Why is there this exact sequence

76 Views Asked by At

On page 49 in Hartshornes "Deformation Theory", in the proof of theorem 6.4 he states that the following exact sequence $0\to J\otimes O_{X}\to O_{X_{1}} \to O_{X}\to 0$ gives rise to an exact sequence of abelian groups $0\to J\otimes O_{X}\to O^*_{X_{1}} \to O^*_{X}\to 0$, so that the left group is just the additive group and the groups in the middle and on the right are the corresponding groups of units. I would like to know why is there this exact sequence, as well as when in general you can do something like this.

Setting: $X$ is a scheme flat over a local artin ring $C$. $C'$ is another local artin ring that surjects to $C$ with kernel $J$ (that has square 0). $X_{1}$ is an extension of $X$ over $C'$, that is: there is a closed imbedding $X\to X_{1}$ where $X_{1}$ is flat over $C'$ and it holds that $X=X'\times_{C'} C$.

1

There are 1 best solutions below

1
On

Let me do this for rings, other case being similar. Let $A$ be any ring (commutative with identity) and $J\subset A$ an ideal with $J^2=0$. Then, I claim that there is an exact sequence $0\to J\to A^*\to (A/J)^*\to 0$ of abelian groups. The only observation is that if $j\in J$, $1+j$ is a unit in $A$, since $(1+j)(1-j)=1$.

The right side map is obvious and the left side map is $j\mapsto 1+j$. Since $1+(j+k)=(1+j)(1+k)$, we see that this indeed a map of abelian groups. Surjectivity on the right is clear, since any lift of a unit in $A/J$ is a unit in $A$. Exactness is clear too, since if $a\mapsto 1\in (A/J)^*$, this means $a=1+j$ with $j\in J$.