Abstract
The affine Temperley–Lieb algebra a TLN(β) is an infinite-dimensional algebra over \(\mathbb {C}\) parametrized by a number \(\beta \in \mathbb {C}\) and an integer \(N\in \mathbb {N}\). It naturally acts on \((\mathbb {C}^{2})^{\otimes N}\) to produce a family of representations labeled by an additional parameter \(z\in \mathbb C^{\times }\). The structure of these representations, which were first introduced by Pasquier and Saleur (Nucl. Phys., 330, 523 1990) in their study of spin chains, is here made explicit. They share their composition factors with the cellular aTLN(β)-modules of Graham and Lehrer (Enseign. Math., 44, 173 1998), but differ from the latter by the direction of about half of the arrows of their Loewy diagrams. The proof of this statement uses a morphism introduced by Morin-Duchesne and Saint-Aubin (J. Phys. A, 46, 285207 2013) as well as new maps that intertwine various aTLN(β)-actions on the periodic chain and generalize applications studied by Deguchi et al. (J. Stat. Phys., 102, 701 2001) and after by Morin-Duchesne and Saint-Aubin (J. Phys. A, 46, 494013 2013).
Similar content being viewed by others
Data Availability
Data sharing is not applicable to this article as no datasets were generated or analysed during the study.
Abbreviations
- a T L N(β):
-
the affine Temperley–Lieb algebra (Section 3.1)
- T L N(β):
-
the (usual) Temperley–Lieb algebra (Section 3.1)
- β :
-
complex parameter of the algebras a TLN(β) and TLN(β) (β = −q − q− 1)
- λ,λ N,Λ,ΛN,Δ,ΔN :
-
sets of pairs (d,z) parametrizing aTLN-modules (Section 2)
- \(\preceq , \unlhd \) :
-
partial orders on λ and Δ respectively (Section 2)
- problematic pairs:
-
the pairs (0,±q) when q + q− 1 = 0 and N is even (Section 3.1)
- (m,n)-diagrams:
-
elements of diagrammatic basis of a TLN(β) (Section 3.2)
- \(\mathfrak {B}_{m,n}\) :
-
diagrammatic basis of the vector space of (m,n)-diagrams (Section 3.1)
- |v|:
-
rank of the diagram v (Section 3.1)
- W N;d,z :
-
cellular modules over the algebra a TLN(β) (Section 3.1)
- 〈 , 〉N;d,z :
-
\(\mathbb {C}\)-valued bilinear form on \(\mathsf {W}_{N;d,z}\times \mathsf {W}_{N;d,z^{-1}}\) (Section 3.3)
- L N;d,z :
-
simple a TLN(β)-module associated to the pair (d,z) ∈ΛN (Sections 2 and 3.1)
- 𝜃 (d,z);(t,x) :
-
Graham–Lehrer morphisms between cellular modules (Section 3.3)
- H XXZ :
-
Hamiltonian acting on the periodic chain (Section 3.2)
- z :
-
twist parameter \(\in \mathbb {C}^{\times }\) (Section 3.2)
- S z :
-
z-component of the spin operator (Section 3.2)
- \({\mathsf {X}}_{N;z}^{\pm }\) :
-
periodic chains seen as modules over aTLN(β) (Section 3.2) or \({{\mathscr{L}}\mathcal {U}}_{q^{\pm 1}}\) (Section 4.2)
- \({\mathsf {X}}_{N;d,z}^{\pm }\) :
-
aTLN(β)-submodule of \({\mathsf {X}}_{N;z}^{\pm }\) seen as a eigenspace of Sz with eigenvalue d (Section 3.2)
- i N;d,z :
-
aTLN(β)-morphism \(\mathsf {W}_{N;d,z}\to \mathsf {X}_{N;d,z}^{+}\) (Section 3.2)
- G P N;d,z :
-
generic part of the cellular module WN;d,z (Section 3.5)
- \(\mathfrak {B}^{\mathsf {X}_{N}}\) :
-
spin basis for \(\mathsf {X}_{N}{}=(\mathbb C^{2})^{\otimes N}\) (Section 3.2)
- ⋆- and ∘-dualities:
-
functors a TLN-mod →aTLN-mod (Section 3.4)
- ℓ :
-
smallest positive integer (if it exists) such that q2ℓ = 1 (Section 2)
- \({{\mathscr{L}}\mathcal {U}}_q\) :
-
Lusztig’s quantum group (Section 4.1)
- L q(i),Δq(i),P q(i):
-
simple, Weyl and projective modules over \({{\mathscr{L}}\mathcal {U}}_q\) (Section 4.1)
- \({E}_{(t,x);(d,z)}^{\pm } \text { and } {F}_{(d,z);(t,x)}^{\pm }\) :
-
aTLN(β)-morphisms between eigenspaces \({\mathsf {X}}_{N;d,z}^{\pm }\) and \({\mathsf {X}}_{N;t,x}^{\pm }\) (Section 4.2)
- μ (n) and ν (n) :
-
maps \({\mathsf {X}}_{N;d+2n\ell ,zq^{n\ell }}^{+}\rightarrow {\mathsf {X}}_{N;d,z}^{+}\) and \({\mathsf {X}}_{N;d,z}^{+}\rightarrow {\mathsf {X}}_{N;d+2n\ell ,zq^{n\ell }}^{+}\) (Section 4.2)
References
Alcaraz, F., Grimm, U., Rittenberg, V.: The XXZ Heisenberg chain, conformal invariance and the operator content of c < 1 systems. Nucl. Phys. B 316, 735–768 (1989)
Andersen, H.: Simple modules for Temperley–Lieb algebras and related algebras. J. Algebra 520, 276–308 (2019)
Andersen, H., Polo, P., Wen, K.: Representations of quantum algebras. Inventiones mathematicae 120(1), 409–410 (1995)
Andersen, H., Stroppel, C., Tubbenhauer, D.: Semisimplicity of Hecke and (walled) Brauer algebras. J. Aust. Math. Soc. 103(1), 1–44 (2017)
Andersen, H., Tubbenhauer, D.: Diagram categories for Uq-tilting modules at roots of unity. Transform. Groups 1, 29–89 (2017). arXiv:1409.2799
Belletête, J., Gainutdinov, A., Jacobsen, J., Saleur, H., Tavares, T.: Topological defects in lattice models and affine Temperley-Lieb algebras. arXiv:1811.02551 (2018)
Belletête, J., Gainutdinov, A., Jacobsen, J., Saleur, H., Tavares, T.: Topological defects in the affine Temperley-Lieb algebra: The critical cases in preparation (2022)
Belletête, J., Saint-Aubin, Y.: On the computation of fusion over the affine Temperley-Lieb algebra. Nucl. Phys. B 937, 333–370 (2018). arXiv:1802.03575
Bushlanov, P., Feigin, B., Gainutdinov, A., Tipunin, I.: Lusztig limit of quantum sl(2) at root of unity and fusion of (1, p) Virasoro logarithmic minimal models. Nuclear Phys. B 818(3), 179–195 (2009)
Chari, V., Pressley, A.: A guide to quantum groups. Cambridge University Press, Cambridge (1995)
Cox, A.: Ext1 for Weyl modules for q-GL(2,k). In: Mathematical Proceedings of the Cambridge Philosophical Society, vol. 124, pp 231–251. Cambridge University Press (1998)
Deguchi, T., Fabricius, K., McCoy, B.: The sl2 loop algebra symmetry of the six-vertex model at roots of unity. J. Stat. Phys. 102, 701–736 (2001). arXiv:cond-mat/9912141
Feigin, B., Fuks, D.: Invariant skew-symmetric differential operators on the line and Verma modules over the Virasoro algebra. Funct. Anal. Applic. 16 (2), 114–126 (1982)
Gainutdinov, A., Read, N., Saleur, H.: Bimodule structure in the periodic gℓ(1|1) spin chain. Nucl. Phys. B 871, 289–329 (2013). arXiv:1112.3407
Gainutdinov, A., Read, N., Saleur, H.: Associative algebraic approach to logarithmic CFT in the bulk: The continuum limit of the gℓ(1|1) periodic spin chain, Howe duality and the interchiral algebra. Comm. Math. Phys. 341, 35–103 (2016). arXiv:1207.6334
Graham, J., Lehrer, G.: Cellular algebras. Inv. Math. 123, 1–34 (1996)
Graham, J., Lehrer, G.: The representation theory of affine Temperley-Lieb algebras. Enseign. Math. 44, 173–218 (1998)
Graham, J., Lehrer, G.: Cellular algebras and diagram algebras in representation theory. In: Representation Theory of Algebraic Groups and Quantum Groups, pp. 141–173 (2004)
Green, R.: On representations of affine Temperley-Lieb algebras. In: Conference Proceedings, Canadian Mathematical Society, vol. 24. American Mathematical Society (1998)
Jantzen, J.: Lectures on Quantum Groups, Volume 6 of Graduate Studies in Mathematics. American Mathematical Society, Providence (1995)
Jimbo, M.: A q-difference analogue of \({U}(\frak g)\) and the Yang-Baxter equation. Lett. Math. Phys. 10, 63–69 (1985)
Jimbo, M.: A q-analogue of \({U}(\frak {gl}({N}+1))\), Hecke algebra, and the Yang-Baxter equation. Lett. Math. Phys. 11, 247–252 (1986)
Klimyk, A, Schmüdgen, K.: Quantum Groups and Their Representations. Texts and Monographs in Physics. Springer, Berlin (1997)
König, S., Xi, C.: Affine cellular algebras. Adv. Math. 229, 139–182 (2012)
Koo, W., Saleur, H.: Representations of the Virasoro algebra from lattice models. Nucl. Phys. B 426(3), 459–504 (1994)
Lentner, S.: The unrolled quantum group inside Lusztig’s quantum group of divided powers. Lett. Math. Phys. 109(7), 1665–1682 (2019)
Lusztig, G.: Introduction to Quantum Groups. Springer Science & Business Media (2010)
Martin, P.: On Schur-Weyl duality, An Hecke algebras and quantum sl(N) on \(\otimes ^{n+1}\mathbb {C}^{N}\). Int. J. Modern Phys. 07, 645–673 (1992)
Martin, P., Saleur, H.: On an algebraic approach to higher dimensional statistical mechanics. Comm. Math. Phys. 158, 1555–190 (1993). arXiv:hep-th/9208061
Martin, P., Saleur, H.: The blob algebra and the periodic Temperley-Lieb algebra. Lett. Math. Phys. 30(3), 189–206 (1994)
Morin-Duchesne, A., Rasmussen, J., Ridout, D.: Boundary algebras and Kac modules for logarithmic minimal models. Nucl. Phys. B 899, 677–769 (2015)
Morin-Duchesne, A., Saint-Aubin, Y.: A homomorphism between link and XXZ modules over the periodic Temperley-Lieb algebra. J. Phys. A 46, 285207 (2013). arXiv:1203.4996
Morin-Duchesne, A., Saint-Aubin, Y.: Jordan cells of periodic loop models. J. Phys. A 46, 494013 (2013). arXiv:1302.5483
Pasquier, V, Saleur, H: Common structures between finite systems and conformal field theories through quantum groups. Nucl. Phys. B 330, 523–553 (1990)
Ridout, D., Saint-Aubin, Y.: Standard modules, induction and the structure of the Temperley-Lieb algebra. Adv. Theor. Math. Phys. 18(5), 957–1041 (2014)
Stum, B.L., Quirós, A.: On quantum integers and rationals. In: Trends in Number Theory, Volume 649 of Contemporary Mathematics. arXiv:1310.8143, pp 107–131. American Mathematical Society, Providence (2015)
Acknowledgements
We thank Jonathan Belletête and Alexi Morin-Duchesne for their interest in the project and their useful comments. We are particularly indebted to Jonathan Belletête for telling us the result generalizing Lemma 5.2 with its proof. Useful discussions with Ibrahim Assem and Loïc Poulain d’Andecy are gratefully acknowledged. The reviewer is also thanked for their useful and enthusiastic comments. TP held scholarships from the Natural Sciences and Engineering Research Council of Canada and the Fonds de recherche Nature et technologies (Québec) while this work was done. TP also received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 945322. YSA holds a grant from the Natural Sciences and Engineering Research Council of Canada. This support is gratefully acknowledged.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of Interests
There are no conflicts of interest associated with this work.
Additional information
Presented by: Alistair Savage
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix A: Some Properties of q-numbers
This appendix recalls the definitions of q-numbers, factorials and binomials while stating some of their properties.
The q-numbers [n]q and q-factorial [n]q! are, for \(n\in \mathbb Z\),
The q-binomial coefficient, for \(m,n\in \mathbb Z_{\geq 0}\), is defined by
if m ≥ n and is 0 otherwise. Here are two propositions that generalise properties of the (usual) binomial coefficients.
Proposition A.1 (Pascal identity, see e.g. [20, p. 6])
Let \(q\in \mathbb C^{\times }\) and \(m,n\in \mathbb N\) with n ≤ m − 1. Then,
Proposition A.2 (q-binomial theorem, see e.g. [27, p. 9])
Let \(q\in \mathbb C^{\times }\) and \(m\in \mathbb Z_{\geq 0}\). Then,
The remaining results describe the behavior of q-numbers when q is a root of unity. Strictly speaking, the left-hand side of the next identity should be understood within a limit \(\lim _{q\to q_{c}}\) where qc is the root of unity considered.
Proposition A.3 (Lucas q-theorem, see e.g. [36, Proposition 2.13])
Let \(q\in \mathbb C^{\times }\) be such that q2 be a primitive ℓ-th root of unity. Write \(m,n\in \mathbb Z_{\geq 0}\) as m = m1ℓ + m2 and n = n1ℓ + n2 with m1,n1 ≥ 0 and 0 ≤ m2,n2 < ℓ. Then,
with \(\big (\begin {smallmatrix}m_{1}\\ n_{1}\end {smallmatrix}\big )=0\) if m1 < n1. In particular \(\big [\begin {smallmatrix}m\\ n\end {smallmatrix}\big ]_{q}=0\) if and only if m1 < n1 or m2 < n2.
As above, the next (easily shown) identities are to be understood multiplied by other terms and within limit signs.
Proposition A.4
If \(q\in \mathbb C^{\times }\) is such that q2 is a primitive ℓ-th root of unity and if \(m,n\in \mathbb Z\), then \(q^{\ell ^{2}} = (-1)^{\ell +1}q^{\ell }\),
Appendix B: Representation Theory of Lusztig’s Extension \(\mathcal LU_q\)
This appendix is devoted to the proof of Theorem 4.6. This theorem is purely representation-theoretic in nature and is of interest in its own right. Throughout \(q \in \mathbb {C}^{\times }\) is such that q2 is a ℓ th-primitive root of unity with ℓ ≥ 2. We start by stating some basic facts. The first one can be easily deduced from the definition of the divided powers of Ures. The last one is well-known (see e.g. [10, Proposition 9.3.5]) but a proof is given for completeness.
Lemma B.1
Let \(k,n\in \mathbb {Z}_{\geq 0}\). Then, in Ures,
Lemma B.2
[10, p. 297-299] Let \(n,c \in \mathbb {Z}\) with n ≥ 0. Then, the element \({\begin {bmatrix}{K;c}\\{n}\end {bmatrix}}_{t}\) of the rational form \(U_{t}\mathfrak {sl}_{2}\) defined by
also belongs to Ures. Moreover the subalgebra generated by \(\{K^{\pm 1}, {\begin {bmatrix}{K;c}\\{n}\end {bmatrix}}_{t} | n,c\in \mathbb {Z}\) \(\text {with } n\geq 0\}\subseteq U_{\text {res}}\) is commutative.
Proposition B.3
In \(\mathcal LU_q\), the following identities hold: Eℓ = Fℓ = K2ℓ −id = 0.
Proof
The first two assertions follow from the specialization of the relations [ℓ]t!E(ℓ) = Eℓ and [ℓ]t!F(ℓ) = Fℓ of Ures. For the last one, note that K2ℓ −id can be factorised as \({\prod }_{m=0}^{\ell -1} (K^{2}-q^{2m}\text {id})\). The definition of \({\begin {bmatrix}{K;0}\\{\ell }\end {bmatrix}}_{t}\) also gives
whose right-hand side contains a factor (tℓ − t−ℓ) that is zero once specialized at q. The left-hand side of the definition becomes \({\prod }_{m=1}^{\ell }(Kq^{1-m}-K^{-1}q^{m-1})=K^{-\ell }q^{\ell (1-\ell )/2}{\prod }_{m=0}^{\ell -1}(K^{2}-q^{2m}\text {id})\) upon specialization and, since K is invertible in \(\mathcal LU_q\), the product \({\prod }_{m=0}^{\ell -1}(K^{2}-q^{2m}\text {id})\), and thus K2ℓ −id, must vanish. □
Fix now integers \(i,r,s \in \mathbb {Z}_{\geq 0}\) such that i = rℓ + s and s < ℓ − 1. Let also j = i + 2(ℓ − s − 1) as in Eq. ??. The following result is well-known (see e.g. [11]), but we still provide a proof.
Lemma B.4
As vector spaces, \(\text {Ext}^{1}_{\mathcal LU_q}({\Delta }_{q}(i),{\Delta }_{q}(j))\simeq \mathbb {C}\).
Proof
The short exact sequence of Proposition 4.4 yields the long exact sequence in cohomology
where the last written term is zero by projectivity of Pq(i) in the category of finite-dimensional (type I) \(\mathcal LU_q\)-modules. (Every object in the above sequence and in the rest of the proof should carry a subscript \(\mathcal LU_q\) which is dropped for brevity.) Hence, the extension group \(\text {Ext}^{1}({\Delta }_{q}(i),{\Delta }_{q}(j))\) is a quotient of EndΔq(j) and, as the exact sequence of Proposition 4.4 is non-split, it is enough to show that EndΔq(j) has dimension at most one to conclude the proof of the lemma. For that goal, apply the left-exact covariant functor Hom(Δq(j),−) on the sequence given in Proposition 4.3 to get the exact sequence
and note that Hom(Δq(j),Lq(i)) = 0 as the simple module Lq(j) = top Δq(j) is not a subquotient of Lq(i). Also, Schur’s lemma gives \(\text {Hom}({\Delta }_{q}(j),L_{q}(j)) \simeq \text {End} L_{q}(j) \simeq \mathbb {C}\) so Eq. B.2 forces \(\dim \text {End}{\Delta }_{q}(j) \leq 1\) as desired. □
To provide an explicit realization for the indecomposable \(\mathcal LU_q\)-module Pq(i), it thus suffices to construct a non-trivial extension of the module Δq(j) by Δq(i). This extension will be obtained as the specialization (at t = q) of the module over the rational form \(U_{t}\mathfrak {sl}_{2}\) given in the next proposition.
Proposition B.5
The \(\mathbb {Q}(t)\)-vector space \(\mathcal {T}_{t}(i)\) with basis \(\{m_{0},\dots ,m_{j},n_{0},\dots ,n_{i}\}\) is a \(U_{t}\mathfrak {sl}_{2}\)-module for the action
where the coefficients γp,v(t) and ωp,v(t) are defined by
with γp,v(t) = ωp,v(t) = 0 if p < 0 or p > i.
Proof
It is not difficult to verify that the given action is compatible with the defining relations of \(U_{t}\mathfrak {sl}_{2}\). The only slightly difficult check, which is of the relation (t − t− 1)[E,F]np = (K − K− 1)np, is now detailed:
where the coefficient of mℓ−s+p− 1 is
(Note that γp+ 1,1(t) and ωp− 1,1(t) respectively vanish if p = i and p = 0.) Furthermore, if v = 1, ωp,v(t) takes a simpler form, namely \(\omega _{p,1}(t)=\delta _{p,i}{\begin {bmatrix}{1}{i+\ell -s-1}\\{i}\end {bmatrix}}_{t}\), as p must satisfy 1 = v > i − p while remaining in the range 0 ≤ p ≤ i. In particular, the term [i − p + 1]tωp− 1,1(t) vanishes. The expressions for the γ’s and ω’s then give
where the factor (1 − δp,i) was added to handle the vanishing of γp+ 1,1(t) when p = i. This forces as claimed
The only remaining thing to check is that the action of the divided powers agrees with \(E^{(v)} = \frac {1}{[v]_{t}!}E^{v}\) and \(F^{(v)} = \frac {1}{[v]_{t}!}F^{v}\) for any \(v\in \mathbb {N}\). This is done by induction on v. There is nothing to prove when v = 1. In preparation for v ≥ 2, note that \(\gamma _{p,1}(t)={\begin {bmatrix}{1}{l-s+p-2}\\{p}\end {bmatrix}}_{t}\) so that the following recursive formula for the γ’s is easily obtained:
The induction hypothesis thus gives as claimed (all γ’s are evaluated at t)
This hypothesis also expresses \(\frac 1{[v-1]_{t}!} F^{v}n_{p}\) as a linear combination of np+v and mℓ−s+p+v− 1. A quick computation brings the coefficient of np+v to be the expected one for the action of [v]tF(v). The coefficient of mℓ−s+p+v− 1 is in turn
The vanishing of ωp+v− 1,1(t) and ωp,v− 1(t) depends on whether i − p is strictly smaller than v − 1 (then ωp+v− 1,1(t) = 0), equal to v − 1 (then ωp,v− 1(t) = 0) or strictly larger than v − 1 (and then both ω’s are zero and so is ωp,v(t)). A direct computation in the first two cases (< and = v − 1) shows that the non-vanishing term equals [v]tωp,v(t), as required. □
From now on, we view \(\mathcal {T}_{t}(i)\) as a \(\mathbb {Z}[t,t^{-1}]\)-module and we restrict the above \(U_{t}\mathfrak {sl}_{2}\)-action to a Ures-action. The Ures-module thus obtained is not finitely-generated but will still give rise to a finitely-generated \(\mathcal LU_q\)-module by specialization. In order to specialize this Ures-module, we need to show that the functions γp,v(t) and ωp,v(t) of Proposition B.5 have a well-defined limit as t tends to q. This is the purpose of the next two lemmas. As before, \(q\in \mathbb C^{\times }\) is such that q2 is a ℓ-th primitive root of unity. Note that a t-number [n]t can be written as t− 2(n− 1)(t2n − 1)/(t2 − 1) and that the polynomial (t2n − 1)/(t2 − 1) has order 2(n − 1) with all its roots distinct. In other words, the poles of 1/[n]t are all simple. Moreover, by Lucas q-theorem A.3, all q-binomial coefficients have a limit in \(\mathbb C\) as t → q.
Lemma B.6
The limit \(\gamma _{p,v} = \lim _{t\rightarrow q}\gamma _{p,v}(t)\) exists and belongs to \(\mathbb C\) for all \(p\in \{0,\dots ,i\}\) and \(v \in \mathbb {N}\).
Proof
By the comments made before the statement, the denominator 1/[v]t! is the only possible source of singularities in γp,v(t). Also, the result holds trivially if v < ℓ as, in this case, [v]t! does not vanish when t → q. If v = ℓ, then the zero in [v]t is simple and the limit \(\lim _{t\to q}\gamma _{\ell ,v}(t)\) needs to be studied. It turns out that this is the only case to consider. Indeed, for v > ℓ, Lemma B.1 ties the two divided powers E(v) and E(ℓ) by \(\kappa E^{(v)}=(E^{(\ell )})^{v_{1}}E^{(v_{2})}\) where
and where v has been written as v = v1ℓ + v2 with \(v_{1},v_{2}\in \mathbb Z_{\geq 0}\) and v2 < ℓ. Note that \(\lim _{t\to q}\kappa \) exists and is non-zero by Theorem A.3. Hence, for v > ℓ, the limit \(\lim _{t\to q}\gamma _{p,v}(t)\) can be obtained as the limit of a sum of \(\gamma _{p^{\prime },\ell }(t)\) and \(\gamma _{p^{\prime \prime },v^{\prime }}(t)\) with \(v^{\prime }<\ell \) and \(p^{\prime }, p^{\prime \prime }\in \{0,\dots , i\}\). The lemma thus rests on the study of \(\lim _{t\to q}\gamma _{p,\ell }(t)\). The rest of the proof is devoted to it.
From now on v = ℓ and the goal is to prove that the sum appearing in γp,ℓ(t) has a vanishing limit. The difference i − p is written as p1ℓ + p2 with \(p_{1},p_{2}\in \mathbb Z_{\geq 0}\) and p2 < ℓ. The summand in γp,ℓ(t) is also written as
so that \(\gamma _{p,\ell }(t) = \frac {1}{[\ell ]_{t}!}{\sum }_{u=0}^{\ell -1} \lambda _{u}(t)\). We prove first that \(\lim _{t\to q} \lambda _{u}(t)=0\) when u is neither ℓ − p2 − 1 nor s − p2.
Case 1: ℓ − p2 ≤ u. Then, [(p1 + 1)ℓ]t appears in the first product of λu(t) when a = ℓ − p2, so the result follows.
Case 2: s − p2 < u < ℓ − p2 − 1. Then the q-binomial in λu(t) can be written as
whose limit is zero by Theorem A.3 as 0 ≤ ℓ − p2 − u − 2 < ℓ − p2 − u + s < ℓ.
Case 3: u ≤ s − (p2 + 1). Then, [(p1 + 1)ℓ]t appears in the second product of λu(t) when b = s − p2.
Two values of u have escaped the above cases, namely u = s − p2 and u = ℓ − p2 − 1. If s < p2, only \(\lambda _{\ell -p_{2}-1}(t)\) of these two possible λ’s will contribute to the sum \({\sum }_{u=0}^{\ell -1} \lambda _{u}(t)\) as u takes non-negative values. However, in this case, [(p1 + 2)ℓ]t appears in the second product defining \(\lambda _{\ell -p_{2}-1}(t)\) when b = ℓ − p2 + s and the aforementioned sum tends to zero. Hence only the case s ≥ p2 remains. A direct computation gives
The q-binomials in these two limits can be simplified using Theorem A.3 and Proposition A.4:
Gathering the common factors in
the limit \(\lim _{t\to q}{\sum }_{u=0}^{\ell -1}\lambda _{u}(t)\) is then equal to
that gives, after changing the index in the first product to c = ℓ + s − 2p2 − b,
which is zero by Theorem A.4. This closes the argument. □
Lemma B.7
The limit \(\omega _{p,v} = \lim _{t\rightarrow q}\omega _{p,v}(t)\) exists and belongs to \(\mathbb C\) for all \(p\in \{0,\dots ,i\}\) and \(v \in \mathbb {N}.\)
Proof
By Lucas q-theorem A.3, the only potential singularity of ωp,v comes from the t-number [ℓ − s + i]t = [(r + 1)ℓ]t appearing in its denominator. As the zero of this t-number at t = q is of order one, it suffices to show that the product \({\begin {bmatrix}{\ell -s+p+v-1}\\{p+v}\end {bmatrix}}_{q}{\begin {bmatrix}{p+v}\\{v}\end {bmatrix}}_{q}\) appearing in ωp,v vanishes when mℓ−s+p+v − 1≠ 0 and i − p < v. Write v and p in the usual forms (that is v = v1ℓ + v2 and p = p1ℓ + p2 with \(v_{1},v_{2},p_{1},p_{2}\in \mathbb Z_{\geq 0}\) and v2, p2 < ℓ) and consider the three following cases.
Case 1: ℓ ≤ p2 + v2. Then, \({\begin {bmatrix}{p+v}\\{v}\end {bmatrix}}_{q} = {\begin {bmatrix}{(p_{1}+v_{1}+1)\ell +p_{2}+v_{2}-\ell }\\{v_{1}\ell +v_{2}}\end {bmatrix}}_{q} = 0\) by Theorem A.3 as 0 ≤ p2 + v2 − ℓ < v2 < ℓ.
Case 2: s < p2 + v2 < ℓ. Then, \({\begin {bmatrix}{\ell -s+p+v-1}\\{p+v}\end {bmatrix}}_{q} = \begin {bmatrix}{(p_{1}+v_{1}+1)\ell +p_{2}+v_{2}-s-1}\\{(p_{1}+v_{1})\ell + p_{2}+v_{2}}\end {bmatrix}_{q} = 0\) again by Theorem A.3.
Case 3: p2 + v2 ≤ s. Then, i < p + v gives r + 1 ≤ p1 + v1 so that mℓ−s+p+v− 1 = 0 as
This concludes the proof as every possible case has been studied. □
Corollary B.8
The \(\mathbb {C}\)-vector space \(\mathcal {T}_{q}(i)\) with basis {m0,..., mj, n0,...,ni} is a \(\mathcal LU_q\)-module for the action given by
Also, the unrolled element H acts on \(\mathcal {T}_{q}(i)\) as H mk = (j − 2k)mk and Hnp = (i − 2p)np.
Proof
The first statement follows from Proposition B.5 and Lemmas B.6 and B.7. The last one is easily proved using the definition of the unrolled element H and Proposition B.5. □
As \(\mathcal {T}_{q}(i)\) is clearly an extension of Δq(j) by Δq(i), Lemma B.4 shows that it is enough to show the following result in order to finish the proof Theorem 4.6. Recall that the construction of \(\mathcal {T}_{q}(i)\) and Theorem 4.6 both assume that s < ℓ − 1.
Lemma B.9
The \(\mathcal LU_q\)-module \(\mathcal {T}_{q}(i)\) is an indecomposable extension of Δq(j) by Δq(i).
Proof
Suppose ad absurdum that the extension is trivial and fix a \(\mathcal LU_q\)-linear isomorphism \(\varphi :{\Delta }_{q}(i)\oplus {\Delta }_{q}(j)\rightarrow \mathcal {T}_{q}(i)\). Let \(\{\overline m_{k}\}_{k=0}^{j}\subseteq {\Delta }_{q}(j)\) and \(\{\overline n_{p}\}_{p=0}^{i}\subseteq {\Delta }_{q}(i)\) be the bases given in Definition 4.1 and observe that the restriction of φ to q(j) must satisfy \(\varphi (\overline m_{k}) = {\varsigma } m_{k}\) for some \({\varsigma }\in \mathbb {C}^{\times }\) as \(\text {End}_{\mathcal LU_q}({\Delta }_{q}(j)) \simeq \mathbb {C}\) by the proof of Lemma B.4.
Write \(\varphi (\overline n_{0}) = {\sum }_{k = 0}^{j} \alpha _{k}^{(m)} m_{k}+{\sum }_{p=0}^{i} \alpha _{p}^{(n)} n_{p}\) where \(\alpha _{0}^{(m)},\dots ,\alpha _{j}^{(m)},\alpha _{0}^{(n)},\dots ,\alpha _{i}^{(n)}\in \mathbb {C}\). Then,
gives \(\varphi (\overline n_{0})=\alpha _{0}^{(n)}n_{0}+\alpha _{\ell -s-1}^{(m)}m_{\ell -s-1}\). However, this implies
and \(\alpha _{0}^{(n)}=0\) as s < ℓ − 1. Hence \({\varsigma }\varphi (\overline n_{0}) = {\varsigma }\alpha _{\ell -s-1}^{(m)}m_{\ell -s-1} = \alpha _{\ell -s-1}^{(m)}\varphi (\overline m_{\ell -s-1})\) and we contradict the injectivity of φ. □
Appendix C: Remaining Cases and Problematic Pairs
In this appendix, we finish the analysis done in Section 4 by letting q = ± 1 and by replacing Lusztig’s quantum group \(\mathcal LU_q\) by the universal envelopping algebra \(U\mathfrak {sl}_{2}\). We also show some results involving the problematic pairs (0,±i) ∈ λN (with N even, see Section 2) which were left unproven in Sections 3 and 5.
1.1 C.1 Periodic Chain Morphisms for ℓ = 1
Throughout this subsection, the parameter q is set to + 1 or − 1 and the algebra \(U_{q}\mathfrak {sl}_{2}\) is understood to be the universal enveloping algebra \(U\mathfrak {sl}_{2}\). This algebra is generated by elements e,f,h satisfying the defining relations
For \(n\in \mathbb {Z}_{\geq 0}\), we denote by L(n) the unique simple \(U\mathfrak {sl}_{2}\)-module with \(\dim L(n) = n+1\). This representation has a basis \(\{v_{0},{\dots } ,v_{n}\}\) with explicit action defined by evi = i(n − i + 1)vi− 1, fvi = vi+ 1 and hvi = (n − 2i)vi (where v− 1 = vn+ 1 = 0). Let \({\Delta }:U\mathfrak {sl}_{2}\rightarrow U\mathfrak {sl}_{2}\otimes U\mathfrak {sl}_{2}\) be the algebra morphism given by
This morphism for q = 1 is the usual coproduct used to define the tensor product of representations of \(\mathfrak {sl}_{2}\). A quick check shows that it is also an algebra morphism at q = − 1.
Define recursively a family of maps \(\{{\Delta }_{n}:U\mathfrak {sl}_{2}\rightarrow U\mathfrak {sl}_{2}^{\otimes n}\}_{n\geq 2}\) with Δ2 = Δ and Δn+ 1 = (id ⊗Δ) ∘Δn where id denotes the identity operator on \(U\mathfrak {sl}_{2}^{\otimes (n-1)}\). (Remark that Δ3≠(Δ ⊗id) ∘Δ when q = − 1 so that Δ is then not coassociative. Therefore the pullback of a threefold tensor product of \(U\mathfrak {sl}_{2}\)-modules by (Δ ⊗id) ∘Δ and its pullback by Δ3 may be non-isomorphic and care must be taken when defining a \(U\mathfrak {sl}_{2}\)-action on the periodic chains with q = − 1.) Identify moreover L(1) with \(\mathbb {C}^{2}\) through the change of basis given by v0↦|+〉 with v1↦|−〉 and associate the periodic chain \(\mathsf {X}_{N;z}^{+}\) (\(=\mathsf {X}_{N;z}^{-}\)) to the pullback of (Lq(1))⊗N by ΔN. Then, an easy computation gives the following formulas for the \(U\mathfrak {sl}_{2}\)-action on \(\mathsf {X}_{N;z}^{+}\)
and \(h|x_{1}{\dots } x_{N}\rangle _{z}^{+} = 2S^{z} |x_{1}{\dots } x_{N}\rangle _{z}^{+} = d|x_{1}{\dots } x_{N}\rangle _{z}^{+}\) where \(d = {\sum }_{j=1}^{N} x_{j}\). Remark that f acts on \(\mathsf {X}_{N;z}^{+}\) as the conjugation of e with the spin-flip s of Section 3.2. This will be used without mention later.
Theorem C.1
Let \(d \in \mathbb {Z}\) be of the parity of N and fix \(z \in \mathbb {C}^{\times }\) such that qd = z2. Set dm = d + 2m with zm = zqm for \(m\in \mathbb {Z}\). Then \(F_{d_{m},z_{m}}^{(1)}:\mathsf {X}_{N;d_{m},z_{m}}^{+}\rightarrow \mathsf {X}_{N;d_{m-1},z_{m-1}}^{+}\) given by \(|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+}\mapsto f|x_{1}{\dots } x_{N}\rangle _{z_{m-1}}^{+}\) is aTLN>-linear.
Proof
Fix a basis element \(|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+}\) of \(\mathsf {X}_{N;d_{m},z_{m}}^{+}\). Then, a straightforward computation using Eq. 3.6 gives
where, as q2 = 1,
Thus, \(e_{1}F_{d_{m},z_{m}}^{(1)}|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+} =F_{d_{m},z_{m}}^{(1)}e_{1}|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+}\) as desired. We also have
so \({\Omega }_{N} F^{(1)}_{d_{m},z_{m}}|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+} = F^{(1)}_{d_{m},z_{m}}{\Omega }_{N}|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+}\) as \(z_{m-1}^{-x_{1}} = q^{-1}z_{m}^{-x_{1}}\) and \(z_{m-1} = q^{d-1}z_{m}^{-1} = q^{N-1}z_{m}^{-1}\) by hypothesis. □
Suppose that qd = z2. Then, q−d = (z− 1)2 and Theorem C.1 produces a morphism of a TLN-modules from \(\mathsf {X}_{N;-d_{m-1},z_{m-1}^{-1}}^{+}\) to \(\mathsf {X}_{N;-d_{m},z_{m}^{-1}}^{+}\). Denote by \(E^{(1)}_{d_{m},z_{m}}:\mathsf {X}_{N;d_{m-1},z_{m-1}}^{+}\rightarrow \mathsf {X}_{N;d_{m},z_{m}}^{+}\) the aTLN-linear morphism obtained by conjugating this last map with the spin flip. (Recall that \(\mathsf {X}_{N;d,z}^{+}\) and \(\mathsf {X}_{N;d,z}^{-}\) are equal as q2 = 1.) Note also that the assumption qd = z2 amounts here to saying that \((d_{m-1},z_{m-1})\unlhd (d_{m},z_{m})\) whenever \(m\in \mathbb {Z}\) where \(\unlhd \) is the partial order on \({\Delta }=\mathbb {Z}\times \mathbb {C}^{\times }\) given through Eq. 4.3. Fix \(m\in \mathbb {N}\) and let \(\mu ^{(m)}:\mathsf {X}_{N;d_{m},z_{m}}^{+}\rightarrow \mathsf {X}_{N;d,z}^{+}\) and \(\nu ^{(m)}:\mathsf {X}_{N;d,z}^{+}\rightarrow \mathsf {X}_{N;d_{m},z_{m}}^{+}\) be the a TLN-morphisms defined by
Then \(\mu ^{(m)}|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+} = f^{m}|x_{1}{\dots } x_{N}\rangle _{z}^{+}\) and \(\nu ^{(m)}|x_{1}{\dots } x_{N}\rangle _{z}^{+} =e^{m}|x_{1}{\dots } x_{N}\rangle _{z_{m}}^{+}\) for \(x_{1},{\dots } ,x_{N} \in \{+,-\}\) since the \(U\mathfrak {sl}_{2}\)-action on \(\mathsf {X}_{N;z}^{+}\) is independent of z. The morphisms μ(m) and ν(m) satisfy properties similar to those described in Theorem 4.20. To prove this fact, observe that, since the category of finite-dimensional \(U\mathfrak {sl}_{2}\)-modules is semisimple, we can decompose the \(U\mathfrak {sl}_{2}\)-module \(\mathsf {X}_{N;z}^{+}\) as a sum of simple modules L(n) with \(n\in \mathbb {Z}_{\geq 0}\). Within L(n), the action of ejfj on the basis vector vi gives a non-zero multiple of vi if i + j ≤ n. This well-known fact and its analogue for fjej lead to the following lemma.
Lemma C.2
Let \(n,m\in \mathbb {Z}_{\geq 0}\) and \(d \in \mathbb {Z}\) with |d|≤ n and d ≡ n modulo 2. Then the action of emfm (or fmem) on L(n) induces a \(\mathbb {C}\)-linear automorphism of the eigenspace L(n)|h=d if |d − 2m|≤|d| (or |d + 2m|≤|d|, respectively).
The analogue of Theorem 4.20 follows directly from this lemma as the action of \(h\in U\mathfrak {sl}_{2}\) on the periodic chain \(\mathsf {X}_{N;z}^{+}\) coincide with the action of the diagonal spin operator 2Sz defining the eigenspaces \(\mathsf {X}_{N;d,z}^{+}\).
Theorem C.3
With the above notation, the maps ν(m)μ(m) and μ(m)ν(m) are respectively one-to-one if |d|≤|d + 2m| and |d|≥|d + 2m|. In particular, μ(m) is injective (or surjective) if |d|≤|d + 2m| (or if |d|≥|d + 2m|, respectively) and ν(m) is injective (or surjective) if |d|≥|d + 2m| (or if |d|≤|d + 2m|, respectively).
1.2 C.2 Problematic Pairs
This section studies the problematic pairs, that is the pairs (0,q) and (0,q− 1) in the case where q + q− 1 = 0 and N is even. Throughout the section q and N will be assumed to satisfy these conditions. Of course, this means that q is either i or − i. Note that \((0,q)\sim (0,q^{-1})\) in Λ and remark that theorem 2.4 of [17] implies that the head of the module WN;0,q does not provide any new simple module that is not already labeled in ΛN (see Eq. 2.1). The main purpose of this subsection is thus to determine the content and structure of these cellular modules. Here it is.
Proposition C.4
Let \(N\in 2\mathbb N\) and q be either i or − i. The Loewy diagram of the cellular module WN;0,q is
Proof
The first step here is to compute the direct successors in λN of (0,q). These are precisely the nodes of the proposed diagram, with arrows pointing toward the successors. These nodes are labelled by an integer \(i\in \mathbb {N}\) which is bounded above by n = N/2 and are given explicitly by (2i,yi) for the left column and by (2i,−yi) for the right where yi = q1−i. The direct successors of (0,q) through condition A and B are respectively (2,1) and (2,− 1). Proposition 3.2 applies as the pairs (0,q), (2,1) and (2,− 1) all belong to λN. There are hence two injective a TLN-morphisms WN;2,1 →WN;0,q and WN;2,− 1 →WN;0,q and the Loewy diagram of WN;0,q must contain the diagrams of WN;2,1 and WN;2,− 1. Fortunately, the latter diagrams may be deduced from Theorem 2.2: they are precisely those obtained by removing, from the diagram drawn in the statement, the node (2,− 1) for WN;2,1 and (2,1) for the other. It is thus sufficient to show that the dimension of WN;0,q is equal to that of the sum of the composition factors appearing in the statement.
As said above, WN;2,1 and WN;2,− 1 share all their composition factors except for their respective simple heads LN;2,1 and LN;2,− 1. Since \(\dim \mathsf {W}_{N;2,1}=\dim \mathsf {W}_{N;2,-1}\) (the dimension Eq. 3.3 of WN;d,z is independent of z), the two simple modules LN;2,1 and LN;2,− 1 have the same dimension. This argument can be repeated for each pair of simple modules \(\mathsf {L}_{N;2i,y_{i}}\) and \(\mathsf {L}_{N;2i,-y_{i}}\). We thus want to prove that \(\dim \mathsf {W}_{N;0,q}={\sum }_{1\leq i\leq n}2d_{i}\) where \(d_{i}=\dim \mathsf {L}_{N;2i,y_{i}}\). The dimensions di satisfy
The third equality was obtained by gathering up all dj’s in the sum but one of the the two di+ 1’s and the last one follows from Eq. 3.3. This recurrence ends at \(d_{n}=\dim \mathsf {L}_{N;N,y_{n}}=\dim \mathsf {W}_{N;N,y_{n}}=1\). Its solution is
Hence,
Note that the alterning sum contains the n first terms of the binomial expansion of (1 − 1)2n. Thus,
as desired. This ends the demonstration. □
We can now adapt the proofs of Theorem 2.4 and of Corollary 3.11 to the case of problematic pairs.
Proof Proof (Corollary 3.11, problematic case)
The proof given for Proposition C.4 shows that \(\mathsf {GP}_{N;0,q} \simeq \mathsf {L}_{N;2,-1} = \text {top } \mathsf {W}_{N;2,-1}\) which is a quotient of imiN;0,q by Proposition 3.9. To repeat this argument for the other problematic pair (0,q− 1), note that \(\mathsf {W}_{N;0,q}\simeq \mathsf {W}_{N;0,q^{-1}}\) by definition of the cellular modules as the cokernel of some fz (see Section 3.1). Note also that (0,q− 1) ≼ (2,− 1) and (0,q− 1) ≼ (2, 1) directly through condition A and B (respectively) by Lemma 3.10. Proposition C.4 thus gives \(\mathsf {GP}_{N;0,q^{-1}} \simeq \mathsf {L}_{N;2,1} = \text {top } \mathsf {W}_{N;2,1}\) which is a quotient of \(\text {im} i_{N;0,q^{-1}}\) by Proposition 3.9. □
Note that the generic part of WN;0,q and \(\mathsf {W}_{N;0,q^{-1}}\) are not isomorphic even though the modules WN;0,q and \(\mathsf {W}_{N;0,q^{-1}}\) are.
In the following proof (which follows almost exactly the steps of Section 5.3) we will often use implicitly the Loewy diagram of Proposition C.4 in order to identify the successors of the problematic pair (0,q) for the order ≼.
Proof Proof (Theorem 2.4, subcase (iii), problematic case)
Fix \(N\in 2\mathbb {N}\) and \(q\in \mathbb {C}\) with q2 = − 1. Then, the reasoning of Section 5.3 shows that \(\mathsf {X}_{N;0,q}^{+}\) admits a submodule M such that \(\mathsf {M}\simeq \mathsf {X}_{N;2,-1}^{+}/\text {im} F^{+}_{(2,-1);(4,-q)}\) (namely \(\mathsf {M}=\text {im} F^{+}_{(0,q);(2,-1)}\)). However, (D,Z) = (2,− 1) is not a problematic pair and is such that qD∉{Z2,Z− 2}. The proof of Section 5.3 thus shows that \(\mathsf {X}_{N;2,-1}^{+}\) has the Loewy diagram
where the image \(\text {im} F^{+}_{(2,-1);(4,-q)}\) has been identified with red arrows. The submodule \(\mathsf {M}\subseteq \mathsf {X}_{N;0,q}^{+}\) hence has the structure:
Using the above analysis for the problematic pair (0,q− 1) and applying the ⋆-duality (with the help of Proposition 3.4), we can moreover deduce that \(\mathsf {X}_{N;0,q}^{+}\) has a quotient with the Loewy diagram:
Also, the proof of Section 5.3 shows that \(\mathsf {X}_{N;0,q}^{+}\) has another submodule, namely \(\text {im} s F^{-}_{(0,-q);(2,-1)}\), which is isomorphic to \((\mathsf {X}_{N;2,-1}^{+}/\text {im} F^{+}_{(2,-1);(4,-q)})^{\circ } \simeq \mathsf {M}^{\circ }\). By the above analysis and Proposition 3.5, the structure of this submodule is
Repeating this argument for the pair (0,q− 1) and applying once again the ⋆-duality (with the help of Proposition 3.4), we deduce that \(\mathsf {X}_{N;0,q}^{+}\) has in addition a quotient with the following Loewy diagram:
We can now finish the proof precisely as in Section 5.3. Indeed, we have shown (see Proposition C.4) that all composition factors of WN;0,q appear in some subquotient of \(\mathsf {X}_{N;0,q}^{+}\). Since \(\dim \mathsf {X}_{N;0,q}^{+} = \dim \mathsf {W}_{N;0,q}\), we have found all composition factors of the eigenspace \(\mathsf {X}_{N;0,q}^{+}\) and its Loewy diagram must be the one announced in Theorem 2.4, that is
with perhaps some missing arrows and where the submodule \(M=\text {im} F^{+}_{(0,q);(2,-1)}\) has been identified with red arrows. The fact that there is no missing arrows is shown exactly like in Section 5.3 if N > 2. When N = 2, the precedent diagram becomes the one of the module W2;0,q (see Proposition C.4). It must therefore be missing an arrow since \(\mathsf {W}_{2;0,q}\not \simeq \mathsf {X}_{2;0,q}^{+}\) (as an easy computation shows that the generator e1 ∈a TLN acts as zero on W2;0,q but not of \(\mathsf {X}_{2;0,q}^{+}\)). However, the only arrow that can be added while still producing a diagram compatible with the submodules and quotients found above is \((2,1)\rightarrow (2,-1)\). The Loewy diagram of \(\mathsf {X}_{2;0,q}^{+}\) must then be given by this arrow, that is
The structure of the other problematic eigenspace \(\mathsf {X}_{N;0,q^{-1}}^{+} \simeq (\mathsf {X}_{N;0,q}^{+})^{\star }\) can also be deduced from Eq. C.2 with the help of Proposition 3.4. The resulting diagram for N = 2 is given by the arrow \((2,-1)\rightarrow (2,1)\). For N > 2, we rather obtain
where the image imF(0,−q);(2,1) is again identified with red arrows (this image is the ⋆-dual of the quotient Eq. C.1 of \(\mathsf {X}_{N;0,q}^{+}\)). The diagrams obtained are compatible with Theorem 2.4 and we can thus conclude the proof. □
We conclude this appendix by noting that a comparison between the diagram Eq. C.2 (or the corresponding diagram for N = 2) and the one given in Proposition C.4 shows that, for the problematic pair (0,q), the image imiN;0,q is exactly the generic part \(\mathsf {GP}_{N;0,q} \simeq \mathsf {L}_{N;2,-1}\). This is also true for the other problematic pair (0,−q) as the generic part GPN;0,−q is then isomorphic to the simple module LN;2,1 (see the proof of Corollary 3.11 for the problematic case) which is a submodule of \(\mathsf {X}_{N;0,-q}^{+}\) by diagram Eq. C.3 (or by the corresponding diagram for N = 2). This is a special case of Corollary 5.4.
Rights and permissions
Springer Nature or its licensor holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Pinet, T., Saint-Aubin, Y. Spin Chains as Modules over the Affine Temperley–Lieb Algebra. Algebr Represent Theor 26, 2523–2584 (2023). https://doi.org/10.1007/s10468-022-10171-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10468-022-10171-0
Keywords
- Affine Temperley–Lieb algebra
- Temperley–Lieb algebra
- Quantum groups
- Uqsl2
- Feigin–Fuchs module
- Schur–Weyl duality
- Indecomposable projective
- XXZ chain
- Periodic XXZ chain