1. Introduction
In 2003, Rodriguez-Villegas [Reference Rodriguez-Villegas, Yui and Lewis5] investigated hypergeometric families of Calabi–Yau manifolds. He observed numerically some remarkable supercongruences between the values of the truncated hypergeometric series and expressions derived from the number of
$\mathbb {F}_p$
-points of the associated Calabi–Yau manifolds. For manifolds of dimension
$d=1$
, he conjectured four interesting supercongruences associated to certain elliptic curves, one of which is
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn1.png?pub-status=live)
where
$p\ge 5$
is a prime. The conjectural supercongruence (1.1) was first proved by Mortenson [Reference Mortenson4].
For polynomials
$A_1(q), A_2(q),P(q)\in \mathbb {Z}[q]$
, the q-congruence
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu1.png?pub-status=live)
is understood as
$A_1(q)$
is divisible by
$P(q)$
, and
$A_2(q)$
is coprime with
$P(q)$
. In general, for rational functions
$A(q),B(q)\in \mathbb {Q}(q)$
and polynomial
$P(q)\in \mathbb {Z}[q]$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu2.png?pub-status=live)
Guo and Zeng [Reference Guo and Zeng3, Corollary 2.2] established a q-analogue of (1.1) as follows:
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu3.png?pub-status=live)
Here and in what follows, the q-analogue of the natural number n is defined by
$[n]=(1-q^n)/(1-q)$
, and for
$n\ge 1$
, the q-shifted factorial is defined by
$(a;q)_n=(1-a)(1-aq)\cdots (1-aq^{n-1})$
with
$(a;q)_0=1$
.
In 2011, Sun [Reference Sun7, Conjecture 5.5] conjectured a supercongruence related to (1.1): modulo
$p^2$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn2.png?pub-status=live)
which was proved by Tauraso [Reference Tauraso8] and Sun [Reference Sun6, Theorem 2.2].
Guo and Zeng [Reference Guo and Zeng3, Corollary 2.7] established a partial q-analogue of (1.2):
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn3.png?pub-status=live)
for all primes
$p\equiv 3\pmod {4}$
.
To continue the q-story of (1.2), we recall some q-series notation. The basic hypergeometric series is defined by
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu4.png?pub-status=live)
where
$(a_1,a_2,\ldots ,a_{m};q)_k=(a_1;q)_k(a_2;q)_k\cdots (a_m;q)_k$
. The nth cyclotomic polynomial is given by
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu5.png?pub-status=live)
where
$\zeta $
denotes a primitive nth root of unity.
The motivation for this paper is to extend the q-congruence (1.3) of Guo and Zeng, and establish a complete q-analogue of (1.2).
Theorem 1.1. Let n be an odd positive integer. Then, modulo
$\Phi _n(q)^2$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn4.png?pub-status=live)
The important ingredients in the proof of (1.4) include Andrews’
$_4\phi _3$
terminating identity [Reference Gasper and Rahman2, (II.17), page 355]:
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn5.png?pub-status=live)
The rest of the paper is organised as follows. In the next section, we shall explain why (1.4) is a q-analogue of (1.2). The proof of Theorem 1.1 will be presented in Section 3.
2. Why (1.4) is a q-analogue of (1.2)
Let p be an odd prime. It is clear that
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu6.png?pub-status=live)
Setting
$n\to p$
and
$q\to 1$
on both sides of (1.4) gives, modulo
$p^2$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn6.png?pub-status=live)
By a result due to Chowla et al. [Reference Chowla, Dwork and Evans1],
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu7.png?pub-status=live)
where
$p\equiv 1\pmod {4}$
and
$p=x^2+y^2$
with
$4\mid (x-1)$
. It follows that
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn7.png?pub-status=live)
where we have used the fact [Reference Sun6, page 1918]:
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu8.png?pub-status=live)
Combining (2.1) and (2.2), we arrive at (1.2). Thus, (1.4) is indeed a q-analogue of (1.2).
3. Proof of Theorem 1.1
Let n be an odd positive integer. Setting
$n\to (n-1)/2,q\to q^2,a\to 1$
on both sides of (1.5) gives
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn8.png?pub-status=live)
Letting
$c\to 0$
on both sides of (3.1) and noting that for
$n\equiv 1\pmod {4}$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu9.png?pub-status=live)
and
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu10.png?pub-status=live)
we obtain
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn9.png?pub-status=live)
Since
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu11.png?pub-status=live)
and
$1-q^n\equiv 0\pmod {\Phi _n(q)}$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu12.png?pub-status=live)
It follows that
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu13.png?pub-status=live)
Thus,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqn10.png?pub-status=live)
Finally, substituting (3.3) into the left-hand side of (3.2) gives, modulo
$\Phi _n(q)^2$
,
![](https://static.cambridge.org/binary/version/id/urn:cambridge.org:id:binary:20240829012901882-0004:S0004972724000467:S0004972724000467_eqnu14.png?pub-status=live)
as desired.