On the motivic Peterson conjecture (Q721502)

The graded polynomial algebra \(P_n:=\mathbb F_2[x_1,x_2,\ldots,x_n]\) over the two-element field \(\mathbb{F}_2\) with \(|x_j|=1\) for all \(j\), has a natural \(\mathcal{A}\)-module structure (where \(\mathcal{A}\) is the mod \(2\) Steenrod algebra), because it is the mod \(2\) cohomology algebra of the product of \(n\) copies of the infinite projective space \(\mathbb R\mathrm P^\infty\). Let \(P_n^d\) be the subspace of \(P_n\) spanned by monomials of degree \(d\), and let \(QP_n^d=P_n^d/(\mathcal A_+(P_n)\cap P_n^d)\), where \(\mathcal A_+(P_n)\) is the subspace of \(P_n\) consisting of \textit{hit polynomials} -- the finite sums \(\sum_{i>0}Sq^i(f_i)\), \(f_i\in P_n\). The \textit{hit problem} is the problem of finding a basis for \(QP_n^d\) (for all \(n\) and \(d\)). In [\textit{R. M. W. Wood}, Math. Proc. Camb. Philos. Soc. 105, No. 2, 307--309 (1989; Zbl 0692.55011)] the author resolved positively the Peterson conjecture: if \(\beta(d)>n\), then \(QP_n^d=\{0\}\), where \(\beta(d)\) is the minimal \(s\) such that \(d=(2^{i_1}-1)+(2^{i_2}-1)+\cdots+(2^{i_s}-1)\) for some \(i_1,i_2,\ldots,i_s\in\mathbb N\). In the present paper the author shows that the analogue of the Peterson conjecture does not hold in the motivic cohomology. The mod \(2\) motivic cohomology of \((\mathbb R\mathrm P^\infty)^n\) is \(M_n=\mathbb F_2[\tau,x_1,\ldots,x_n,y_1,\ldots,y_n]/(x_1^2+\tau y_1,\ldots,x_n^2+\tau y_n)\), where \(\deg\tau=(0,1)\), \(\deg x_j=(1,1)\) and \(\deg y_j=(2,1)\). Similarly as above, \(QM_n^{d,*}=M_n^{d,*}/(\mathcal A_+^{*,*}(M_n)\cap M_n^{d,*})\), where \(M_n^{d,*}\) is the subspace of \(M_n\) spanned by elements of degree \((d,*)\), and \(\mathcal A_+^{*,*}\) is the set of elements of the mod \(2\) motivic Steenrod algebra \(\mathcal A^{*,*}\) with nonzero degree. The main theorem of the paper reads as follows: if \(n\) is such that \(\alpha(n-2)\geq3\) (as usual, \(\alpha(n-2)\) is the number of ones in the binary expansion of \(n-2\)), \(k=n-3\) and \(d=(n-1)(2^{k+1}-2)+k\), then \(\beta(d)>n\) but \(QM_n^{d,*}\neq\{0\}\).