1.
Introduction
In 1920, Ramanujan introduced 17 mock theta functions of odd order in his last letter to Hardy. In addition Ramanujan also gave the mock theta functions of order 6 that are listed in the Lost Notebook. In the study of the arithmetic properties of mock theta functions, many authors have found some congruence properties for their coefficients. For instance, Andrews et al. [1] found several congruences for the partition functions pω(n) and pν(n) corresponding to the mock theta functions ω(q) and ν(q) respectively, defined as:
where
for some positive integer n and
with |q|<1.They proved congruences for modulo 2 and some infinite families of congruences for pω(n) and pν(n). In 2017, Fathima and Pore [4] obtained a number of congruences for pω(n) and pν(n) modulo 20 and some infinite families of congruences modulo 2. In the sequel, Baruah and Begum [2] in 2019 established many congruences for the same partition functions modulo powers of 5.
Zhang in 2018 proved some congruences for the sixth order mock theta function β(q) shown below and also gave some conjectures in [8].
Brietzke, Silva, and Sellers [3] in 2019 found many arithmetic properties satisfied by the coefficients of the eighth order mock theta function V0(q) given as:
Silva and Sellers in [7] proved some congruence relations for the third order mock theta function ξ(q) given by Gordon and McIntosh given below:
The main purpose of this paper is to study the arithmetic properties of the sixth order mock theta functions λ(q) and ρ(q) given by Ramanujan where the two mock theta functions are defined as:
Ramanujan also listed linear relations connecting the sixth order mock theta functions with each other as:
where ψ6(q) is the sixth order mock theta function
The proof technique of all the congruences for mock theta functions involves applying identities on the coefficients in arithmetic progressions, we use the same idea to prove infinite family of congruences modulo certain numbers of the form 2α⋅3β for pλ(n) and pρ(n). The main results are found in Theorem 3.2–3.7 given in Section 3. Before proceeding towards the main theorems, we need some preliminary results given in Section 2 for proving the results in Section 3.
2.
Preliminaries
To shorten the notations, we define
for some positive integer l.
Now, we define Ramanujan's general theta function
Jacobi's triple product identity is defined as:
The special cases for f(a,b) are:
In some of the proofs, we make use of the following identities:
The following lemma exhibits the 3-dissection of ψ(q) and 1/φ(−q).
Lemma 2.1. We have
Proof. Identity (12) is equation (14.3.3) of [5]. Identity (13) comes from [6] shown in equation (2).
Lemma 2.2. We have
Proof. The above identity (14) is (22.1.13) in [5]. And (15) follows from (14) by replacing q by −q and using
From the binomial theorem, for any positive integer l and prime p, we have
3.
Congruence relations for λ(q) and ρ(q)
We first prove the 2-dissection of λ(q).
Theorem 3.1. We have
Proof. To prove the above dissections, consider (5) and replacing q by −q in (5), we have
where the last equality follows from (14). Extracting even and odd terms from above equation, we get
Replace q2 by q in (18) to arrive at (16). Divide (19) by q and replace q2 by q to obtain (17).
Theorem 3.2. We have
Proof. According to (16), we have
Therefore, we complete the proof.
Theorem 3.3. We have
Proof. According to (16), we have
Therefore, we complete the proof.
Corollary 1. Let p>3 be a prime and r an integer such that 8r+1 is a quadratic non-residue modulo p. Then for all n≥0,
Proof. As
Thus, 2r≡k2+k(modp) or 8r+1≡(2k+1)2(modp) which contradicts the fact that 8r+1 is a quadratic non-residue modulo p. Therefore, from Theorem 3.2,
Similarly,
then 2r≡9k2−3k(modp) or 8r+1≡(6k−1)2(modp) which contradicts the fact that 8r+1 is a quadratic non-residue modulo p. Therefore, from Theorem 3.3,
From (22) and (23), we readily arrive at the main result.
Since gcd(6, p) = 1, among the p−1 residues modulo p, there are (p−1)/2 residues r for which 8r+1 is a quadratic non-residue modulo p. So the above result leads us to a number of congruences for different primes p>3 as shown below:
Theorem 3.4. We have
Proof. From (16), we get
Using (9) and (10), we have
Using Lemma 2.1, we obtain
Extracting the terms involving q3n,q3n+1,q3n+2 from above equation, we have
To prove (25), dividing (31) by q2 and replacing q3 by q, we have
which can also be written as
The above equation readily implies (25). Consider (30), dividing by q and replacing q3 by q, we have
which proves (24). Now using (13) in above equation,
Extracting the terms involving q3n,q3n+1,q3n+2 from (33), we have
Dividing (35) and (36) by q and q2 respectively, replacing q3 by q, we arrive at
From the above equations, we get (26) and (27).
Corollary 2. We have
Now we present the infinite families of congruences modulo 12.
Theorem 3.5. For prime p≥5, we have
where i=1,2,⋯,p−1.
Proof. From (29), replacing q3 by q, we have
Reducing modulo 4,
By (21), we have
Extracting the terms involving q3n and replacing q3 by q, we have
From (39) and (40), we have
This implies that pλ(6k)≡0(mod12) unless k is a pentagonal number, or equivalently, unless 24k+1 is a square. Now letting k=p2n+pi+(p2−1)/24 where i=1,2,⋯,(p−1) and p is a prime, we have that 24k+1=24p2n+24pi+p2, and this is evidently not a square since p2 divides the first and third terms but not the middle term. Thus pλ(6k)=pλ(6p2n+6pi+(p2−1)/4)≡0(mod12).
Theorem 3.6. For m≥1, we have
Theorem 3.7. For m≥1, we have
Proof of Theorems 3.6 and 3.7. The proof for the above theorems follows by induction. Let us first prove the first step of induction, for m=1.
From (32), we have
Reducing modulo 12, we have
which proves (41) for m=1. Using (13), we get
or
Extracting the terms involving q3n,q3n+1,q3n+2 from above, we have
Dividing (52) and (53) by q and q2 respectively then replacing q3 by q, we have
Here (54) and (55) proves (46) and (49), respectively for m=1. Replacing q3 by q in (51), we have
The above equation proves (43) for m=1 and it can also be written as:
Using Lemma 2.1, we have
Extracting the terms involving q3n,q3n+1,q3n+2 from above, we get
Replacing q3 by q in (57), we have
which proves (47) for m=1. Now, dividing (58) and (59) by q and q2 respectively then replacing q3 by q,
The above congruences imply (42) and (45) for m=1. Consider
or
Extracting the terms involving q3n,q3n+1,q3n+2 from above, we arrive at
Replacing q3 by q in (60), we obtain
which proves (44) for m=1.
Similar to (56), extracting the terms involving q3n,q3n+1,q3n+2, we ultimately get
Here (63) is the case when m=1 in (48). For the next step of induction, let us suppose that (41)-(49) holds true for m=k. Then, for m=(k+1), we prove the relations in similar manner mentioned above starting from (50) (taking m=k) and obtain (64) (m=k+1). Same process follows for other parts.
Corollary 3. For m≥1, we have
Now, we prove the 2-dissection of ρ(q).
Theorem 3.8. We have
Proof. From (6), we have
Substituting the value from (15), we have
Extracting even and odd terms from above equation, we easily arrive at (69) and (70).
Finally on comparing (16) and (69), we arrive at the following theorem.
Theorem 3.9. We have
The above theorem yields that all the results shown above for λ(q) also hold for ρ(q).
4.
Conclusion and discussions
We have provided elementary proofs of numerous infinite family of congruences satisfied by pλ(n) and pρ(n). We did not carry out a computer search for congruences, and so we are unaware whether other congruences hold beyond the ones we prove in this paper, but certainly there is a possibility to explore more in this direction.
Acknowledgments
The first author is supported by UGC, under grant Ref No. 971/ (CSIR-UGC NET JUNE 2018) and the second author is supported by SERB-MATRICS grant MTR/2019/000123. We would like to thank the two referees for carefully reading our paper and offering corrections and helpful suggestions.