I am trying to understand the difference between the "Drinfeld" and the "Lusztig" theory of quantum groups, more specifically with respect to the problem of quantization of Lie bialgebras/Poisson Lie groups.

There is a very well understood (and by now classical) theory of Etingof and Kazhdan on quantization of Lie bialgebras. For any (finite dimensional) Lie bialgebra $(\mathfrak{g}, \delta)$, one can construct a deformation Hopf algebra $U_h(\mathfrak{g})$ defined over $\mathbb{C}[[h]]$ quantizing $(\mathfrak{g}, \delta)$, and any quasitriangular $r$-matrix $r \in \mathfrak{g} \otimes \mathfrak{g}$ can be quantized to a quasitriangular $R$-matrix $R \in U_h(\mathfrak{g}) \hat{\otimes} U_h(\mathfrak{g})$.

My question is whether a similar quantization theory exists over $\mathbb{C}[q, q^{-1}]$. Can one always construct a Hopf algebra $U_q(\mathfrak{g})$ defined over $\mathbb{C}[q, q^{-1}]$ (or another similar looking polynomial ring) quantizing $(\mathfrak{g}, \delta)$ in a precise sense? Can any quasitriangular $r$-matrix be quantized to a universal $R$-matrix $R \in U_q(\mathfrak{g}) \hat{\otimes} U_q(\mathfrak{g})$?

My understanding is that the latter theory, if it exists would be better, as one would get an actual family of deformations, as opposed to a formal one.


It all depends on what you mean by a well defined quantization theory.

In the case of standard semimsimple Lie bialgebra you can define elements $K_i=e^{\hbar d_i H_i}$ for each Cartan generator $H_i$, where $d_i$ are the natural numbers symmetrizing the Cartan matrix. Then you can show that letting $q=e^\hbar$ you have a $\mathbb C[q,q^{-1}]$ rational form $U_q(\mathfrak g)$ inside $U_\hbar(\mathfrak g)$. Finally, inside this rational form you still have an $R$-matrix of the form you considered.

All this is not functorial, however. In general you can ask for the existence of a suitable rational form inside the local deformation. Personally I have never seen an existence theorem of the form: inside each Etingof-Kazhdan $\hbar$-deformation there is a $q$-rational form. But I wouldn't be too much surprised if such a result holds true. Usually what is difficult is to identify one such rational form, since you cannot expect any uniqueness, and I guess that's why you rather go by generators and relations.

However, for the case of compact Poisson-Lie groups, since there you have a classification of Lie bialgebra structures, you can verify that any Lie bialgebra can be $q$-quantized and the $r$-matrix survives quantization. Best reference for this is the book by Korogodskii and Soibelman: Algebra of Functions on Quantum Groups I.

This discussion may be of help as well:

difference between $q$-deformations and $\hbar$-deformations

  • $\begingroup$ Chari and Pressley also discuss integral/rational forms of QUE algebras in the chapter 9 of their book. $\endgroup$ – Victor Mouquin Jul 11 '18 at 2:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.