Intersections on Surfaces

The quadratic and an axis.

(Taken from my notes on intersection theory)

Let X=P2X= \mathbb P^2, CC be the vanishing set of the global section x2yzx^2-yz of O(2)\mathcal O(2), and DD be the vanishing set of the global section yy of O(1)\mathcal O(1) (here by vanishing of ff we mean the set of all points PP where fmPf\in \mathfrak m_P). The intersection CDC\cap D contains all the points where yy and x2yzx^2-yz are in mP\mathfrak m_P. Consider the affine patch {z0}\{z\neq0\}, and suppose that CDC\cap D has a point PP outside that open set. Clearly the point P=[0:0:1]P=[0:0:1] in this patch is in CDC\cap D. We seek to compute the intersection multiplicity at this point. For this it will suffice to work in the affine plane, with the equations x2y=0x^2-y = 0 and y=0y = 0. Then we compute OX,P=k[x,y](x,y)\mathcal O_{X, P} = k[x, y]_{(x, y)} and

OX,P/(x2y,y)=k[x]x/(x2)kk,\mathcal O_{X,P}/(x^2-y, y) = k[x]_{x}/(x^2) \cong k\oplus k,

and so (CD)P=2(C\cdot D)_P=2. This confirms our intuition from calculus that the multiplicity of the quadratic’s zero is 2. But are there no other points in the intersection? There are no other intersections in this affine patch, so it remains only to check the points at infinity. If zmPz\in \mathfrak m_P, then x2yzmPx^2-yz\in \mathfrak m_P implies xmPx\in\mathfrak m_P by primality, so if ymPy\in \mathfrak m_P we have that x,y,zmPx, y, z\in \mathfrak m_P. But (x,y,z)∉Projk[x,y,z](x, y, z)\not\in \mathop{\mathrm{Proj}}k[x, y, z] as we exclude the irrelevant ideal, so there is no intersection at infinity and we have verified Bézout’s theorem in this case.