Higher-modulus congruences for the 2-color overpartition function

Less than 1 year old · traced to

Let p‾6(n)\overline{p}_{6}(n) and p‾8(n)\overline{p}_{8}(n) denote the 22-color overpartition functions with indicated color parameters, and let n≥0n\ge0. Higher-modulus congruence conjecture. For all n≥0n\ge0,

p‾8(32n+28)≡0(mod32),p‾6(24n+17)≡0(mod64),p‾6(24n+23)≡0(mod64).\begin{aligned} \overline{p}_{8}(32n+28)&\equiv0\pmod{32},\\ \overline{p}_{6}(24n+17)&\equiv0\pmod{64},\\ \overline{p}_{6}(24n+23)&\equiv0\pmod{64}. \end{aligned}

These congruences are suggested by computations as higher-modulus versions of congruences modulo 88; the paper states them as conjectures, with no proof given.

References

Primary source

H. S. Sumanth Bharadwaj, N. Sujatha and S. Chandankumar, “Arithmetic properties of the 2-color overpartition function p_(n)”, arXiv:2607.16608 (2026).

Progress summary

Refreshed
Claimed solved

A reader-written complete proof claims all three conjectures are true, but no independent verification of that proof has appeared.

Bharadwaj, Sujatha, and Chandankumar’s 2026 paper records three higher-divisibility conjectures for the 22-color overpartition function: two values of p‾6(n)\overline{p}_6(n) should vanish modulo 6464, and one value of p‾8(n)\overline{p}_8(n) should vanish modulo 3232. The paper presents them as conjectures rather than proved results.

Known results

  • The paper proves related Ramanujan-type congruences modulo powers of 22, including p‾8(8n+7)≡0(mod64)\overline{p}_8(8n+7)\equiv0\pmod{64} (Bharadwaj, Sujatha, and Chandankumar, 2026).

Posted attempt

A reader-written argument claims a complete proof using eta products, character projections, Sturm bounds, and finite coefficient recurrences; it states that the required checks establish all three congruences for every n≥0n\ge0. The attempt has not been independently verified.

Current status (as of August 2026): The original paper leaves all three higher-modulus congruences open, while a complete proof has been claimed publicly but remains unverified.

Sources

Solutions 1

ProofThis solution needs a summarySee full solutionHide full solution

All three congruences hold for every n≥0n\ge0. We prove them by holomorphic eta products and complete, explicitly reproducible Sturm certificates.

Put

fr=(qr;qr)∞,∑n≥0p‾ℓ(n)qn=f2f2ℓf12fℓ2.f_r=(q^r;q^r)_\infty,\qquad \sum_{n\ge0}\overline p_\ell(n)q^n =\frac{f_2f_{2\ell}}{f_1^2f_\ell^2}.

Repeated use of the binomial theorem gives

fr2a≡f2r2a−1(mod2a).(1)f_r^{2^a}\equiv f_{2r}^{2^{a-1}}\pmod{2^a}. \tag{1}

The two ℓ=6\ell=6 congruences. The source's exact equation (3.5) gives

∑n≥0p‾6(3n+2)qn=4G(q),G(q)=f4f63f16.\sum_{n\ge0}\overline p_6(3n+2)q^n = 4G(q), \qquad G(q)=\frac{f_4f_6^3}{f_1^6}.

The two required progressions correspond to coefficients of GG in residue classes 5,7(mod8)5,7\pmod8, and these coefficients must vanish modulo 1616. By (1),

G(q)f810≡f158f4f63f82(mod16).G(q)f_8^{10} \equiv f_1^{58}f_4f_6^3f_8^2 \pmod{16}.

Since f8f_8 is a unit in Z[[q8]]\mathbb Z[[q^8]], multiplying by f810f_8^{10} preserves vanishing within each residue class. Form

E6(τ)=η(τ)58η(4τ)η(6τ)3η(8τ)2=q4f158f4f63f82.E_6(\tau) = \eta(\tau)^{58}\eta(4\tau)\eta(6\tau)^3\eta(8\tau)^2 = q^4f_1^{58}f_4f_6^3f_8^2.

The eta exponents satisfy

∑rδ=64,∑δrδ=96,∑(24/δ)rδ=1416.\sum r_\delta=64,\quad \sum\delta r_\delta=96,\quad \sum(24/\delta)r_\delta=1416.

Thus E6E_6 is holomorphic of weight 3232, level 2424, and character (6/⋅)(6/\cdot); all exponents are nonnegative, so every cusp is holomorphic.

For r=1,3r=1,3, project using the four Dirichlet characters modulo 88:

S6,r=∑χ mod 8χ(r)(E6⊗χ)=4∑j≡r (8)[qj]E6 qj.S_{6,r} = \sum_{\chi\bmod8}\chi(r)(E_6\otimes\chi) = 4\sum_{j\equiv r\ (8)}[q^j]E_6\,q^j.

These are holomorphic forms of weight 3232, level 15361536, with index 30723072 and Sturm bound

B6=32⋅307212=8192.B_6=\frac{32\cdot3072}{12}=8192.

The principal-character twist is the oldform E6−(E6∣U2)∣V2E_6-(E_6|U_2)|V_2; the other twists have conductor dividing 88.

The exact coefficient recurrence below verifies both original progressions modulo 6464 through original index

3(B6−4)+2=24566:3(B_6-4)+2=24566:

there are exactly 10231023 checks for each progression. Therefore S6,1≡S6,3≡0(mod64)S_{6,1}\equiv S_{6,3}\equiv0\pmod{64} through their Sturm bound, hence identically. Dividing the displayed character projections by 44 proves

p‾6(24n+17)≡p‾6(24n+23)≡0(mod64)(n≥0).\boxed{ \overline p_6(24n+17) \equiv \overline p_6(24n+23) \equiv0\pmod{64} \quad(n\ge0).}

The ℓ=8\ell=8 congruence. The source's exact equation (4.54) gives

A(q)=∑n≥0p‾8(4n)qn=f217f114f45.A(q) = \sum_{n\ge0}\overline p_8(4n)q^n = \frac{f_2^{17}}{f_1^{14}f_4^5}.

Equation (1) implies

A(q)f836≡f1114f217f427f84(mod32).A(q)f_8^{36} \equiv f_1^{114}f_2^{17}f_4^{27}f_8^4 \pmod{32}.

Set

E8(τ)=η(τ)114η(2τ)17η(4τ)27η(8τ)4=q12f1114f217f427f84.E_8(\tau) = \eta(\tau)^{114}\eta(2\tau)^{17} \eta(4\tau)^{27}\eta(8\tau)^4 = q^{12}f_1^{114}f_2^{17}f_4^{27}f_8^4.

Its eta exponent sums are 162,288,4152162,288,4152, giving a holomorphic form of weight 8181, level 3232, and character (−2/⋅)(-2/\cdot). The desired residue class becomes 3(mod8)3\pmod8 after the shift by 1212. Thus

S8,3=∑χ mod 8χ(3)(E8⊗χ)=4∑j≡3 (8)[qj]E8 qjS_{8,3} = \sum_{\chi\bmod8}\chi(3)(E_8\otimes\chi) = 4\sum_{j\equiv3\ (8)}[q^j]E_8\,q^j

has weight 8181, level 20482048, index 30723072, and Sturm bound

B8=81⋅307212=20736.B_8=\frac{81\cdot3072}{12}=20736.

The exact recurrence verifies the original progression modulo 3232 through index

4(B8−12)=82896,4(B_8-12)=82896,

giving 25902590 checks. Therefore S8,3≡0(mod128)S_{8,3}\equiv0\pmod{128} through the Sturm bound and hence identically. Dividing by 44 and undoing the invertible f8f_8 factor proves

p‾8(32n+28)≡0(mod32)(n≥0).\boxed{\overline p_8(32n+28)\equiv0\pmod{32} \quad(n\ge0).}

Reproducible exact Sturm certificate. All finite checks use only the following division-free integer recurrences:

a(0)=1,a(r)=2∑j2≤r(−1)j+1a(r−j2),a(0)=1,\qquad a(r)= 2\sum_{j^2\le r}(-1)^{j+1}a(r-j^2),

where a(r)=p‾(r)a(r)=\overline p(r), and

p‾ℓ(r)=∑j=0⌊r/ℓ⌋a(r−ℓj)a(j).\overline p_\ell(r) = \sum_{j=0}^{\lfloor r/\ell\rfloor} a(r-\ell j)a(j).

Compute a(r)(mod64)a(r)\pmod{64} for 0≤r≤828960\le r\le82896. Direct substitution gives

p‾6(24j+17)≡0(mod64)(0≤j≤1022),p‾6(24j+23)≡0(mod64)(0≤j≤1022),p‾8(32j+28)≡0(mod32)(0≤j≤2589).\begin{aligned} \overline p_6(24j+17)&\equiv0\pmod{64} && (0\le j\le1022),\\ \overline p_6(24j+23)&\equiv0\pmod{64} && (0\le j\le1022),\\ \overline p_8(32j+28)&\equiv0\pmod{32} && (0\le j\le2589). \end{aligned}

These cover the complete bounds above; Sturm's theorem supplies the all-nn conclusions.

Source: H. S. Sumanth Bharadwaj, N. Sujatha and S. Chandankumar, arXiv:2607.16608, equations (3.5), (4.54), and Conjecture 2, equations (5.4)–(5.6).