On some arithmetical properties of solutions of decomposable form equations (Q5315876)

A homogeneous polynomial \(F(\underline x)=F(x_1,\dots,x_n)\in \mathbb{Z}[x_1,\dots, x_n]\) is said to be a decomposable form if it factorizes over \(\mathbb{C}\) as a product of linear forms. It is known that, for a decomposable form \(F\) as above, there exist number-fields \(M_1,\dots, M_t\), linear forms \(\varphi_i(\underline x) \in M_i[\underline x]\) \((i=1, \dots,t)\) and \(q\in\mathbb{Q}^*\) such that \(F (\underline x)\) factorizes as \(F(\underline x)=q\prod^t_{i=1}N_{M_i/\mathbb{Q}} (\varphi(\underline x))\). If, moreover, for every \(i=1,\dots,t\) the number field \(M_i\) has no subfield of unit rank 1 and is either totally real or a totally imaginary extension of a totally real field, then \(F\) is said to be of CM type. The Diophantine equation \[ F(\underline x)=a,\quad \underline x\in\mathbb{Z}^n \tag{1} \] has already been studied by a number of authors, the present ones included among them. In this paper the authors assume that the following two conditions are satisfied: (i) \(F\) contains \(n\) linearly independent forms among its factors. (ii) Equation (1) has infinitely many solutions. For a ``large'' positive real number \(T\) they consider the set \[ F(a,T)=\bigl\{\underline x\text{ is a solution to (1) and }|\underline x |<t\bigr\}, \] where \(|\underline x|\) denotes the max-norm of \(\underline x\). Moreover, let \(S\) be the set of solutions to (1). For a non-negative real number \(\delta\leq 1\) the subset \(V\) of \(S\) is said to have density \(\delta\) if \[ \lim_{T\to\infty}\frac{|\{\underline x \in V:| \underline x|<T\}|}{|\{\underline x\in S:|\underline x|<T\}|}=\delta. \] The first result of the paper is the following theorem \((P(\cdot)\) denotes greatest prime factor): Let the decomposable form \(F\) be of CM type and fix some index \(i\in\{1,\dots,n\}\). There exists an effectively computable constant \(C_i>0\) such that the set of solutions \(\underline x\) to (1) with \(|x_i|\to \infty\) and \(P(x_i)>C_i\log\log|x_i|\) has density 1. A byproduct of the proof of this theorem is the following: If \(V\) is the set of solutions to (1) with \(|x_i| \to\infty\) and \(P(x_i)\leq C_i\log\log |x_i|\) then \[ \frac{|\{\underline x\in V:| \underline x|<T\}|}{| \{\underline x\in S:|\underline x|<T\}|}=O\left(\frac{1} {\log\log T} \right) \] as \(T\to\infty\). The third result proved in this paper is the following: Fix a non-zero integer \(w\). Let \(F\) be a decomposable form, as before, of CM type, fix an index \(i\in\{1,\dots,n\}\) and let \(V\) be the set of solutions to (1) with \(|x_i|\to\infty\) and with the extra property that \(x_i=wy^k\) for some \(y\in\mathbb{Z}\) and some integer \(k>1\). Then \(V\) has density 1.