies that and sowhere , and so the summand of S2 corresponding to vanishes modulo . Suppose now that we are in the second case, . Since for each , we have and in particular , so that by Lemma 4.2, we have(4.3) (4.3)
From (4.2) and (4.3), we obtainwhere , and so the summand of S2 corresponding to vanishes modulo in this case also. All in all, .
Finally, we deal with S3. We have shown that S1 and S2 vanish modulo . By (4.1), it suffices to show that . If then clearly since forces . So from now on we suppose that p = 3, and shall show .
We first demonstrate that each summand of S3 is divisible by . As this is vacuous for k = 1, we assume . The condition implies , which forces one of not being divisible by 3, say . Next, at least one of a2 and c2 is also indivisible by 3 (say, a2) since otherwise the condition gives a contradiction modulo 3. The equality implies that indeed each summand of S3 is divisible by . Assuming still that , we shall show that(4.4) (4.4) finishing the proof for this case. For each in the summation range we define and . The condition is equivalent to . Moreover, we have . This forces being a permutation of , and so if is in the summation range, then and are two additional distinct tuples in the range, and evidently cyclic permutations of contribute the same value to the right-hand side of (4.4), proving the congruence (4.4) holds.
It is left to deal with the case k = 1, which we deal with differently. We must prove that for . By the formula , dating back to Zagier’s paper [Citation57], we have and . For n = 1, it is easy to check that both expression are . For (with ), both expressions are 0 modulo 9 by e.g., Kummer’s [Citation32] theorem . □
5 Legendrian and hypergeometric sequences
Zagier [Citation57] and Almkvist and Zudilin [Citation5] have found the following Legendrian and hypergeometric Apéry-like sequences:for , where if and if . It is relatively easy to find constant-term representations for the hypergeometric sequences, several are listed in the OEIS. For the second-order Legendrian sequence we have found and verified the representation with
The only interior point of Λ’s Newton polytope is the origin, and Λ becomes a ‘good’ Laurent polynomial upon substituting x2, y2 in place of x, y. Similarly, for the second-order Legendrian sequence we have a representation withwhich has no interior points other than the origin, and which factorizes upon substituting x3, y3 for x, y.
For the third-order Legendrian sequence , we have found a constant-term representation withwith no interior points other than the origin. It would be interesting to find useful representations for all the Legendrian solutions.
6 Future directions
Suppose an integer sequence is given in terms of a binomial sum, or as the diagonal of rational function. Is there an algorithm deciding whether for some Laurent polynomial Λ, and for finding such a Λ? Is there an efficient algorithm which produces Λ such that the only integral point in its interior is the origin, in case such Λ exists?
Positive answers to these questions would lead to an alternative proof of Theorem 1.1. The first question is a special case of a question raised by Zagier [58, p. 769, Q1], asking how to recognize that a differential equation has geometric origin.
Fix . Can one classify integer sequences such that satisfies (1.6) (if d = 2) or (1.7) (if d = 3), and in addition for some Laurent polynomial Λ in d variables?
Such a classification might give support to the conjecture that the 15 sporadic sequences that were discovered are the only integral solutions to (1.6) and (1.7).
Can one find new Calabi-Yau equations by considering the annihilating operator of for various good Laurent polynomials Λ? Here “good” is used as in §3.4.
Straub generalized the third-order Gauss congruences for the Apéry nmbers as follows. He proved [51, Thm. 1.2] that for every holds for all and primes . Here is the coefficient of in the Taylor expansion of (3.4). Since , this recovers .
Can one prove (6.1) for the Laurent polynomials in Proposition 3.3, with k = 2 or k = 3?
Acknowledgments
In addition to useful comments, corrections and suggestions, I would like to thank Armin Straub for crucial computational help in the proof of Proposition 3.3 for , and Wadin Zudilin for letting me know of the survey [Citation58]. I am grateful for the referee’s helpful comments regarding the exposition.
Declaration of Interest
No potential conflict of interest was reported by the author(s).
Notes
Additional information
Funding
Notes
1 An example of a binomial sum is (1.1). The interested reader may find a rigorous definition of “binomial sums” in [Citation11].
2 Version 4.1, available from https://specfun.inria.fr/chyzak/mgfun.html.
3 Version 1.7.3, available from https://www3.risc.jku.at/research/combinat/software/ergosum/installation.html.
References
- Apéry, R. (1979). Irrationalité de ζ(2) et ζ(3). Astérisque, 61:11–13.
- Apéry, R. (1981). Interpolation de fractions continues et irrationalité de certaines constantes. In Mathematics, CTHS: Bull. Sec. Sci., III. Paris: Bib. Nat.; pp. 37–53.
- Amdeberhan, T., Tauraso, R. (2016). Supercongruences for the Almkvist-Zudilin numbers. Acta Arith. 173(3):255–268.
- Almkvist, G., van Straten, D., Zudilin, W. Generalizations of Clausen’s formula and algebraic transformations of Calabi-Yau differential equations. Proc. Edinb. Math. Soc. (2), 54(2):273–295. doi:10.1017/S0013091509000959
- Almkvist, G., Zudilin, W. (2006). Differential equations, mirror maps and zeta values. In Mirror Symmetry. V, Vol. 38 of AMS/IP Stud. Adv. Math. Providence, RI: Amer. Math. Soc.; pp. 481–515.
- Bostan, A., Boukraa, S., Christol, G., Hassani, S., Maillard, J.-M. (2013). Ising n-fold integrals as diagonals of rational functions and integrality of series expansions. J. Phys. A. 46(18):185202, 44.
- Bostan, A., Boukraa, S., Maillard, J.-M., Weil, J.-A. (2015).Diagonals of rational functions and selected differential Galois groups. J. Phys. A 48(50):504001, 29.
- Bostan, A., Chyzak, F., van Hoeij, M., Pech, L. (2011/12). Explicit formula for the generating series of diagonal 3D rook paths. Sém. Lothar. Combin. 66:Art. B66a, 27.
- Beukers, F. (1987). Another congruence for the Apéry numbers. J. Number Theory 25(2):201–210. doi:10.1016/0022-314X(87)90025-4
- Beukers, F. (1987). Irrationality proofs using modular forms. Astérisque (147/148):271–283, 345. Journées arithmétiques de Besançon (Besançon, 1985).
- Bostan, A., Lairez, P., Salvy, B. (2017).Multiple binomial sums. J. Symbolic Comput. 80(part 2): 351–386. doi:10.1016/j.jsc.2016.04.002
- Borwein, J. M., Nuyens, D., Straub, A., Wan, J. (2011). Some arithmetic properties of short random walk integrals. Ramanujan J. 26(1):109–132. doi:10.1007/s11139-011-9325-y
- Viggo Brun, J., Stubban, O., Fjeldstad, J. E., Tambs Lyche, R., Aubert, K. E., Ljunggren, W., Jacobsthal, E. (1952). On the divisibility of the difference between two binomial coefficients. In Den 11te Skandinaviske Matematikerkongress, Trondheim, 1949. Johan Grundt Tanums Forlag, Oslo; pp. 42–54.
- Borwein, J. M., Straub, A., Vignat, C. (2016). Densities of short uniform random walks in higher dimensions. J. Math. Anal. Appl., 437(1):668–707.
- Borwein, J. M., Straub, A., Wan, J. (2013).Three-step and four-step random walk integrals. Exp. Math. 22(1):1–14.
- Chowla, S., Cowles, J., Cowles, M. (1980). Congruence properties of Apéry numbers. J. Number Theory 12(2):188–190. doi:10.1016/0022-314X(80)90051-7
- Chan, H. H., Cooper, S., Sica, F. (2010). Congruences satisfied by Apéry-like numbers. Int. J. Number Theory 6(1):89–97. doi:10.1142/S1793042110002879
- Christol, G. (1984). Diagonales de fractions rationnelles et equations différentielles. In Amice, Y., Christol, G., Robba, P., eds. Study group on ultrametric analysis, 10th year: 1982/83, No. 2, pages Exp. No. 18, 10. Paris: Inst. Henri Poincaré.
- Christol, G. (1985). Diagonales de fractions rationnelles et équations de Picard-Fuchs. In Amice, Y., Christol, G., Robba, P., eds. Study group on ultrametric analysis, 12th year, 1984/85, No. 1, pages Exp. No. 13, 12. Paris: Secrétariat Math.
- Chyzak, F. (1997). An extension of Zeilberger’s fast algorithm to general holonomic functions. Vol. 217, pages 115–134. 2000. Formal power series and algebraic combinatorics (Vienna).
- Cooper, S. (2012). Sporadic sequences, modular forms and new series for 1/π. Ramanujan J. 29(1–3):163–183.
- Cooper, S. (2017). Ramanujan’s Theta Functions. Cham: Springer.
- Coster, M. J. (1988). Supercongruences. PhD thesis, University of Leiden.
- Chan, H. H., Tanigawa, Y., Yang, Y., Zudilin, W. (2011). New analogues of Clausen’s identities arising from the theory of modular forms. Adv. Math. 228(2):1294–1314.
- Chan, H. H., Zudilin, W. (2010). New representations for Apéry-like sequences. Mathematika 56(1):107–117. doi:10.1112/S0025579309000436
- Delaygue, É. (2018). Arithmetic properties of Apéry-like numbers. Compos. Math. 154(2):249–274.
- Deutsch, E., Sagan, B. E. (2006). Congruences for Catalan and Motzkin numbers and related sequences. J. Number Theory, 117(1):191–215. doi:10.1016/j.jnt.2005.06.005
- Dwork, B. (1969). p-adic cycles. Inst. Hautes Études Sci. Publ. Math. (37):27–115. doi:10.1007/BF02684886
- Gessel, I. (1982). Some congruences for Apéry numbers. J. Number Theory 14(3):362–368. doi:10.1016/0022-314X(82)90071-3
- Gorodetsky, O. (2019). q-congruences, with applications to supercongruences and the cyclic sieving phenomenon. Int. J. Number Theory 15(9):1919–1968.
- Kazandzidis, G. S. (1969). On congruences in number-theory. Bull. Soc. Math. Grèce (N.S.), 10(fasc., fasc. 1):35–40.
- Kummer, E. E. (1852). Über die Ergänzungssätze zu den allgemeinen Reciprocitätsgesetzen. J. Reine Angew. Math. 44:93–146.
- Lipshitz, L. (1988). The diagonal of a D-finite power series is D-finite. J. Algebra 113(2):373–378. doi:10.1016/0021-8693(88)90166-4
- Lucas, E. (1988). Sur les congruences des nombres eulériens et les coefficients différentiels des functions trigonométriques suivant un module premier. Bull. Soc. Math. Fr. 6:49–54.
- MacMahon, P. A. (1960). Combinatory analysis. Two volumes (bound as one). New York: Chelsea Publishing Co.
- McIntosh, R. J. (1992). A generalization of a congruential property of Lucas. Amer. Math. Monthly 99(3):231–238. doi:10.1080/00029890.1992.11995840
- Mimura, Y. (1983). Congruence properties of Apéry numbers. J. Number Theory 16(1):138–146. doi:10.1016/0022-314X(83)90038-0
- Moy, R. (2013). Congruences among power series coefficients of modular forms. Int. J. Number Theory 9(6):1447–1474. doi:10.1142/S1793042113500371
- Malik, A., Straub, A. (2016). Divisibility properties of sporadic Apéry-like numbers. Res. Number Theory 2:Art. 5, 26.
- Mellit, A., Vlasenko, M. (2016). Dwork’s congruences for the constant terms of powers of a Laurent polynomial. Int. J. Number Theory 12(2):313–321. doi:10.1142/S1793042116500184
- Osburn, R., Sahu, B. (2011). Congruences via modular forms. Proc. Amer. Math. Soc. 139(7):2375–2381.
- Osburn, R., Sahu, B. (2011). Supercongruences for Apéry-like numbers. Adv. in Appl. Math. 47(3):631–638.
- Osburn, R., Sahu, B. (2013). A supercongruence for generalized Domb numbers. Funct. Approx. Comment. Math. 48(part 1):29–36. doi:10.7169/facm/2013.48.1.3
- Osburn, R., Sahu, B., Straub, A. (2016). Supercongruences for sporadic sequences. Proc. Edinb. Math. Soc. (2), 59(2):503–518.
- Peters, C., Stienstra, J. (1989). A pencil of K3-surfaces related to Apéry’s recurrence for ζ(3) and Fermi surfaces for potential zero. In Barth, W. P., Lange, H., eds. Arithmetic of Complex Manifolds (Erlangen, 1988), Vol. 1399 of Lecture Notes in Math. Berlin: Springer; pp. 110–127.
- Robert, A. M. (2000). A Course in p-adic Analysis, Vol. 198 of Graduate Texts in Mathematics. New York: Springer-Verlag.
- Stienstra, J., Beukers, F. (1985). On the Picard-Fuchs equation and the formal Brauer group of certain elliptic K3-surfaces. Math. Ann. 271(2):269–304.
- Schmidt, A. L. (1995). Legendre transforms and Apéry’s sequences. J. Austral. Math. Soc. Ser. A 58(3):358–375.
- Sloane, N. J. A. (1996). The on-line encyclopedia of integer sequences. Available at: https://oeis.org.
- Strehl, V. (1994). Binomial identities—combinatorial and algorithmic aspects. Discret. Math. 136(1–3):309–346. Trends in Discrete Mathematics. doi:10.1016/0012-365X(94)00118-3
- Straub, A. (2014). Multivariate Apéry numbers and supercongruences of rational functions. Algebra Number Theory 8(8):1985–2007.
- Samol, K., van Straten, D. (2015). Dwork congruences and reflexive polytopes. Ann. Math. Qué. 39(2):185–203.
- Straub, A., Zudilin, W. (2015). Positivity of rational functions and their diagonals. J. Approx. Theory 195:57–69. doi:10.1016/j.jat.2014.05.012
- Verrill, H. A. (1999). Some congruences related to modular forms. Max Planck Institute for Mathematics. Available at: http://www.mpim-bonn.mpg.de/preprints?title=%22Some+congruences+related+to+modular+forms%22
- Verrill, H. A. (2001). Picard-Fuchs equations of some families of elliptic curves. In Proceedings on Moonshine and related topics (Montréal, QC, 1999), Vol. 30 of CRM Proc. Lecture Notes, Providence, RI: Amer. Math. Soc.; pp. 253–268.
- Verrill, H. A. (2010).Congruences related to modular forms. Int. J. Number Theory 6(6):1367–1390. doi:10.1142/S1793042110003587
- Zagier, D. (2009). Integral solutions of Apéry-like recurrence equations. In Harnad, J., Winternitz, P., eds. Groups and Symmetries, Vol. 47 of CRM Proc. Lecture Notes. Providence, RI: Amer. Math. Soc.; pp. 3349–3663.
- Zagier, D. (2018). The arithmetic and topology of differential equations. In Mehrmann, V., Skutella, M., eds. European Congress of Mathematics. Zürich: Eur. Math. Soc.; pp. 717–776.
- Zeilberger, D. (1991). The method of creative telescoping. J. Symbolic Comput. 11(3):195–204. doi:10.1016/S0747-7171(08)80044-2