Abstract
It is shown that the \(q\)-difference Noether problem for all classical Weyl groups has a positive solution, simultaneously generalizing well known results on multisymmetric functions of Mattuck (Proc Am Math Soc 19:764–765, 1968) and Miyata (Nagoya Math J 41:69–73, 1971) in the case \(q=1\), and \(q\)-deforming the noncommutative Noether problem for the symmetric group (Futorny et al. in Adv Math 223:773–796, 2010). It is also shown that the quantum Gelfand–Kirillov conjecture for \(\mathfrak gl _N\) (for a generic \(q\)) follows from the positive solution of the \(q\)-difference Noether problem for the Weyl group of type \(D_n\). The proof is based on the theory of Galois rings (Futorny and Ovsienko in J Algebra 324:598–630, 2010). From here we obtain a proof of the quantum Gelfand–Kirillov conjecture for \(\mathfrak gl _N\), and for a certain extension of \(\mathfrak sl _N\). Previously, the case of \(\mathfrak sl _N\) was shown by Fauquant-Millet (J Algebra 218:93–116, 1999) and by Alev and Dumas (J Algebra 170:229–265, 1994) (for \(N=2,3\)). Moreover, we give an explicit description of the skew fields of fractions for \(U_q(\mathfrak gl _N)\) and \(U_q^\mathrm{ext}(\mathfrak sl _N)\) which generalizes the results of Alev and Dumas (J Algebra 170:229–265, 1994).
Similar content being viewed by others
References
Alev, J., Dumas, F.: Sur le corps des fractiones de certaines algebres quantiques. J. Algebra 170, 229–265 (1994)
Alev, J., Dumas, F.: Opérateurs différentiels invariants et problème de Noether, Studies in Lie theory, 2150, Progr. Math., 243, Birkhäuser Boston, Boston, MA (2006)
Alev, J., Ooms, A., Van den Bergh, M.: A class of counterexamples to the Gelfand–Kirillov conjecture. Trans. Am. Math. Soc. 348, 1709–1716 (1996)
Alev, J., Ooms, A., Van den Bergh, M.: The Gelfand–Kirillov conjecture for Lie algebras of dimension at most eight. J. Algebra 227, 549–581 (2000). Corrigendum. J. Algebra 230, 749 (2000)
Borho, W., Gabriel, P., Rentschler, R.: Primideale in Einhullenden auflosbarer Lie-Algebren. Lecture Notes in Math., vol. 357. Springer, Berlin and New York (1973)
Bracken, A., Gould, M., Zhang, R.: Quantum group invariants and link polynomials. Commun. Math. Phys. 137, 13–27 (1991)
Brown, K.A., Goodearl, K.R.: Lectures on Algebraic Quantum Groups, Advance Course in Math. CRM Barcelona, vol. 2. Birkhauser, Basel (2002)
Caldero, P.: On the Gelfand–Kirillov conjecture for quantum algebras. Proc. AMS 128, 943–951 (1999)
Cauchon, G.: Effacement des dérivations et spectres premiers des algèbres quantiques. J. Algebra 260, 476–518 (2003)
Dumas, F.: An Introduction to Noncommutative Polynomial Invariants. Lecture Notes, Homological Methods and Representations of Non-commutative Algebras. Mar del Plata, Argentina March 6–17 (2006)
Farkas, D., Schofield, A., Snider, R., Stafford, J.: The isomorphism question for division rings of group rings. Proc. AMS 85, 327–330 (1982)
Fauquant-Millet, F.: Quantification de la localisation de de Dixmier de \(U(sl_{n+1}(\mathbb{C}))\). J. Algebra 218, 93–116 (1999)
Futorny, V., Molev, A., Ovsienko, S.: The Gelfand–Kirillov Conjecture and Gelfand-Tsetlin modules for finite \(W\)-algebras. Adv. Math. 223, 773–796 (2010)
Futorny, V., Ovsienko, S.: Galois orders in skew monoid rings. J. Algebra 324, 598–630 (2010)
Goodearl K.R.: Prime spectra of quantized coordinate rings. In: Interactions Between Ring Theory and Representations of Algebras (Murcia). Lecture Notes in Pure and Appl. Math., vol. 210, pp. 205–237. Dekker, New York (2000)
Gelfand, IMet, Kirillov, A.A.: Sur les corps liés cor aux algèbres enveloppantes des algèbres de Lie. Publ. Inst. Hautes Sci. 31, 5–19 (1966)
Iohara, K., Malikov, F.: Rings of skew polynomials and Gelfand–Kirillov conjecture for quantum groups. Commun. Math. Phys. 164, 217–237 (1994)
Joseph, A.: Proof of the Gelfand–Kirillov conjecture for solvable Lie algebras. Proc. Am. Math. Soc. 45, 1–10 (1974)
Joseph, A.: Sur une conjecture de Feigin. C. R. Acad. Sci. Paris Ser.I Math. 320(12) , 1441–1444 (1995)
Jordan, D.: A simple localization of the quantized Weyl algebra. J. Algebra 174, 267–281 (1995)
Klimyk, A., Schmudgen, K.: Quantum Groups and Their Representations. Springer, Berlin (1997)
Li, J.: The quantum Casimir operators of \(U_q(\mathfrak{gl}_n)\) and their eigenvalues. J. Phys. A: Math. Theor. 43, 345202 (2010) (9pp)
Maltsiniotis, G.: Calcul Differentiel Quantique. Groupe de travail, Universite Paris VII, Paris (1992)
Mattuck, A.: The field of multisymmetric functions. Proc. Am. Math. Soc. 19, 764–765 (1968)
Mazorchuk, V., Turowska, L.: On Gelfand-Zetlin modules over \(U_q(\mathfrak{gl}_n)\). Czechoslovak J. Physics 50, 139–144 (2000)
McConnell, J.C.: Representations of solvable Lie algebras and the Gelfand–Kirillov conjecture. Proc. Lond. Math. Soc. 29, 453–484 (1974)
Miyata, T.: Invariants of certain groups. 1. Nagoya Math. J. 41, 69–73 (1971)
Panov, A.N.: Skew field of rational functions on \(GL_q(n, K)\). Funktsional. Anal. i Prilozhen. 28, 7577 (1994)
Panov, A.: Fields of fractions of quantum solvable algebras. J. Algebra 236, 110–121 (2001)
Premet, A.: Modular Lie algebras and the Gelfand–Kirillov conjecture. Invent. Math. 181, 395–420 (2010)
Richard, L.: Sur les endomorphismes des tores quantiques. Commun. Algebra 30, 5283–5306 (2002)
Ueno, K., Takebayashi, T., Shibukawa, Y.: Gelfand–Tsetlin basis for \(U_q(gl(N+1))\). Lett. Math. Phys. 18, 215–221 (1989)
Zhelobenko, D.: Compact Lie groups and their representations. Translations of Mathematical Monographs, AMS (1973)
Acknowledgments
The authors are grateful to Michel Van den Bergh, Fedor Malikov, Eugene Mukhin and Alan Weinstein for encouraging discussions.
Author information
Authors and Affiliations
Corresponding author
Additional information
The first author is grateful to the Max Planck Institute for Mathematics in Bonn for support and hospitality during his visit. The first author is supported in part by the CNPq Grant (301743/2007-0) and by the Fapesp Grant (2010/50347-9).
Appendix
Appendix
1.1 Proof of Proposition 3.1
The statement is equivalent to proving that
in \(C_n^q[X,Y]\). Put
so that \(P(X)=\sum _{i=1}^n Q_i(X)\). Observe that
Thus, to prove (8.1), it is enough to show the following two identities:
Since \(y_1 x_i = q^{\delta _{i1}} x_i y_1\) we have
which is symmetric in \(X,Y\). This proves (8.4).
Next we prove (8.5). Let
Then
and
We have
This shows (8.5) and completes the proof that \([t_i,t_j]=0\) for all \(i,j\).
1.2 Proof of Proposition 3.7
The relation (3.15) holds by Proposition 3.1, while (3.16) holds by the definition, (3.11), of \(e_d\). Relation (3.17) is trivial for \(k=0\).
Using (3.12) and that \(w(e_k)=e_k\) for any \(w\in S_n\) we have, for any \(j,k\in \{1,\ldots ,n\}\),
Substituting \(y_n=t_1+x_nt_2+\cdots +x_n^{n-1}t_n\) and using that \(w(t_i)=t_i\) for all \(w\in S_n\) we get
Write \(e_{n-j}^{\prime }\) as a sum of monomials \(x_{i_1}\cdots x_{i_{n-j}}\) and \(1\le i_1<\cdots <i_{n-j}\le n-1\). We claim that the only way to get a nonzero contribution is when \(i_r=r\) for all \(r\). Indeed, suppose \(i_r>r\) for some \(r\) chosen minimal. Then the product
will be fixed by the transposition \((i_r-1\;\;i_r)\). Therefore, after anti-symmetrization, the term will cancel out. In other words, the substitution \(w\mapsto w\cdot (i_r-1\;\; i_r)\) in the sum
gives the same expression with opposite sign, proving it is zero. Thus, noting also that
we have
The term \(i=j\): Write \(e_k^{\prime }= \sum _{1\le i_1<\cdots <i_k\le n-1} x_{i_1}\cdots x_{i_k}\). Consider
An expression like this containing factors \((x_rx_{r^{\prime }})^s\) (\(r\ne r^{\prime }\)) will become zero after anti-symmetrization. If \(n-j\ge k\) there is a unique way to get a nonzero result, namely to choose \((i_1,\ldots ,i_k)=(1,2,\ldots ,k)\). If \(n-j<k\) there is no way to get nonzero result. Thus
where \(a(i_1,\ldots ,i_n):=\sum _{w\in S_n}{{\mathrm{sgn}}}(w)w(x_1^{i_1}\cdots x_n^{i_n})\). Use that \(w(a(i_1,\ldots ,i_n))={{\mathrm{sgn}}}(w)a\) \((i_1,\ldots ,i_n)\) with
which is a cycle of length \(j\), to get
Using that the Schur function
defined for a partition \(\lambda =(\lambda _1,\ldots ,\lambda _n)\), \(\lambda _1\ge \cdots \ge \lambda _n\ge 0\), satisfies \(s_{1^k0^{n-k}}=e_k\) and that \(\Delta =a(n-1,n-2,\ldots ,0)\) we get that
Similarly, if we look at the term containing \(qx_ne_{k-1}^{\prime }\), there is at most one tuple \((i_1,\ldots ,i_{k-1})\), \(1\le i_1<\cdots <i_{k-1}\le n-1\) such that the antisymmetrization of
is nonzero, namely \((i_1,\ldots ,i_{k-1})=(1,\ldots ,k-1)\) and this time, due to the presence of \(x_n^j\), it gives nonzero result if and only if \(k-1\ge n-j\) i.e. \(j+k>n\). Thus
To get a descending sequence inside the parenthesis we apply the cyclic permutation which places \(j\) between \(j-1\) and \(j+1\). This cycle has length \(j\), giving a factor \((-1)^{j-1}\). As before, this gives
Combining (8.7) and (8.8) yields
The terms where \(i>j\): We first look at the \(e_k^{\prime }\) term in (8.9). That \(i>j\) means the exponent \(i-1\) of \(x_n\) occurs in one of the exponents in \(x_1^{n-1}x_2^{n-2}\cdots x_{n-j}^j\), namely in \(x_{n-(i-1)}^{i-1}\). Therefore \(x_{i_1}\cdots x_{i_k}\) must contain \(x_1x_2\cdots x_{n-(i-1)}\). In particular \(k\ge n-(i-1)\). The remaining factors must be \(x_{n-j+1}x_{n-j+2}\cdots \) and they cannot continue beyond \(x_{n-1}\) meaning that \(k-(n-i+1) + (n-j) \le n-1\). Thus the following inequalities are necessary conditions in order to avoid having two variables with the same exponent:
i.e.
If these inequalities hold there is a unique tuple
with \(1\le i_1<\cdots <i_k\le n-1\) such that
is nonzero. With this choice we get
where we applied the cyclic permutation \((n-i+2\;\; n-i+3\;\; \cdots \;\; n-1\;\; n)\) of length \(i-1\) in the second equality.
The argument for the term containing \(qx_1e_{k-1}^{\prime }\) is analogous, but gives an extra minus sign. Together with (8.10) one obtains that for \(i>j\) we have
The terms where \(i<j\): We look at the \(e_k^{\prime }\) term in (8.9). Necessary conditions for nonzero contribution are \(k\ge j-i\) and \(k-(j-i)\le n-j\), i.e.
After a similar computation as the \(i>j\) case we obtain
Combining (8.13), (8.11) and (8.9) we obtain
Making the change of summation variables \(i\mapsto i+j\) we get
In the first sum, the condition \(i\le n-k\) is redundant since, by the notational convention, \(e_{k+i}=0\) for \(i>n-k\). Similarly, \(-k\le i\) is superfluous in the second sum. Thus we obtain (3.17).
1.3 Proof of Proposition 3.9
First note that (3.24) implies that
We now prove (3.29). Let \(j\in [\![{1},{n-1}]\!]\) and \(k\in [\![{0},{n-1}]\!]\). Then the left hand side of (3.29) equals
By (3.19) with \((j,k)\) replaced by \((k+1,n-j)\), the term (8.20) equals
Similarly, applying (3.20) with \((j,k)\) replaced by \((k+1,n+1-j)\) shows that (8.21) is equal to
Adding together (8.22), (8.23) and (8.19) gives the right hand side of (3.29). This proves (3.29).
In particular, taking \(k=0\) and \(k=n-1\) in (3.29) we get
for all \(j\in [\![{1},{n-1}]\!]\). Using these identities, together with \([T_j,E_0]=0\) and \([T_j,E_n]_q=0\) which follow from (3.24), one can check that
proving (3.30).
That (3.28) holds is trivial from the assumption (3.22).
We now prove (3.27). Let \(j,k\in [\![{1},{n-1}]\!]\). We will bring \(\widetilde{T}_j\widetilde{T}_k\) to the normal form where all the \(E\)’s are to the left of all the \(T\)’s and prove that the resulting expression is symmetric in \(j,k\). We may assume \(j\ne k\). Using (8.24), (8.25) and (3.29), we have
We prove that all parts of this expression are symmetric in \(j,k\). The first term, containing \(E_jE_k\), is trivially symmetric.
The terms containing \(E_0^2T_1^2\). There are two terms in (8.26) containing \(E_0^2T_1^2\):
Applying (3.19) with \((j,k)\) replaced by \((n-j,n-k)\) we get that (8.27) equals
which is symmetric in \(j,k\).
The terms containing \(E_n^2T_n^2\).
Here we can apply (3.20) with \((j,k)\) replaced by \((n+1-j,n+1-k)\) to see that (8.28) equals
which is symmetric in \(j,k\).
The terms containing \(E_0T_1^2T_n\).
The parenthesis equals
If \(j\ge k\), we can include \((-1)^k(q-1)E_jT_{n-k}\) as the term \(i=j-k\) in the sum. If \(j<k\), the term \((-1)^k(q-1)E_jT_{n-k}\) cancels the term \(i=j-k\) in the sum, and \((-1)^j(q^{\delta _{k>j}}-1)E_kT_{n-j}\) may be included in the sum as \(i=0\). Thus (8.30) can be written
Making the change of variables \(i\mapsto i+j-k\) in this sum gives the same expression but with \(j\) and \(k\) interchanged. Thus it is symmetric in \(j\) and \(k\).
The terms containing \(E_nT_1T_n^2\).
Similarly to the previous case, the expression inside the parenthesis can be written as
Substituting \(i\mapsto i-k+j\) one checks this is symmetric in \(j\) and \(k\).
The terms containing \(E_0E_nT_1T_n\). Finally, there are four terms in (8.26) containing \(E_0E_nT_1T_n\):
Applying (3.19) with \((j,k)\) replaced by \((n-j+1,n-k)\) and (3.20) with \((j,k)\) replaced by \((n-j,n-k+1)\) we obtain that the parenthesis in (8.34) equals
which is symmetric in \(j\) and \(k\). This completes the proof that (8.26) is symmetric in \(j\) and \(k\). Thus (3.27) holds.
The last statement about generators follows from the fact that (3.25) and (3.26) can be used to express \(E_j\) for \(j\in [\![{1},{n-1}]\!]\) and \(T_k\) for \(k\in [\![{1},{n}]\!]\), in terms of the new generators \(\{E_0,E_n\}\cup \{\widetilde{T}_j\}_{j=1}^{n-1}\cup \{\widetilde{E}_k\}_{k=0}^{n-1}\).
1.4 Example: the case \(n=2\)
If \(n=2\) then (3.3) becomes
and from this, or using (3.12), we get
By definition (3.11), we have
By Corollary 3.5, \(\mathbb{k }_q(\bar{x},\bar{y})^{S_2}\) is generated as a skew field over \(\mathbb{k }\) by \(e_1,e_2, t_1, t_2\). By Proposition 3.7 we have the following relations:
Using the notation in (3.32) and (3.31) we have
By (3.48) or direct computations,
Thus, \((Z_1,Z_2,Z_3,Z_4)=(X_1,Y_1,X_2,Y_2)\) satisfy \(Z_iZ_j=q^{s_{ij}}Z_jZ_i\) with
Using the definition (3.33),
As in the proof of Theorem 3.10, \(\widehat{X}_1, \widehat{X}_2, \widehat{Y}_1,\widehat{Y}_2\) generate \(\mathbb{k }(\bar{x},\bar{y})^{S_2}\) as a skew field, the following relations hold:
and we have an isomorphism \(\mathbb{k }_q(x_1,x_2,y_1,y_2)^{S_2}\simeq \mathbb{k }_q(x_1,x_2,y_1,y_2)\).
1.5 Example: the case \(n=3\)
The elementary symmetric polynomials \(e_d\) are
By (3.12) we have
where
By Corollary 3.5, \(\mathbb{k }_q(\bar{x},\bar{y})^{S_3}\) is generated as a skew field over \(\mathbb{k }\) by \(e_1,e_2,e_3,t_1,t_2,t_3\) and by Proposition 3.7 or direct computations, we have the following relations:
By (3.32) and (3.31),
By (3.48),
Thus, if we let \((Z_1,Z_2,\ldots ,Z_6)=(X_1,Y_1,X_2,Y_2,X_3,Y_3)\), then \(Z_iZ_j=q^{s_{ij}}Z_jZ_i\) with
By performing simultaneous elementary row and column transformations, this matrix can be brought to the skew normal form
As in (3.33), changing generators to
one can also verify directly that
which means that there is an isomorphism of skew fields
Rights and permissions
About this article
Cite this article
Futorny, V., Hartwig, J.T. Solution of a \(q\)-difference Noether problem and the quantum Gelfand–Kirillov conjecture for \(\mathfrak gl _N\) . Math. Z. 276, 1–37 (2014). https://doi.org/10.1007/s00209-013-1184-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00209-013-1184-3