Structure of certain Weyl modules for the spin groups (Q1645008)

Let $K$ be an algebraically closed field of positive characteristic $p$ and let $G$ be a simple, simply connected linear algebraic group over $K$. The structure of irreducible $KG$-modules is still unknown. Fixing a Borel subgroup $B$ and a maximal torus $T$ contained in it, the irreducible, rational modules $L_G(\lambda)$ for $G$ are in bijective correspondence with the dominant weights $\lambda \in X^+(T)$. Then, for each $\lambda \in X^+(T)$, the irreducible module $L_G(\lambda)$ is isomorphic to $V_G(\lambda)$ by its radical, where the Weyl module $V_G(\lambda)$ is a universal highest weight module of highest weight $\lambda$. One knows formulae to obtain the dimension etc. of $V_G(\lambda)$ but the composition factors seem out of reach at the moment. This is equivalent to finding weight multiplicities in simple modules for $G$, and has been done for large $p$ and small $\lambda$ in a recursive fashion using Kazhdan-Lusztig polynomials. In small rank, Jantzen's $p$-sum formula has proved useful. In the paper under review, the author determines the structure of Weyl modules for $G =\mathrm{Spin}(W)$ with $W$, a $K$-vector space of dimension $\geq 5$ and $p \neq 2$. He adopts an idea of McNinch, generalizes and uses it here. \textit{G. McNinch} [Proc. Lond. Math. Soc. (3) 76, No. 1, 1 (1998; Zbl 0891.20032)] had proved: \par Let $V$ be a finite-dimensional, rational $KG$-module and let $\lambda$ be a dominant weight of $V$. Assume that $V$ is a tilting module. Then, there exists an injection $i \in \mathrm{Hom}_{KG}(V_G(\tau),V)$ and a surjection $\phi \in\mathrm{Hom}_{KG}(V,H^0(\tau))$. \par Generalizing McNinch's result [loc. cit.], the author proves: \par Let $V$ be as above. Identify $V_G(\tau)$ with its image under the injection $i$ above, and writing \[\operatorname{ch}(V) = \chi(\tau) + \sum_{\mu \in X^+(T)} a_{\mu} \operatorname{ch}(L_G(\mu)) \] we have for every $\mu \neq \tau$ that \[V_G(\tau):L_G(\mu)] \leq a_{\mu}. \] The author uses the above generalization to determine the composition factors of a carefully-chosen tilting module in order to reduce the list of possible composition factors for $V_G(\lambda)$. Then, he uses a truncated version of Jantzen's $p$-sum formula in a suitable form. This enables him to obtain the dimensions of the corresponding irreducible $KG$-modules. The author states the results for the types $B_n$ ($n \geq 2$) and $D_n$ ($n \geq 3$) in terms of tables that give the radicals corresponding to certain highest weights and the corresponding dimensions of the irreducible module. In fact, as $\operatorname{char} K \neq 2$, for $B_n$ as well as $D_n$, any Weyl module of highest weight in $\Lambda(G)$ is irreducible. The certain weights referred to above are those in $\omega_1 + \Lambda(G)$. McNinch [loc. cit.] has dealt with the special case $\omega_1 + \omega_2$. The methods used in the paper cannot work for the groups of type $C_n$ because in this case the Weyl modules having fundamental weights as highest weights are not necessarily irreducible, even when $\operatorname{char}K \neq 2$. \par The paper is written very clearly.