Loading [MathJax]/jax/output/HTML-CSS/jax.js

Wednesday, April 16, 2014

Computation of Algebraic Sets and Coordinate Rings (15.1.24)

MathJax TeX Test Page Let k be an infinite field, and let V=Z(xyz). Prove that V is isomorphic to A2 and provide an explicit isomorphism φ from V to A2 and associated k-algebra isomorphism Φ from k[V] to k[A2] and their inverses. Is V=Z(xyz2) isomorphic to A2?

Proof: We see V is the set of all points of the form (x,y,xy), so there is an isomorphism φ : VW given by φ(x,y,xy)=(x,y) with inverse ψ : WV given by ψ(x,y)=(x,y,xy). We see (xyz)I(V) naturally and by reducing modulo (xyz) any fI(V) to a polynomial in k[x,y] we see the remainder must be 0 as it would annihilate all of A2, hence I(V)(xyz). Thus this induces the isomorphism Φ : k[W]k[V] where k[V]=k[x,y,z]/(xyz) and k[W]=k[x,y] given by Φ(f(x,y))=¯f(x,y) with inverse Ψ : k[V]k[W] given by Ψ(¯f(x,y,z))=f(x,y).

For the second part, we shall show k[V]k[W] where W=A2 so that VW as algebraic sets. First, we have (xyz2)I(V). Assume I(V)(xyz2) and let fI(V)(xyz2) be minimal with respect to the lexicographic ordering z>y>x, so f=z·g1(x,y)+g2(x,y).

Let k2 be the subset of nonzero squares in k, necessarily infinite since each square has only a finite number of square roots, and for each nonzero bk let k2/b={v/b | vk2}. Note V is the set of points of the form (x,y,xy) whenever xy is square, so for each nonzero bk we have an infinite number of ak2/b such that f(a,b,ab)=f(a,b,ab)=(ab)g1(a,b)+g2(a,b)=(ab)g1(a,b)+g2(a,b)=0 Since ab is nonzero, for each such b we have g1(x,b)k[x] has an infinite number of zeros x=a, so g1(x,b)=0 for these b. Since there are an infinite number of such b, we have g1(x,y)k(x)[y] has an infinite number of roots y=b in kk(x), hence g1(x,y)=0 and by the same argument we find g2(x,y)=0. Hence f=0 and we have shown I(V)=(xyz2).

We see k[W]=k[x,y] is a UFD (cf. 9.3.8), and we shall prove k[V]k[W] by showing k[V]=k[x,y,z]/(xyz2) is not a UFD. Assume otherwise and observe the element xk[V], which we shall prove is prime; we may write any element in k[V] uniquely as a polynomial in k[x,y][z] of degree 1, so assume x=(g1+z·g2)(g3+z·g4). Then x=g1g3+xy·g2g4+z(g1g4+g2g3) is the unique reduced form, so g1g3+xy·g2g4=x and g1g4+g2g3=0. Taking the first relation modulo x we may assume without loss that x | g1, and now taking the second relation modulo x necessarily either x | g2 or x | g3; the former implies x | (g1+z·g2), and the latter that the relation g1g3+xy·g2g4 involves adding monomial terms of degree strictly larger than x, hence the expression does not equal x. Thus x must be prime. But xy=z2 would imply x | z, a contradiction by observing unique representation. Hence k[V] is not a UFD and not isomorphic to k[W], and so neither is V isomorphic to A2. 

No comments:

Post a Comment