V. Chari et A. Pressley, MINIMAL AFFINIZATIONS OF REPRESENTATIONS OF QUANTUM GROUPS - THE IRREGULAR CASE, letters in mathematical physics, 36(3), 1996, pp. 247-266
Let g be a finite-dimensional complex simple Lie algebra and U-q(g) th
e associated quantum group (q is a nonzero complex number which we ass
ume is transcendental). If V is a finite-dimensional irreducible repre
sentation of U-q(g), an affinization of V is an irreducible representa
tion (V) over cap of the quantum affine algebra U-q((q) over cap) whic
h contains V with multiplicity one and is such that all other irreduci
ble U-q(g)-components of (V) over cap have highest weight strictly sma
ller than the highest weight lambda of V. There is a natural partial o
rder on the set of U-q(g)-isomorphism classes of affinizations, and we
look for the minimal one(s). In earlier papers, we showed that (i) if
g is of type A, B, C, F or G, the minimal affinization is unique up t
o U-q(g)-isomorphism; (ii) if g is of type D or E and lambda is not or
thogonal to the triple node of the Dynkin diagram of g, there are eith
er one or three minimal affinizations (depending on lambda). In this p
aper, we show, in contrast to the regular case, that if U-q(g) is of t
ype D-4 and lambda is orthogonal to the triple node, the number of min
imal affinizations has no upper bound independent of lambda. As a by-p
roduct of our methods, we disprove a conjecture according to which, if
g is of type A(n), every affinization is isomorphic to a tensor produ
ct of representations of U-q((g) over cap) which are irreducible under
U-q(g) (in an earlier paper, we proved this conjecture when n = 1).