{"id":1811,"date":"2022-06-22T11:25:43","date_gmt":"2022-06-22T14:25:43","guid":{"rendered":"http:\/\/localhost\/?page_id=1811"},"modified":"2022-06-22T14:25:42","modified_gmt":"2022-06-22T17:25:42","slug":"lying-over-going-down","status":"publish","type":"page","link":"http:\/\/localhost\/index.php\/ensino\/algebra-comutativa\/lying-over-going-down\/","title":{"rendered":"Lying over, going down"},"content":{"rendered":"
Seja \\(S\\)<\/span> uma \\(R\\)<\/span>-\u00e1lgebra, com morfismo \\(\\varphi:R\\to S\\)<\/span>, e seja \\(\\mathfrak p\\in\\mbox{Spec}(S)\\)<\/span>. Ent\u00e3o \\(\\varphi^{-1}(\\mathfrak p)\\)<\/span> pertence ao \\(\\mbox{Spec}(R)\\)<\/span>. Dizemos que o ideal \\(\\mathfrak p\\)<\/span> est\u00e1 acima de<\/em><\/span> \\(\\varphi^{-1}(\\mathfrak p)\\)<\/span>. Quando \\(R\\subseteq S\\)<\/span> e \\(\\varphi\\)<\/span> \u00e9 a inclus\u00e3o, temos que \\(\\varphi^{-1}(\\mathfrak p)=R\\cap \\mathfrak p\\)<\/span>.<\/p>\n Se \\(\\mathfrak q\\)<\/span> est\u00e1 acima de \\(\\mathfrak p\\)<\/span> ent\u00e3o \\(\\mathfrak q\\)<\/span> \u00e9 maximal se e somente se \\(\\mathfrak p\\)<\/span> \u00e9 maximal.<\/p>\n<\/li>\n Assuma que \\(\\mathfrak q_1\\)<\/span> e \\(\\mathfrak q_2\\)<\/span> est\u00e3o acima de \\(\\mathfrak p\\)<\/span>. Ent\u00e3o \\(\\mathfrak q_1=\\mathfrak q_2\\)<\/span>.<\/p>\n<\/li>\n Existe \\(\\mathfrak r\\)<\/span> que est\u00e1 acima de \\(\\mathfrak p\\)<\/span>.<\/p>\n<\/li>\n Assuma que \\(\\mathfrak a\\cap R\\subseteq \\mathfrak p\\)<\/span>. Ent\u00e3o existe \\(\\mathfrak r\\)<\/span> em (3) tal que \\(\\mathfrak a\\subseteq \\mathfrak r\\)<\/span>.<\/p>\n<\/li>\n<\/ol>\n<\/div>\n (2) Assuma primeiro que \\((R,\\mathfrak p)\\)<\/span> \u00e9 um anel local. Ent\u00e3o \\(\\mathfrak q_1\\)<\/span> e \\(\\mathfrak q_2\\)<\/span> s\u00e3o maximais por item (1) e segue que \\(\\mathfrak q_1=\\mathfrak q_2\\)<\/span>.<\/p>\n No caso geral, seja \\(U=R\\setminus\\mathfrak p\\)<\/span> e considere as localiza\u00e7\u00f5es \\(R_{(\\mathfrak p)}=U^{-1}R\\)<\/span> e \\(U^{-1}S\\)<\/span>. O mapa de mergulho \\(U^{-1}R\\to U^{-1}S\\)<\/span> \u00e9 injetivo, pois se \\(r\/u\\in U^{-1}R\\)<\/span> tal que \\(r\/u=0\/1\\in U^{-1}S\\)<\/span> ent\u00e3o \\(rw=0\\)<\/span> com algum \\(w\\in R\\setminus \\mathfrak p\\)<\/span> e \\(r\/u=0\/1\\in U^{-1}R\\)<\/span>. Ent\u00e3o \\(U^{-1}R\\subseteq U^{-1}S\\)<\/span> \u00e9 uma extens\u00e3o de an\u00e9is. Afirmamos que, a extens\u00e3o \\(U^{-1}R\\subseteq U^{-1}S\\)<\/span> \u00e9 integral. De fato, se \\(s\/u\\in U^{-1}S\\)<\/span>, ent\u00e3o \\(U^{-1}R[s\/u]=U^{-1}R[s]\\)<\/span>. Mas \\(s\\)<\/span> \u00e9 integral sobre \\(R\\)<\/span> e satisfaz uma equa\u00e7\u00e3o integral \\[s^n+\\beta_1s^{n+1}+\\cdots+\\beta_n=0\\]<\/span> com coeficientes \\(\\beta_i\\in R\\)<\/span>. A mesma equa\u00e7\u00e3o ser\u00e1 uma equa\u00e7\u00e3o integral para \\(s\/1\\)<\/span> com coeficientes em \\(U^{-1}R\\)<\/span>. Portanto \\(U^{-1}R[s\/u]= Como \\(U^{-1}R=R_{(\\mathfrak p)}\\)<\/span> \u00e9 local com \u00fanico ideal maximal \\(\\mathfrak pU^{-1}R\\)<\/span> e como \\(\\mathfrak pU^{-1}R\\subseteq \\mathfrak q_i U^{-1}S\\cap U^{-1}R\\)<\/span> para \\(i\\in\\{1,2\\}\\)<\/span>, e o lema anterior implica que \\(\\mathfrak q_i U^{-1}S\\)<\/span> est\u00e1 acima de \\(\\mathfrak pU^{-1}R\\)<\/span>. Ora, o par\u00e1grafo anterior implica que \\(\\mathfrak q_1U^{-1}S=\\mathfrak q_2U^{-1}S\\)<\/span>. Como \\(\\mathfrak q_i\\cap U=\\emptyset\\)<\/span>, e a correspond\u00eancia entre os primos em \\(\\mbox{Spec}(S)\\)<\/span> que interceptam trivialmente com \\(U\\)<\/span> e os primos de \\(U^{-1}S\\)<\/span> implica que \\(\\mathfrak q_1=\\mathfrak q_2\\)<\/span>.<\/p>\n (3) Assuma primeiro que \\((R,\\mathfrak p)\\)<\/span> \u00e9 um anel local. O anel \\(S\\)<\/span> tem um ideal maximal \\(\\mathfrak q\\)<\/span> e \\(\\mathfrak q\\cap R\\)<\/span> \u00e9 maximal por afirma\u00e7\u00e3o\u00a0(1). Logo \\(\\mathfrak q\\cap R\\)<\/span> deve ser igual a \\(\\mathfrak p\\)<\/span>.<\/p>\n No caso geral, considere as localiza\u00e7\u00f5es \\(U^{-1}R\\)<\/span> e \\(U^{-1}S\\)<\/span> como na demonstra\u00e7\u00e3o da afirma\u00e7\u00e3o\u00a0(2). Assuma que \\(\\mathfrak r'\\)<\/span> \u00e9 um ideal maximal de \\(U^{-1}S\\)<\/span>. Ent\u00e3o \\(\\mathfrak r'\\cap U^{-1}R\\)<\/span> maximal em \\(U^{-1}R\\)<\/span> e portanto \\(\\mathfrak r'\\cap U^{-1}R=\\mathfrak pU^{-1}R\\)<\/span>. Definimos \\(\\mathfrak r=\\varphi_S^{-1}(\\mathfrak r')\\)<\/span>. J\u00e1 monstramos da demonstra\u00e7\u00e3o da afirma\u00e7\u00e3o\u00a0(2) que o mapa can\u00f4nico \\(U^{-1}R\\to U^{-1}S\\)<\/span> \u00e9 uma inclus\u00e3o e assim temos o seguinte diagrama comutativo: (4) Aplique (3) para o extens\u00e3o \\(R\/(\\mathfrak a\\cap R)\\subseteq S\/\\mathfrak a\\)<\/span>.<\/div>\n Existe uma extens\u00e3o \\(T\\)<\/span> de \\(R\\)<\/span> tal que \\(f(x)=\\prod (x-x_i)\\)<\/span> em \\(T[x]\\)<\/span>. Al\u00e9m disso, \\(T\\)<\/span> \u00e9 \\(R\\)<\/span>-m\u00f3dulo livre de posto \\(d!\\)<\/span> onde \\(d=\\deg f\\)<\/span>.<\/p>\n<\/li>\n \\(h\\)<\/span> \u00e9 m\u00f4nico e os coeficientes de \\(g\\)<\/span> e \\(h\\)<\/span> s\u00e3o integrais sobre \\(R\\)<\/span>.<\/p>\n<\/li>\n<\/ol>\n<\/div>\n (2) O coeficiente l\u00edder de \\(f\\)<\/span> \u00e9 o produto dos coeficientes l\u00edderes de \\(f\\)<\/span> e \\(g\\)<\/span>. Como \\(f\\)<\/span> \u00e9 m\u00f4nico, \\(g\\)<\/span> \u00e9 m\u00f4nico, obtemos que \\(h\\)<\/span> \u00e9 m\u00f4nico. Aplicando (1), obtemos uma extens\u00e3o \\(Q_1\\)<\/span> de \\(S\\)<\/span> tal que \\(g=\\prod (x-y_i)\\)<\/span> e uma extens\u00e3o \\(Q_2\\)<\/span> de \\(S\\)<\/span> tal que \\(h=\\prod(x-z_i)\\)<\/span>. Os elementos \\(y_i\\)<\/span> e \\(z_i\\)<\/span> s\u00e3o integrais sobre \\(R\\)<\/span>, pois s\u00e3o ra\u00edzes de \\(f\\)<\/span>. Mas os coeficientes de \\(g\\)<\/span> e \\(h\\)<\/span> s\u00e3o polin\u00f4mios em \\(y_i\\)<\/span> e \\(z_i\\)<\/span> e eles tamb\u00e9m s\u00e3o integrais sobre \\(R\\)<\/span>.<\/div>\n\n
\n &(s+\\mathfrak q)^n+(\\beta_1+\\mathfrak p)(s+\\mathfrak q)^{n-1}+\\cdots+(\\beta_n+\\mathfrak p)\\\\&=
\n (s+\\mathfrak q)^n+(\\beta_1+\\mathfrak q)(s+\\mathfrak q)^{n-1}+\\cdots+(\\beta_n+\\mathfrak q)\\\\&=
\n (s^n+\\beta_1s^{n-1}+\\cdots+\\beta_n)+\\mathfrak q=0.
\n \\end{aligned}\\]<\/span> Portanto, \\(R\/\\mathfrak p\\subseteq S\/\\mathfrak q\\)<\/span> \u00e9 integral. Por um lema anterior, um \u00e9 corpo se e somente se outro \u00e9.<\/p>\n
\n U^{-1}[s]\\)<\/span> \u00e9 finitamente gerado como \\(U^{-1}R\\)<\/span>-m\u00f3dulo e \\(s\/u\\)<\/span> \u00e9 integral sobre \\(U^{-1}R\\)<\/span>.<\/p>\n
\n\\[
\n\\begin{CD}
\n R @>>> S \\\\
\n\t\t@V{\\varphi_R}VV @VV{\\varphi_S}V\\\\
\n U^{-1}R @>>> U^{-1}S
\n \\end{CD}\\]<\/span> Calculando a pr\u00e9-imagem em \\(R\\)<\/span> de \\(\\mathfrak r'\\subseteq U^{-1}S\\)<\/span> pela composi\u00e7\u00e3o dos mapas das duas maneiras no diagrama, obtemos que \\[R\\cap \\mathfrak r=R\\cap \\varphi_S^{-1}(\\mathfrak r')=\\varphi_R^{-1}(\\mathfrak r'\\cap U^{-1}R)=\\varphi_{R}^{-1}(\\mathfrak pU^{-1}R)=\\mathfrak p\\]<\/span><\/p>\n\n