Polarized endomorphisms of normal projective threefolds in arbitrary characteristic (Q2200779)

Let \(f: X \rightarrow X\) be a surjective endomorphism of a projective variety over a fixed algebraically closed field k of arbitrary characteristic \( p \geq 0\). \(f\) is said to be \( q\)-polarized (resp. numerically \(q\)-polarized) by \(H\), if there is an ample Cartier divisor \(H\) such that \(f^*H \sim q H\), a linear equivalence (resp. \(f^*H \equiv q H\), a numerical equivalence) for some integer \( q > 1\). In this paper, the authors prove the following results. 1. If \(\mathrm{char} k = 0\), in 2010, \textit{N. Nakayama} and \textit{D.-Q. Zhang} [Math. Ann. 346, No. 4, 991--1018 (2010; Zbl 1189.14043)] showed that the numerically polarized condition is equivalent to the polarized condition after replacing \(H\). If \( X\) is normal and \(f\) is separable, the authors generalize this result to arbitrary characteristic. 2. The authors generalize the result of \textit{S. Boucksom} et al. [Duke Math. J. 161, No. 8, 1455--1520 (2012; Zbl 1251.14026)] for \(\mathbb{Q}\)-Gorenstein normal projective variety \(X\) and prove: Let \(f: X \rightarrow X\) be a polarized separable endomorphism of a \(\mathbb{Q}\)-Gorenstein normal projective variety \(X\) of dimension \(n \geq 0\) over the field \(k\) of arbitrary characteristic. Then \(-K_X\) is numerically equivalent to an effective \(\mathbb{Q}\)-Cartier divisor. In particular, if \(\mathrm{Alb}(X)\) is trivial, then the Iitaka \(D\)-dimension \(\kappa (X, -K_X ) \geq 0\). 3. They give an affirmative answer to a question of \textit{H. Krieger} and \textit{P. Reschke} [Bull. Soc. Math. Fr. 145, No. 3, 449--468 (2017; Zbl 1387.32022)] and with a very different proof, generalize a theorem of \textit{S. Meng} and \textit{D.-Q. Zhang} [Adv. Math. 325, 243--273 (2018; Zbl 1387.14057)] to arbitrary characteristics. 4. They prove: If \(f: X \rightarrow X\) is a numerically polarized separable endomorphism of a normal projective variety \(X\) with \( K_X\) being pseudo-effective and \(\mathbb{Q}\)-Cartier, then \(f\) is quasi-étale and \(K_X\sim_\mathbb{Q} 0 \). 5. Let \(f^{\mathrm{Gal}}: \bar{X} \rightarrow X\) be the Galois closure of \( f\). Nakayama's result on a normal projective surface \(X\) defined over a field of characteristic 0 is generalized to char \(p>5\) and \(p\nmid\) deg \(f^{\mathrm{Gal}}\): If \(f: X \rightarrow X\) is a polarized endomorphism of a normal projective surface \(X\) over the field \(k\) of characteristic \(p > 5\). Suppose \(p\nmid\) deg \( f^{\mathrm{Gal}}\) and \(K_X\) is pseudo-effective, then \(X\) is a \(\mathbb{Q}\)-abelian surface. In particular, \(X\) is \(\mathbb{Q}\)-factorial and klt. 6. They show: If \( p > 5\) and co-prime to degree \(f^{\mathrm{Gal}}\), then one can run the minimal model program (MMP) \(f\)-equivariantly, after replacing \(f\) by a positive power, for a mildly singular threefold \(X\) and reach a variety \(Y\) with torsion canonical divisor (and also with \( Y\) being a quasi-étale quotient of an abelian variety when dim\((Y ) \leq 2\)). They also show that a power of \(f\) acts as a scalar multiplication on the Néron-Severi group of \( X\) (modulo torsion) when \(X\) is a smooth and rationally chain connected projective variety of dimension at most three.