Iwahori Subgroups: Structure & Applications

• Iwahori subgroups are minimal parahoric subgroups attached to chambers in the Bruhat–Tits building, serving as the p-adic analogue of Borel subgroups.
• They yield canonical double coset decompositions that underpin the structure of affine flag varieties and the formulation of Iwahori–Hecke algebras.
• Their study extends to applications in pro-p groups, Whittaker models, and diverse geometric frameworks in both finite and infinite-dimensional settings.

Searching arXiv for recent and foundational papers on Iwahori subgroups and related structures. Searching arXiv for recent and foundational papers on Iwahori subgroups and related structures. Searching arXiv for recent and foundational papers on Iwahori subgroups and related structures. Iwahori subgroups are parahoric subgroups attached to chambers, or alcoves, in the Bruhat–Tits building of a reductive group over a non-Archimedean local field. In standard split integral models they can also be realized as inverse images of Borel subgroups under reduction modulo the maximal ideal. They are minimal parahorics, serve as the $p$-adic analogue of Borel subgroups, and organize Iwahori–Bruhat decompositions, affine flag varieties, Iwahori–Hecke algebras, affine Deligne–Lusztig varieties, Whittaker models, and pro-$p$ and loop-group structures (Barbasch et al., 2018, Ganapathy et al., 2021).

1. Definition and geometric realization

For a connected reductive group over a non-Archimedean local field, a parahoric subgroup is the stabilizer of a facet in the Bruhat–Tits building, and an Iwahori subgroup is the parahoric attached to a chamber. In particular, if $F$ is a facet of maximal dimension, then its parahoric $S_F$ is an Iwahori subgroup, and any two Iwahori subgroups are conjugate. In this sense, Iwahori subgroups are “minimal” among parahorics (Barbasch et al., 2018).

The intrinsic building-theoretic definition has a concrete integral model. For a split group with fixed Borel $B$, one standard Iwahori subgroup is the inverse image of $B(\mathbb F_q)$ under reduction $G(\mathcal O)\to G(\mathbb F_q)$. An opposite choice also occurs naturally: in the metaplectic setting for split $\GL_{r+1}$, the subgroup $J$ is defined as the preimage of $B^-(\mathbb F_q)$ under the reduction map $p$0 (Naprienko, 2021). In the symmetric-space setting $p$1, the Iwahori subgroup used is the stabilizer $p$2 of a chosen chamber $p$3 in the building (Broussous, 2024).

A standard geometric normalization is to work over $p$4, fix a $p$5-stable maximal torus $p$6, a $p$7-stable Borel $p$8, and a $p$9-stable Iwahori subgroup $F$0 defined by an alcove opposite to the dominant cone given by $F$1. With this choice, the parahoric of $F$2, $F$3, lies inside $F$4, and the extended affine Weyl group is identified as

$F$5

This choice controls how Newton points are described and how $F$6 acts on the building (2305.00683).

Rank-one examples make the definition especially concrete. For $F$7, the Bruhat–Tits building is a tree, maximal compact subgroups stabilize vertices, and the Iwahori subgroup stabilizes an edge; equivalently, it is the intersection of two adjacent maximal parahorics. For $F$8, the standard Iwahori in $F$9 is

$S_F$0

which reduces modulo $S_F$1 to upper-triangular matrices in $S_F$2 (Eicher, 13 May 2026).

A common misconception is to identify Iwahori subgroups only with “upper-triangular modulo $S_F$3” matrices. That description is standard and useful, but only after choosing a split integral model and a Borel. The intrinsic definition is the stabilizer of a chamber in the Bruhat–Tits building, and this formulation persists in quasi-split, loop-group, and hovel settings (Barbasch et al., 2018).

2. Double cosets, affine Weyl groups, and Hecke algebras

Fixing an Iwahori subgroup $S_F$4 produces a canonical double-coset combinatorics. For $S_F$5, the corresponding Iwahori double coset is

$S_F$6

and every double coset $S_F$7 is uniquely represented by some $S_F$8. In the affine symmetric-space setting one likewise has the Iwahori–Bruhat decomposition

$S_F$9

with $B$0 the affine Weyl group attached to the chosen apartment and chamber (2305.00683, Broussous, 2024).

This combinatorics is the basis of affine flag geometry. The affine flag variety is $B$1, and Iwahori orbits are indexed by the affine Weyl group. For $B$2, the affine flag variety is $B$3, and the Iwahori orbits are the finite-dimensional Schubert cells $B$4 and $B$5, each explicitly described as an affine space (Eicher, 13 May 2026). More generally, characteristic functions of double cosets $B$6 provide the standard basis of the Iwahori–Hecke algebra (Broussous, 2024).

The Iwahori–Hecke algebra $B$7 is the convolution algebra of compactly supported bi-$B$8-invariant functions. If $B$9 denotes the characteristic function of $B(\mathbb F_q)$0, then the generators attached to simple reflections satisfy the usual quadratic relation

$B(\mathbb F_q)$1

together with braid relations reflecting the Coxeter structure (Broussous, 2024). In another normalization, one writes generators $B(\mathbb F_q)$2 with

$B(\mathbb F_q)$3

and the corresponding braid relations (Neretin, 2021).

This Hecke algebra is the algebraic avatar of the chamber-stabilizer viewpoint. In the finite-field analogues $B(\mathbb F_q)$4 with $B(\mathbb F_q)$5, the algebra $B(\mathbb F_q)$6 is the finite Iwahori–Hecke algebra of type $B(\mathbb F_q)$7, and its direct limit

$B(\mathbb F_q)$8

is generated by countably many $B(\mathbb F_q)$9 satisfying the same quadratic and braid relations (Neretin, 2021). This suggests that the chamber-stabilizer formalism is robust under both finite-rank and infinite-rank limits.

A second misconception is to treat the affine Weyl group parametrization as merely combinatorial. In the cited works it governs not only the algebra basis and the orbit stratification of $G(\mathcal O)\to G(\mathbb F_q)$0, but also the geometry of affine Deligne–Lusztig varieties, Whittaker models, and deeper-level Hecke algebras.

3. Pro-$G(\mathcal O)\to G(\mathbb F_q)$1 Iwahori groups, filtrations, and algebraic variants

Inside an Iwahori subgroup $G(\mathcal O)\to G(\mathbb F_q)$2 sits its maximal pro-$G(\mathcal O)\to G(\mathbb F_q)$3 subgroup, the pro-$G(\mathcal O)\to G(\mathbb F_q)$4 Iwahori. For $G(\mathcal O)\to G(\mathbb F_q)$5 or $G(\mathcal O)\to G(\mathbb F_q)$6 over $G(\mathcal O)\to G(\mathbb F_q)$7, it is the subgroup of $G(\mathcal O)\to G(\mathbb F_q)$8 that is upper triangular and unipotent modulo $G(\mathcal O)\to G(\mathbb F_q)$9. In root-theoretic form, for split $\GL_{r+1}$0 one has

$\GL_{r+1}$1

and in the unramified split setting the same factorization appears as

$\GL_{r+1}$2

(Kongsgaard, 2022, Ariaz et al., 13 Jan 2026).

A distinct but related refinement is the Moy–Prasad filtration of an Iwahori subgroup. If $\GL_{r+1}$3 is the barycenter of the defining alcove, then for $\GL_{r+1}$4 the subgroup $\GL_{r+1}$5 is generated by

$\GL_{r+1}$6

This gives a decreasing filtration

$\GL_{r+1}$7

by open compact subgroups, and the Hecke algebras $\GL_{r+1}$8 admit presentations generalizing Iwahori–Matsumoto. In the unramified case, the refined “Howe–Tits presentation” depends on a Tits group lifting the Iwahori–Weyl group; in ramified cases such a Tits group may fail to exist (Ganapathy et al., 2021).

Larger parahorics containing a fixed Iwahori also produce Hecke-type algebras. If $\GL_{r+1}$9 is a parahoric, the Peter–Weyl idempotent $J$0 is defined as the sum of primitive central idempotents of those irreducible representations of $J$1 that have nonzero $J$2-fixed vectors. The associated algebra

$J$3

is a Peter–Weyl Iwahori algebra. A central theorem shows that $J$4 is Morita equivalent to the usual Iwahori–Hecke algebra $J$5, and that this equivalence preserves irreducible Hermitian and unitary modules for both the convolution $J$6-involution and the Barbasch–Ciubotaru $J$7-involution (Barbasch et al., 2018).

The pro-$J$8 case brings in homological and Iwasawa-theoretic structures. For $J$9 with $B^-(\mathbb F_q)$0, Sørensen’s spectral sequence

$B^-(\mathbb F_q)$1

collapses at $B^-(\mathbb F_q)$2, yielding explicit descriptions of $B^-(\mathbb F_q)$3 and all cup products (Kongsgaard, 2022). For split connected reductive $B^-(\mathbb F_q)$4 over an unramified extension of $B^-(\mathbb F_q)$5, the graded mod $B^-(\mathbb F_q)$6 Iwasawa algebra of a pro-$B^-(\mathbb F_q)$7 Iwahori subgroup is determined via the graded Lie algebra $B^-(\mathbb F_q)$8, and its maximal commutative quotient is a polynomial algebra (Ariaz et al., 13 Jan 2026). The same work shows that if one expects large Gelfand–Kirillov dimensions from global constructions, the action of $B^-(\mathbb F_q)$9 on the relevant graded module cannot factor through this maximal commutative quotient (Ariaz et al., 13 Jan 2026).

4. Newton stratification, Levi subgroups, and affine Deligne–Lusztig geometry

For $p$00, the Iwahori double coset $p$01 carries a Newton stratification indexed by $p$02-conjugacy classes in

$p$03

The relevant subset is

$p$04

and the corresponding affine Deligne–Lusztig variety is

$p$05

Non-emptiness of $p$06 is equivalent to $p$07 (2305.00683).

A central class of double cosets is defined by $p$08-alcove elements. If $p$09 is $p$10-stable, $p$11 is the corresponding standard Levi, and $p$12, then $p$13 is a $p$14-alcove element if

$p$15

and for every $p$16,

$p$17

Following Viehmann, $p$18 is normalized when $p$19 has minimal length in $p$20. These are the quasi-split generalization of $p$21-alcoves (2305.00683).

The main structural theorem sharpens earlier work of Görtz–He–Nie. If $p$22 is a normalized $p$23-alcove element and $p$24, then the natural map

$p$25

is a bijection, and for every corresponding class one has

$p$26

Since the embedding $p$27 is in general neither injective nor surjective, this result isolates an Iwahori locus on which those pathologies disappear (2305.00683).

The same theorem has geometric consequences. There is a canonical isomorphism of affine Deligne–Lusztig varieties

$p$28

and numerical invariants such as dimension and the number of top-dimensional irreducible components modulo the $p$29-centralizer agree. A corollary states that if $p$30 for a $p$31-alcove element, then

$p$32

Using this congruence, the paper proves Dong-Gyu Lim’s conjectural criterion for emptiness of basic affine Deligne–Lusztig varieties in terms of spherical $p$33-support and the existence of a proper $p$34-alcove structure (2305.00683).

Conceptually, this is a precise Levi reduction principle at Iwahori level. The Newton stratification of $p$35 behaves as if it were induced from the double coset $p$36 inside the Levi subgroup, and the ambient group contributes no additional Newton slopes on this locus.

5. Iwahori-spherical representations, Whittaker theory, and model spaces

The category of smooth representations generated by Iwahori-fixed vectors is controlled by the Iwahori–Hecke algebra. If $p$37 is a smooth representation of $p$38, then

$p$39

is its Iwahori component. Borel and Casselman established an equivalence between the full subcategory of representations generated by $p$40-fixed vectors and the category of $p$41-modules; in the notation of one paper,

$p$42

(Broussous, 2024).

This equivalence supports a relative distinction theory at Iwahori level. For the Galois symmetric space $p$43 attached to an unramified quadratic extension $p$44, there exists a subgroup $p$45 such that for any irreducible Iwahori-spherical representation $p$46 with Hecke module $p$47,

$p$48

Equivalently, $p$49 is $p$50-distinguished if and only if $p$51 is $p$52-distinguished. The subgroup $p$53 is generated by ratios $p$54 attached to pairs of $p$55-admissible galleries with the same terminal chamber, so the criterion translates distinction into a path-independence condition inside the building (Broussous, 2024).

Iwahori fixed vectors also support refined Whittaker theories. For metaplectic covers of $p$56, the opposite Iwahori $p$57 is used, the fixed space $p$58 has dimension $p$59, and a standard Iwahori basis $p$60 is defined by support on Bruhat–Iwahori cells $p$61. An explicit Iwahori decomposition of the maximal unipotent subgroup yields cells $p$62 indexed by valuation data and “colorings”, and these are shown to parametrize generalized Mirković–Vilonen cycles in the affine flag variety. The resulting theorem evaluates the Iwahori Whittaker integrals $p$63 as sums over colored Lusztig data, colored Gelfand–Tsetlin patterns, or colored lattice states (Naprienko, 2021).

For generalized Steinberg representations, the Iwahori-fixed space is one-dimensional. The Iwahori–Hecke algebra acts on it through the sign character, and the associated Whittaker function $p$64, normalized by $p$65, is determined explicitly: $p$66 This generalizes earlier $p$67 results to arbitrary split reductive groups (Karameris, 2024).

Model spaces provide another application. For special orthogonal groups, the Iwahori component of a Bessel model space is computed and identified with an explicit projective module over the Iwahori Hecke algebra (Chan et al., 2018). This places Bessel models alongside Whittaker and distinction problems as examples where the Iwahori component is not an auxiliary construction but the primary algebraic object.

6. Infinite-rank, Kac–Moody, and affine-flag generalizations

Iwahori-type structures persist well beyond finite-dimensional reductive $p$68-adic groups. For the group $p$69 of invertible infinite matrices with only finitely many nonzero entries below the diagonal, the compact open subgroup

$p$70

plays the role of an Iwahori subgroup, and

$p$71

is the direct limit of the finite Iwahori–Hecke algebras. Vershik–Kerov classified its indecomposable positive traces, and Neretin realized the corresponding GNS representations as irreducible representations of the double $p$72 and of $p$73 (Neretin, 2021).

In the affine current-group setting, the Iwahori subgroup becomes

$p$74

with Lie algebra

$p$75

The algebra of functions $p$76 is studied via categories of graded bounded $p$77-modules. These categories admit a stratified structure, their standard and costandard objects are identified with generalized Weyl modules, and the characters of proper standard and proper costandard objects are expressed through specialized nonsymmetric Macdonald polynomials. The associated Peter–Weyl theorem states that $p$78 decomposes as a sum of $p$79, providing a loop-group analogue of the classical Peter–Weyl theorem (Feigin et al., 2023).

For almost split Kac–Moody groups over local fields, the Bruhat–Tits building is replaced by a hovel. The fixer $p$80 of a chamber $p$81 in the standard apartment is the Iwahori subgroup analogue, and one has a positive semigroup decomposition

$p$82

The corresponding Iwahori–Hecke algebra is defined by $p$83-bi-invariant functions with finite support in $p$84, admits a Bernstein–Lusztig type presentation, and in the affine case contains Cherednik’s double affine Hecke algebra (Bardy-Panse et al., 2014). This shows that the chamber-fixer formalism extends from reductive groups to hovels and Kac–Moody geometry, although the global decomposition must be restricted to a semigroup because two chambers are not always contained in one apartment (Bardy-Panse et al., 2014).

Recent affine-flag geometry for $p$85 illustrates how far these chamber-based structures can be refined. Starting from the standard Iwahori subgroup $p$86, one may remove successive affine root subgroups and obtain a chain

$p$87

together with $p$88. Each finite-dimensional Schubert cell in the affine flag variety decomposes into orbits for these subgroups, and at each step an orbit either stays intact or splits into an open orbit and a hyperplane orbit in explicit affine coordinates (Eicher, 13 May 2026). This suggests that, even in rank one, small-codimension subgroups of an Iwahori can generate a rich secondary stratification inside the standard Schubert decomposition.

Across these settings, the recurrent pattern is stable: an Iwahori subgroup is the subgroup attached to a chamber, its double cosets encode a Weyl-type combinatorics, and its fixed vectors or bi-invariant functions produce the algebraic structures that control representation theory, geometry, and homological invariants.