Coefficients of mock theta functions

DefinitionThe eighth-order mock theta function \(v(q)\)#✓fully proved

The mock theta function \(v(q) = \sum_{n \ge 0} q^{(n+1)^2} (-q; q^2)_n / (q; q^2)_{n+1}\), where \((-q; q^2)_n = \prod_{k=0}^{n-1} (1 + q^{2k+1})\) and \((q; q^2)_{n+1} = \prod_{k=0}^{n} (1 - q^{2k+1})\), is taken as a power series in \(q\) with integer coefficients.

The \(n\)th summand has order \((n+1)^2\), so the coefficient of \(q^m\) only involves the summands with \(n \le m\).

public noncomputable def mockThetaV : ℤ⟦X⟧ :=
  PowerSeries.mk fun m => coeff m (∑ n ∈ Finset.range (m + 1),
    X ^ ((n + 1) ^ 2) * (∏ k ∈ Finset.range n, (1 + X ^ (2 * k + 1))) *
      invOfUnit (∏ k ∈ Finset.range (n + 1), (1 - X ^ (2 * k + 1))) 1)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (4).
Used by 7

DefinitionThe sixth-order mock theta function \(\sigma(q)\)#✓fully proved

The mock theta function \(\sigma(q) = \sum_{n \ge 0} q^{(n+1)(n+2)/2} (-q; q)_n / (q; q^2)_{n+1}\), where \((-q; q)_n = \prod_{k=1}^{n} (1 + q^k)\) and \((q; q^2)_{n+1} = \prod_{k=0}^{n} (1 - q^{2k+1})\), is taken as a power series in \(q\) with integer coefficients.

The \(n\)th summand has order \((n+1)(n+2)/2 > n\), so the coefficient of \(q^m\) only involves the summands with \(n \le m\).

public noncomputable def mockThetaSigma : ℤ⟦X⟧ :=
  PowerSeries.mk fun m => coeff m (∑ n ∈ Finset.range (m + 1),
    X ^ ((n + 1) * (n + 2) / 2) * (∏ k ∈ Finset.range n, (1 + X ^ (k + 1))) *
      invOfUnit (∏ k ∈ Finset.range (n + 1), (1 - X ^ (2 * k + 1))) 1)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (2).
Used by 4

DefinitionThe sixth-order mock theta function \(\lambda\)#✓fully proved

With \((a; b)_n = \prod_{k=0}^{n-1} (1 - a b^k)\), the mock theta function \(\lambda\) is the integer power series

\[ \lambda(q) = \sum_{n \ge 0} \frac{(-1)^n q^n (q; q^2)_n}{(-q; q)_n}. \]

The \(n\)th summand has order \(n\), so the coefficient of \(q^m\) only involves the summands with \(n \le m\).

public noncomputable def mockThetaLambda : ℤ⟦X⟧ :=
  PowerSeries.mk fun m => coeff m (∑ n ∈ Finset.range (m + 1),
    (-1) ^ n * X ^ n * (∏ k ∈ Finset.range n, (1 - X ^ (2 * k + 1))) *
      invOfUnit (∏ k ∈ Finset.range n, (1 + X ^ (k + 1))) 1)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (3).
Used by 3

TheoremFirst coefficients of the mock theta function \(v\)#✓fully proved

Write \(v(q) = \sum_{n \ge 0} P_v(n) q^n\) for the eighth-order mock theta function \(v\). Then \(P_v(0) = 0\) and \(P_v(1) = 1\).

public theorem coeff_mockThetaV_zero_one :
    coeff 0 mockThetaV = 0 ∧ coeff 1 mockThetaV = 1
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (4).
Uses 1

TheoremFirst coefficients of the mock theta function \(\sigma\)#✓fully proved

Write \(\sigma(q) = \sum_{n \ge 0} P_\sigma(n) q^n\) for the sixth-order mock theta function \(\sigma\). Then \(P_\sigma(0) = 0\) and \(P_\sigma(1) = 1\).

public theorem coeff_mockThetaSigma_zero_one :
    coeff 0 mockThetaSigma = 0 ∧ coeff 1 mockThetaSigma = 1
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (2).
Uses 1

TheoremFirst coefficients of the mock theta function \(\lambda\)#✓fully proved

Write \(\lambda(q) = \sum_{n \ge 0} P_\lambda(n) q^n\) for the sixth-order mock theta function \(\lambda\). Then \(P_\lambda(0) = 1\) and \(P_\lambda(1) = -1\).

public theorem coeff_mockThetaLambda_zero_one :
    coeff 0 mockThetaLambda = 1 ∧ coeff 1 mockThetaLambda = -1
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (3).
Uses 1

DefinitionRamanujan's theta function at monomial arguments#✓fully proved

For an integer \(\varepsilon\) and natural numbers \(a\) and \(b\), \(f(\varepsilon q^a, \varepsilon q^b)\) is the integer power series in \(q\) whose coefficient of \(q^n\) is the sum of \(\varepsilon^{m^2}\) over the integers \(m\) with \(|m| \le n + 1\) and \(a \cdot m(m+1)/2 + b \cdot m(m-1)/2 = n\).

When \(a + b \ge 1\) the exponent \(a \cdot m(m+1)/2 + b \cdot m(m-1)/2\) is at least \(|m|(|m|-1)/2\), so the bound \(|m| \le n + 1\) excludes no term and this is Ramanujan's \(f(c, d) = \sum_{m \in \mathbb{Z}} c^{m(m+1)/2} d^{m(m-1)/2}\) at \(c = \varepsilon q^a\) and \(d = \varepsilon q^b\); the definition is meant for \(a + b \ge 1\). Special cases are \(\psi(q) = f(q, q^3)\) and \(f(-q) = f(-q, -q^2)\), which equals \((q; q)_\infty = \prod_{k \ge 1} (1 - q^k)\).

public def ramanujanThetaMonomial (ε : ℤ) (a b : ℕ) : ℤ⟦X⟧ :=
  PowerSeries.mk fun n => ∑ m ∈ Finset.Icc (-(n : ℤ) - 1) ((n : ℤ) + 1),
    if (a : ℤ) * (m * (m + 1) / 2) + (b : ℤ) * (m * (m - 1) / 2) = n then
      ε ^ (m.natAbs ^ 2) else 0
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Section 2 (definition of f(c, d), Equations (10) and (11)).
Used by 3

TheoremFirst coefficients of Ramanujan's theta function \(\psi\)#✓fully proved

Let \(\psi(q) = f(q, q^3) = \sum_{m \in \mathbb{Z}} q^{m(m+1)/2 + 3m(m-1)/2}\), where \(f\) is Ramanujan's general theta function. The coefficients of \(q^0\), \(q^1\), \(q^2\), \(q^3\) in \(\psi(q)\) are \(1\), \(1\), \(0\), \(1\).

public theorem coeff_ramanujanThetaMonomial_one_one_three :
    coeff 0 (ramanujanThetaMonomial 1 1 3) = 1 ∧ coeff 1 (ramanujanThetaMonomial 1 1 3) = 1 ∧
      coeff 2 (ramanujanThetaMonomial 1 1 3) = 0 ∧
      coeff 3 (ramanujanThetaMonomial 1 1 3) = 1
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Equation (10).
Uses 1

TheoremThe p-dissection of Ramanujan's psi function#●ready to prove

Let \(p\) be an odd prime, let \(f(c, d)\) be Ramanujan's theta function and \(\psi(q) = f(q, q^3)\). Then, as integer power series, \(\psi(q) = \sum_{m=0}^{(p-3)/2} q^{(m^2+m)/2} f(q^{(p^2+(2m+1)p)/2}, q^{(p^2-(2m+1)p)/2}) + q^{(p^2-1)/8} f(q^{p^2}, q^{3p^2})\), where the last factor is \(\psi(q^{p^2})\). Moreover \((m^2+m)/2 \not\equiv (p^2-1)/8 \pmod p\) for every integer \(m\) with \(0 \le m \le (p-3)/2\).

The paper quotes this from Cui and Gu, Theorem 2.1.

public theorem_wanted ramanujanPsi_p_dissection (p : ℕ) (hp : p.Prime) (hp2 : p ≠ 2) :
    ramanujanThetaMonomial 1 1 3 =
        ∑ m ∈ Finset.range ((p - 3) / 2 + 1),
          X ^ ((m ^ 2 + m) / 2) *
            ramanujanThetaMonomial 1 ((p ^ 2 + (2 * m + 1) * p) / 2)
              ((p ^ 2 - (2 * m + 1) * p) / 2) +
          X ^ ((p ^ 2 - 1) / 8) * ramanujanThetaMonomial 1 (p ^ 2) (3 * p ^ 2) ∧
      ∀ m ≤ (p - 3) / 2, ¬ (m ^ 2 + m) / 2 ≡ (p ^ 2 - 1) / 8 [MOD p]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Lemma 1.
Uses 1

TheoremThe \(p\)-dissection of \(f_1\)#●ready to prove

Let \(p \ge 5\) be a prime, let \(f_k = \prod_{m \ge 1} (1 - q^{km})\) for \(k \ge 1\), and let \(f(a, b) = \sum_{m \in \mathbb{Z}} a^{m(m+1)/2} b^{m(m-1)/2}\) be Ramanujan's general theta function. Put \(c = (p-1)/6\) if \(p \equiv 1 \pmod 6\) and \(c = (-p-1)/6\) otherwise. Then, as integer power series,

\[ f_1 = \sum_{t=-(p-1)/2,\ t \ne c}^{(p-1)/2} (-1)^t q^{(3t^2+t)/2} f\bigl(-q^{(3p^2+(6t+1)p)/2}, -q^{(3p^2-(6t+1)p)/2}\bigr) + (-1)^c q^{(p^2-1)/24} f_{p^2}. \]

Moreover, \((3t^2 + t)/2 \not\equiv (p^2 - 1)/24 \pmod p\) for every integer \(t\) with \(-(p-1)/2 \le t \le (p-1)/2\) and \(t \ne c\).

public theorem_wanted eulerProduct_p_dissection (p : ℕ) (hp : p.Prime) (hp5 : 5 ≤ p) :
    let c : ℤ := if p % 6 = 1 then ((p : ℤ) - 1) / 6 else (-(p : ℤ) - 1) / 6
    𝔣 1 =
        ∑ t ∈ (Finset.Icc (-(((p : ℤ) - 1) / 2)) (((p : ℤ) - 1) / 2)).erase c,
          (-1) ^ t.natAbs * X ^ ((3 * t ^ 2 + t) / 2).toNat *
            ramanujanThetaMonomial (-1) ((3 * (p : ℤ) ^ 2 + (6 * t + 1) * p) / 2).toNat
              ((3 * (p : ℤ) ^ 2 - (6 * t + 1) * p) / 2).toNat +
          (-1) ^ c.natAbs * X ^ ((p ^ 2 - 1) / 24) * 𝔣 (p ^ 2) ∧
      ∀ t ∈ (Finset.Icc (-(((p : ℤ) - 1) / 2)) (((p : ℤ) - 1) / 2)).erase c,
        ¬ (3 * t ^ 2 + t) / 2 ≡ ((p : ℤ) ^ 2 - 1) / 24 [ZMOD p]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Lemma 2.
Uses 2

TheoremThe \(p\)-dissection of \(f_1^3\)#●ready to prove

Let \(p\) be an odd prime and, for \(k \ge 1\), let \(f_k = \prod_{m \ge 1} (1 - q^{km})\). Then, as integer power series,

\[ f_1^3 = \sum_{k=0,\ k \ne (p-1)/2}^{p-1} (-1)^k q^{k(k+1)/2} \sum_{n \ge 0} (-1)^n (2pn + 2k + 1) q^{pn(pn+2k+1)/2} + p (-1)^{(p-1)/2} q^{(p^2-1)/8} f_{p^2}^3. \]

Moreover, \((k^2 + k)/2 \not\equiv (p^2 - 1)/8 \pmod p\) for every integer \(k\) with \(0 \le k \le p - 1\) and \(k \ne (p-1)/2\).

public theorem_wanted eulerProduct_cube_p_dissection (p : ℕ) (hp : p.Prime) (hp2 : p ≠ 2) :
    𝔣 1 ^ 3 =
        ∑ k ∈ (Finset.range p).erase ((p - 1) / 2),
          (-1) ^ k * X ^ (k * (k + 1) / 2) *
            PowerSeries.mk (fun e => ∑ n ∈ Finset.range (e + 1),
              if p * n * (p * n + 2 * k + 1) / 2 = e then
                (-1) ^ n * ((2 * p * n + 2 * k + 1 : ℕ) : ℤ) else 0) +
          (p : ℤ⟦X⟧) * (-1) ^ ((p - 1) / 2) * X ^ ((p ^ 2 - 1) / 8) * 𝔣 (p ^ 2) ^ 3 ∧
      ∀ k ∈ (Finset.range p).erase ((p - 1) / 2),
        ¬ (k ^ 2 + k) / 2 ≡ (p ^ 2 - 1) / 8 [MOD p]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Lemma 3.
Uses 1

TheoremGenerating function for the odd-indexed coefficients of \(v(q)\)#●ready to prove

Let \(P_v(n)\) be the coefficient of \(q^n\) in the eighth-order mock theta function \(v(q)\), and for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\). Then, as integer power series, \(f_1 f_2 \sum_{n \ge 0} P_v(2n+1) q^n = f_4^3\), that is, \(\sum_{n \ge 0} P_v(2n+1) q^n = f_4^3 / (f_1 f_2)\).

public theorem_wanted mockThetaV_odd_generatingFunction :
    𝔣 1 * 𝔣 2 * PowerSeries.mk (fun n => coeff (2 * n + 1) mockThetaV) = 𝔣 4 ^ 3
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 1.
Uses 2

TheoremThe coefficients of \(q^{6n+5}\) in \(v(q)\) are divisible by 3#●ready to prove

Let \(P_v(n)\) be the coefficient of \(q^n\) in the eighth-order mock theta function \(v(q)\). Then \(P_v(6n+5) \equiv 0 \pmod 3\) for every \(n \ge 0\).

public theorem_wanted mockThetaV_coeff_six_mul_add_five_mod_three (n : ℕ) :
    coeff (6 * n + 5) mockThetaV ≡ 0 [ZMOD 3]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 2 (i).
Uses 1

TheoremInfinite family of congruences modulo 2 for \(v(q)\)#●ready to prove

Let \(P_v(n)\) be the coefficient of \(q^n\) in the eighth-order mock theta function \(v(q)\), and for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\). Let \(p \ge 3\) be a prime whose Legendre symbol \(\left(\frac{-2}{p}\right)\) equals \(-1\), and let \(\alpha \ge 0\) be an integer. Then the integer power series \(f_1 \sum_{n \ge 0} P_v\bigl(2 p^{2\alpha} n + (3 p^{2\alpha} + 1)/4\bigr) q^n\) and \(f_2 f_4^2\) have congruent coefficients modulo \(2\). Moreover, for every \(n \ge 0\) and every \(j\) with \(1 \le j \le p - 1\), \(P_v\bigl(2 p^{2\alpha+2} n + 2 p^{2\alpha+1} j + (3 p^{2\alpha+2} + 1)/4\bigr) \equiv 0 \pmod 2\).

The divisions by \(4\) are exact since \(p^2 \equiv 1 \pmod 8\). Since \(f_1\) has an inverse with integer coefficients, the first congruence says that \(\sum_{n \ge 0} P_v(2 p^{2\alpha} n + (3 p^{2\alpha} + 1)/4) q^n \equiv \psi(q) \psi(q^2) \pmod 2\), where \(\psi(q) = \sum_{n \ge 0} q^{n(n+1)/2}\), so that \(\psi(q) \psi(q^2) = f_2 f_4^2 / f_1\).

public theorem_wanted mockThetaV_congruences_mod_two (p α : ℕ) [Fact p.Prime] (hp : 3 ≤ p)
    (hleg : legendreSym p (-2) = -1) :
    (∀ i, coeff i (𝔣 1 * PowerSeries.mk (fun n =>
        coeff (2 * p ^ (2 * α) * n + (3 * p ^ (2 * α) + 1) / 4) mockThetaV)) ≡
      coeff i (𝔣 2 * 𝔣 4 ^ 2) [ZMOD 2]) ∧
    ∀ j n, 1 ≤ j → j ≤ p - 1 →
      coeff (2 * p ^ (2 * α + 2) * n + 2 * p ^ (2 * α + 1) * j + (3 * p ^ (2 * α + 2) + 1) / 4)
        mockThetaV ≡ 0 [ZMOD 2]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 2 (ii), Equations (21) and (22).
Uses 2

TheoremInfinite family of congruences modulo 6 for \(v(q)\)#●ready to prove

Let \(P_v(n)\) be the coefficient of \(q^n\) in the eighth-order mock theta function \(v(q)\), and for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\). Let \(p \ge 5\) be a prime whose Legendre symbol \(\left(\frac{-18}{p}\right)\) equals \(-1\), and let \(\alpha \ge 0\) be an integer. Then for every \(n \ge 0\), \(P_v\bigl(6 p^{2\alpha} n + (19 p^{2\alpha} + 1)/4\bigr)\) is congruent modulo \(6\) to the coefficient of \(q^n\) in \(3 f_1 f_6^3\). Moreover, for every \(n \ge 0\) and every \(j\) with \(1 \le j \le p - 1\), \(P_v\bigl(6 p^{2\alpha+2} n + 6 p^{2\alpha+1} j + (19 p^{2\alpha+2} + 1)/4\bigr) \equiv 0 \pmod 6\).

The divisions by \(4\) are exact since \(p^2 \equiv 1 \pmod 8\).

public theorem_wanted mockThetaV_congruences_mod_six (p α : ℕ) [Fact p.Prime] (hp : 5 ≤ p)
    (hleg : legendreSym p (-18) = -1) :
    (∀ i, coeff (6 * p ^ (2 * α) * i + (19 * p ^ (2 * α) + 1) / 4) mockThetaV ≡
      coeff i (3 * 𝔣 1 * 𝔣 6 ^ 3) [ZMOD 6]) ∧
    ∀ j n, 1 ≤ j → j ≤ p - 1 →
      coeff (6 * p ^ (2 * α + 2) * n + 6 * p ^ (2 * α + 1) * j +
        (19 * p ^ (2 * α + 2) + 1) / 4) mockThetaV ≡ 0 [ZMOD 6]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 2 (iii), Equations (23) and (24).
Uses 2

TheoremSum formula for the odd-indexed coefficients of \(v(q)\)#●ready to prove

Let \(P_v(n)\) be the coefficient of \(q^n\) in the eighth-order mock theta function \(v(q)\), for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\), and let \(b_4^1(k)\) be the coefficient of \(q^k\) in the integer power series \(f_4 / f_1\). Then for every \(n \ge 0\), \(P_v(2n+1) = \sum_{m \ge 0,\ m(m+1) \le n} b_4^1\bigl(n - m(m+1)\bigr)\).

public theorem_wanted mockThetaV_odd_eq_sum_regular (n : ℕ) :
    coeff (2 * n + 1) mockThetaV =
      ∑ m ∈ (Finset.range (n + 1)).filter (fun m => m * (m + 1) ≤ n),
        coeff (n - m * (m + 1)) (𝔣 4 * invOfUnit (𝔣 1) 1)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 3.
Uses 2

TheoremSum formula for the coefficients of \(q^{6n+5}\) in \(v(q)\)#●ready to prove

Let \(P_v(n)\) be the coefficient of \(q^n\) in the eighth-order mock theta function \(v(q)\), for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\), and let \(b_6^3(k)\) be the coefficient of \(q^k\) in the integer power series \((f_6 / f_1)^3\). Then for every \(n \ge 0\), \(P_v(6n+5) = 3 b_6^3(n) + 3 \sum_{m \ge 1,\ 2m(3m+1) \le n} (-1)^m b_6^3\bigl(n - 2m(3m+1)\bigr) + 3 \sum_{m \ge 1,\ 2m(3m-1) \le n} (-1)^m b_6^3\bigl(n - 2m(3m-1)\bigr)\).

public theorem_wanted mockThetaV_six_mul_add_five_eq_sum_regular (n : ℕ) :
    coeff (6 * n + 5) mockThetaV =
      3 * coeff n ((𝔣 6 * invOfUnit (𝔣 1) 1) ^ 3) +
        3 * ∑ m ∈ (Finset.Icc 1 n).filter (fun m => 2 * m * (3 * m + 1) ≤ n),
          (-1) ^ m * coeff (n - 2 * m * (3 * m + 1)) ((𝔣 6 * invOfUnit (𝔣 1) 1) ^ 3) +
        3 * ∑ m ∈ (Finset.Icc 1 n).filter (fun m => 2 * m * (3 * m - 1) ≤ n),
          (-1) ^ m * coeff (n - 2 * m * (3 * m - 1)) ((𝔣 6 * invOfUnit (𝔣 1) 1) ^ 3)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 4.
Uses 2

TheoremGenerating function for the odd-indexed coefficients of \(\sigma(q)\)#●ready to prove

Let \(P_\sigma(n)\) be the coefficient of \(q^n\) in the sixth-order mock theta function \(\sigma(q)\), and for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\). Then, as integer power series, \(f_1^2 f_3 \sum_{n \ge 0} P_\sigma(2n+1) q^n = f_2^2 f_6^2\), that is, \(\sum_{n \ge 0} P_\sigma(2n+1) q^n = f_2^2 f_6^2 / (f_1^2 f_3)\).

public theorem_wanted mockThetaSigma_odd_generatingFunction :
    𝔣 1 ^ 2 * 𝔣 3 * PowerSeries.mk (fun n => coeff (2 * n + 1) mockThetaSigma) =
      𝔣 2 ^ 2 * 𝔣 6 ^ 2
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 5.
Uses 2

TheoremCongruences modulo 2 for the mock theta function \(\sigma\)#●ready to prove

Write \(\sigma(q) = \sum_{n \ge 0} P_\sigma(n) q^n\) for the sixth-order mock theta function \(\sigma\) and let \(f_k = \prod_{m \ge 1} (1 - q^{km})\). Let \(p \ge 5\) be a prime with Legendre symbol \(\left(\frac{-2}{p}\right) = -1\) and let \(\alpha \ge 0\) be an integer. Then, coefficientwise,

\[ f_3 \sum_{n \ge 0} P_\sigma\Bigl(2 p^{2\alpha} n + \frac{11 p^{2\alpha} + 1}{12}\Bigr) q^n \equiv f_2 f_6^2 \pmod 2, \]

and for all integers \(n \ge 0\) and all integers \(j\) with \(1 \le j \le p - 1\),

\[ P_\sigma\Bigl(2 p^{2\alpha+2} n + 2 p^{2\alpha+1} j + \frac{11 p^{2\alpha+2} + 1}{12}\Bigr) \equiv 0 \pmod 2. \]

Since \(f_2 f_6^2 / f_3 = f_2 \psi(q^3)\) with \(\psi(q) = \sum_{n \ge 0} q^{n(n+1)/2}\), the first congruence says that the series of \(P_\sigma(2 p^{2\alpha} n + (11 p^{2\alpha} + 1)/12)\) is congruent to \(f_2 \psi(q^3)\) modulo \(2\). The division by \(12\) is exact since \(p^2 \equiv 1 \pmod{24}\).

public theorem_wanted mockThetaSigma_congruences_mod_two (p α : ℕ) [Fact p.Prime] (hp : 5 ≤ p)
    (hleg : legendreSym p (-2) = -1) :
    (∀ i, coeff i (𝔣 3 * PowerSeries.mk (fun n =>
        coeff (2 * p ^ (2 * α) * n + (11 * p ^ (2 * α) + 1) / 12) mockThetaSigma)) ≡
      coeff i (𝔣 2 * 𝔣 6 ^ 2) [ZMOD 2]) ∧
    ∀ j n, 1 ≤ j → j ≤ p - 1 →
      coeff (2 * p ^ (2 * α + 2) * n + 2 * p ^ (2 * α + 1) * j +
        (11 * p ^ (2 * α + 2) + 1) / 12) mockThetaSigma ≡ 0 [ZMOD 2]
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 6.
Uses 2

TheoremSum formula for the odd-indexed coefficients of \(\sigma(q)\)#●ready to prove

Let \(P_\sigma(n)\) be the coefficient of \(q^n\) in the sixth-order mock theta function \(\sigma(q)\), for \(k \ge 1\) let \(f_k = \prod_{m \ge 1} (1 - q^{km})\), and let \(b_2^2(k)\) be the coefficient of \(q^k\) in the integer power series \((f_2 / f_1)^2\). Then for every \(n \ge 0\), \(P_\sigma(2n+1) = \sum_{m \ge 0,\ (3m^2+3m)/2 \le n} b_2^2\bigl(n - (3m^2+3m)/2\bigr)\).

public theorem_wanted mockThetaSigma_odd_eq_sum_regular (n : ℕ) :
    coeff (2 * n + 1) mockThetaSigma =
      ∑ m ∈ (Finset.range (n + 1)).filter (fun m => (3 * m ^ 2 + 3 * m) / 2 ≤ n),
        coeff (n - (3 * m ^ 2 + 3 * m) / 2) ((𝔣 2 * invOfUnit (𝔣 1) 1) ^ 2)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 7.
Uses 2

TheoremEven-index coefficients of the mock theta function \(\lambda\)#●ready to prove

Write \(\lambda(q) = \sum_{n \ge 0} P_\lambda(n) q^n\) for the sixth-order mock theta function \(\lambda\), let \(f_k = \prod_{m \ge 1} (1 - q^{km})\), and let \(b_2^3(m)\) be the coefficient of \(q^m\) in \((f_2/f_1)^3\). Then for every \(n \ge 0\),

\[ P_\lambda(2n) = b_2^3(n) + 2 \sum_{k \ge 1,\ 3k^2 \le n} (-1)^k b_2^3(n - 3k^2). \]
public theorem_wanted mockThetaLambda_even_eq_sum_regular (n : ℕ) :
    coeff (2 * n) mockThetaLambda =
      coeff n ((𝔣 2 * invOfUnit (𝔣 1) 1) ^ 3) +
        2 * ∑ k ∈ (Finset.Icc 1 n).filter (fun k => 3 * k ^ 2 ≤ n),
          (-1) ^ k * coeff (n - 3 * k ^ 2) ((𝔣 2 * invOfUnit (𝔣 1) 1) ^ 3)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 8.
Uses 2

TheoremCoefficients of the mock theta function \(\lambda\) at indices \(6n+2\)#●ready to prove

Write \(\lambda(q) = \sum_{n \ge 0} P_\lambda(n) q^n\) for the sixth-order mock theta function \(\lambda\), let \(f_k = \prod_{m \ge 1} (1 - q^{km})\), and let \(\overline{p}_3(j)\) be the coefficient of \(q^j\) in \((f_2/f_1^2)^3\). Then for every \(n \ge 0\),

\[ P_\lambda(6n+2) = 3 \sum_{m \ge 0,\ 3m(m+1)/2 \le n} (-1)^m (2m+1)\, \overline{p}_3\bigl(n - 3m(m+1)/2\bigr) + 6 \sum_{m \ge 0} \sum_{k \ge 1,\ 3k^2 + 3m(m+1)/2 \le n} (-1)^{m+k} (2m+1)\, \overline{p}_3\bigl(n - 3k^2 - 3m(m+1)/2\bigr). \]
public theorem_wanted mockThetaLambda_six_mul_add_two_eq_sum_overpartition (n : ℕ) :
    coeff (6 * n + 2) mockThetaLambda =
      3 * ∑ m ∈ (Finset.range (n + 1)).filter (fun m => 3 * m * (m + 1) / 2 ≤ n),
          (-1) ^ m * (2 * m + 1 : ℤ) *
            coeff (n - 3 * m * (m + 1) / 2) ((𝔣 2 * invOfUnit (𝔣 1 ^ 2) 1) ^ 3) +
        6 * ∑ m ∈ Finset.range (n + 1), ∑ k ∈ (Finset.Icc 1 n).filter
            (fun k => 3 * k ^ 2 + 3 * m * (m + 1) / 2 ≤ n),
          (-1) ^ (m + k) * (2 * m + 1 : ℤ) *
            coeff (n - 3 * k ^ 2 - 3 * m * (m + 1) / 2) ((𝔣 2 * invOfUnit (𝔣 1 ^ 2) 1) ^ 3)
Sources
  • Sabi Biswas and Nipen Saikia, Arithmetic Properties of the Coefficients of Some Mock Theta Functions, INTEGERS 25 (2025), Article A72, Theorem 9.
Uses 2
8 fully proved14 ready to prove

The dependency explorer needs JavaScript and its graph data.

Download the complete graph data.