Let $F$ be a smooth complete intersection of $r$ hypersurfaces of degree $d_{1},\dots,d_{r}$ in $\mathbb{P}^{n+r}$ over an algebraic closed field. A classical result of A. Roitman says that the group of zero-cycles of degree zero of $F$ is trivial: $$\text{CH}_{0}^{\text{deg 0}}(F)=0$$ if $\sum d_{i}\leq n+r$.

Now, for an equidimensional quasi projective scheme $X$, Bloch defines Higher Chow groups, see here for a quick introduction, $\text{CH}^{q}(X,m)$. Let $q=\dim F+n$ and consider $\text{CH}^{q}(F,n)$.

I have the following (very vague) questions:

  • What should be the right analogue of the result given by Roitman in this context?

  • Are there known results in this direction?

    For example, is $\ker(\text{CH}^{q}(F,n)\to \text{CH}^{n}(k,n))$ finitely generated? Note that if $n=0$ and $\sum d_{i}\leq n+r$, we obtain the above result.

Thank you.


Edit. What I originally wrote was wrong, because I was ignoring the (huge) K-theory of the ground field $k$. I corrected this below.

Let $k$ be an algebraically closed field. Let $X$ be a proper $k$-scheme of pure dimension $d$. Denote by $\pi$ the following unique $k$-morphism, $$\pi:X\to \text{Spec}\ k,$$ and denote by $i$ any $k$-point, $$i:\text{Spec}\ k \to X.$$ Denote the image of $i$ by $\widetilde{x}$, a specified base point. Because the higher Chow groups are covariant for proper morphisms, there are induced maps, $$\text{CH}^{q}(i,n):\text{CH}^q(\text{Spec}\ k,n) \to \text{CH}^{d+q}(X,n),$$ $$\text{CH}^{d+q}(\pi,n):\text{CH}^{d+q}(X,n)\to \text{CH}^q(\text{Spec}\ k,n).$$ The composition $\text{CH}^{d+q}(\pi,n)\circ \text{CH}^q(i,n)$ is $\text{CH}^q(\text{Id}_{\text{Spec}\ k},n)$, which is the identity map.

Definition. A $k$-scheme $X$ is strongly $\mathbb{A}^1$-connected if for every pair $(x_0,x_1)$ in $X\times_{\text{Spec}\ k}X(k),$ there exists a $k$-morphism, $$u:(\mathbb{A}^1_k,0,1)\to (X,x_0,x_1).$$

Proposition. For every strongly $\mathbb{A}^1$-connected, proper $k$-scheme $X$ of dimension $d$, for every integer $n$, the morphisms $\text{CH}^{d+n}(\pi,n)$ and $\text{CH}^n(i,n)$ are inverse isomorphisms.

Proof. Via covariance, the composition $\text{CH}^{d+n}(\pi,n)\circ \text{CH}^n(i,n)$ equals the identity map. Thus, it suffices to prove that $\text{CH}^n(i,n)$ is surjective. Since $Z^{d+n}(X,n)$ has as $\mathbb{Z}$-basis the classes of $k$-points of $X\times \Delta^n$ that are in the complement of the face maps, it suffices to prove that every such class is cohomologous to the class of a $k$-point in the image of $i\times \text{Id}_{\Delta_n}$.

Let $(x,a)$ be a $k$-point of $X\times \Delta^n$ that is not in the image of the face maps, i.e.,
$$a=(a_0,\dots,a_{n-1},a_n)\in \Delta_{n} \setminus \cup_i \text{Image}(\delta^i_{n-1}), \ \ a_0+\dots+a_n=1, \ \ a_0\cdots a_n \neq 0.$$ For a the $k$-points $x$ and $\widetilde{x}$ of $X$, since $X$ is strongly $\mathbb{A}^1$-connected, there exists a morphism $u:(\mathbb{A}^1,0,1)\to (X,x,\widetilde{x})$. Now extend this to a morphism $$u:\mathbb{A}^1\to X\times \Delta_{n+1},\ \ \ t\mapsto (u(t),a_0,\dots,a_{n-1},a_nt,a_n(1-t)). $$ This is a closed immersion; in fact the composition with the projection to the last two factors is already a closed immersion. By construction, this does intersect transversally the faces of $X\times \Delta_{n+1}.$ Thus the class of the image is an element of $Z^{d+n}(X,n+1)$. Up to a sign, the differential of this class equals $[(x,a)]-[(\widetilde{x},a)].$ Thus, the class $[(x,a)]$ is cohomologous to the class $[(\widetilde{x},a)]$. Since $\widetilde{x}$ is the image of $i$, the class $[(\widetilde{x},a)]$ is in the image of $\text{CH}^n(i,n).$ QED

Note. For $n$ equal to $0,$ the Chow group $\text{CH}^0(\text{Spec}\ k,0)$ is isomorphic to $\mathbb{Z}$ via the degree homomorphism. Thus, for $n$ equal to $0$, the Proposition is Roitman's result. Based on the works of Yi Zhu, specifically the following, I also suspect that the argument above works for quasi-projective $k$-schemes $X$ that are strongly $\mathbb{A}^1$-connected, but where the higher Chow groups are replaced by Suslin's groups.

Yi Zhu.
$\mathbb{A}^1$-equivalence of zero cycles on surfaces.
To appear: Transactions of the AMS.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.