betaFeedback

WorksMathematics

An analogue of the Erdos-Kac theorem for signed Iwasawa invariants of modular forms

Abstract

Erdos and Kac proved that the number of distinct prime factors of an integer has a Gaussian limiting distribution. We prove an analogue for the signed Iwasawa invariants of non-ordinary weight-two modular forms. Under the hypotheses of this paper, we construct infinitely many forms congruent to a given form gg that realize every common nonnegative increase of its two signed λ\lambda-invariants. We count the levels realizing each increase. In the resulting family, the common increase is asymptotically Gaussian, with mean and variance δΠlog⁡log⁡X+Og(1)\delta_\Pi\log\log X+O_g(1) for a positive constant δΠ\delta_\Pi.

Introduction

Background and motivation

A central theme in arithmetic geometry is to understand how arithmetic invariants vary in infinite families. Iwasawa theory encodes the growth of class groups, Selmer groups and related arithmetic objects in pp-adic towers in terms of a small collection of algebraic invariants.

The theory originated in Iwasawa’s study of ideal class groups in infinite towers of number fields. Let pp be a prime number and K∞/KK_{\infty}/K be the cyclotomic Zp\mathbb{Z}_p-extension. Let KnK_n be the unique intermediate field satisfying [Kn:K]=pn[K_n:K]=p^n. If AnA_n denotes the pp-primary part of the ideal class group of KnK_n, then the inverse limit of the groups AnA_n, with respect to the norm maps, is a module over the Iwasawa algebra Λ=Zp[[T]]\Lambda=\mathbb{Z}_p[[T]]. Writing #An=pen\#A_n=p^{e_n}, Iwasawa proved that there are integers μ,λ≥0\mu,\lambda\geq0 and ν\nu such that

en=μpn+λn+ν(1)e_n=\mu p^n+\lambda n+\nu \tag*{(1)}

for all sufficiently large nn; see [12]. This essentially leverages the structure theory of finitely generated and torsion modules over Λ\Lambda. The integers μ\mu and λ\lambda are the Iwasawa invariants of the tower.

Mazur extended this point of view to elliptic curves [20]. Let E/QE/\mathbb{Q} be an elliptic curve and let Q∞/Q\mathbb{Q}_{\infty}/\mathbb{Q} be the cyclotomic Zp\mathbb{Z}_p-extension. The role of the class group is played by the pp-primary Selmer group Sel⁡p∞(E/Q∞)\operatorname{Sel}_{p^\infty}(E/\mathbb{Q}_{\infty}). For every finite extension F/QF/\mathbb{Q} contained in Q∞\mathbb{Q}_{\infty}, the Selmer group fits into the fundamental exact sequence

0⟶E(F)⊗ZQp/Zp⟶Sel⁡p∞(E/F)⟶X(E/F)[p∞]⟶0.0\longrightarrow E(F)\otimes_{\mathbb{Z}}\mathbb{Q}_p/\mathbb{Z}_p\longrightarrow\operatorname{Sel}_{p^\infty}(E/F)\longrightarrow X(E/F)[p^\infty]\longrightarrow0.

When its Pontryagin dual is torsion over the cyclotomic Iwasawa algebra, it has well defined invariants μp(E)\mu_p(E) and λp(E)\lambda_p(E) which arise from the structure theory of Λ\Lambda-modules. These invariants describe the growth of the Selmer groups in the finite layers of Q∞/Q\mathbb{Q}_{\infty}/\mathbb{Q} and are closely related to the growth of Mordell–Weil groups and Tate–Shafarevich groups.

The same construction applies to the Galois representation attached to a modular form. Let hh be a normalized newform and let VhV_h be its pp-adic Galois representation. A choice of stable lattice Th⊂VhT_h\subset V_h gives a discrete representation Ah=Vh/ThA_h=V_h/T_h, and one defines a Selmer group over Q∞\mathbb{Q}_{\infty} by imposing suitable local conditions on AhA_h. In the ordinary case, the local condition at pp is obtained from the ordinary filtration of VhV_h. Kato’s Euler-system shows that the Pontryagin dual of the resulting Selmer group is a finitely generated torsion module over the Iwasawa algebra [13]. Its characteristic power series therefore defines the associated μ\mu- and λ\lambda-invariants.

At a non-ordinary prime there is no ordinary filtration giving the required local condition. For elliptic curves with supersingular reduction, Kobayashi introduced the plus and minus Selmer groups [16], while Pollack constructed the corresponding bounded pp-adic LL-functions when ap=0a_p=0 [21]. Signed Coleman maps and signed Selmer groups for modular forms were subsequently developed by Lei, Lei–Loeffler–Zerbes and Sprung [17, 18, 19, 26, 27]. In the weight-two setting considered here, these constructions give two signed Selmer groups, indexed by ♮∈{♯,♭}\natural\in\{\sharp,\flat\}, and hence two pairs of Iwasawa invariants (μ♮(h),λ♮(h))(\mu^\natural(h),\lambda^\natural(h)) and (μ♭(h),λ♭(h))(\mu^\flat(h),\lambda^\flat(h)).

A fundamental question is to understand how Iwasawa invariants vary with respect to congruences of Galois representations. Greenberg and Vatsal proved that, under suitable hypotheses, the vanishing of the Iwasawa μ\mu-invariant is preserved under congruences between ordinary elliptic curves. When the μ\mu-invariants vanish, they expressed the difference of the corresponding λ\lambda-invariants as a sum of explicit local correction terms [9]. Emerton–Pollack–Weston extended this study to ordinary modular forms in Hida families [8]. They showed that if the algebraic or analytic μ\mu-invariant vanishes for one form, it vanishes throughout the family. They also described how the λ\lambda-invariants vary between branches and proved that they are constant along each branch. The algebraic and analytic variation formulae agree; consequently, the main conjecture for one modular form implies it for the other modular forms that are congruent to it. Kim extended the Greenberg–Vatsal comparison to the plus and minus Selmer groups of congruent supersingular elliptic curves [14, 15]. Pollack–Weston used congruences to prove Kida-type formulas, including one for supersingular elliptic curves [22]. Hatley and Lei then treated congruent non-ordinary modular forms of the same weight. For each sign, they showed that the μ\mu-invariant vanishes for one form if and only if it vanishes for the other; when it vanishes, the difference of the λ\lambda-invariants is determined by local terms at primes away from pp [10].

In [23], we combined the ordinary congruence formula with level raising results of Diamond and Taylor [5]. Starting from an ordinary weight-two newform gg with μp(g)=0\mu_p(g)=0, we showed that, for every n≥0n \ge0, there are infinitely many newforms ff congruent to gg with μp(f)=0\mu_p(f)=0 and λp(f)=λp(g)+n\lambda_p(f)=\lambda_p(g)+n. The present paper extends this construction to non-ordinary modular forms and develops a probabilistic model for the variation of their signed λ\lambda-invariants as the tame level varies. In the non-ordinary setting there are two signed λ\lambda-invariants. We show that level raising increases them by the same amount, so their difference remains unchanged. We then study the distribution of the common increment as the tame level varies.

Main results

Let p≥5p \ge5 be a prime number. Let

g=∑n≥1an(g)qn∈S2(Γ0(Ng))g=\sum_{n\ge1}a_n(g)q^n\in S_2\left(\Gamma_0(N_g)\right)

be a normalized newform of trivial nebentypus, where p∤Ngp\nmid N_g, and suppose that gg is non-ordinary at pp. Write GQ=Gal⁡(Q‾/Q)G_{\mathbb{Q}}=\operatorname{Gal}(\overline{\mathbb{Q}}/\mathbb{Q}) and let

ρ‾g:GQ⟶GL⁡2(k)\overline{\rho}_g:G_{\mathbb{Q}}\longrightarrow\operatorname{GL}_2(k)

be the residual Galois representation attached to gg at a fixed pp-adic place of its coefficient field, where kk is the corresponding residue field. Let χ‾p\overline{\chi}_p denote the mod-pp cyclotomic character, and let μpr\mu_{p^r} denote the group of prp^r-th roots of unity. Put

Hg=ρˉg(GQ)⊆GL⁡2(k).H_g=\bar{\rho}_g(G_{\mathbb{Q}})\subseteq\operatorname{GL}_2(\mathbf{k}).

Let LgL_g be the fixed field of ker⁡(ρˉg)\ker(\bar{\rho}_g). Since gg has weight two and trivial nebentypus, det⁡ρˉg=χˉp\det\bar{\rho}_g=\bar{\chi}_p. In particular, det⁡(Hg)=Fp×\det(H_g)=\mathbb{F}_p^\times and Q(μp)⊆Lg\mathbb{Q}(\mu_p)\subseteq L_g. For a prime q∤Ngpq\nmid N_gp, we write Frob⁡q\operatorname{Frob}_q for an arithmetic Frobenius element at qq.

For h∈Hgh\in H_g, put d(h)=det⁡(h)d(h)=\det(h) and define the conjugacy-stable subsets

Cg+={h∈Hg:d(h)∉{±1},Tr⁡(h)=1+d(h)},C_g^+=\{h\in H_g:d(h)\notin\{\pm1\},\operatorname{Tr}(h)=1+d(h)\},
Cg−={h∈Hg:d(h)∉{±1},Tr⁡(h)=−(1+d(h))}.C_g^-=\{h\in H_g:d(h)\notin\{\pm1\},\operatorname{Tr}(h)=-(1+d(h))\}.

Thus the characteristic polynomial of an element of Cg+C_g^+ is (X−1)(X−d(h))(X-1)(X-d(h)), whereas that of an element of Cg−C_g^- is (X+1)(X+d(h))(X+1)(X+d(h)).

For a non-ordinary modular form hh and ♮∈{♯,♭}\natural\in\{\sharp,\flat\}, let X♮(h)X^\natural(h) denote the Pontryagin dual of its ♮\natural-signed Selmer group over the cyclotomic Zp\mathbb{Z}_p-extension of Q\mathbb{Q}. When X♮(h)X^\natural(h) is torsion, we denote its Iwasawa invariants by μ♮(h)\mu^\natural(h) and λ♮(h)\lambda^\natural(h). We impose the following hypotheses.

(H1) The representation ρˉg\bar{\rho}_g is absolutely irreducible, both Cg+C_g^+ and Cg−C_g^- are nonempty, and

Lg∩Q(μp2)=Q(μp).L_g\cap\mathbb{Q}(\mu_{p^2})=\mathbb{Q}(\mu_p).

(H2) The integer NgN_g is the prime-to-pp Artin conductor of ρˉg\bar{\rho}_g.

(H3) For each ♮∈{♯,♭}\natural\in\{\sharp,\flat\}, the signed Selmer dual X♮(g)X^\natural(g) is torsion over the cyclotomic Iwasawa algebra and μ♮(g)=0\mu^\natural(g)=0.

The residual-image hypothesis (H1) is weaker than requiring k=Fp\mathbf{k}=\mathbb{F}_p and Hg=GL⁡2(Fp)H_g=\operatorname{GL}_2(\mathbb{F}_p). In particular, Proposition 3.3 shows that it holds whenever SL⁡2(k)⊆Hg\operatorname{SL}_2(\mathbf{k})\subseteq H_g.

Let Πg\Pi_g be the set of primes q∤Ngpq\nmid N_gp such that ρˉg(Frob⁡q)∈Cg+\bar{\rho}_g(\operatorname{Frob}_q)\in C_g^+ and qp−1≢1(modp2)q^{p-1}\not\equiv1\pmod{p^2}. Let Ωg\Omega_g be the set of primes ℓ∤Ngp\ell\nmid N_gp such that ρˉg(Frob⁡ℓ)∈Cg−\bar{\rho}_g(\operatorname{Frob}_\ell)\in C_g^-. Write δΠ\delta_\Pi and δΩ\delta_\Omega for the natural densities of Πg\Pi_g and Ωg\Omega_g, respectively. These densities exist by the Chebotarev density theorem, and Proposition 3.4 gives

References

  1. [1]P. N. Balister and S. Howson. “Note on Nakayama’s lemma for compact Λ-modules”. In: Asian Journal of Mathematics 1.2 (1997), pp. 224–229. DOI: 10.4310/AJM.1997.v1.n2.a2.DOI
  2. [2]P. Billingsley. Probability and Measure. 3rd ed. Wiley Series in Probability and Mathematical Statistics. New York: John Wiley & Sons, Inc., 1995. ISBN: 0-471-00710-2.
  3. [3]L. Berger, H. Li, and H. J. Zhu. “Construction of some families of 2-dimensional crystalline representations”. In: Mathematische Annalen 329.2 (2004), pp. 365–377. DOI: 10.1007/s00208-004-0529-y.DOI
  4. [4]H. Carayol. “Sur les représentations galoisiennes modulo ℓ attachées aux formes modulaires”. In: Duke Mathematical Journal 59.3 (1989), pp. 785–801. DOI: 10.1215/S0012-7094-89-05937-1.DOI
  5. [5]F. Diamond and R. Taylor. “Non-optimal levels of mod l modular representations”. In: Inventiones Mathematicae 115.3 (1994), pp. 435–462. DOI: 10.1007/BF01231768.DOI
  6. [6]B. Edixhoven. “The weight in Serre’s conjectures on modular forms”. In: Inventiones Mathematicae 109.3 (1992), pp. 563–594. DOI: 10.1007/BF01232041.DOI
  7. [7]P. Erdős and M. Kac. “The Gaussian law of errors in the theory of additive number theoretic functions”. In: American Journal of Mathematics 62.1 (1940), pp. 738–742. DOI: 10.2307/2371483.DOI
  8. [8]M. Emerton, R. Pollack, and T. Weston. “Variation of Iwasawa invariants in Hida families”. In: Inventiones Mathematicae 163.3 (2006), pp. 523–580. DOI: 10.1007/s00222-005-0467-7.DOI
  9. [9]R. Greenberg and V. Vatsal. “On the Iwasawa invariants of elliptic curves”. In: Inventiones Mathematicae 142.1 (2000), pp. 17–63. DOI: 10.1007/s002220000080.DOI
  10. [10]J. Hatley and A. Lei. “Arithmetic properties of signed Selmer groups at non-ordinary primes”. In: Annales de l’Institut Fourier 69.3 (2019), pp. 1259–1294. DOI: 10.5802/aif.3270.DOI
  11. [11]G. H. Hardy and S. Ramanujan. “The normal number of prime factors of a number n”. In: Quarterly Journal of Pure and Applied Mathematics 48 (1917), pp. 76–92.
  12. [12]K. Iwasawa. “On Γ-extensions of algebraic number fields”. In: Bulletin of the American Mathematical Society 65.4 (1959), pp. 183–226. DOI: 10.1090/S0002-9904-1959-10317-7.DOI
  13. [13]K. Kato. “p-adic Hodge theory and values of zeta functions of modular forms”. In: Astérisque 295 (2004), pp. 117–290.
  14. [14]B. D. Kim. “The Iwasawa invariants of the plus/minus Selmer groups”. In: Asian Journal of Mathematics 13.2 (2009), pp. 181–190. DOI: 10.4310/AJM.2009.v13.n2.a2.DOI
  15. [15]B. D. Kim. “The plus/minus Selmer groups for supersingular primes”. In: Journal of the Australian Mathematical Society 95.2 (2013), pp. 189–200. DOI: 10.1017/S1446788713000165.DOI
  16. [16]S. Kobayashi. “Iwasawa theory for elliptic curves at supersingular primes”. In: Inventiones Mathematicae 152.1 (2003), pp. 1–36. DOI: 10.1007/s00222-002-0265-4.DOI
  17. [17]A. Lei. “Iwasawa theory for modular forms at supersingular primes”. In: Compositio Mathematica 147.3 (2011), pp. 803–838. DOI: 10.1112/S0010437X10005130.DOI
  18. [18]A. Lei, D. Loeffler, and S. L. Zerbes. “Wach modules and Iwasawa theory for modular forms”. In: Asian Journal of Mathematics 14.4 (2010), pp. 475–528. DOI: 10.4310/AJM.2010.v14.n4.a2.DOI
  19. [19]A. Lei, D. Loeffler, and S. L. Zerbes. “Coleman maps and the p-adic regulator”. In: Algebra & Number Theory 5.8 (2011), pp. 1095–1131. DOI: 10.2140/ant.2011.5.1095.DOI
  20. [20]B. Mazur. “Rational points of abelian varieties with values in towers of number fields”. In: Inventiones Mathematicae 18.3–4 (1972), pp. 183–266. DOI: 10.1007/BF01389815.DOI
  21. [21]R. Pollack. “On the p-adic L-function of a modular form at a supersingular prime”. In: Duke Mathematical Journal 118.3 (2003), pp. 523–558. DOI: 10.1215/S0012-7094-03-11835-9.DOI
  22. [22]R. Pollack and T. Weston. “Kida’s formula and congruences”. In: Documenta Mathematica Extra Vol. (2006), pp. 615–630. DOI: 10.4171/DMS/4/18.
  23. [23]A. Ray. “Constructing Galois representations with prescribed Iwasawa λ-invariant”. In: Bulletin of the London Mathematical Society 56.5 (2024), pp. 1624–1642. DOI: 10.1112/blms.13015.DOI
  24. [24][Rib94] K. A. Ribet. “Report on mod ℓ representations of Gal(̅ℚ/ℚ)”. In: Motives (Seattle, WA, 1991). Vol. 55, Part 2. Proceedings of Symposia in Pure Mathematics. Providence, RI: American Mathematical Society, 1994, pp. 639–676.
  25. [25][Ser81] J.-P. Serre. “Quelques applications du théorème de densité de Chebotarev”. In: Publications Mathématiques de l’IHÉS 54 (1981), pp. 123–201. DOI: 10.1007/BF02698692.DOI
  26. [26][Spr12] F. E. I. Sprung. “Iwasawa theory for elliptic curves at supersingular primes: a pair of main conjectures”. In: Journal of Number Theory 132.7 (2012), pp. 1483–1506. DOI: 10.1016/j.jnt.2011.11.003.DOI
  27. [27][Spr17] F. Sprung. “On pairs of p-adic L-functions for weight-two modular forms”. In: Algebra & Number Theory 11.4 (2017), pp. 885–928. DOI: 10.2140/ant.2017.11.885.DOI
  28. [28][Ten15] G. Tenenbaum. Introduction to Analytic and Probabilistic Number Theory. Third. Vol. 163. Graduate Studies in Mathematics. Providence, RI: American Mathematical Society, 2015.DOI
  29. [29][Tur34] P. Turán. “On a theorem of Hardy and Ramanujan”. In: Journal of the London Mathematical Society s1-9.4 (1934), pp. 274–276. DOI: 10.1112/jlms/s1-9.4.274.DOI
  30. [30][Zyw23] D. Zywina. “Modular forms and some cases of the inverse Galois problem”. In: Canadian Mathematical Bulletin 66.2 (2023), pp. 568–586. DOI: 10.4153/S0008439522000534.DOI

Paper details

vlnc:W_vasr3k3h
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents