Symmetric tensor categories in characteristic 2 (Q2001595)

One of classical results of \textit{P. Deligne} [Prog. Math. 87, 111--195 (1990; Zbl 0727.14010)] states that every symmetric tensor category \(\mathcal{C}\) with moderate growth admits a fiber functor to finite dimensional super vector spaces if \(\mathcal{C}\) is defined over an algebraically closed field of characteristic zero. What about the positive characteristic case? This is the main motivation of the paper under review. In this paper, the authors consider the question in the case of characteristic 2. Now let \(\mathrm{k}\) be an algebraically closed field with characteristic 2 and take a symmetric tensor category \(\mathcal{C}\). It is known that if \(\mathcal{C}\) is finite and semisimple, then answer is positive by a theorem of \textit{V. Ostrik} [``On symmetric fusion categories in positive characteristic'', Preprint, \url{arXiv:1503.01492}] which states that \(\mathcal{C}\) admits a fiber functor to the category \(\mathsf{Ver}_{p}(\mathrm{k})\) which is the symmetric tensor category obtained from Rep\(_{\mathrm{k}}(\mathbb{Z}/p)\) by the process of semisimplification (see [the second author and \textit{V. Ostrik}, ``On semisimplification of tensor categories'', Preprint, \url{arXiv:1801.04409}]). Of course, if \(p=2\) then \(\mathsf{Ver}_{p}(\mathrm{k})\) is just \(\mathsf{Ver}_{\mathrm{k}}\). But if we remove the semisimplicity assumption, then such a fiber functor may be not exist. But we still can ask the following question: do we have a fiber functor to other nice symmetric tensor categories, say, \(\mathcal{C}_1:=\) Rep-\(\mathrm{k}[d]/d^2\)? The answer is still no since \textit{V. Ostrik} [private communication (2017)] constructed a symmetric tensor category \(\mathcal{C}_2\) which does not admit a fiber functor to \(\mathcal{C}_1\). So one may ask whether any finite symmetric tensor category \(\mathcal{C}\) has a fiber functor to \(\mathcal{C}_2\)? The striking fact that the authors prove is: there in \textit{no} finite symmetric tensor category over \(\mathrm{k}\) such that any finite symmetric tensor category \(\mathcal{D}\) over \(\mathrm{k}\) admits a fiber functor to \(\mathcal{D}\). To prove this fact, the authors construct an infinite ascending chain of finite symmetric categories \[\mathsf{Ver}_{\mathrm{k}}=\mathcal{C}_0\subset \mathcal{C}_1\subset \mathcal{C}_2\subset \mathcal{C}_3\subset \cdots\] such that \(\mathcal{C}_{2n}\) is incompressible, i.e., does not admit a tensor functor to a finite tensor category of smaller Frobenius-Perron dimension. The construction is inductive: 1) starting from \(\mathcal{C}_{2n}\), one can get a typical algebra \(A_n\) which is a commutative and cocommutative Hopf algebra and \(\mathcal{C}_{2n+1}\) is defined to be \(A_n\)-mod; 2) starting from \(\mathcal{C}_{2n+1}\), we define \(\mathcal{C}_{2n+2}:=\mathcal{C}_{2n+1}\oplus \mathcal{M}_n\) where \(\mathcal{M}_n=\mathcal{C}_{2n}\) as a \(\mathcal{C}_{2n+1}\)-module category. The properties and structures of the categories \(\mathcal{C}_n\) are studied in the paper too. The authors give a question for guiding further research: Consider the infinite category \(\mathcal{C}_{\infty}:=\bigcup_{n\geq 0}\mathcal{C}_{n} \). Does any finite symmetric tensor category over \(\mathrm{k}\) admit a fiber functor to \(\mathcal{C}_{\infty}\)?