06.08.2013 Views

Commutative algebra - Department of Mathematical Sciences - old ...

Commutative algebra - Department of Mathematical Sciences - old ...

Commutative algebra - Department of Mathematical Sciences - old ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

ELEMENTARY<br />

COMMUTATIVE ALGEBRA<br />

LECTURE NOTES<br />

H.A. NIELSEN<br />

DEPARTMENT OF MATHEMATICAL SCIENCES<br />

UNIVERSITY OF AARHUS<br />

2005


Elementary <strong>Commutative</strong> Algebra<br />

H.A. Nielsen


Contents<br />

Prerequisites 7<br />

1. A dictionary on rings and ideals 9<br />

1.1. Rings 9<br />

1.2. Ideals 11<br />

1.3. Prime ideals 13<br />

1.4. Chinese remainders 14<br />

1.5. Unique factorization 15<br />

1.6. Polynomials 16<br />

1.7. Roots 18<br />

1.8. Fields 19<br />

1.9. Power series 20<br />

2. Modules 21<br />

2.1. Modules and homomorphisms 21<br />

2.2. Submodules and factor modules 23<br />

2.3. Kernel and cokernel 25<br />

2.4. Sum and product 28<br />

2.5. Homomorphism modules 30<br />

2.6. Tensor product modules 33<br />

2.7. Change <strong>of</strong> rings 36<br />

3. Exact sequences <strong>of</strong> modules 39<br />

3.1. Exact sequences 39<br />

3.2. The snake lemma 43<br />

3.3. Exactness <strong>of</strong> Hom 48<br />

3.4. Exactness <strong>of</strong> Tensor 49<br />

3.5. Projective modules 50<br />

3.6. Injective modules 52<br />

3.7. Flat modules 54<br />

4. Fraction constructions 57<br />

4.1. Rings <strong>of</strong> fractions 57<br />

4.2. Modules <strong>of</strong> fractions 58<br />

4.3. Exactness <strong>of</strong> fractions 60<br />

4.4. Tensor modules <strong>of</strong> fractions 62<br />

4.5. Homomorphism modules <strong>of</strong> fractions 63<br />

4.6. The polynomial ring is factorial 64<br />

5. Localization 65<br />

5.1. Prime ideals 65<br />

5.2. Localization <strong>of</strong> rings 67<br />

5


6 CONTENTS<br />

5.3. Localization <strong>of</strong> modules 68<br />

5.4. Exactness and localization 70<br />

5.5. Flat ring homomorphisms 71<br />

6. Finite modules 73<br />

6.1. Finite Modules 73<br />

6.2. Free Modules 75<br />

6.3. Cayley-Hamilton’s theorem 77<br />

6.4. Nakayama’s Lemma 78<br />

6.5. Finite Presented Modules 80<br />

6.6. Finite ring homomorphisms 83<br />

7. Modules <strong>of</strong> finite length 85<br />

7.1. Simple Modules 85<br />

7.2. The Length 85<br />

7.3. Artinian Rings 87<br />

7.4. Localization 90<br />

7.5. Local artinian ring 91<br />

8. Noetherian modules 93<br />

8.1. Modules and submodules 93<br />

8.2. Noetherian rings 94<br />

8.3. Finite type rings 95<br />

8.4. Power series rings 97<br />

8.5. Fractions and localization 98<br />

8.6. Prime filtrations <strong>of</strong> modules 98<br />

9. Primary decomposition 101<br />

9.1. Support <strong>of</strong> modules 101<br />

9.2. Ass <strong>of</strong> modules 103<br />

9.3. Primary modules 106<br />

9.4. Decomposition <strong>of</strong> modules 106<br />

9.5. Decomposition <strong>of</strong> ideals 108<br />

10. Dedekind rings 111<br />

10.1. Principal ideal domains 111<br />

10.2. Discrete valuation rings 112<br />

10.3. Dedekind domains 113<br />

Bibliography 117<br />

Index 119


Prerequisites<br />

The basic notions from <strong>algebra</strong>, such as groups, rings, fields and their homomorphisms<br />

together with some linear <strong>algebra</strong>, bilinear forms, matrices and determinants.<br />

Linear <strong>algebra</strong>: Fraleigh & Beauregard, Linear <strong>algebra</strong>, New York 1995.<br />

Algebra: Niels Lauritzen, Concrete abstract <strong>algebra</strong>, Cambridge 2003.<br />

Also recommended: Jens Carsten Jantzen, Algebra 2, Aarhus 2004.<br />

The propositions are stated complete and precise, while the pro<strong>of</strong>s are quite short.<br />

No specific references to the literature are given. But lacking details may all be<br />

found at appropriate places in the books listed in the bibliography.<br />

Nielsen, University <strong>of</strong> Aarhus, Winter 2004<br />

7


1<br />

A dictionary on rings and ideals<br />

1.1. Rings<br />

1.1.1. Definition. An abelian group is a set A with an addition A×A → A, (a, b) ↦→<br />

a + b and a zero 0 ∈ A satisfying<br />

(1) associative: (a + b) + c = a + (b + c)<br />

(2) zero: a + 0 = a = 0 + a<br />

(3) negative: a + (−a) = 0<br />

(4) commutative: a + b = b + a<br />

for all a, b, c ∈ A. A subset B ⊂ A is a subgroup if 0 ∈ B and a − b ∈ B for<br />

all a, b ∈ B. The factor group A/B is the abelian group whose elements are the<br />

cosets a + B = {a + b|b ∈ B} with addition (a + B) + (b + B) = (a + b) + B. A<br />

homomorphism <strong>of</strong> groups φ : A → C respects addition φ(a + b) = φ(a) + φ(b).<br />

The projection π : A → A/B, a ↦→ a + B is a homomorphism. If φ(b) = 0 for all<br />

b ∈ B, then there is a unique homomorphism φ ′ : A/B → C such that φ = φ ′ ◦ π.<br />

1.1.2. Definition. A ring is an abelian group R, addition (a, b) ↦→ a + b and zero<br />

0, together with a multiplication R × R → R, (a, b) ↦→ ab and an identity 1 ∈ R<br />

satisfying<br />

(1) associative: (ab)c = a(bc)<br />

(2) distributive: a(b + c) = ab + ac, (a + b)c = ac + bc<br />

(3) identity : 1a = a = a1<br />

(4) commutative : ab = ba<br />

for all a, b, c ∈ R. If (4) is not satisfied then R is a noncommutative ring. A<br />

subring R ′ ⊂ R is an additive subgroup such that 1 ∈ R ′ and ab ∈ R ′ for all<br />

a, b ∈ R ′ . The inclusion R ′ ⊂ R is a ring extension. A homomorphism <strong>of</strong> rings<br />

φ : R → S is an additive group homomorphism respecting multiplication and<br />

identity<br />

φ(a + b) = φ(a) + φ(b), φ(ab) = φ(a)φ(b), φ(1) = 1<br />

An isomorphism is a homomorphism φ : R → S having an inverse map φ −1 :<br />

S → R which is also a homomorphism. The identity isomorphism is denoted<br />

1R : R → R.<br />

1.1.3. Remark. (1) A bijective ring homomorphism is an isomorphism.<br />

(2) Recall the usual formulas: a + (−b) = a − b, 0a = 0, (−1)a = −a.<br />

(3) The identity 1 is unique.<br />

(4) A ring R is nonzero if and only if the elements 0 = 1.<br />

(5) If φ : R → S is a ring homomorphism, then φ(0) = 0 and φ(R) is a subring<br />

<strong>of</strong> S.<br />

(6) The unique additive group homomorphism Z → R, 1 ↦→ 1 is a ring homomorphism.<br />

9


10 1. A DICTIONARY ON RINGS AND IDEALS<br />

1.1.4. Proposition. Let R1, R2 be rings. The product ring is the product <strong>of</strong> additive<br />

groups R1×R2, ((a1, a2), (b1, b2)) ↦→ (a1+b1, a2+b2), with coordinate multiplication<br />

((a1, a2), (b1, b2)) ↦→ (a1b1, a2b2). The element (1, 1) is the identity. The<br />

projections R1 × R2 → R1, (a1, a2) ↦→ a1 and R1 × R2 → R2, (a1, a2) ↦→ a2 are<br />

ring homomorphisms.<br />

1.1.5. Lemma. In a ring R the binomial formula is true<br />

(a + b) n n<br />

<br />

n<br />

= a<br />

k<br />

n−k b k<br />

a, b ∈ R and n a positive integer.<br />

k=0<br />

Pro<strong>of</strong>. The multiplication is commutative, so the usual pro<strong>of</strong> for numbers works.<br />

Use the binomial identity<br />

<br />

n<br />

+<br />

k − 1<br />

together with induction on n.<br />

<br />

n<br />

=<br />

k<br />

<br />

n + 1<br />

1.1.6. Definition. a ∈ R is a nonzero divisor if ab = 0 for all b = 0 otherwise a<br />

zero divisor. a is a unit if there is a b such that ab = 1.<br />

1.1.7. Remark. (1) A unit is a nonzero divisor.<br />

(2) If ab = 1 then b is uniquely determined by a and denoted b = a −1 .<br />

1.1.8. Definition. A nonzero ring R is a domain if every nonzero element is a<br />

nonzero divisor and a field if every nonzero element is a unit. Clearly a field is a<br />

domain.<br />

1.1.9. Example. The integers Z is a domain. The units in Z are {±1}. The rational<br />

numbers Q, the real numbers R and the complex numbers C are fields. The natural<br />

numbers N is not a ring.<br />

1.1.10. Example. The set <strong>of</strong> n × n-matrices with entries from a commutative ring<br />

is an important normally noncommutative ring.<br />

1.1.11. Exercise. (1) Show that the product <strong>of</strong> two domains is never a domain.<br />

(2) Let R be a ring. Show that the set <strong>of</strong> matrices<br />

<br />

a<br />

U2 =<br />

0<br />

a, <br />

b<br />

b ∈ R<br />

a<br />

with matrix addition and matrix multiplication is a ring.<br />

(3) Show that the set <strong>of</strong> matrices with real number entries<br />

a, <br />

a −b<br />

b ∈ R<br />

b a<br />

with matrix addition and multiplication is a field isomorphic to C.<br />

(4) Show that the composition <strong>of</strong> two ring homomorphisms is again a ring homomorphism.<br />

(5) Show the claim 1.1.3 that a bijective ring homomorphism is a ring isomorphism.<br />

(6) Let φ : 0 → R be a ring homomorphism from the zero ring. Show that R is itself the<br />

zero ring.<br />

k


1.2. IDEALS 11<br />

1.2. Ideals<br />

1.2.1. Definition. Let R be a ring. An ideal I is an additive subgroup <strong>of</strong> R such<br />

that ab ∈ I for all a ∈ R, b ∈ I. A proper ideal is an ideal I = R.<br />

1.2.2.<br />

<br />

Lemma. Let {Iα} be a family <strong>of</strong> ideals in R. Then the additive subgroups<br />

α Iα and <br />

α Iα are ideals.<br />

Pro<strong>of</strong>. The claim for the intersection is clear. Use the formulas bα + cα =<br />

(bα + cα) and a bα = abα to conclude the claim for the sum.<br />

1.2.3. Definition. The intersection, 1.2.2, <strong>of</strong> all ideals containing a subset B ⊂ R<br />

is the ideal generated by B and denoted (B) = RB = BR. It is the smallest ideal<br />

containing B. A principal ideal (b) = Rb is an ideal generated by one element. A<br />

finite ideal (b1, . . . , bn) is an ideal generated by finitely many elements. The zero<br />

ideal is (0) = {0}. The ring itself is a principal ideal, (1) = R. The ideal product<br />

<strong>of</strong> two ideals I, J is denoted IJ and is the ideal generated by all ab, a ∈ I, b ∈ J.<br />

This generalizes to the product <strong>of</strong> finitely many ideals. The power <strong>of</strong> an ideal is<br />

denoted I n . The colon ideal I : J is the ideal <strong>of</strong> elements a ∈ R such that aJ ⊂ I.<br />

1.2.4. Example. (1) Every ideal in Z is principal.<br />

(2) In a field (0), (1) are the only ideals.<br />

(3) A subring is normally not an ideal.<br />

(4) Let K be a field. In K × K there are four ideals (0), (1), ((1, 0)), ((0, 1)).<br />

1.2.5. Proposition. Let R be a ring and B a subset, then RB = <br />

b∈B Rb<br />

A principal ideal is<br />

A finite ideal is<br />

RB = {a1b1 + · · · + anbn|ai ∈ R, bi ∈ B}<br />

Rb = {ab|a ∈ R}<br />

(b1, . . . , bn) = Rb1 + · · · + Rbn<br />

Pro<strong>of</strong>. The righthand side is contained in the ideal RB. Moreover the righthand<br />

side is an ideal containing B, so equality.<br />

1.2.6. Definition. Let φ : R → S be a ring homomorphism. For an ideal J ⊂ S<br />

the contracted ideal is φ −1 (J) ⊂ R and denoted J ∩ R. The kernel is the ideal<br />

Ker φ = φ −1 (0). For an ideal I ⊂ R the extended ideal is the ideal φ(I)S ⊂ S<br />

and denoted IS. Note that (J ∩ R)S ⊂ J and I ⊂ (IS) ∩ R.<br />

1.2.7. Lemma. Let I ⊂ R be an ideal and let R/I be the additive factor group.<br />

The multiplication<br />

R/I × R/I → R/I, (a + I, b + I) ↦→ ab + I<br />

is well defined. Together with the addition the conditions <strong>of</strong> 1.1.2 are satisfied.<br />

Pro<strong>of</strong>. If a + I = a ′ + I and b + I = b ′ + I then a − a ′ , b − b ′ ∈ I and so<br />

ab − a ′ b ′ = a(b − b ′ ) + b ′ (a − a ′ ) ∈ I. Therefore ab + I = a ′ b ′ + I and the<br />

multiplication is well defined. Clearly 1.1.2 are satisfied.<br />

1.2.8. Definition. Let R be a ring and I an ideal, then the factor ring is the additive<br />

factor group R/I with addition (a + I, b + I) ↦→ (a + b) + I and multiplication,<br />

1.2.7, (a + I, b + I) ↦→ ab + I. The projection π : R → R/I, a ↦→ a + I is a ring<br />

homomorphism.


12 1. A DICTIONARY ON RINGS AND IDEALS<br />

1.2.9. Proposition. Let φ : R → S be a ring homomorphism.<br />

(1) Let I ⊂ Ker φ be an ideal. Then there is a unique ring homomorphism<br />

φ ′ : R/I → S such that φ = φ ′ ◦ π.<br />

R<br />

π<br />

<br />

R/I<br />

φ<br />

(2) The homomorphism φ ′ : R/ Ker φ → S is a ring isomorphism onto the<br />

subring φ(R) <strong>of</strong> S.<br />

R<br />

π<br />

<br />

R/ Ker φ<br />

φ<br />

φ ′<br />

φ ′<br />

<br />

<br />

S<br />

<br />

φ(R)<br />

<br />

(3) For any ideal J ⊂ S, I = φ −1 (J) ⊂ R is an ideal and the map φ ′ : R/I →<br />

S/J is an injective ring homomorphism.<br />

Pro<strong>of</strong>. The statements are clear for the addition and the factor map φ ′ (a + I) =<br />

φ(a) is clearly a ring homomorphism.<br />

1.2.10. Corollary. Let π : R → R/I be the projection. The map I ′ ↦→ J =<br />

π −1 (I ′ ) gives a bijective correspondence between ideals I ′ in R/I and ideals J in<br />

R containing I. Also I ′ = π(J) = J/I. This correspondence preserves inclusions,<br />

sums and intersections <strong>of</strong> ideals.<br />

1.2.11. Corollary. Let I ⊂ J ⊂ R be ideals. Then there is a canonical isomorphism<br />

R/J → (R/I)/(J/I)<br />

Pro<strong>of</strong>. The kernel <strong>of</strong> the surjective east-south composite<br />

R<br />

<br />

R/J<br />

<br />

R/I<br />

<br />

<br />

(R/I)/(J/I)<br />

is J. By 1.2.9 the horizontal lower factor map gives the isomorphism.<br />

1.2.12. Example. For any integer n the ideals in the factor ring Z/(n) correspond<br />

to ideals (m) ⊂ Z where m divides n.<br />

1.2.13. Definition. Let R be a ring. The additive kernel <strong>of</strong> the unique ring homomorphism<br />

Z → R is a principal ideal generated by a natural number char(R), the<br />

characteristic <strong>of</strong> R. Z/(char(R)) is isomorphic to the smallest subring <strong>of</strong> R.<br />

1.2.14. Proposition. If the characteristic char(R) = p is a prime number, then<br />

the Frobenius homomorphism R → R, a ↦→ a p is a ring homomorphism.<br />

Pro<strong>of</strong>. By the binomial formula 1.1.4<br />

(a + b) p p<br />

<br />

p<br />

=<br />

k<br />

k=0<br />

a p−k b k = a p + b p


1.3. PRIME IDEALS 13<br />

since a prime number p divides p<br />

k , 0 < k < p. Clearly (ab) p = apbp .<br />

1.2.15. Exercise. (1) Let I, J be ideals in R. Show that the ideal product<br />

IJ = {a1b1 + · · · + anbn|ai ∈ I, bi ∈ J}<br />

(2) Let I ⊂ R be an ideal. Show that I = I : R.<br />

(3) Show that a ∈ R is a unit if and only if (a) = R.<br />

(4) Show that a ring is a field if and only if (0) = (1) are the only two ideals.<br />

(5) Show that a nonzero ring K is a field if and only if any nonzero ring homomorphism<br />

φ : K → R is injective.<br />

(6) Let m, n be natural numbers. Determine the ideals in Z<br />

(m, n), (m) + (n), (m) ∩ (n), (m)(n)<br />

as principal ideals.<br />

(7) Show that a additive cyclic group has a unique ring structure.<br />

(8) Let p be a prime number. What is the Frobenius homomorphism on the ring Z/(p)?<br />

1.3. Prime ideals<br />

1.3.1. Definition. Let R be a ring and P = R a proper ideal.<br />

(1) P is a prime ideal if for any product ab ∈ P either a ∈ P or b ∈ P . This<br />

amounts to: if a, b /∈ P then ab /∈ P .<br />

(2) P is a maximal ideal if no proper ideal = P contains P .<br />

1.3.2. Proposition. Let P be a prime ideal and I1, . . . In ideals such that I1 . . . In ⊂<br />

P , then some Ik ⊂ P .<br />

Pro<strong>of</strong>. If there exist ak ∈ Ik\P for all k, then since P is prime a1 . . . an ∈<br />

I1 . . . In\P contradicting the inclusion I1 . . . In ⊂ P .<br />

1.3.3. Proposition. Let R be a ring and P an ideal.<br />

(1) P is a prime ideal if and only if R/P is a domain.<br />

(2) P is a maximal ideal if and only if R/P is a field.<br />

Pro<strong>of</strong>. Remark P = R ⇔ R/P = 0. (1) Assume a + P, b + P are nonzero in<br />

R/P . Then a, b /∈ P . So if P is prime then by 1.3.1 ab /∈ P and ab + P is nonzero<br />

in R/P . It follows that R/P is a domain. The converse is similar.<br />

(2) Assume R/P is a field and a /∈ P . Then a + P is a unit in R/P and there is<br />

b such that ab − 1 ∈ P . It follows that the ideal (a) + P = R and therefore P is<br />

maximal. The converse is similar.<br />

1.3.4. Corollary. (1) A maximal ideal is a prime ideal.<br />

(2) A ring is an domain if and only if the zero ideal is a prime ideal.<br />

(3) A ring a field if and only if the zero ideal is a maximal ideal.<br />

1.3.5. Corollary. (1) If φ : R → S is a ring homomorphism and Q ⊂ S a prime<br />

ideal then φ −1 (Q) is a prime ideal <strong>of</strong> R.<br />

(2) Let I ⊂ R be an ideal. An ideal I ⊂ P is a prime ideal in R if and only if<br />

P/I is a prime ideal in R/I.<br />

(3) Let I ⊂ R be an ideal. An ideal I ⊂ P is a maximal ideal in R if and only if<br />

P/I is a maximal ideal in R/I.<br />

Pro<strong>of</strong>. (1) By 1.2.9 R/φ −1 (Q) is a subring <strong>of</strong> the domain S/Q. (2) (3) By 1.2.11<br />

R/P (R/I)/(P/I).


14 1. A DICTIONARY ON RINGS AND IDEALS<br />

1.3.6. Example. An ideal in Z is a prime ideal if it is generated by 0 or a prime<br />

number. Any nonzero prime ideal is a maximal ideal.<br />

1.3.7. Definition. For an ideal I in a ring R the radical is<br />

√ I = {a ∈ R|a n ∈ I for some n}<br />

a is nilpotent is a n = 0 for some positive integer n. A ring is reduced if the<br />

nilradical √ 0 = 0, that is if 0 is the only nilpotent element.<br />

1.3.8. Proposition. (1) The radical <strong>of</strong> an ideal is an ideal.<br />

(2) The nilradical is contained in any prime ideal.<br />

(3) A domain is reduced.<br />

Pro<strong>of</strong>. (1) If am , bn ∈ I then by the binomial formula<br />

(a + b) m+n m+n <br />

<br />

m + n<br />

=<br />

a<br />

k<br />

m+n−k b k ∈ I<br />

k=0<br />

and the radical is an ideal. (2) (3) Clearly a nilpotent element is contained in any<br />

prime ideal.<br />

1.3.9. Exercise. (1) Show that the characteristic <strong>of</strong> a domain is either 0 or a prime<br />

number.<br />

(2) Let m, n be a natural numbers. Show that n + (m) ∈ Z/(m) is a unit if and only if<br />

m, n are relative prime.<br />

(3) Let m be a natural number. Show that Z/(m) is reduced if m is square free.<br />

(4) Show that a product <strong>of</strong> reduced rings is reduced.<br />

(5) Let a be nilpotent. Show that 1 − a is a unit.<br />

(6) Let I, J be ideals. Show that √ IJ = √ I ∩ J = √ I ∩ √ J.<br />

(7) Assume an ideal I is contained in a prime ideal P . Show that √ I ⊂ P .<br />

1.4. Chinese remainders<br />

1.4.1. Definition. Ideals I, J ⊂ R are comaximal ideals if I + J = R.<br />

1.4.2. Proposition (Chinese remainder theorem). Let I1, . . . , Ik be pairwise comaximal<br />

ideals in a ring R. Then<br />

(1) For a1, . . . , ak ∈ R there is a a ∈ R, such that a−am ∈ Im for m = 1, . . . , k<br />

(2)<br />

(3) The product <strong>of</strong> projections<br />

is an isomorphism.<br />

Pro<strong>of</strong>. (1) For each m<br />

I1 · · · Ik = I1 ∩ · · · ∩ Ik<br />

R/I1 · · · Ik → R/I1 × · · · × R/Ik<br />

R = <br />

(Im + In) = Im + <br />

n=m<br />

n=m<br />

So choose um ∈ Im and vm ∈ <br />

n=m In with um + vm = 1. Put a = a1v1 + · · · +<br />

akvk. Then a − am = · · · + amum + · · · ∈ Im. (2) For a in the intersection assume<br />

by induction that a ∈ I2 · · · Ik. From the pro<strong>of</strong> <strong>of</strong> (1) a = u1a + av1 ∈ I1 · · · Ik.<br />

(3) Subjectivity follows from (1). The kernel is the intersection which by (2) is the<br />

product. 1.2.9 gives the isomorphism.<br />

In


1.5. UNIQUE FACTORIZATION 15<br />

1.4.3. Corollary. Let P1, . . . , Pk be pairwise different maximal ideals in a ring R.<br />

Then<br />

and<br />

P n1<br />

1 · · · P nk n1<br />

nk<br />

k = P1 ∩ · · · ∩ Pk R/P n1<br />

1 · · · P nk<br />

n1<br />

nk<br />

k → R/P1 × · · · × R/Pk is an isomorphism.<br />

1.4.4. Definition. An element e in a ring R is idempotent if e = e 2 . A nontrivial<br />

idempotent is an idempotent e = 0, 1.<br />

1.4.5. Proposition. A ring R is a product <strong>of</strong> two nonzero rings if and only if it<br />

contains a nontrivial idempotent e.<br />

Pro<strong>of</strong>. Use that the ideals Re and R(1 − e) are proper and comaximal.<br />

1.4.6. Exercise. (1) Show that for a prime number p the rings Z/(p2 ) and Z/(p) ×<br />

Z/(p) are not isomorphic.<br />

(2) Let n = p n1<br />

1 be a factorization into different primes. Show that<br />

. . . pnk<br />

k<br />

Z/(p n1<br />

1<br />

· · · pnk<br />

k<br />

) → Z/(pn1 1 ) × · · · × Z/(pnk k )<br />

is an isomorphism.<br />

(3) Let elements e1 + e2 = 1 with e1e2 = 0 be given in a ring R. Show that R <br />

R/(e1) × R/(e2).<br />

(4) Let I, J be ideals such that √ I, √ J are comaximal. Show that I, J are comaximal.<br />

1.5. Unique factorization<br />

1.5.1. Lemma. Let R be a domain and (a) = (b) principal ideals. Then there is a<br />

unit u ∈ R such that b = ua.<br />

Pro<strong>of</strong>. b = ua and a = vb giving b = uvb. If b = 0 then uv = 1 showing that u is<br />

a unit.<br />

1.5.2. Definition. Let R be a domain and P the set <strong>of</strong> principal ideals different<br />

from (0) and R. By 1.5.1 multiplication <strong>of</strong> generators gives a well defined multiplication<br />

<strong>of</strong> principal ideals on P. An element in P maximal for inclusion is an<br />

irreducible principal ideal. A generator <strong>of</strong> an irreducible element is an irreducible<br />

element in R.<br />

1.5.3. Definition. A domain R is a unique factorization domain if<br />

(1) Every irreducible element in P is a prime ideal.<br />

(2) Every element in P is a product <strong>of</strong> irreducible elements.<br />

1.5.4. Proposition. In a unique factorization domain the factorization <strong>of</strong> elements<br />

in P into irreducibles is unique up to order.<br />

Pro<strong>of</strong>. Proceed by induction on the shortest factorization <strong>of</strong> an element in P. Let<br />

(a1) . . . (am) = (b1) . . . (bn) be factorizations into irreducibles. (a1) is a prime<br />

ideal, so by 1.3.2 and reordering (b1) = (a1). By cancellation m − 1 = n − 1 and<br />

(ai) = (bi) after a reordering.<br />

1.5.5. Definition. A domain R is a principal ideal domain if every ideal is principal.<br />

1.5.6. Proposition. A principal ideal domain R is a unique factorization domain.


16 1. A DICTIONARY ON RINGS AND IDEALS<br />

Pro<strong>of</strong>. Let (a) be irreducible and x, y /∈ (a). Then (a, x), (a, y) are principal ideals<br />

properly containing (a) giving (a, x) = (a, y) = R. Let ba + cx = da + ey = 1<br />

and look at (ba+cx)(da+ey) = 1 to see that xy /∈ (a). It follows that (a) is prime,<br />

part (1) <strong>of</strong> 1.5.3. If (b) is not irreducible, then (b) = (a1)(b1) for some irreducible<br />

(a1) and (b) ⊂ (b1). Continue this process to get (b) = (a1) . . . (an)(bn) for some<br />

irreducibles (a1), . . . (an). The chain <strong>of</strong> ideals (b) ⊂ (b1) ⊂ · · · ⊂ (bn) has a<br />

union <br />

n (bn) which is a principal ideal. The generator <strong>of</strong> the union must be in<br />

some (bn). Therefore (bn) = (bn+1) giving that (bn) is irreducible. This gives a<br />

factorization required in 1.5.3 part (2).<br />

1.5.7. Example. The integers Z is a principal ideal domain and therefore a unique<br />

factorization domain.<br />

1.5.8. Definition. The supremum <strong>of</strong> a set <strong>of</strong> elements in P is the greatest common<br />

divisor and an infimum is the least common multiple.<br />

1.5.9. Corollary. In a unique factorization domain the greatest common divisor<br />

and the least common multiple <strong>of</strong> a finite set <strong>of</strong> elements exist.<br />

If (a) = (p m1<br />

1 ) . . . (pmk k ) and (b) = (pn1 1 ) . . . (pnk k ) with mi, ni ≥ 0 then greatest<br />

common divisor is (p min(m1,n1)<br />

1 ) . . . (p min(mk,nk)<br />

k ) and least common multiple is<br />

).<br />

(p max(m1,n1)<br />

1<br />

) . . . (p max(mk,nk)<br />

k<br />

1.5.10. Exercise. (1) Show that an irreducible element in a principal ideal domain generates<br />

a maximal ideal.<br />

(2) Show that there are infinitely many prime numbers.<br />

(3) Let Z[ √ −1] be the smallest subring <strong>of</strong> C containing √ −1. Show that Z[ √ −1] is a<br />

principal ideal domain.<br />

(4) Let Z[ √ −5] be the smallest subring <strong>of</strong> C containing √ −5. Show that Z[ √ −5] is not<br />

a unique factorization domain.<br />

1.6. Polynomials<br />

1.6.1. Definition. Let R be a ring. The polynomial ring R[X] is the additive<br />

group given by the direct sum <br />

n RXn , n = 0, 1, 2, . . . consisting <strong>of</strong> all finite<br />

sums f = a0 + a1X + . . . amX m , a polynomial with an ∈ R being the n ′ th<br />

coefficient. Multiplication is given by XiX j = Xi+j extended by linearity. If<br />

g = b0 + b1X + . . . bnX n is an other polynomial, then<br />

f + g = (a0 + b0) + (a1 + b1)X + · · · + (ak + bk)X k + . . .<br />

fg = a0b0 + (a0b1 + a1b0)X + · · · + (a0bk + a1bk−1 + · · · + akb0)X k + . . .<br />

A monomial is polynomial <strong>of</strong> form aX n . The construction may be repeated to<br />

give the polynomial ring in n-variables R[X1, . . . , Xn] or even in infinitely many<br />

variables.<br />

1.6.2. Definition. The degree, deg(f), <strong>of</strong> a polynomial 0 = f ∈ R[X] is the<br />

index <strong>of</strong> the highest nonzero coefficient, the leading coefficient. A polynomial<br />

with leading coefficient the identity is a monic polynomial.<br />

1.6.3. Remark. (1) R is identified with the subring <strong>of</strong> constants in the polynomial<br />

ring R[X1, . . . , Xn].<br />

(2) The nonzero constants are the polynomials <strong>of</strong> degree 0.<br />

(3) The constant polynomial 1 is the unique monic polynomial <strong>of</strong> degree 0 and<br />

the identity in the polynomial ring.


1.6. POLYNOMIALS 17<br />

1.6.4. Proposition. Let 0 = f, g ∈ R[X].<br />

(1) If fg = 0 then deg(fg) ≤ deg(f) + deg(g).<br />

(2) If the leading coefficient <strong>of</strong> f or g is a nonzero divisor in R, then fg = 0 and<br />

deg(fg) = deg(f) + deg(g)<br />

Pro<strong>of</strong>. (1) This is clear. (2)Clearly the leading coefficient <strong>of</strong> the product is the<br />

product <strong>of</strong> the leading coefficients.<br />

1.6.5. Corollary. Let R be a domain.<br />

(1) The polynomial ring R[X] is a domain.<br />

(2) The units in R[X] are the constants, which are units in R.<br />

1.6.6. Proposition. Let R be a domain, 0 = f, d ∈ R[X] polynomials with d<br />

monic. Then there are a unique q, r ∈ R[X] such that<br />

f = qd + r, r = 0 or deg(r) < deg(d)<br />

Pro<strong>of</strong>. Induction on deg(f). If deg(f) < deg(d) then q = 0, r = f. Otherwise<br />

if a is the leading coefficient <strong>of</strong> f, then f − adX deg(f)−deg(d) has degree less than<br />

deg(f). By induction f − adX deg(f)−deg(d) = qd + r giving the claim.<br />

1.6.7. Proposition. Let φ : R → S be a ring homomorphism. For any element<br />

b ∈ S there is a unique ring homomorphism R[X] → S extending φ and mapping<br />

X ↦→ b.<br />

Pro<strong>of</strong>.<br />

a0 + a1X + . . . amX m ↦→ φ(a0) + φ(a1)b + . . . φ(am)b m<br />

is clearly the one and only choice.<br />

1.6.8. Definition. The homomorphism in 1.6.7 is the evaluation map at b in S.<br />

The image <strong>of</strong> a polynomial f ∈ R[X] is denoted f(b) ∈ S.<br />

1.6.9. Proposition. Let I ⊂ R be an ideal. Then there is a canonical isomorphism.<br />

(R/I)[X] R[X]/IR[X]<br />

Pro<strong>of</strong>. There is an obvious pair <strong>of</strong> inverse homomorphisms constructed by 1.2.9<br />

and 1.6.7.<br />

1.6.10. Corollary. If P ⊂ R is a prime ideal, then P R[X] ⊂ R[X] is a prime<br />

ideal.<br />

1.6.11. Definition. Let φ : R → S be a ring homomorphism and B ⊂ S a subset.<br />

The ring generated over R by B is<br />

R[B] = φ(R)[B] ⊂ S<br />

the smallest subring <strong>of</strong> S containing φ(R) ∪ B. If there is a finite subset B such<br />

that R[B] = S then S is a finite type ring or a finitely generated ring over R.<br />

1.6.12. Corollary. Let φ : R → S be a ring homomorphism.<br />

(1) If bα ∈ S and Xα is a family <strong>of</strong> variables, then there is a surjective ring<br />

homomorphism<br />

R[Xα] → R[bα], Xα ↦→ bα<br />

making R[bα] a factor ring <strong>of</strong> the polynomial ring R[Xα].


18 1. A DICTIONARY ON RINGS AND IDEALS<br />

(2) If S is a finite type ring over R, then S is a factor ring <strong>of</strong> a polynomial ring<br />

in finitely many variables over R.<br />

1.6.13. Exercise. (1) Let K be a field. Show that there are infinitely many prime ideals<br />

in K[X].<br />

(2) What are the units in the ring Z[X]/(1 − 2X)?<br />

(3) Determine the prime ideals in Q[X]/(X − X 2 ).<br />

(4) Show that the ring Z[X] is not a principal ideal domain.<br />

(5) Show that the ring Q[X, Y ] is not a principal ideal domain.<br />

1.7. Roots<br />

1.7.1. Definition. Let φ : R → S be a ring homomorphism and f ∈ R[X] a<br />

polynomial. An element b ∈ S is a root <strong>of</strong> f (in S) if the evaluation f(b) = 0.<br />

1.7.2. Proposition. Let R be a domain. An element a ∈ R is a root <strong>of</strong> the polynomial<br />

f ∈ R[X] if and only if there is a q ∈ R[X] such that<br />

f = q(X − a)<br />

Pro<strong>of</strong>. By 1.6.6 f = q(X −a)+r. It follows that a is a root if and only if r = 0.<br />

1.7.3. Corollary. Let R is a domain. There are at most deg(f) roots in a nonzero<br />

polynomial f ∈ R[X].<br />

1.7.4. Definition. The multiplicity <strong>of</strong> a root a <strong>of</strong> a nonzero polynomial f ∈ R[X]<br />

is highest m such that<br />

f = q(X − a) m<br />

A root <strong>of</strong> multiplicity m = 1 is a simple root.<br />

1.7.5. Corollary. Let R is a domain. If m1, . . . , mk are the multiplicities <strong>of</strong> the<br />

roots <strong>of</strong> a nonzero polynomial f ∈ R[X], then m1 + · · · + mk ≤ deg(f).<br />

1.7.6. Definition. The derivative <strong>of</strong> a polynomial f = anX n ∈ R[X] is<br />

1.7.7. Lemma. The derivative satisfies<br />

(1) (f + g) ′ = f ′ + g ′ .<br />

(2) (fg) ′ = f ′ g + fg ′<br />

(3) If f is constant, then f ′ = 0.<br />

f ′ = nanX n−1<br />

1.7.8. Proposition. Let R is a domain. An element a ∈ R is a root <strong>of</strong> multiplicity<br />

m > 1 <strong>of</strong> a nonzero f ∈ R[X] if and only if a is a root <strong>of</strong> f and f ′ .<br />

Pro<strong>of</strong>. By 1.6.6 f = q(X − a) 2 + cX + d and by 1.7.7 f ′ = q ′ (X − a) 2 + 2q(X −<br />

a) + c. I follows that a is a root <strong>of</strong> multiplicity m > 1 if and only if c = d = 0.<br />

1.7.9. Exercise. (1) Let a1, . . . ak be roots with multiplicities m1, . . . , mk in a polynomial<br />

f. Show that m1 + · · · + mk ≤ deg(f).<br />

(2) Let K be a field and let a1, . . . , an ∈ K. Show that the ideal (X1 −a1, . . . , Xn −an)<br />

is maximal in K[X1, . . . , Xn].<br />

(3) Let the characteristic char(R) = n > 0. What is (X n ) ′ in R[X].


1.8. FIELDS 19<br />

1.8. Fields<br />

1.8.1. Definition. Let p ∈ Z be a prime number. The factor ring Fp = Z/(p) is a<br />

field with p elements. Together with Q they constitute the prime fields.<br />

1.8.2. Proposition. Let K be a field then the polynomial ring K[X] is a principal<br />

ideal domain.<br />

Pro<strong>of</strong>. Let d = 0 be a polynomial <strong>of</strong> lowest degree in an ideal I. Given f ∈ I then<br />

by 1.6.6 f = qd + r with r = f − qd ∈ I. By degree considerations r = 0 and<br />

I = (d).<br />

1.8.3. Corollary. Let K be a field then the polynomial ring K[X] is a unique<br />

factorization domain.<br />

Pro<strong>of</strong>. Follows from 1.5.6.<br />

1.8.4. Definition. A subfield is a subring, which is a field. A field extension is the<br />

inclusion <strong>of</strong> a subfield K ⊂ L in a field. A finite field extension K ⊂ L is an<br />

extension, where L is finite dimensional as a vector space over K.<br />

1.8.5. Example. (1) Let K be a field and f an irreducible polynomial in K[X].<br />

Then K ⊂ K[X]/(f) is a finite field extension.<br />

(2) Let K ⊂ R ⊂ L be a subring in a finite field extension. Then R is a field.<br />

Namely multiplication on R with a nonzero element <strong>of</strong> R is a K-linear map<br />

on the finite dimensional K-vector space R and therefore an isomorphism.<br />

1.8.6. Proposition. (1) Let K be a field and f a polynomial in K[X]. Then there<br />

is a finite field extension K ⊂ L such that f factors in linear factors in L[X].<br />

(2) If K ⊂ L1 and K ⊂ L2 are finite field extensions then there is a finite field<br />

extension K ⊂ L such that L1 ∪ L2 ⊂ L.<br />

Pro<strong>of</strong>. (1) Assume f irreducible. In L = K[X]/(f) the class X + (f) is a root <strong>of</strong><br />

f. In general proceed adjoining roots <strong>of</strong> irreducible factors <strong>of</strong> f. (2) An element x<br />

in a finite field extension K ⊂ K ′ is the root <strong>of</strong> the irreducible monic polynomial<br />

f generating the kernel <strong>of</strong> the evaluation homomorphism K[X] → K ′ , X ↦→ x.<br />

Now proceed by (1) adjoining elements in L2 to L1.<br />

1.8.7. Proposition. Let p ∈ Z be a prime number. For any power q = p n there is<br />

a field Fq with q elements, unique up to isomorphism.<br />

Pro<strong>of</strong>. Let Fp ⊂ K be a field extension, where X q − X factors into linear factors,<br />

1.8.6 (1). The subset <strong>of</strong> roots is the set <strong>of</strong> elements fixed under n-times the<br />

Frobenius and therefore a subring being a subfield by 1.8.5 (2). The derivative<br />

(X q − X) ′ = −1 so by 1.7.8 there are q elements in this subfield. Uniqueness<br />

follows from 1.8.6 (2).<br />

1.8.8. Exercise. (1) Show that the ring R[X]/(X 2 + 1) is isomorphic to the field <strong>of</strong><br />

complex numbers.<br />

(2) Show that the ring F2[X]/(X 2 + X + 1) is a field with 4 elements.<br />

(3) Show that the ring F2[X]/(X 3 + X + 1) is a field with 8 elements.<br />

(4) Let K ⊂ L be a field extension <strong>of</strong> fields <strong>of</strong> characteristic 0 and let a ∈ L be a root <strong>of</strong><br />

an irreducible polynomial f ∈ K[X]. Show that a is a simple root.<br />

(5) Let p be a prime number. Show that Fp is the only ring with p elements.<br />

(6) Let p be a prime number. Show that a ring with p 2 elements is isomorphic to one <strong>of</strong><br />

four non isomorphic Z/(p 2 ), Fp × Fp, Fp[X]/(X 2 ), F p 2.


20 1. A DICTIONARY ON RINGS AND IDEALS<br />

1.9. Power series<br />

1.9.1. Definition. Let R be a ring. The power series ring R[[X]] is the additive<br />

group <br />

n RXn , n = 0, 1, 2, . . . <strong>of</strong> all power series anX n with n ′ th coefficient<br />

an ∈ R. Multiplication is given by X i X j = X i+j extended by linearity. For<br />

another power series bnX n the sum and product are<br />

anX n + bnX n = (an + bn)X n<br />

anX n · bnX n = ( <br />

k<br />

an−kbk)X n<br />

The construction may be repeated to give the power series ring in n-variables<br />

R[[X1, . . . , Xn]] or even in infinitely many variables.<br />

1.9.2. Remark. The polynomial ring is identified as a subring R[X] ⊂ R[[X]] <strong>of</strong><br />

power series with only finitely many nonzero terms.<br />

1.9.3. Definition. The order, o(f), <strong>of</strong> a power series 0 = f ∈ R[[X]] is the index<br />

<strong>of</strong> the least nonzero coefficient.<br />

1.9.4. Proposition. If R is a domain, then R[[X]] is a domain and for 0 = f, g ∈<br />

R[X]<br />

o(fg) = o(f) + o(g)<br />

Pro<strong>of</strong>. Clearly the lowest nonzero coefficient in the product is the product <strong>of</strong> the<br />

two lowest nonzero coefficients.<br />

1.9.5. Proposition. A power series f = anX n is a unit if and only if a0 is a<br />

unit.<br />

Pro<strong>of</strong>. It suffices to look at a power series f = 1 − gX. Then the power series<br />

1/f = 1 + gX + g 2 X 2 + · · · + g n X n + . . . is well defined and f · 1/f = 1.<br />

1.9.6. Proposition. If K is a field, then K[[X]] is a principal ideal domain. and<br />

(X) is the only nonzero prime ideal.<br />

Pro<strong>of</strong>. If the lowest order <strong>of</strong> an element in an ideal I is n. Then clearly I =<br />

(X n ).<br />

1.9.7. Corollary. If K is a field, then K[[X]] is a unique factorization domain.<br />

1.9.8. Proposition. Let I ⊂ R be an ideal. Then there is a canonical surjective<br />

homomorphism<br />

R[[X]]/IR[[X]] → R/I[[X]]<br />

1.9.9. Corollary. If Q ⊂ R[[X]] is a maximal ideal, then P = Q ∩ R ⊂ R is a<br />

maximal ideal and Q = (P, X).<br />

Pro<strong>of</strong>. X ∈ Q so R/Q ∩ R R[[X]]/Q.<br />

1.9.10. Exercise. (1) Show that the ring Z[[X]] is not a principal ideal domain.<br />

(2) Show that the ring Q[[X, Y ]] is not a principal ideal domain.<br />

(3) Let K be a field. Show that (X1, . . . , Xn) is the unique maximal ideal in the power<br />

series ring K[[X1, . . . , Xn]].<br />

(4) Let a ∈ R be nilpotent. Show that the ring R[[X]]/(X − a) is isomorphic to R.<br />

(5) What is R[[X]]/(X − a) if a ∈ R is a unit?<br />

(6) Let I ⊂ R be an ideal. Show that IR[[X]] ⊂ R[[X]] is not a maximal ideal.


2<br />

Modules<br />

2.1. Modules and homomorphisms<br />

2.1.1. Definition. Let R be a ring. A module (R-module) is an abelian group<br />

M, addition (x, y) ↦→ x + y and zero 0, together with a scalar multiplication<br />

R × M → M, (a, x) ↦→ ax satisfying<br />

(1) associative : (ab)x = a(bx)<br />

(2) bilinear : a(x + y) = ax + ay, (a + b)x = ax + bx<br />

(3) identity: 1x = x<br />

for all a, b ∈ R, x, y ∈ M. A submodule M ′ ⊂ M is an additive subgroup such<br />

that ax ∈ M ′ for all a ∈ R, x ∈ M ′ . A homomorphism is an additive group<br />

homomorphism f : M → N respecting scalar multiplication<br />

f(x + y) = f(x) + f(y), f(ax) = af(y)<br />

for all a ∈ R, x, y ∈ M. An isomorphism is a homomorphism f : M → N<br />

having an inverse map f −1 : N → M which is also a homomorphism. The<br />

identity isomorphism is denoted 1M : M → M.<br />

2.1.2. Lemma. Let R be a ring and M a module.<br />

(1) a0 = 0 = 0x<br />

(2) (−1)x = −x<br />

(3) (−a)x = −(ax) = a(−x)<br />

for all a ∈ R, x ∈ M.<br />

Pro<strong>of</strong>. (1) Calculate a0 = a(0 + 0) = a0 + a0 and cancel to get a0 = 0. Similarly<br />

0x = 0. (2) By (1) 0 = 0x = (1 + (−1))x = x + (−1)x, so conclude −x =<br />

(−1)x. (3) Calculate (−a)x = ((−1)a)x = (−1)(ax) = −(ax). Similarly<br />

a(−x) = −(ax).<br />

2.1.3. Lemma. Let R be a ring and f : M → N a homomorphism <strong>of</strong> modules.<br />

(1) f(0) = 0.<br />

(2) f(ax + by) = af(x) + bf(y).<br />

(3) f(−x) = −f(x) for all x ∈ M.<br />

for all a, b ∈ R, x, y ∈ M.<br />

Pro<strong>of</strong>. (1) Calculate f(0) = f(0 + 0) = f(0) + f(0) and conclude f(0) = 0.<br />

(2) Calculate f(ax + by) = f(ax) + f(by) = af(x) + bf(y). (3) By 2.1.2<br />

f(−x) = f((−1)x) = (−1)f(x) = −f(x).<br />

2.1.4. Example. (1) The zero group is the zero module.<br />

(2) Over the zero ring the zero module is the only module.<br />

(3) The zero subgroup <strong>of</strong> a module is the zero submodule.<br />

(4) The ring R is a module under multiplication. An ideal is a submodule.<br />

21


22 2. MODULES<br />

(5) If R is a field, a module is a vector space and a homomorphism is a linear<br />

map.<br />

(6) A module over Z is an abelian group and an additive map <strong>of</strong> abelian groups<br />

is a homomorphism.<br />

2.1.5. Proposition. A bijective homomorphism is an isomorphism.<br />

Pro<strong>of</strong>. Let f : M → N be a bijective homomorphism <strong>of</strong> R-modules and let g :<br />

N → M be the inverse map. For x, y ∈ N write x = f(g(x)), y = f(g(y)) and<br />

get additivity <strong>of</strong> g, g(x + y) = g(f(g(x)) + f(g(y))) = g(f(g(x) + g(y))) =<br />

g(x) + g(y). Similarly for a ∈ R g(ax) = g(af(g(x))) = g(f(ag(x))) = ag(x),<br />

so g respects scalar multiplication and is a homomorphism.<br />

2.1.6. Lemma. Let a ∈ R and M be a module. The map M → M, x ↦→ ax is a<br />

homomorphism.<br />

Pro<strong>of</strong>. Let f(x) = ax and calculate f(x+y) = a(x+y) = ax+ay = f(x)+f(y)<br />

and f(bx) = a(bx) = (ab)x = (ba)x = b(ax) = bf(x) to get that f is a<br />

homomorphism. Remark that the last calculation uses that R is commutative 1.1.2<br />

(4).<br />

2.1.7. Definition. Let a ∈ R and M be a module.<br />

(1) The scalar multiplication with a is the homomorphism, 2.1.6,<br />

aM : M → M, x ↦→ ax<br />

(2) a ∈ R is a nonzero divisor on M if scalar multiplication aM is injective, i.e.<br />

ax = 0 for all 0 = x ∈ M. Otherwise a is a zero divisor.<br />

2.1.8. Remark. The two notions <strong>of</strong> nonzero divisor on R : 1.1.6 as element in the<br />

ring and 2.1.4 as scalar multiplication on the ring coincide.<br />

2.1.9. Example. If R is a field, then scalar multiplication on a vector space is either<br />

zero or an isomorphism.<br />

2.1.10. Lemma. Let φ : R → S be a ring homomorphism and N an S-module.<br />

The map<br />

R × N → N, (a, x) ↦→ φ(a)x<br />

is an R-scalar multiplication on N, 2.1.1.<br />

Pro<strong>of</strong>. Let a, b ∈ R, x, y ∈ N and µ(a, x) = φ(a)x. Calculate µ(a + b, x) =<br />

φ(a + b)x = φ(a)x + φ(b)x = µ(a, x) + µ(b, x), µ(a, x + y) = φ(a)(x + y) =<br />

φ(a)x + φ(a)y = µ(a, x) + µ(a, y), µ(1, x) = φ(1)x = 1x = x and µ(ab, x) =<br />

φ(ab)x = φ(a)φ(b)x = µ(a, µ(bx)) showing the conditions 2.1.1.<br />

2.1.11. Definition. Let φ : R → S be a ring homomorphism. The restriction <strong>of</strong><br />

scalars <strong>of</strong> an S-module N is the same additive group N viewed as an R-module<br />

through φ. The scalar multiplication is 2.1.10<br />

R × N → N, (a, x) ↦→ ax = φ(a)x<br />

An S-module homomorphism g : N → N ′ is also an R-module homomorphism.<br />

2.1.12. Example. (1) The scalar multiplication with a Restriction <strong>of</strong> scalars for<br />

the unique ring homomorphism Z → R give just the underlying abelian group<br />

<strong>of</strong> a module, 2.1.4 (6).


2.2. SUBMODULES AND FACTOR MODULES 23<br />

(2) Let I ⊂ R be an ideal. Restriction <strong>of</strong> scalars along the projection R → R/I<br />

gives any R/I-module as an R-module.<br />

2.1.13. Proposition. Let R be a ring. There is a dictionary:<br />

(1) To an R[X]-module N associate the pair (N, f) consisting <strong>of</strong> N as R-module<br />

through restriction <strong>of</strong> scalars and f = XN : N → N, f(y) = Xy scalar<br />

multiplication with X as an R-module homomorphism. An R[X]-homomorphism<br />

g : N → N ′ gives an R-homomorphism such that g ◦ f = f ′ ◦ g.<br />

(2) To a pair (N, f) <strong>of</strong> an R-module and a homomorphism f : N → N associate<br />

the R[X]-module with abelian group N and scalar multiplication determined<br />

by Xy = f(y) for y ∈ N. Note<br />

( anX n )y = anf ◦n (y)<br />

An R-homomorphism g : N → N ′ such that g ◦ f = f ′ ◦ g is an R[X]homomorphism.<br />

Pro<strong>of</strong>. The statement is an algorithm to follow.<br />

2.1.14. Proposition. Let R be a ring and M a module. The abelian group R ⊕ M<br />

with multiplication<br />

(a + x)(b + y) = ab + (ay + bx)<br />

is a ring. R is a subring and M is an ideal.<br />

Pro<strong>of</strong>. Simple calculations show that the conditions 1.1.2 are satisfied.<br />

2.1.15. Exercise. (1) Show that a composition <strong>of</strong> homomorphisms is a homomorphism.<br />

(2) Show that composition <strong>of</strong> scalar multiplications with a, b ∈ R on a module M is a<br />

scalar multiplication with the product, aM ◦ bM = (ab)M .<br />

(3) Let φ : R → S be a ring homomorphism. Show that φ is an R-module homomorphism,<br />

when S is viewed as R-module through restriction <strong>of</strong> scalars 2.1.11.<br />

(4) Fill out the dictionary 2.1.13.<br />

2.2. Submodules and factor modules<br />

2.2.1. Lemma. Let R be a ring and M a module. Let Nα be a family <strong>of</strong> submodules.<br />

Then the additive subgroups <br />

α Nα and <br />

α Nα are submodules.<br />

Pro<strong>of</strong>. Use the formulas xα + yα = (xα + yα) and a xα = axα to<br />

conclude that Nα is a submodule. If x, y ∈ Nα for all α, then x + y, ax ∈ Nα<br />

for all α, so Nα is a submodule.<br />

2.2.2. Definition. Let R be a ring and M a module. The intersection <strong>of</strong> all submodules<br />

containing a subset Y ⊂ M is the submodule generated by Y and denoted<br />

RY . This is the smallest submodule, 2.2.1, <strong>of</strong> M containing Y . The module M is<br />

generated by Y if RY = M. Let I be an ideal. The submodule generated by all<br />

products ax, a ∈ I, x ∈ M is denoted IM.<br />

2.2.3. Proposition. Let R be a ring and M a module. If Y ⊂ M, then RY =<br />

<br />

y∈Y Ry,<br />

RY = {a1y1 + · · · + anyn|ai ∈ R, yi ∈ Y }<br />

Pro<strong>of</strong>. The righthand side is contained in the submodule RY . Moreover the righthand<br />

side is a submodule containing Y , so equality.<br />

2.2.4. Corollary. Let I ⊂ R be an ideal and M a module.


24 2. MODULES<br />

(1)<br />

IM = {a1y1 + · · · + anyn|ai ∈ I, yi ∈ M}<br />

(2) If I = (a) is principal, then<br />

aM = (a)M = {ay|y ∈ M}<br />

Pro<strong>of</strong>. (1) is clear. (2) By (1) an element in aM is biayi = a biyi = ay for<br />

y = biyi ∈ M as claimed.<br />

2.2.5. Lemma. Let R be a ring, M a module and N ⊂ M a submodule. Let M/N<br />

be the abelian factor group, then the map<br />

R × M/N → M/N, (a, x + N) ↦→ ax + N<br />

is well defined and a scalar multiplication, 2.1.1.<br />

Pro<strong>of</strong>. If x+N = y+N then x−y ∈ N and so a(x−y) = ax−ay ∈ N. Therefore<br />

ax + N = ay + N and the multiplication is well defined. Since representatives<br />

may be chosen such that (x + N) + (y + N) = x + y + N, a(x + N) = ax + N,<br />

the laws for scalar multiplication are satisfied.<br />

2.2.6. Definition. Let R be a ring, M a module and N ⊂ M a submodule, then the<br />

factor module is the additive factor group M/N with, 2.2.5, scalar multiplication<br />

a(x + N) = ax + N. The projection p : M → M/N, x ↦→ x + N is a surjective<br />

homomorphism.<br />

2.2.7. Lemma. Let R be a ring, N ⊂ M a submodule and p : M → M/N<br />

the projection. p is surjective and if Y ⊂ M generates M, then p(Y ) ⊂ M/N<br />

generates the factor module.<br />

Pro<strong>of</strong>. Clearly if RY = M then Rp(Y ) = p(RY ) = M/N.<br />

2.2.8. Example. (1) A submodule <strong>of</strong> R is the same as an ideal.<br />

(2) Both an ideal I ⊂ R and a factor ring R/I are modules.<br />

(3) The module structure on R/I as a factor module and the structure by restriction<br />

<strong>of</strong> scalars through the projection R → R/I are identical.<br />

2.2.9. Proposition. Let R = R1×R2 be the product ring 1.1.4. There is a bijective<br />

(up to natural isomorphism) correspondence.<br />

(1) If M1 is an R1-module and M2 is an R2-module, then M = M1 × M2 is an<br />

R-module with coordinate scalar multiplication. A pair <strong>of</strong> homomorphisms<br />

induce a homomorphism on the product.<br />

(2) If M is an R-module then M1 = (1, 0)M is an R1-module and M2 =<br />

(0, 1)M is an R2-module. A homomorphism induces a pair <strong>of</strong> homomorphisms.<br />

2.2.10. Remark. The correspondence 2.2.9 indicates that the structure <strong>of</strong> modules<br />

and homomorphisms over a product ring is identified with the structure <strong>of</strong> pairs <strong>of</strong><br />

modules and homomorphisms over each component ring in the product.<br />

2.2.11. Exercise. (1) Give an example <strong>of</strong> two submodules N, L ⊂ M such that the<br />

union N ∪ L is not a submodule.<br />

(2) Let R be a ring and a ∈ R. Show that the R-module R[X]/(X − a) is isomorphic<br />

to R.<br />

(3) Show that the projection p 2.2.6 is a homomorphism.<br />

(4) Fill in the details in the dictionary 2.2.9


2.3. KERNEL AND COKERNEL 25<br />

2.3. Kernel and cokernel<br />

2.3.1. Lemma. Let R be a ring and f : M → N a homomorphism <strong>of</strong> modules.<br />

Given submodules M ′ ⊂ M, N ′ ⊂ N, then f −1 (N ′ ) ⊂ M and f(M ′ ) ⊂ N are<br />

submodules.<br />

Pro<strong>of</strong>. If x, y ∈ f −1 (N ′ ) then f(x + y) = f(x) + f(y) ∈ N ′ and for a ∈ R<br />

f(ax) = af(x) ∈ N ′ so x+y, ax ∈ f −1 (N ′ ) proving f −1 (N ′ ) to be a submodule.<br />

The same equations prove that f(M ′ ) is a submodule.<br />

2.3.2. Definition. Let f : M → N be a homomorphism <strong>of</strong> modules. Then there<br />

are submodules, 2.3.1.<br />

(1) The kernel Ker f = f −1 (0).<br />

(2) The image Im f = f(M).<br />

(3) The cokernel Cok f = N/ Im f.<br />

2.3.3. Proposition. Let f : M → N be a homomorphism <strong>of</strong> modules.<br />

(1) f is injective if and only if Ker f = 0.<br />

(2) f is surjective if and only if Cok f = 0.<br />

(3) f is an isomorphism if and only if Ker f = 0 and Cok f = 0.<br />

Pro<strong>of</strong>. (1) If f is injective and x ∈ Ker f then f(x) = 0 = f(0) so x = 0.<br />

Conversely if Ker f = 0 and f(x) = f(y) then f(x − y) = 0 so x = y. (2) The<br />

factor module N/ Im f = 0 if and only if Im f = N. (3) This follows from (1)<br />

and (2).<br />

2.3.4. Example. Let a ∈ R give scalar multiplication aM : M → M, x ↦→ ax.<br />

(1) Im aM = aM = {ax ∈ M|x ∈ M}.<br />

(2) Ker aM = {x ∈ M|ax = 0}.<br />

(3) Cok aM = M/aM.<br />

2.3.5. Proposition. Let f : M → N be a homomorphism <strong>of</strong> modules.<br />

(1) Let L ⊂ Ker f be a submodule. Then there is a unique homomorphism f ′ :<br />

M/L → N such that f = f ′ ◦ p.<br />

M<br />

p<br />

<br />

M/L<br />

f<br />

(2) The homomorphism f ′ : M/ Ker f → N is a module isomorphism onto the<br />

submodule Im f <strong>of</strong> N.<br />

M<br />

p<br />

<br />

M/ Ker f<br />

f<br />

f ′<br />

f ′<br />

<br />

<br />

N<br />

<br />

Im f<br />

<br />

Pro<strong>of</strong>. (1) If x+L = x ′ +L then x−x ′ ∈ L giving f(x) = f(x ′ ). The factor map<br />

f ′ : M/L → N, x + L ↦→ f(x) is therefore well defined and f = f ′ ◦ p. Since<br />

f, p are homomorphisms and p is surjective it follows that f ′ is a homomorphism.<br />

(2) The kernel <strong>of</strong> f ′ is Ker f/ Ker f = 0 so by 2.3.3 it is an isomorphism.


26 2. MODULES<br />

2.3.6. Corollary. Let p : M → M/N be the projection onto a factor module. The<br />

map L ′ ↦→ L = p −1 (L ′ ) gives a bijective correspondence between submodules<br />

in M/N and submodules in M containing N. Also L ′ = p(L) = L/N. This<br />

correspondence preserves inclusions, additions and intersections <strong>of</strong> submodules.<br />

Pro<strong>of</strong>. If L ′ is a submodule <strong>of</strong> M/N then clearly p(p−1 (L ′ )) = L ′ . If N ⊂ L<br />

is a submodule <strong>of</strong> M then clearly L ⊂ p−1 (p(L)). Moreover if x ∈ p−1 (p(L))<br />

then p(x) = p(y) for some y ∈ L and therefore y − x ∈ N. It follows that<br />

L = p−1 (p(L)) and the correspondence is bijective. Inclusions are easily seen to<br />

be preserved. Also easily p(L1 + L2) = p(L1) + p(L2) and p−1 (L ′ 1 ∩ L′ 2 ) =<br />

p−1 (L ′ 1 ) ∩ p−1 (L ′ 2 ) h<strong>old</strong>. The two resting equalities are consequences <strong>of</strong> this and<br />

bijectivity <strong>of</strong> the correspondence.<br />

2.3.7. Corollary. Let L ⊂ N ⊂ M be submodules. Then there is a canonical<br />

isomorphism<br />

M/N → (M/L)/(N/L)<br />

Pro<strong>of</strong>. The kernel <strong>of</strong> the surjective east-south composite<br />

M<br />

<br />

M/N<br />

<br />

M/L<br />

<br />

<br />

(M/L)/(N/L)<br />

is N. By 2.3.5 the horizontal lower factor map gives the isomorphism.<br />

2.3.8. Corollary. Let L, N ⊂ M be submodules. Then there is a canonical isomorphism<br />

N/N ∩ L → N + L/L<br />

given by x + N ∩ L ↦→ x + L.<br />

Pro<strong>of</strong>. The kernel <strong>of</strong> the east-south composite<br />

N<br />

<br />

N/N ∩ L<br />

<br />

N + L<br />

<br />

<br />

N + L/L<br />

is N ∩ L. Since x + y + L = x + L for x ∈ N, y ∈ L this composite is also<br />

surjective. By 2.3.5 the horizontal lower factor map gives the isomorphism.<br />

2.3.9. Proposition. Let f : M → N and g : N → L be homomorphisms such that<br />

Im f ⊂ Ker g. Then there is a unique homomorphism g ′ : Cok f → L such that<br />

g = g ′ ◦ p.<br />

Pro<strong>of</strong>. This follows from 2.3.5.<br />

M f<br />

<br />

N<br />

g<br />

<br />

<br />

p<br />

<br />

<br />

L<br />

<br />

g ′<br />

Cok f<br />

2.3.10. Lemma. Let R be a ring and M a module. For x ∈ M the map R →<br />

M, a ↦→ ax is the unique homomorphism such that 1 ↦→ x.


2.3. KERNEL AND COKERNEL 27<br />

Pro<strong>of</strong>. Let f(a) = ax. f(ab) = (ab)x = a(bx) = af(b) shows that f is a<br />

homomorphism. This argument does not use the commutativity 1.1.2 <strong>of</strong> R.<br />

2.3.11. Definition. Denote the homomorphism 2.3.10<br />

The annihilator <strong>of</strong> x is the ideal<br />

1x : R → M, a ↦→ ax<br />

Ann(x) = Ker 1x = {a ∈ R|ax = 0}<br />

For a subset Y ⊂ M the annihilator Ann(Y ) = <br />

y∈Y Ann(y) is the ideal <strong>of</strong><br />

elements<br />

Ann(Y ) = {a ∈ R|ay = 0, for all y ∈ Y }<br />

2.3.12. Proposition. Let R be a ring and Y a subset <strong>of</strong> a module M.<br />

(1) Ann(Y ) = Ann(RY ).<br />

(2) a ∈ Ann(M) if and only if aM = 0.<br />

(3) If modules M M ′ then Ann(M) = Ann(M ′ ).<br />

(4) The induced homomorphism<br />

is an isomorphism.<br />

1 ′ x : R/ Ann(x) → Rx, a + Ann(x) ↦→ ax<br />

Pro<strong>of</strong>. (1) If a ∈ Ann(Y ) then a biyi = biayi = 0 giving the not so obvious<br />

Ann(Y ) ⊂ Ann(RY ). (2) Clear since aM(x) = ax, 2.1.7. (3) Let f : M → M ′<br />

be an isomorphism and a ∈ R. Then af(x) = f(ax) expresses that f ◦ aM =<br />

aM ′ ◦ f. Since f is bijective, aM = 0 if and only if aM ′ = 0. By (2) Ann(M) =<br />

Ann(M ′ ). (4) This follows from 2.3.5.<br />

2.3.13. Corollary. Let I, J ⊂ R be ideals.<br />

(1) Ann(R/I) = I<br />

(2) If R/I R/J then I = J.<br />

2.3.14. Lemma. Let I ⊂ R be an ideal and M an R-module. If I ⊂ Ann(M)<br />

then M is an R/I-module with the scalar multiplication<br />

That is, M is an R/ Ann(M)-module.<br />

R/I × M → M, (a + I, x) ↦→ ax<br />

2.3.15. Example. Let a ∈ R give scalar multiplication aM : M → M, x ↦→ ax.<br />

(1) a ∈ Ann(Ker aM).<br />

(2) a ∈ Ann(Cok aM).<br />

(3) Ker aM and Cok aM are modules over the factor ring R/(a), 2.3.14.<br />

2.3.16. Definition. Let R be a ring and L, N ⊂ M submodules. The colon ideal<br />

N : L is the ideal <strong>of</strong> elements a ∈ R such that aL ⊂ N.<br />

2.3.17. Proposition. The colon ideal <strong>of</strong> submodules L, N ⊂ M is<br />

N : L = Ann(N + L/N)<br />

Pro<strong>of</strong>. If a ∈ N : L then aL ⊂ N. Therefore a(N + L) ⊂ N and a ∈ Ann(N +<br />

L/N). Conversely if a ∈ Ann(N + L/N) then aL ⊂ N and therefore a ∈ N :<br />

L.<br />

2.3.18. Exercise. (1) Give an example <strong>of</strong> a homomorphism f : M → N submodules<br />

M1, M2 ⊂ M such that f(M1 ∩ M2) = f(M1) ∩ f(M2).


28 2. MODULES<br />

(2) Give an example <strong>of</strong> a homomorphism f : M → N submodules N1, N2 ⊂ N such<br />

that f −1 (N1 + N2) = f −1 (M1) + f −1 (M2).<br />

(3) Let R be a ring and M a module. Show that M may be regarded as an R/ Ann(M)module<br />

in a natural way.<br />

(4) Let L, N ⊂ M be submodules. Show that Ann(L + N) = Ann(L) ∩ Ann(N).<br />

(5) Let f : M → N be a surjective homomorphism. Show that Ann(M) ⊂ Ann(N).<br />

(6) Let f : M → N be an injective homomorphism. Show that Ann(N) ⊂ Ann(M).<br />

2.4. Sum and product<br />

2.4.1.<br />

<br />

Lemma. Let R be a ring and (Mα) a family <strong>of</strong> modules. The product<br />

α Mα is the abelian group <strong>of</strong> all families (xα), xα ∈ Mα with term wise addition.<br />

The setting<br />

r(xα) = (rxα)<br />

is a scalar multiplication on <br />

α Mα. The direct sum <br />

α Mα is the subgroup<br />

<strong>of</strong> <br />

α Mα consisting <strong>of</strong> families with only finitely many nonzero terms. This is a<br />

submodule.<br />

Pro<strong>of</strong>. The laws in 2.1.1 are true for each factor and therefore trivially verified for<br />

the product and sum.<br />

2.4.2. Definition. Let R be a ring and Mα a family <strong>of</strong> modules. By 2.4.1 there are<br />

modules and homomorphisms<br />

(1) The direct product is <br />

α Mα.<br />

(2) The projections pβ : <br />

α Mα → Mβ are the homomorphisms pβ((xα)) =<br />

xβ.<br />

(3) The direct sum is <br />

α Mα. Elements in <br />

α Mα <br />

are written as finite sums<br />

xα.<br />

(4) The injections iβ : Mβ → <br />

α Mα are the homomorphisms given by iβ(x) =<br />

(xα), where xβ = x and xα = 0, α = β.<br />

2.4.3. Proposition. Let R be a ring and Mα a family <strong>of</strong> modules.<br />

(1) Given a family <strong>of</strong> homomorphisms fα : L → Mα, then there exists a unique<br />

homomorphism f : L → Mα such that fα = pα ◦ f.<br />

L <br />

<br />

fα <br />

<br />

f<br />

Mα<br />

pα<br />

<br />

<br />

<br />

α Mα<br />

(2) Given a family <strong>of</strong> homomorphisms gα : Mα → L, then there exists a unique<br />

homomorphism g : Mα → L such that gα = g ◦ iα.<br />

<br />

α Mα<br />

g<br />

<br />

<br />

<br />

<br />

<br />

iα <br />

gα<br />

<br />

Mα<br />

Pro<strong>of</strong>. (1) f(y) = (fα(y)) is the unique homomorphism. (2) g( xα) = gα(xα)<br />

is well defined since only finitely many xα = 0 and a homomorphism.<br />

L


2.4. SUM AND PRODUCT 29<br />

2.4.4. Definition. A family <strong>of</strong> submodules Mα ⊂ M constitutes a direct sum if<br />

any element x ∈ <br />

α Mα has a unique finite representation<br />

x = <br />

xα, xα ∈ Mα<br />

2.4.5. Proposition. The following conditions are equivalent<br />

(1) The family Mα ⊂ M constitutes a direct sum.<br />

(2) The natural homomorphism<br />

<br />

Mα → <br />

is an isomorphism.<br />

(3) For all β<br />

α<br />

α<br />

α<br />

Mα<br />

Mβ ∩ <br />

Mα = 0<br />

α=β<br />

Pro<strong>of</strong>. (1) and (2) are equivalent. If (1) is true and xβ = <br />

α=β xα ∈ Mβ ∩<br />

<br />

α=β Mα, then xβ − <br />

α=β xα = 0. Therefore by uniqueness xβ = 0 and (3)<br />

is true. Conversely if (3) is true and <br />

α xα = 0, then xβ = − <br />

α=β xα ∈<br />

Mβ ∩ <br />

α=β Mα = 0. for any β. This shows uniqueness and therefore (1) is<br />

true.<br />

2.4.6. Definition. Let R be a ring. A module isomorphic to a direct sum <br />

α<br />

R <strong>of</strong><br />

copies <strong>of</strong> the ring R is a free module.<br />

A basis <strong>of</strong> a module, is a subset Y such that any element admits a unique finite<br />

representation <br />

α aαyα, where aα ∈ R, yα ∈ Y .<br />

The standard basis <strong>of</strong> <br />

α R consists <strong>of</strong> the elements eα, where each is a family<br />

with 0 for β = α and exactly 1 at index α.<br />

2.4.7. Proposition. A module is free if and only if it admits a basis. If yα is a basis<br />

<strong>of</strong> F then there is an isomorphism<br />

f : <br />

R → F<br />

where f( aα) = aαyα.<br />

α<br />

Pro<strong>of</strong>. Given a free module f : <br />

α R → F then yα = f(eα) is a basis. Conversely<br />

given a basis yα ∈ F then 1yα : R → F 2.3.11 is a family <strong>of</strong> homomorphisms<br />

giving a homomorphism f : <br />

α R → F by 2.4.3. As f( aα) = aαyα<br />

it follows that f is bijective and therefore by 2.1.5 an isomorphism.<br />

2.4.8. Remark. The polynomial ring R[X1, . . . , Xn] is free as R-module.<br />

2.4.9. Example. (1) A nonzero ideal is a free module if and only if it has a basis<br />

consisting <strong>of</strong> a nonzero divisor. Namely if x1 = x2 where in a basis then the<br />

product x1x2 has two different representations.<br />

(2) Let I ⊂ R be an ideal. The module R/I is free if and only if I = 0 or I = R.<br />

2.4.10. Proposition. Any module over a field is free. Conversely if any module<br />

over a nonzero ring is free, then the ring is a field<br />

Pro<strong>of</strong>. By Zorn’s lemma any vector space admits a basis. If I ⊂ R is an ideal and<br />

R/I is free, then I = Ann(R/I) is either 0 or R. So R is a field.


30 2. MODULES<br />

2.4.11. Proposition. Let F be a free module with basis yα. For a module M and<br />

a family <strong>of</strong> elements xα ∈ M there is a unique homomorphism g : F → M such<br />

that g(yα) = xα given by g( aαyα) = aαxα.<br />

Pro<strong>of</strong>. The basis yα ∈ F gives the isomorphism 2.4.7 f : <br />

α R → F . The<br />

family 1xα : R → M gives a homomorphism g ′ : <br />

R → M by 2.4.3. Then<br />

g = g ′ ◦ f −1 .<br />

2.4.12. Corollary. Let M be an R-module and <br />

M R the free module with basis<br />

ex indexed by x ∈ M. The homomorphism<br />

<br />

R → M, axex ↦→ ax x<br />

M<br />

is surjective identifying M as a factor module <strong>of</strong> a free module in a natural way.<br />

2.4.13. Definition. A module is indecomposable if it is not isomorphic to a direct<br />

sum <strong>of</strong> two nonzero submodules, otherwise decomposable.<br />

2.4.14. Example. Q is an indecomposable Z-module. Namely if m1<br />

n1<br />

nonzero numbers in two submodules, then n1m2 m1<br />

n1<br />

ber in the intersection.<br />

α<br />

= n2m1 m2<br />

n2<br />

, m2<br />

n2 are<br />

is a nonzero num-<br />

2.4.15. Exercise. (1) Show that if a ring is decomposable as a module, then it is the<br />

product <strong>of</strong> two nonzero rings.<br />

(2) Let Mα be a finite family <strong>of</strong> modules. Show that Mα = Mα,<br />

(3) Let Nα ⊂ Mα be a family <strong>of</strong> submodules modules. Show that<br />

Mα/ Nα Mα/Nα<br />

and that Mα/ Nα Mα/Nα<br />

2.5. Homomorphism modules<br />

2.5.1. Lemma. Let R be a ring and f, g : M → N homomorphisms.<br />

(1) (f + g)(x) = f(x) + g(x) is a homomorphism.<br />

(2) If a ∈ R, then (af)(x) = af(x) is a homomorphism.<br />

Pro<strong>of</strong>. Calculate according to 2.1.1. (1) (f + g)(x + y) = f(x + y) + g(x + y) =<br />

f(x) + f(y) + g(x) + g(y) = (f + g)(x) + (f + g)(y) and (f + g)(ax) =<br />

f(ax) + g(ax) = a(f(x) + g(x)) = a(f + g)(x). (2) (af)(x + y) = af(x + y) =<br />

af(x) + af(y) = (af)(x) + (af)(y) and (af)(bx) = af(bx) = abf(x) =<br />

baf(x) = b(af)(x). The last calculation uses that R is commutative 1.1.2 (4).<br />

2.5.2. Definition. Let R be a ring and M, N modules. By 2.5.1, the homomorphism<br />

module HomR(M, N) is the additive group <strong>of</strong> all homomorphism with<br />

scalar multiplication<br />

R × HomR(M, N) → HomR(M, N), (a, f) ↦→ af = [x ↦→ af(x)]<br />

2.5.3. Definition. Let a ∈ R be a ring and f : M → M ′ , g : N → N ′ , h, k :<br />

M ′ → N homomorphisms <strong>of</strong> modules. By 2.5.1<br />

(1) (h + k) ◦ f = h ◦ f + k ◦ f.<br />

(2) (ah) ◦ f = a(h ◦ f).<br />

(3) g ◦ (h + k) = g ◦ h + g ◦ k.<br />

(4) g ◦ (ah) = a(g ◦ h).


There is induced a homomorphism<br />

<strong>of</strong> R-modules.<br />

2.5. HOMOMORPHISM MODULES 31<br />

Hom(f, g) : HomR(M ′ , N) → HomR(M, N ′ )<br />

(h : M ′ → N) ↦→ (g ◦ h ◦ f : M → N ′ )<br />

2.5.4. Definition. Let R, S be a rings. A functor is a construction T , which to<br />

R-modules M, N associates S-modules T (M), T (N) and a homomorphism<br />

such that<br />

HomR(M, N) → HomS(T (M), T (N)), f ↦→ T (f)<br />

(1) T (1M) = 1 T (M)<br />

(2) T (g ◦ f) = T (g) ◦ T (f)<br />

In case the homomorphism goes<br />

and<br />

HomR(M, N) → HomS(T (N), T (M)), f ↦→ T (f)<br />

(1) T (1M) = 1 T (M)<br />

(2) T (g ◦ f) = T (f) ◦ T (g)<br />

the functor is contravariant. Clearly functors transform isomorphisms into isomorphism.<br />

Given functors T, T ′ a natural homomorphism is a family νM : T (M) → T ′ (M)<br />

<strong>of</strong> homomorphisms, such that for each f : M → N the following diagram commutes<br />

T (M)<br />

T (f)<br />

<br />

νN <br />

T (N)<br />

In the contravariant case the diagram is<br />

T (M)<br />

<br />

T (f)<br />

T (N)<br />

νM <br />

T ′ (M)<br />

νM <br />

T ′ (f)<br />

<br />

T ′ (N)<br />

T ′ (M)<br />

<br />

T ′ (f)<br />

νN <br />

T ′ (N)<br />

A natural isomorphism is a natural homomorphism such that each νM is an isomorphism.<br />

2.5.5. Proposition. Let R be a ring.<br />

(1) The construction<br />

is a functor.<br />

(2) The construction<br />

is a contravariant functor<br />

N ↦→ HomR(M, N), g ↦→ Hom(1M, g)<br />

M ↦→ HomR(M, N), f ↦→ Hom(f, 1N)


32 2. MODULES<br />

Pro<strong>of</strong>. By 2.5.3 the construction on homomorphisms are homomorphisms. Given<br />

also homomorphisms f ′ : M ′ → M ′′ and g ′ : N ′ → N ′′ . Then<br />

and<br />

Hom(1M, g ′ ◦ g) = Hom(1M, g ′ ) ◦ Hom(1M, g)<br />

Hom(f ′ ◦ f, 1N) = Hom(f, 1N) ◦ Hom(f ′ , 1N)<br />

showing the conditions on compositions.<br />

2.5.6. Corollary. Let a ∈ R give scalar multiplications aM, aN.<br />

(1)<br />

Hom(aM, 1N) = Hom(1M, aN) : HomR(M, N) → HomR(M, N) f ↦→ af<br />

is scalar multiplication a HomR(M,N).<br />

(2) The map R → HomR(M, M), a ↦→ aM is a homomorphism.<br />

2.5.7. Example. Let R = R1 × R2 be the product ring. The constructions in 2.2.9<br />

is: (1) A functor which to a pair <strong>of</strong> an R1-module and an R2-module associates an<br />

R-module.(2) A functor which to an R-module associates a pair <strong>of</strong> an R1-module<br />

and an R2-module.<br />

2.5.8. Proposition. Let R be a ring and Mα a family <strong>of</strong> modules. For any module<br />

N there are natural isomorphisms<br />

(1) HomR( <br />

α Mα, N) <br />

α HomR(Mα, N)<br />

(2) HomR(N, <br />

α Mα) <br />

α HomR(N, Mα)<br />

Pro<strong>of</strong>. This is 2.4.3 reformulated. (1) A homomorphism g : <br />

α Mα → N is<br />

uniquely determined by the family gα = g ◦ iα : Mα → N. (2) A homomorphism<br />

f : N → <br />

α Mα is uniquely determined by the family fα = pα ◦f : N → Mα.<br />

2.5.9. Lemma. Let R be a ring and M, N modules. For x ∈ M there is a homomorphism<br />

HomR(M, N) → N, f ↦→ f(x).<br />

Pro<strong>of</strong>. Calculate according to 2.1.1 (f + g) ↦→ (f + g)(x) = f(x) + g(x) and<br />

(af) ↦→ (af)(x) = af(x).<br />

2.5.10. Definition. The natural homomorphism 2.5.9<br />

is the evaluation at x.<br />

evx : HomR(M, N) → N, f ↦→ f(x)<br />

2.5.11. Lemma. There is a natural homomorphism<br />

M → HomR(HomR(M, N), N), x ↦→ evx<br />

Pro<strong>of</strong>. Calculate according to 2.1.1 evx+y(f) = f(x + y) = f(x) + f(y) =<br />

evx(f) + evy(f). and evax(f) = f(ax) = af(x) = aevx(f) to see that the map<br />

is a homomorphism.<br />

2.5.12. Proposition. Let R be a ring and M a module. The evaluation<br />

ev1 : HomR(R, M) M, f ↦→ f(1)<br />

is a natural isomorphism. x ↦→ 1x 2.3.11 is the inverse.<br />

Pro<strong>of</strong>. Calculate the composite ev1(x ↦→ 1x) = 1x(1) = x and 1 f(1)(a) =<br />

af(1) = f(a) proving the claims.


2.6. TENSOR PRODUCT MODULES 33<br />

2.5.13. Definition. Let R be a ring and M a module. The dual module is<br />

M ∨ = HomR(M, R)<br />

If f : M → N is a homomorphism, then the dual homomorphism is<br />

f ∨ = Hom(f, 1R) : N ∨ → M ∨<br />

This construction is a contravariant functor.<br />

2.5.14. Lemma. There is a natural homomorphism<br />

where evx(f) = f(x) 2.5.10.<br />

M → M ∨∨ = HomR(HomR(M, R), R), x ↦→ evx<br />

Pro<strong>of</strong>. This is a special case <strong>of</strong> 2.5.11<br />

2.5.15. Definition. A module M is a reflexive module if the homomorphism 2.1.14,<br />

is an isomorphism.<br />

2.5.16. Example. Let R be a ring.<br />

M → M ∨∨<br />

(1) The module R is reflexive.<br />

(2) If (a) = R, (0) in a domain, then R/(a) is not reflexive.<br />

2.5.17. Exercise. (1) Show that HomZ(Q, Z) = 0.<br />

(2) Calculate HomZ(Z/(m), Z/(n)) = 0 for integers m, n.<br />

(3) Let I ⊂ R be an ideal and M a module. Show that HomR(R/I, M) = {x ∈ M|I ⊂<br />

Ann(x)}.<br />

(4) Let R be a ring. Show that a free module with a finite basis is a reflexive module.<br />

(5) If (n) ⊂ (m) ⊂ Z, then show that (m)/(n) is a reflexive Z/(n)-module.<br />

2.6. Tensor product modules<br />

2.6.1. Definition. Let R be a ring and M, N modules. The tensor product module<br />

M ⊗R N = F/F ′<br />

is the factor module F/F ′ , where F = ⊕M×NR is the free module with basis<br />

(x, y) = e (x,y) 2.4.6 and F ′ is the submodule generated by all elements <strong>of</strong> form<br />

(x1 + x2, y) − (x1, y) − (x2, y), (x, y1 + y2) − (x, y1) − (x, y2)<br />

(ax, y) − a(x, y), (x, ay) − a(x, y)<br />

The projection <strong>of</strong> the basis element (x, y) onto M ⊗R N is x ⊗ y = (x, y) + F ′ .<br />

2.6.2. Remark. The relations are interpreted.<br />

(1) There are identities in M ⊗R N<br />

(x1 + x2) ⊗ y = x1 ⊗ y + x2 ⊗ y, x ⊗ (y1 + y2) = x ⊗ y1 + x ⊗ y2<br />

(2) The map<br />

ax ⊗ y = a(x ⊗ y) = x ⊗ ay<br />

⊗ : M × N → M ⊗R N, (x, y) ↦→ x ⊗ y<br />

has partial maps x ↦→ x⊗y : M → M ⊗R N and y ↦→ x⊗y : N → M ⊗R N<br />

that are all homomorphisms.


34 2. MODULES<br />

(3) The formation <strong>of</strong> partial homomorphism are again homomorphisms.<br />

and<br />

N → HomR(M, M ⊗R N)), y ↦→ (x ↦→ x ⊗ y)<br />

M → HomR(N, M ⊗R N)), x ↦→ (y ↦→ x ⊗ y)<br />

2.6.3. Proposition. Given a map µ : M × N → L such that the partial maps x ↦→<br />

µ(x, y) : M → L and y ↦→ µ(x, y) : N → L are homomorphisms. Then there<br />

exists a unique homomorphism u : M ⊗R N → L such that u(x ⊗ y) = µ(x, y).<br />

⊗<br />

M × N <br />

<br />

M ⊗R N<br />

<br />

<br />

µ u<br />

<br />

<br />

<br />

L<br />

Pro<strong>of</strong>. By 2.6.1 M ⊗R N = F/F ′ . The homomorphism 2.4.11 F → K, (x, y) ↦→<br />

µ(x, y) has F ′ in the kernel. 2.3.5 gives the homomorphism u.<br />

2.6.4. Remark. Two homomorphisms u, v : M ⊗RN → L are equal if u(x⊗y) =<br />

v(x ⊗ y) for all x ∈ M, y ∈ N.<br />

2.6.5. Proposition. Let R be a ring and f : M → M ′ , g : N → N ′ homomorphisms<br />

<strong>of</strong> modules. Then there is induced a homomorphism<br />

Pro<strong>of</strong>. The map south-east<br />

M ⊗R N → M ′ ⊗R N ′ , x ⊗ y ↦→ f(x) ⊗ g(y)<br />

M × N<br />

f×g<br />

<br />

M ′ × N ′<br />

⊗<br />

<br />

M ⊗R N<br />

<br />

⊗<br />

<br />

M ′<br />

⊗R N ′<br />

satisfies the assumptions in 2.6.3 to induce the right vertical map x ⊗ y ↦→ f(x) ⊗<br />

g(y).<br />

2.6.6. Definition. f ⊗ g : M ⊗R N → M ′ ⊗R N ′ is the induced homomorphism<br />

2.6.5.<br />

2.6.7. Proposition. Let R be a ring. The constructions<br />

(1)<br />

(2)<br />

are functors.<br />

M ↦→ M ⊗R N, f ↦→ f ⊗ 1N<br />

N ↦→ M ⊗R N, g ↦→ 1M ⊗ g<br />

Pro<strong>of</strong>. Given also homomorphisms f ′ : M ′ → M ′′ and g ′ : N ′ → N ′′ . Then by<br />

2.6.4<br />

f ′ ◦ f ⊗ g ′ ◦ g = f ′ ⊗ g ′ ◦ f ⊗ g<br />

The rest follows directly from 2.6.5.


2.6. TENSOR PRODUCT MODULES 35<br />

2.6.8. Corollary. Let a ∈ R give scalar multiplications aM, aN. The homomorphisms<br />

aM ⊗ 1N = 1M ⊗ aN : M ⊗R N → M ⊗R N, x ⊗ y ↦→ a(x ⊗ y)<br />

is scalar multiplication aM⊗RN.<br />

2.6.9. Example. Let R be a ring.<br />

(1) Then there is an isomorphism<br />

R ⊗R R R, a ⊗ b ↦→ ab<br />

(2) Let M, N, L be modules. Composition <strong>of</strong> maps gives a homomorphism<br />

HomR(N, L) ⊗R HomR(M, N) → HomR(M, L), g ⊗ f ↦→ g ◦ f<br />

by 2.5.3. This is a natural homomorphism in each variable.<br />

(3) For a module M composition<br />

HomR(M, M) ⊗R HomR(M, M) → HomR(M, M), g ⊗ f ↦→ g ◦ f<br />

gives HomR(M, M) a structure <strong>of</strong> a normally noncommutative ring. The<br />

map R → HomR(M, M), a ↦→ aM is a ring homomorphism.<br />

2.6.10. Proposition. Let R be a ring and M, N, L modules. Then there are natural<br />

isomorphisms<br />

(1) M ⊗R R M, x ⊗ a ↦→ ax<br />

(2) (M ⊗R N) N ⊗R M, x ⊗ y ↦→ y ⊗ x<br />

(3) (M ⊗R N) ⊗R L M ⊗R (N ⊗R L), (x ⊗ y) ⊗ z ↦→ x ⊗ (y ⊗ z)<br />

Pro<strong>of</strong>. (1) M × R → M, (x, a) ↦→ ax induces the homomorphism M ⊗R R →<br />

M, x ⊗ a ↦→ ax by 2.6.3. The map M → R ⊗R M, x ↦→ 1 ⊗ x is the inverse. (2)<br />

M × N → N ⊗R M, (x, y) ↦→ y ⊗ x induces the homomorphism (M ⊗R N) <br />

N ⊗R M, x ⊗ y ↦→ y ⊗ x by 2.6.3. The inverse is constructed similarly and the<br />

composites are the identities by the uniqueness statement in 2.6.3. (3) For a fixed<br />

z ∈ L the map M × N → M ⊗R (N ⊗R L), (x, y) ↦→ x ⊗ (y ⊗ z) induces the<br />

homomorphism µz : M ⊗R N → M ⊗R (N ⊗R L), x ⊗ y ↦→ x ⊗ (y ⊗ z) by 2.6.3.<br />

Finally the map (M⊗RN)×L → M⊗R(N⊗RL), (x⊗y, z) ↦→ µz(x⊗y). induces<br />

the homomorphism (M ⊗RN)⊗RL M ⊗R(N ⊗RL), (x⊗y)⊗z ↦→ x⊗(y⊗z)<br />

by 2.6.3. The inverse is constructed similarly and the composites are the identities<br />

by the uniqueness statement in 2.6.3.<br />

2.6.11. Proposition. Let R be a ring and Mα a family <strong>of</strong> modules. For any module<br />

N there is a natural isomorphism<br />

( <br />

Mα) ⊗R N <br />

(Mα ⊗R N)<br />

α<br />

giving the identification ( xα) ⊗ y = (xα ⊗ y).<br />

Pro<strong>of</strong>. 2.4.3 and 2.6.3 give a pair <strong>of</strong> inverse homomorphisms. Fix y ∈ N. The<br />

family gα : Mα → (Mα ⊗R N), xα ↦→ xα ⊗ y induces by 2.4.3 a homomorphism<br />

gy : Mα → (Mα ⊗R N). The map ( Mα) × N → (Mα ⊗R<br />

N), ( xα, y) ↦→ gy( xα) induces by 2.6.3 the homomorphism ( Mα) ⊗R<br />

N → (Mα⊗RN), ( xα)⊗y ↦→ xα⊗y. The family iα⊗1N : Mα⊗RN →<br />

( Mα) ⊗R N induces by 2.4.3 the inverse.<br />

α


36 2. MODULES<br />

2.6.12. Example. Let R be a ring, F a free module with basis yα and G a free<br />

module with basis zβ. Then F ⊗R G is a free module with basis yα ⊗ zβ.<br />

2.6.13. Proposition. Let R be a ring and M, N, L modules. Then there is a natural<br />

isomorphism<br />

HomR(M ⊗R N, L) HomR(M, HomR(N, L))<br />

(x ⊗ y ↦→ g(x)(y)) ← g<br />

f ↦→ (x ↦→ [y ↦→ f(x ⊗ y)])<br />

Pro<strong>of</strong>. A given f : M ⊗R N → L is mapped to the composite homomorphism<br />

M → HomR(N, M ⊗R N) → HomR(N, L), 2.5.4 and 2.6.2. This is a homomorphism<br />

as map <strong>of</strong> f by 2.5.4. Given g : M → HomR(N, L) the map M × N →<br />

L, (x, y) ↦→ g(x)(y) induces a homomorphism M ⊗R N → L, x ⊗ y ↦→ g(x)(y)<br />

by 2.6.3. Clearly the maps are inverse to each other and therefore giving an isomorphism<br />

by 2.1.5.<br />

2.6.14. Exercise. (1) Show that Q ⊗Z Q/Z 0.<br />

(2) Show that Z/(m) ⊗Z Z/(n) = 0 if (m, n) = Z.<br />

(3) Let P, Q ⊂ R be different maximal ideals and M a module. Show that M/P M ⊗R<br />

M/QM = 0.<br />

2.7. Change <strong>of</strong> rings<br />

2.7.1. Proposition. (1) Let φ : R → S be a ring homomorphism and N an S<br />

module. The restriction scalars 2.1.11 viewing N as an R-module through<br />

φ, R × N → N, (a, x) ↦→ ax = φ(a)x, is a functor from S-modules to<br />

R-modules.<br />

(2) Let I ⊂ R be an ideal. Restriction <strong>of</strong> scalars along R → R/I identifies<br />

R/I-modules M with R-modules such that I ⊂ Ann(M). For R/I-modules<br />

M, N there is a natural isomorphism<br />

HomR(M, N) Hom R/I(M, N)<br />

Pro<strong>of</strong>. This is a restatement <strong>of</strong> 2.1.11 using 2.5.4.<br />

2.7.2. Lemma. Let R → S be a ring homomorphism, M an R-module and N an<br />

S-module. Then<br />

is an S-scalar multiplication.<br />

S × M ⊗R N → M ⊗R N, (b, x ⊗ y) ↦→ x ⊗ by<br />

Pro<strong>of</strong>. For fixed b ∈ S the map M × N → M ⊗R N, (x, y) ↦→ x ⊗ by induces<br />

the homomorphism µb : M ⊗R N → M ⊗R N, x ⊗ y ↦→ x ⊗ by by 2.6.3. This<br />

gives a well defined scalar multiplication S × M ⊗R N → M ⊗R N, (b, x ⊗ y) ↦→<br />

µb(x ⊗ y).<br />

2.7.3. Definition. Let R → S be a ring homomorphism and M an R-module. The<br />

change <strong>of</strong> ring S-module is M ⊗R S with S-scalar multiplication 2.7.2<br />

S × M ⊗R S → M ⊗R S, (b, x ⊗ c) ↦→ x ⊗ bc


2.7.4. Proposition. The construction<br />

2.7. CHANGE OF RINGS 37<br />

M ↦→ M ⊗R S<br />

and<br />

f : M → M ′ ↦→ f ⊗ 1S : M ⊗R S → M ′ ⊗R S<br />

is a functor from R-modules to S-modules.<br />

Pro<strong>of</strong>. This is clear from 2.7.2 and 2.6.7.<br />

2.7.5. Proposition. Let R → S be a ring homomorphism, M an R-module and N<br />

an S-modules. Then there is a natural isomorphism <strong>of</strong> S-modules.<br />

M ⊗R S ⊗S N M ⊗R N, x ⊗ b ⊗ y ↦→ x ⊗ by<br />

Pro<strong>of</strong>. The homomorphism v : S ⊗S N → N, b ⊗ y ↦→ by is an isomorphism,<br />

2.6.10. The homomorphism 1M ⊗ v : M ⊗R S ⊗S N M ⊗R N is an R-module<br />

isomorphism, 2.6.7. The identity x ⊗ bc ⊗ y = x ⊗ b ⊗ cy proves this to be an<br />

S-module homomorphism.<br />

2.7.6. Proposition. Let R → S be a ring homomorphism, M an R-module and N<br />

an S-modules. Then there is a natural isomorphism<br />

HomR(M, N) HomS(M ⊗R S, N), f ↦→ (x ⊗ b ↦→ bf(x))<br />

Pro<strong>of</strong>. A given f is mapped to the composite M⊗R → N ⊗R S → N which is an<br />

S-homomorphism. Given a homomorphism g : M ⊗R S → N then the composite<br />

M → M ⊗R S → N is an R-homomorphism and an inverse to the first given<br />

map.<br />

2.7.7. Lemma. Let R → S be a ring homomorphism, M an R-module and N an<br />

S-module. Then<br />

S × HomR(N, M) → HomR(N, M), (b, f : N → M) ↦→ (y ↦→ f(by))<br />

is an S-scalar multiplication.<br />

Pro<strong>of</strong>. The map (b, f) ↦→ f ◦ bN satisfies the laws 2.1.1.<br />

2.7.8. Definition. Let R → S be a ring homomorphism and M an R-module. The<br />

induced module is the S-module HomR(S, M) with S-scalar multiplication 2.7.7<br />

S × HomR(S, M) → HomR(S, M), (b, f : S → M) ↦→ (c ↦→ f(bc))<br />

2.7.9. Proposition. The induced module<br />

M ↦→ HomR(S, M)<br />

and<br />

f : M → M ′ ↦→ Hom(1S, f) : HomR(S, M) → HomR(S, M ′ )<br />

is a functor from R-modules to S-modules.<br />

Pro<strong>of</strong>. This is clear from 2.7.7 using 2.5.4.<br />

2.7.10. Proposition. Let R → S be a ring homomorphism and M an R-module<br />

and N an S-modules. Then there is a natural isomorphism<br />

HomR(N, M) HomS(N, HomR(S, M)), f ↦→ (y ↦→ [b ↦→ f(by)])<br />

Pro<strong>of</strong>. g ↦→ (y ↦→ g(y)(1)) is an inverse.<br />

2.7.11. Example. Let I ⊂ R be an ideal and R → R/I the projection.


38 2. MODULES<br />

(1) The change <strong>of</strong> ring functor maps an R-module M to the R/I-module M/IM.<br />

The natural isomorphism 2.7.6 is<br />

HomR(M, N) Hom R/I(M/IM, N)<br />

for any R/I-module N.<br />

(2) The induced module functor maps an R-module M to the R/I-module {x ∈<br />

M|I ⊂ Ann(x)}. The natural isomorphism 2.7.10 is<br />

HomR(N, M) Hom R/I(N, HomR(R/I, M))<br />

for any R/I-module N.<br />

2.7.12. Definition. Let R → S, S ′ be ring homomorphisms. The tensor product<br />

ring over R is S ⊗R S ′ with multiplication given by (b ⊗ b ′ )(c ⊗ c ′ ) = bc ⊗ b ′ c ′<br />

extended by linearity. R → S ⊗R S ′ , r ↦→ r ⊗ 1 = 1 ⊗ r is the natural ring<br />

homomorphism.<br />

2.7.13. Proposition. Let φ, φ ′ : R → S, S ′ and ψ, ψ ′ : S, S ′ → T give a commutative<br />

diagram <strong>of</strong> ring homomorphisms, ψ ◦ φ = ψ ′ ◦ φ ′ . Then b ⊗ b ′ ↦→ ψ(b)ψ ′ (b ′ )<br />

is the unique homomorphism making the following diagram commutative.<br />

Pro<strong>of</strong>. This is clear by 2.6.3.<br />

R <br />

S ′<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

S <br />

<br />

<br />

S ⊗R S <br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

′<br />

<br />

T<br />

2.7.14. Example. Let R → S be a ring homomorphism. Then<br />

is an isomorphism.<br />

R[X] ⊗R S S[X]<br />

2.7.15. Exercise. (1) Show that the change <strong>of</strong> rings <strong>of</strong> a free R-module is a free Smodule.<br />

(2) Let φ : R → S be a ring homomorphism. Show that the change <strong>of</strong> rings <strong>of</strong> a<br />

scalar multiplication a : M → M on an R-module is a scalar multiplication φ(a) :<br />

M ⊗R S → M ⊗R S.<br />

(3) Show that the change <strong>of</strong> rings <strong>of</strong> the composition <strong>of</strong> two homomorphisms is the<br />

composition <strong>of</strong> the change <strong>of</strong> rings <strong>of</strong> each homomorphism.<br />

(4) Show the isomorphism<br />

R[X] ⊗R R[Y ] R[X, Y ]


3<br />

Exact sequences <strong>of</strong> modules<br />

3.1. Exact sequences<br />

3.1.1. Definition. Let f : M → N and g : N → L be homomorphisms <strong>of</strong><br />

modules. The sequence<br />

<strong>of</strong> homomorphisms is a<br />

M f<br />

g<br />

<br />

N<br />

<br />

L<br />

(1) 0-sequence: g ◦ f = 0 or equivalently Im f ⊂ Ker g<br />

(2) exact sequence: Im f = Ker g<br />

For a sequence <strong>of</strong> more homomorphisms the conditions should be satisfied for<br />

every consecutive composition. E.g. The sequence<br />

M f<br />

g<br />

<br />

N<br />

<br />

h<br />

L <br />

K<br />

is a 0-sequence if g ◦ f = 0 and h ◦ g = 0. The sequence is exact if Im f = Ker g<br />

and Im g = Ker h.<br />

3.1.2. Remark. An interpretation <strong>of</strong> 2.3.3 gives:<br />

(1) The sequence<br />

is exact if and only if f is injective.<br />

(2) The sequence<br />

0<br />

M f<br />

f<br />

<br />

M<br />

<br />

N<br />

is exact if and only if f is surjective.<br />

(3) The sequence<br />

0<br />

f<br />

<br />

M<br />

<br />

N<br />

is exact if and only if f is an isomorphism.<br />

3.1.3. Proposition. (1) For a homomorphism f : M → N the sequence<br />

0<br />

<br />

Ker f<br />

f<br />

<br />

M<br />

<br />

N<br />

<br />

N<br />

<br />

0<br />

<br />

0<br />

<br />

Cok f<br />

is exact.<br />

(2) For scalar multiplication with a ∈ R on M the sequence<br />

0<br />

is exact.<br />

<br />

Ker aM<br />

aM <br />

M <br />

M<br />

39<br />

<br />

M/aM<br />

<br />

0<br />

<br />

0


40 3. EXACT SEQUENCES OF MODULES<br />

3.1.4. Proposition. The 0-sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

is exact if and only if the following equivalent statements are satisfied.<br />

(1) f is an isomorphism onto Ker g.<br />

(2) Given a homomorphism h : K → N such that g ◦ h = 0 then there is a<br />

unique h ′ : K → M such that h = f ◦ h ′ .<br />

0<br />

f g<br />

<br />

M <br />

N<br />

h<br />

K<br />

′<br />

<br />

<br />

h <br />

<br />

<br />

<br />

Pro<strong>of</strong>. (1) This is clearly equivalent with exactness. (2) Assume the sequence exact.<br />

Im h ⊂ Ker g = Im f, so by (1) put h ′ = f −1 ◦ h. Assume (2) satisfied and<br />

apply it to Ker g → M to see that (1) is satisfied.<br />

3.1.5. Proposition. The 0-sequence<br />

M f<br />

g<br />

<br />

N<br />

is exact if and only if the following equivalent statements are satisfied.<br />

(1) The factor homomorphism 2.3.9 g ′ : Cok f → L induced by g is an isomorphism.<br />

(2) Given a homomorphism k : N → K such that k ◦ f = 0 then there is a<br />

unique k ′ : L → K such that k = k ′ ◦ g.<br />

M f<br />

<br />

L<br />

g<br />

<br />

N <br />

L<br />

<br />

k<br />

<br />

k<br />

<br />

′<br />

<br />

K<br />

Pro<strong>of</strong>. (1) The equivalence follows from 2.3.9. (2) Assume the sequence exact. By<br />

2.3.5 there is k ′′ : Cok f → K such that k ′′ ◦ p = k. By (1) put k ′ = k ′′ ◦ g −1 .<br />

Assume (2) satisfied and apply it to N → Cok f to see that (1) is satisfied.<br />

3.1.6. Proposition. Let<br />

Mα<br />

<br />

Nα<br />

<br />

Lα<br />

<br />

L<br />

<br />

L<br />

<br />

be a family <strong>of</strong> exact sequences. Then there are exact sequences:<br />

(1) The sum<br />

<br />

Mα<br />

<br />

Nα<br />

Lα<br />

(2) The product<br />

Mα<br />

<br />

Nα<br />

<br />

0<br />

<br />

0<br />

<br />

Lα<br />

Pro<strong>of</strong>. Clear since kernel and image is calculated componentwise.<br />

3.1.7. Definition. An exact sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

is a short exact sequence. That is f is injective, Im f = Ker g and g is surjective.<br />

<br />

L<br />

<br />

0


3.1. EXACT SEQUENCES 41<br />

<br />

<br />

<br />

<br />

3.1.8. Proposition.<br />

quence<br />

(1) Let I ⊂ R be an ideal, then there is a short exact se-<br />

0 I R R/I 0<br />

(2) Let M ⊂ N be a submodule, then there is a short exact sequence<br />

0<br />

<br />

M<br />

<br />

N<br />

<br />

N/M<br />

(3) For scalar multiplication with nonzero divisor a ∈ R on M the sequence<br />

0<br />

aM <br />

M <br />

M<br />

<br />

M/aM<br />

is a short exact sequence.<br />

(4) Given a homomorphism f : M → N there are associated two short exact<br />

sequences.<br />

and<br />

0<br />

0<br />

<br />

Ker f<br />

<br />

Im f<br />

f<br />

<br />

M <br />

Im f<br />

<br />

N<br />

<br />

Cok f<br />

(5) For scalar multiplication with any a ∈ R on M there are associated two short<br />

exact sequences.<br />

and<br />

0<br />

0<br />

<br />

Ker aM<br />

<br />

aM<br />

<br />

M<br />

aM <br />

M <br />

aM<br />

<br />

M/aM<br />

3.1.9. Definition. Let f : M → N be a homomorphism.<br />

(1) f has a retraction if there is a homomorphism u : N → M such that u ◦ f =<br />

1M.<br />

(2) f has a section if there is a homomorphism v : N → M such that f ◦v = 1N.<br />

3.1.10. Proposition. Let f : M → N be a homomorphism.<br />

(1) If f has a retraction u : N → M then f is injective, u is surjective and<br />

N = Im f ⊕ Ker u<br />

(2) If f has a section v : N → M then f is surjective, v is injective and<br />

M = Ker f ⊕ Im v<br />

Pro<strong>of</strong>. (1) u(f(x)) = x so f is injective and u is surjective. If y ∈ N then<br />

y = f(u(y)) + (y − f(u(y)) and u(y − f(u(y))) = 0, so N = Im f + Ker u. Let<br />

y ∈ Im f ∩ Ker u. Then y = f(x) gives x = u(f(x)) = u(y) = 0, so y = 0.<br />

Conclude by 2.4.5 that the sum is direct. (2) y = f(v(y)) so f is a retraction <strong>of</strong> v.<br />

Finish by (1).<br />

3.1.11. Lemma. For a short exact sequence<br />

the following are equivalent<br />

0<br />

(1) f has a retraction.<br />

(2) g has a section.<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

<br />

L<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0


42 3. EXACT SEQUENCES OF MODULES<br />

For any retraction u there is a unique section v and wise-verse such that<br />

1N = f ◦ u + v ◦ g<br />

Pro<strong>of</strong>. If u is a retraction <strong>of</strong> f, then Ker g = Im f ⊂ Ker(1N − f ◦ u). By 3.1.5<br />

there is a homomorphism v : L → N such that v ◦ g = 1N − f ◦ u. This is a<br />

section <strong>of</strong> g. Conversely if v is a section <strong>of</strong> g then Im(1N − v ◦ g) ⊂ Ker g, so<br />

there is a homomorphism u : N → M such that f ◦ u = 1N − v ◦ g, 3.1.4. u is a<br />

retraction <strong>of</strong> f. The equation is clearly satisfied.<br />

3.1.12. Definition. Let R be a ring and f : M → N, g : N → L homomorphisms.<br />

A short exact sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

is a split exact sequence if equivalently 3.1.11 f has a retraction or g has a section.<br />

3.1.13. Proposition. A sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

is a split exact sequence if and only if there are homomorphism u : N → M, v :<br />

L → N satisfying<br />

<br />

L<br />

<br />

L<br />

<br />

0<br />

<br />

0<br />

g ◦ f = 0, u ◦ f = 1M, g ◦ v = 1L, f ◦ u + v ◦ g = 1N<br />

If the sequence is split exact then<br />

0<br />

<br />

v<br />

L <br />

u<br />

N <br />

N<br />

is split exact and (x, y) ↦→ f(x)+v(y) and z ↦→ u(z)+g(z) gives the isomorphism<br />

M ⊕ L N<br />

Pro<strong>of</strong>. The sequence is a 0-sequence f is injective and g is surjective. From f ◦<br />

u + v ◦ g = 1N follows that z ∈ Ker g ⊂ Im f, so the sequence is short exact. The<br />

rest is contained in 3.1.10.<br />

3.1.14. Corollary. A (contravariant) functor preserves split exact sequences. If<br />

0<br />

f<br />

<br />

M<br />

is split exact and T a functor, then<br />

is split exact.<br />

0<br />

g<br />

<br />

N<br />

<br />

L<br />

<br />

T (M) T (f) <br />

T (N) T (g) <br />

T (L)<br />

Pro<strong>of</strong>. By 3.1.13 a split exact sequence is characterized by a set <strong>of</strong> equations.<br />

These are preserved by the functor, 2.5.4.<br />

3.1.15. Example. A short exact sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

where L is a free module is a split exact sequence. Namely let xα ∈ L be a<br />

basis and choose yα ∈ N with g(yα) = xα. The define v : L → N by setting<br />

v(xα) = yα, 2.4.11.<br />

<br />

L<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0


3.2. THE SNAKE LEMMA 43<br />

3.1.16. Example. Let Zi is the family <strong>of</strong> modules each a copy <strong>of</strong> Z indexed by the<br />

natural numbers. Then the short exact sequence<br />

0<br />

<br />

<br />

i Zi<br />

<br />

<br />

i Zi<br />

is not split exact.<br />

The element f = (1, 2, 22 , . . . , 2n , . . . ) + <br />

fk = (0, . . . , 0, 2 n−k , . . . ) + <br />

But in <br />

<br />

<br />

i Zi/ i Zi<br />

<br />

0<br />

i Zi is divisible by 2 k for all k. If<br />

i Zi for n ≥ k, then 2 k fk = f in <br />

i Zi the only element divisible with all 2 k is 0, so no section exists.<br />

3.1.17. Exercise. (1) Show that the sequence<br />

0<br />

<br />

Z<br />

is short exact, but not split exact.<br />

(2) Show that the sequence<br />

0<br />

<br />

Q<br />

<br />

n<br />

Z <br />

Z<br />

is exact, but not split exact for n = 0, 1.<br />

(3) Show that the sequence<br />

is exact, but not split exact.<br />

(4) Show that the sequence<br />

is split exact.<br />

0<br />

0<br />

<br />

Z/(2) 1↦→2 <br />

Z/(4)<br />

<br />

Z/(2) 1↦→3 <br />

Z/(6)<br />

<br />

Q/Z<br />

<br />

Z/(n)<br />

3.2. The snake lemma<br />

<br />

Z/(2)<br />

<br />

Z/(3)<br />

3.2.1. Example. Given a commutative diagram <strong>of</strong> homomorphisms<br />

M<br />

u<br />

<br />

M ′<br />

there is induced a commutative diagram<br />

0<br />

0<br />

<br />

Ker f<br />

u<br />

<br />

<br />

Ker f ′ <br />

<br />

M<br />

u<br />

<br />

M ′<br />

where the rows are exact sequences.<br />

The diagram splits into two diagrams<br />

0<br />

0<br />

<br />

Ker f<br />

u<br />

<br />

<br />

Ker f ′ <br />

f<br />

f ′<br />

f<br />

f ′<br />

<br />

M<br />

u<br />

<br />

M ′<br />

<br />

N<br />

v<br />

<br />

<br />

N ′<br />

<br />

N<br />

<br />

0<br />

<br />

0<br />

<br />

Cok f<br />

<br />

0<br />

<br />

0<br />

<br />

<br />

<br />

<br />

<br />

v<br />

v<br />

N ′ Cok f ′<br />

f<br />

<br />

Im f<br />

<br />

f<br />

v<br />

′<br />

<br />

Im f ′<br />

<br />

0<br />

0<br />

i<br />

<br />

0<br />

0<br />

Zi/ <br />

i Zi.


44 3. EXACT SEQUENCES OF MODULES<br />

and<br />

0<br />

0<br />

<br />

Im f<br />

<br />

N<br />

where the rows are short exact sequences.<br />

<br />

Cok f<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

v<br />

v<br />

v<br />

Im f ′<br />

N ′ Cok f ′<br />

3.2.2. Lemma. Given a commutative diagram <strong>of</strong> homomorphisms<br />

0<br />

M<br />

u<br />

<br />

<br />

M ′<br />

f<br />

f ′<br />

<br />

N<br />

v<br />

<br />

<br />

N ′<br />

g<br />

g ′<br />

<br />

L<br />

w<br />

<br />

<br />

L ′<br />

where the rows exact sequences. The snake homomorphism δ : Ker w → Cok u<br />

is well defined by: For z ∈ Ker w choose y ∈ N such that g(y) = z. The<br />

element v(y) ∈ Ker g ′ so there is x ′ ∈ M ′ such that f ′ (x ′ ) = v(y). Then δ(z) =<br />

x ′ + Im u ∈ Cok u.<br />

Pro<strong>of</strong>. Assume g(y ′ ) = z and f ′ (x ′′ ) = v(y ′ ). There is x ∈ M with f(x) = y−y ′ .<br />

Now f ′ (u(x)) = v(f(x)) = v(y − y ′ ) = f ′ (x ′ − x ′′ ) so u(x) = x ′ − x ′′ since f ′<br />

is injective. Then x ′ + Im u = x ′′ + Im u as wanted. The choices made respect<br />

addition and scalar multiplication showing that δ is a homomorphism.<br />

3.2.3. Remark. The snake is<br />

Ker u<br />

f<br />

<br />

Ker v<br />

<br />

0<br />

<br />

0<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

f<br />

g<br />

M N L 0<br />

<br />

u<br />

v<br />

w<br />

<br />

0<br />

<br />

M ′<br />

f ′<br />

<br />

N ′<br />

g ′<br />

<br />

L ′<br />

<br />

<br />

Cok u<br />

The construction <strong>of</strong> δ is schematically<br />

M ′<br />

<br />

Cok u<br />

f ′<br />

N<br />

v<br />

<br />

<br />

N ′<br />

g<br />

f ′<br />

Ker w<br />

<br />

<br />

L<br />

<br />

<br />

Cok v<br />

g<br />

g ′<br />

x ′<br />

<br />

<br />

δ(z)<br />

<br />

Ker w<br />

<br />

<br />

Cok w<br />

0<br />

<br />

<br />

v(y)<br />

z <br />

<br />

y <br />

z


3.2. THE SNAKE LEMMA 45<br />

3.2.4. Theorem (snake lemma). Given a commutative diagram <strong>of</strong> homomorphisms<br />

0<br />

M<br />

u<br />

<br />

<br />

M ′<br />

f<br />

f ′<br />

<br />

N<br />

v<br />

<br />

<br />

N ′<br />

g<br />

g ′<br />

<br />

L<br />

w<br />

<br />

<br />

L ′<br />

where the rows exact sequences. There is induced a six term long exact sequence<br />

Ker u<br />

f<br />

<br />

Ker v<br />

δ<br />

<br />

f ′<br />

<br />

Cok v<br />

Cok u<br />

g<br />

g ′<br />

<br />

Ker w<br />

<br />

Cok w<br />

Pro<strong>of</strong>. By construction <strong>of</strong> δ it is clear that the sequence is a 0-sequence: If y ∈<br />

Ker v then to calculate δ(g(y)) the choice v(y) = 0 gives δ ◦ g = 0. Also<br />

f ′ (δ(z)) = v(y) + Im v shows that f ′ ◦ δ = 0. Exactness at Ker v and Cok v<br />

are clear. Given z ∈ Ker w such that δ(z) = 0. By 3.2.2 choose y, g(y) = z<br />

and x ′ , f ′ (x ′ ) = v(y). Then δ(z) = x ′ + Im u = 0, so choose x, u(x) = x ′ .<br />

Now v(f(x)) = f ′ (u(x)) = v(y) so y − f(x) ∈ Ker v and g(y − f(x)) =<br />

g(y) = z. Therefore exactness at Ker w. Given x ′ + Im u ∈ Cok u such that<br />

f ′ (x ′ ) + Im v = 0 ∈ Cok v. Choose y, v(y) = f ′ (x ′ ) and put z = g(y). Then<br />

w(g(y)) = g ′ (v(y)) = g ′ (f ′ (x ′ )) = 0. Now z ∈ Ker w and δ(z) = x ′ + Im u.<br />

Therefore exactness at Cok u.<br />

3.2.5. Corollary. If f is injective then the f : Ker u → Ker v is injective and the<br />

long exact sequence is<br />

0<br />

<br />

Ker u<br />

f<br />

<br />

Ker v<br />

δ<br />

<br />

f ′<br />

<br />

Cok v<br />

Cok u<br />

g<br />

g ′<br />

<br />

0<br />

<br />

Ker w<br />

<br />

Cok w<br />

If g ′ is surjective then g ′ : Cok v → Cok w is surjective and the long exact sequence<br />

is<br />

Ker u<br />

f<br />

<br />

Ker v<br />

δ<br />

<br />

f ′<br />

<br />

Cok v<br />

Cok u<br />

g<br />

g ′<br />

<br />

Ker w<br />

<br />

Cok w<br />

3.2.6. Corollary. (1) If v is injective and u is surjective, then w is injective.<br />

(2) If v is surjective and w is injective, than u is surjective.<br />

(3) If v is an isomorphism, then w is injective if and only if u is surjective.<br />

3.2.7. Proposition. Given submodules N, L ⊂ M, then there is a short exact<br />

sequence<br />

0<br />

<br />

0<br />

<br />

M/N ∩ L x↦→(x,x)<br />

<br />

M/N ⊕ M/L<br />

(x,y)↦→x−y<br />

<br />

M/N + L<br />

<br />

0


46 3. EXACT SEQUENCES OF MODULES<br />

Pro<strong>of</strong>. The commutative diagram<br />

0<br />

0<br />

<br />

N ∩ L<br />

<br />

N ⊕ L<br />

<br />

N + L<br />

<br />

x↦→(x,x) <br />

(x,y)↦→x−y <br />

<br />

M <br />

M ⊕ M <br />

M<br />

where the rows are short exact sequences, gives by 3.2.4 a five term long exact<br />

sequence<br />

0<br />

<br />

M/N ∩ L<br />

<br />

M/N ⊕ M/L<br />

<br />

M/N + L<br />

3.2.8. Proposition (five lemma). Given a commutative diagram <strong>of</strong> homomorphisms<br />

M1<br />

u1<br />

<br />

M ′ 1<br />

<br />

M2<br />

u2<br />

<br />

<br />

M ′<br />

2<br />

<br />

M3<br />

u3<br />

<br />

<br />

M ′<br />

3<br />

<br />

M4<br />

u4<br />

<br />

<br />

M ′<br />

4<br />

<br />

M5<br />

u5<br />

<br />

<br />

M ′<br />

5<br />

where the rows are exact sequences. If u1 is surjective, u2, u4 are isomorphism<br />

and u5 is injective, then u3 is an isomorphism.<br />

Pro<strong>of</strong>. Let fi : Mi → Mi+1, f ′ i : M ′ i → M ′ i+1 . Split the given diagram in three as<br />

follows<br />

0<br />

0<br />

0<br />

0<br />

0<br />

M1<br />

u1<br />

<br />

<br />

Ker f ′<br />

2<br />

<br />

Cok f2<br />

u ′′<br />

3<br />

<br />

<br />

Ker f ′<br />

4<br />

<br />

Im f2<br />

u ′ 3<br />

<br />

<br />

Im f ′<br />

2<br />

<br />

M2<br />

u2<br />

<br />

<br />

M ′<br />

2<br />

<br />

M4<br />

u4<br />

<br />

<br />

M ′<br />

4<br />

<br />

M3<br />

u3<br />

<br />

<br />

M ′<br />

3<br />

<br />

Cok f1<br />

u ′ 3<br />

<br />

<br />

Im f ′<br />

2<br />

<br />

Im f4<br />

u5<br />

<br />

<br />

M ′<br />

5<br />

<br />

Cok f2<br />

u ′′<br />

3<br />

<br />

<br />

Cok f ′<br />

2<br />

Note that Cok f1 Im f2 and Cok f ′ 2 Ker f ′ 4 . Now use 2.3.3 and the snake<br />

lemma to conclude that Ker u3 = 0 and Cok u3 = 0 and u3 is therefore an isomorphism.<br />

3.2.9. Proposition (windmill lemma). Given homomorphism M f<br />

There is induced an eight term long exact sequence<br />

0<br />

<br />

Ker f<br />

<br />

Ker g ◦ f<br />

<br />

<br />

Cok f<br />

f<br />

g<br />

<br />

Ker g<br />

<br />

<br />

Cok g ◦ f<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

Cok g<br />

<br />

0<br />

g<br />

<br />

N<br />

<br />

0<br />

<br />

L .


Pro<strong>of</strong>. Look at the two diagrams<br />

0<br />

0<br />

3.2. THE SNAKE LEMMA 47<br />

0<br />

<br />

<br />

Ker g<br />

M<br />

f<br />

<br />

<br />

<br />

g◦f<br />

L<br />

1<br />

By the snake lemma the sequences<br />

0<br />

<br />

N<br />

<br />

M<br />

g<br />

<br />

L<br />

<br />

Ker f<br />

δ=f<br />

<br />

<br />

Cok f<br />

Ker g<br />

Ker g ◦ f<br />

f<br />

<br />

g<br />

<br />

N<br />

<br />

Ker g<br />

δ=g<br />

<br />

<br />

Cok g<br />

Cok g ◦ f<br />

1 <br />

M<br />

g◦f<br />

<br />

<br />

L<br />

<br />

Cok f<br />

<br />

<br />

0<br />

are exact and overlap to give the windmill sequence.<br />

3.2.10. Remark. The windmill is<br />

<br />

Ker g ◦ f<br />

<br />

<br />

Cok g ◦ f<br />

<br />

Cok f<br />

<br />

Ker g ◦ f<br />

<br />

Ker g<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

<br />

f<br />

Ker f<br />

<br />

M<br />

<br />

N <br />

<br />

<br />

g◦f <br />

g<br />

<br />

<br />

Cok f<br />

<br />

<br />

0 <br />

L <br />

<br />

<br />

<br />

Cok g <br />

Cok g ◦ f<br />

3.2.11. Exercise. (1) Given a short exact sequence<br />

0<br />

<br />

M<br />

<br />

N<br />

Show that Ann(N) ⊂ Ann(M) ∩ Ann(L).<br />

(2) Give a short exact sequence<br />

0<br />

<br />

M<br />

<br />

N<br />

such that Ann(N) = Ann(M) ∩ Ann(L).<br />

(3) Given ideals I, J ⊂ R. Show that there is a short exact sequence.<br />

0<br />

<br />

R/I ∩ J<br />

<br />

R/I ⊕ R/J<br />

<br />

L<br />

<br />

L<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

0<br />

<br />

R/I + J<br />

<br />

0


48 3. EXACT SEQUENCES OF MODULES<br />

3.3. Exactness <strong>of</strong> Hom<br />

3.3.1. Proposition. Given an exact sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

<br />

L<br />

and an R-module K. Then the following sequence is exact<br />

0<br />

<br />

HomR(K, M)<br />

<br />

HomR(K, N)<br />

<br />

HomR(K, L)<br />

Pro<strong>of</strong>. Given h : K → M such that f ◦ h = 0 then h = 0 since f is injective.<br />

Given k : K → N such that g ◦ k = 0 then by 3.1.4 there is h : K → M such that<br />

f ◦ h = k. So the sequence is exact.<br />

3.3.2. Proposition. Given a sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

<br />

L<br />

Such that for any module K, the following sequence is exact<br />

0<br />

<br />

HomR(K, M)<br />

Then the original sequence is exact.<br />

<br />

HomR(K, N)<br />

<br />

HomR(K, L)<br />

Pro<strong>of</strong>. For K = M the identity 1M is mapped to g ◦ f ◦ 1M = 0 so it is a 0sequence.<br />

Assume f(x) = 0. Take K = R then 1x is mapped to f ◦ 1x = 1 f(x) =<br />

0. Therefore 1x = 0 and so x = 0. That insures that f is injective. Assume<br />

g(y) = 0. Take K = R then 1y is mapped to g ◦ 1y = 1 g(y) = 0. There exists a<br />

homomorphism h : R → M such that f ◦ h = 1y. By 2.5.12 h = 1x and therefore<br />

f(x) = y. The original sequence is now proven exact.<br />

3.3.3. Proposition. Given an exact sequence<br />

M f<br />

g<br />

<br />

N<br />

<br />

L<br />

and a module K. Then the following sequence is exact<br />

0<br />

<br />

HomR(L, K)<br />

<br />

HomR(N, K)<br />

<br />

0<br />

<br />

HomR(M, K)<br />

Pro<strong>of</strong>. Given h : L → K such that h ◦ g = 0 then h = 0 since g is surjective.<br />

Given k : N → K such that k ◦ f = 0 then by 3.1.5 there is h : L → K such that<br />

h ◦ g = k. So the sequence is exact.<br />

3.3.4. Proposition. Given a sequence<br />

M f<br />

g<br />

<br />

N<br />

<br />

L<br />

such that for any module K, the following sequence is exact<br />

0<br />

<br />

HomR(L, K)<br />

Then the original sequence is exact.<br />

<br />

HomR(N, K)<br />

<br />

0<br />

<br />

HomR(M, K)<br />

Pro<strong>of</strong>. For K = L the identity 1L is mapped to 1L ◦g ◦f = 0 so it is a 0-sequence.<br />

Take K = Cok g then pg : L → Cok g has pg ◦ g = 0, but by exactness 0 is the<br />

unique homomorphism satisfying this, so pg = 0. Therefore Cok g = 0 and g is<br />

surjective. Take K = Cok f, p : N → Cok f the projection. p ◦ f = 0 so by<br />

exactness there exists a unique q : L → Cok f such that q ◦ g = p. It follows that<br />

Ker g ⊂ Ker p = Im f. All together the original sequence is exact.


3.3.5. Proposition. Given a sequence<br />

The following are equivalent.<br />

0<br />

3.4. EXACTNESS OF TENSOR 49<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

(1) The sequence is split exact.<br />

(2) For any K the following sequence is exact<br />

0<br />

<br />

HomR(K, M)<br />

<br />

HomR(K, N)<br />

(3) For any K the following sequence is exact<br />

0<br />

<br />

HomR(L, K)<br />

<br />

HomR(N, K)<br />

<br />

L<br />

<br />

0<br />

<br />

HomR(K, L)<br />

<br />

HomR(M, K)<br />

If the conditions are true, then the sequences (2) and (3) are split exact.<br />

Pro<strong>of</strong>. (1) ⇒ (2), (1) ⇒ (3) are clear by 3.1.14 giving that the sequences (2), (3)<br />

are split exact. (2) ⇒ (1): Let K = L, then there is a section to g. By 3.3.2 and<br />

3.1.11 the original sequence is split exact. (3) ⇒ (1): Let K = M, then there is a<br />

retraction to f. By 3.3.4 and 3.1.11 the original sequence is split exact.<br />

3.3.6. Exercise. (1) Show that the sequence<br />

0<br />

<br />

HomZ(Q/Z, Z)<br />

<br />

HomZ(Q/Z, Q)<br />

is exact, but the rightmost map is not surjective.<br />

(2) Show that the sequence<br />

0<br />

<br />

HomZ(Z/(n), Z)<br />

n <br />

HomZ(Z/(n), Z)<br />

is exact, but the rightmost map is not surjective.<br />

3.4. Exactness <strong>of</strong> Tensor<br />

3.4.1. Proposition. Given an exact sequence<br />

M f<br />

g<br />

<br />

N<br />

<br />

L<br />

<br />

0<br />

and an R-module K. Then the following sequence is exact<br />

M ⊗R K<br />

<br />

N ⊗R K<br />

<br />

L ⊗R K<br />

<br />

0<br />

<br />

0<br />

<br />

HomZ(Q/Z, Q/Z)<br />

<br />

HomZ(Z/(n), Z/(n))<br />

Pro<strong>of</strong>. Let K ′ be any module. By 3.3.4 it is enough to see that the sequence<br />

0<br />

<br />

HomR(L ⊗R K, K ′ )<br />

is exact. By 2.6.13 it amounts to see that the sequence<br />

0<br />

<br />

HomR(L, HomR(K, K ′ ))<br />

is exact. This follows from 3.3.3.<br />

<br />

0<br />

<br />

HomR(N ⊗R K, K ′ )<br />

<br />

HomR(M ⊗R K, K ′ )<br />

<br />

HomR(M, HomR(K, K ′ ))<br />

<br />

HomR(N, HomR(K, K ′ ))


50 3. EXACT SEQUENCES OF MODULES<br />

3.4.2. Proposition. Given a split exact sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

<br />

L<br />

and a module K. Then the following sequence is split exact<br />

0<br />

<br />

K ⊗R M<br />

<br />

K ⊗R N<br />

Pro<strong>of</strong>. This follows from the functor properties 3.1.14.<br />

<br />

0<br />

<br />

K ⊗R L<br />

3.4.3. Proposition. Let I ⊂ R be an ideal. For any module M, the homomorphism<br />

is an isomorphism.<br />

M ⊗R R/I → M/IM, x ⊗ a + I ↦→ ax + IM<br />

Pro<strong>of</strong>. x + IM ↦→ x ⊗ 1 is an inverse.<br />

3.4.4. Corollary. Let I, J ⊂ R be ideals. Then<br />

is an isomorphism.<br />

R/I ⊗R R/J → R/(I + J), a + I ⊗ b + J ↦→ ab + I + J<br />

3.4.5. Proposition. Let I1, . . . , Ik ⊂ R be pairwise comaximal ideals and M a<br />

module. Then the product <strong>of</strong> projections<br />

is an isomorphism.<br />

Pro<strong>of</strong>. By Chinese remainders 1.4.2<br />

M/I1 · · · IkM → M/I1M × · · · × M/IkM<br />

R/I1 · · · Ik → R/I1 × · · · × R/Ik<br />

is an isomorphism. Tensor with M and use 3.4.3 to get the isomorphism.<br />

3.4.6. Exercise. (1) Calculate Z/(m) ⊗Z Z/(n) for all integers m, n.<br />

(2) Let I ⊂ R be an ideal. Show that R/I ⊗R R/I R/I.<br />

(3) Let I ⊂ R be an ideal. Show that I ⊗R R/I I/I 2 .<br />

(4) Let 2Z be scalar multiplication. Show that 2Z ⊗ 1 Z/(2) : Z ⊗Z Z/(2) → Z ⊗Z Z/(2)<br />

is not injective.<br />

3.5. Projective modules<br />

3.5.1. Definition. An R-module F is a projective module if for any exact sequence<br />

N → L → 0 the sequence<br />

is exact.<br />

HomR(F, N) → HomR(F, L) → 0<br />

3.5.2. Proposition. A module F is projective if and only if any surjective homomorphism<br />

M → F → 0 has a section.<br />

Pro<strong>of</strong>. Assume F projective and g : M → F surjective. Then HomR(F, M) →<br />

HomR(F, F ) → 0 is exact. So there exists a v : F → M such that g ◦ v = 1F .<br />

v is then a section. Conversely given g : N → L surjective and h : F → L. Let<br />

M = Ker N ⊕F → L, (y, z) ↦→ g(y)−h(z) and pN : M → N, pF : M → F the<br />

<br />

0


3.5. PROJECTIVE MODULES 51<br />

projections, then g ◦ pN = h ◦ pF . Now pF is surjective since g is. Let v : F → M<br />

be a section <strong>of</strong> pF , then h ′ = pN ◦ v satisfies h = g ◦ h ′ .<br />

pN<br />

N g<br />

<br />

<br />

L<br />

h<br />

M <br />

F<br />

′<br />

<br />

<br />

h<br />

v<br />

<br />

3.5.3. Corollary. A short exact sequence<br />

0<br />

f<br />

<br />

M<br />

pF<br />

g<br />

<br />

N<br />

<br />

0<br />

<br />

F<br />

where F is a projective module is a split exact sequence.<br />

3.5.4. Corollary. A free module is projective. Over a field every module is projective.<br />

3.5.5. Example. Let I ⊂ R be an ideal. If R/I is projective, then there is a ring<br />

decomposition R/I × R ′ R.<br />

3.5.6. Proposition. A direct summand in a projective module is projective.<br />

Pro<strong>of</strong>. Let F ⊕ F ′ be projective and g : M → F surjective. By 3.5.2 there is a<br />

section v ′ : F ⊕ F ′ → M ⊕ F ′ to (g, 1F ′). Then v(y) = pM ◦ v ′ (y, 0) is a section<br />

to g and F is projective.<br />

3.5.7. Proposition. A module is projective if and only if it is a direct summand in<br />

a free module.<br />

Pro<strong>of</strong>. By 2.4.12 any module is a factor module <strong>of</strong> a free module. By 3.5.2 a<br />

projective factor module has a section, and is therefore by 3.1.10 a direct summand.<br />

3.5.8. Proposition. Let Fα be a family <strong>of</strong> projective modules, then the direct sum<br />

<br />

α Fα is a projective module.<br />

Pro<strong>of</strong>. Let N → L be surjective. Then by 2.5.8<br />

is the product<br />

<br />

0<br />

HomR( Fα, N) → HomR( Fα, L)<br />

HomR(Fα, N) → HomR(Fα, L)<br />

which is surjective by 3.1.6. So Fα is projective.<br />

3.5.9. Proposition. Let F, F ′ be projective modules. Then F ⊗R F ′ is projective.<br />

Pro<strong>of</strong>. F ⊗R F ′ is clearly a direct summand in a free module.<br />

3.5.10. Proposition. Let R → S be a ring homomorphism and F a projective<br />

module. The change <strong>of</strong> ring module F ⊗R S is a projective S-module.<br />

Pro<strong>of</strong>. A direct summand <strong>of</strong> a free R-module is changed to a direct summand <strong>of</strong> a<br />

free S-module.


52 3. EXACT SEQUENCES OF MODULES<br />

3.5.11. Proposition. Any module M admits an exact sequence<br />

F → M → 0<br />

where F is a projective module. That is, any module is a factor module <strong>of</strong> a projective<br />

module.<br />

Pro<strong>of</strong>. Take F free, 2.4.12.<br />

3.5.12. Exercise. (1) Let R = R1 × R2. Show that R1, R2 are projective ideals in R.<br />

(2) Show that the ideal (2)/(6) in the ring Z/(6) is projective.<br />

(3) Show that the ideal (2)/(4) in the ring Z/(4) is not projective.<br />

3.6. Injective modules<br />

3.6.1. Definition. An R-module E is an injective module if for any exact sequence<br />

0 → M → N the sequence<br />

is exact.<br />

HomR(N, E) → HomR(M, E) → 0<br />

3.6.2. Proposition. A module E is an injective module if and only if any injective<br />

homomorphism 0 → E → L has a retraction.<br />

Pro<strong>of</strong>. Assume E injective and f : E → L injective. Then HomR(L, E) →<br />

HomR(E, E) → 0 is exact. So there exists a u : L → E such that u ◦ f = 1E.<br />

Then u is a retraction. Conversely given f : M → N injective and h : M → E.<br />

Let L = Cok M → E ⊕ N, x ↦→ h(x) − f(x) and iE : E → L, iN : N → L the<br />

injections, then iN ◦ f = iE ◦ h. Now iE is injective since f is. Let u : L → E be<br />

a retraction <strong>of</strong> iE, then h ′ = u ◦ iN satisfies h = h ′ ◦ f.<br />

0<br />

<br />

M<br />

3.6.3. Corollary. A short exact sequence<br />

0<br />

f<br />

<br />

E<br />

f<br />

<br />

N<br />

h<br />

h<br />

<br />

<br />

′<br />

<br />

E iE<br />

<br />

<br />

L<br />

u<br />

g<br />

<br />

N<br />

iN<br />

<br />

L<br />

where E is an injective module is a split exact sequence.<br />

3.6.4. Example. Let I ⊂ R be an ideal. If I is injective, then there is a ring<br />

decomposition R/I × R ′ R.<br />

3.6.5. Proposition. A direct summand in an injective module is injective.<br />

Pro<strong>of</strong>. Let E ⊕ E ′ be an injective module and f : E → L an injective homomorphism.<br />

By 3.6.2 there is a retraction u ′ : L ⊕ E ′ → E ⊕ E ′ to (f, 1E ′). Then<br />

u(y) = pE ◦ u ′ (y, 0) is a retraction to f and E is injective.<br />

3.6.6. Proposition. Let Eα be a family <strong>of</strong> injective modules, then the direct product<br />

<br />

α Eα is an injective module<br />

<br />

0


Pro<strong>of</strong>. Let M → N be injective. Then by 2.5.8<br />

is the product<br />

3.6. INJECTIVE MODULES 53<br />

HomR(N, Eα) → HomR(M, Eα)<br />

HomR(N, Eα) → HomR(M, Eα)<br />

which is surjective by 3.1.6. So Eα is injective.<br />

3.6.7. Proposition. A module E is injective, if for any ideal I ⊂ R<br />

is exact.<br />

HomR(R, E) → HomR(I, E) → 0<br />

Pro<strong>of</strong>. Let f : E → L be an injective homomorphism. The set <strong>of</strong> submodules<br />

f(E) ⊂ L ′ ⊂ L and retractions u ′ : L ′ → E is nonempty and inductively ordered.<br />

By Zorn’s lemma a maximal L ′ , u ′ exists. If L/L ′ = 0 choose a y ∈ L\L ′ . The<br />

homomorphism Ann(y + L ′ ) → E, a ↦→ u ′ (ay) extends to u ′′ : R → E by<br />

hypothesis. The setting L ′ + Ry → E, x + ay ↦→ u ′ (x) + u ′′ (a) is a well defined<br />

retraction. This contradicts maximality.<br />

3.6.8. Definition. Let R be a domain. M is a divisible module if scalar multiplication<br />

with a nonzero a ∈ R is surjective.<br />

3.6.9. Proposition. (1) An injective module over a domain is divisible.<br />

(2) A divisible module over a principal ideal domain is injective.<br />

(3) Over a field any module is injective.<br />

Pro<strong>of</strong>. (1) Let E be an injective module and a = 0. Choose x ∈ E and look<br />

at 1x : R → E. Scalar multiplication aR is injective, so there is an extension<br />

1y : R → E such that 1x = 1y ◦ aR. Then ay = x. (2) Let E be divisible. To<br />

extend f : (a) → E, choose x ∈ E such that ax = f(a), then 1x : R → E, 1 → x<br />

extends f. (3) This is clear.<br />

3.6.10. Proposition. Let R → S be a ring homomorphism and E an injective<br />

R-module. The induced module HomR(S, E) is an injective S-module.<br />

Pro<strong>of</strong>. Let M → N be an injective homomorphism <strong>of</strong> S-modules. Then by<br />

2.7.10 HomS(N, HomR(S, E)) → HomS(M, HomR(S, E)) is HomR(N, E) →<br />

HomR(M, E) being surjective since E is an injective R-module.<br />

3.6.11. Lemma. (1) Q and Q/Z are divisible and therefore injective Z modules.<br />

(2) The homomorphism 2.5.11<br />

is injective.<br />

x ↦→ evx : M → HomZ(HomZ(M, Q/Z), Q/Z)<br />

Pro<strong>of</strong>. (1) Clear by 3.6.9. (2) Let x = 0 and choose h : Z/ Ann(x) → Q/Z<br />

nonzero. Now Z/ Ann(x) Zx ⊂ M so extend h to h ′ : M → Q/Z. Then<br />

evx(h ′ ) = h(1) = 0.<br />

3.6.12. Proposition. Any module M admits an exact sequence<br />

0 → M → E<br />

where E is an injective module. That is, any module is a submodule <strong>of</strong> an injective<br />

module.


54 3. EXACT SEQUENCES OF MODULES<br />

Pro<strong>of</strong>. By 2.4.12 choose a surjection F → HomZ(M, Q/Z) where F is a free<br />

R-module. Then<br />

0 → M → HomZ(F, Q/Z)<br />

is exact by 3.6.11. Since F <br />

α R the module<br />

HomZ(F, Q/Z) <br />

HomZ(R, Q/Z)<br />

is injective, 3.6.6 and 3.6.10.<br />

3.6.13. Proposition. A homomorphism f : M → N is injective if and only if<br />

is surjective for any injective module E.<br />

HomR(N, E) → HomR(M, E) → 0<br />

Pro<strong>of</strong>. Assume that Ker f = 0 and choose 0 → Ker f → E, 3.6.12. The sequence<br />

HomR(N, E) → HomR(M, E) → HomR(Ker f, E) → 0<br />

is exact. So HomR(N, E) → HomR(M, E) is not surjective.<br />

3.6.14. Definition. A submodule N ⊂ M is an essential extension if any nonzero<br />

submodule L ⊂ M has nonzero intersection N ∩ L = 0. An essential extension<br />

M ⊂ E with E injective is an injective envelope <strong>of</strong> M.<br />

3.6.15. Proposition. Any module M has an injective envelope. If M ⊂ E, E ′ are<br />

two injective envelopes, then there is an isomorphism f : E → E ′ fixing M.<br />

Pro<strong>of</strong>. By 3.6.12 choose M ⊂ E ′ with E ′ injective. By Zorn’s lemma choose<br />

M ⊂ E ⊂ E ′ maximal among the essential extensions <strong>of</strong> M. If E = E ′ then the<br />

set <strong>of</strong> modules N ′ ⊂ E ′ such that E ∩ N ′ is nonempty by maximality <strong>of</strong> E. Let<br />

N be maximal among these by Zorn’s lemma. It follows that E ⊕ N = E ′ and E<br />

is injective by 3.6.6. Given two envelopes, let f : E → E ′ be any homomorphism<br />

fixing M. Then f is injective, since M ⊂ E is essential. If f is not surjective, then<br />

E ′ f(E) ⊕ E ′′ contradicting that M ⊂ E ′ is essential.<br />

3.6.16. Exercise. (1) Let R be a domain. Show that the fraction field is an injective<br />

module.<br />

(2) Let R be a domain. Show that the torsion free divisible module injective.<br />

(3) Show that for a ring that all modules are projective if and only all modules are injective.<br />

α<br />

3.7. Flat modules<br />

3.7.1. Definition. An R-module F is a flat module if for any exact sequence 0 →<br />

M → N the sequence<br />

0 → M ⊗R F → N ⊗ F<br />

is exact.<br />

3.7.2. Example. If a ∈ R is a nonzero divisor and M is a flat modules, then aM<br />

is injective and a is a nonzero divisor on M.<br />

3.7.3. Proposition. A direct summand in a flat module is flat.<br />

Pro<strong>of</strong>. Let F ⊕ F ′ be flat and M → N injective. Then M ⊗R (F ⊕ F ′ ) →<br />

N ⊗R (F ⊕ F ′ ) is injective. Conclusion by 2.6.11.


3.7. FLAT MODULES 55<br />

3.7.4. Proposition. Let Fα be family <strong>of</strong> flat modules, then the direct sum <br />

α Fα<br />

is a flat module.<br />

Pro<strong>of</strong>. Let M → N be injective. Then M ⊗R ( Fα) → N ⊗R ( Fα) is<br />

injective by 2.6.11 and 3.1.6, so the product is flat.<br />

3.7.5. Example. A free module is flat.<br />

3.7.6. Corollary. A projective module is flat.<br />

Pro<strong>of</strong>. By 3.5.7 a projective module is a direct summand in a free. Conclusion by<br />

3.7.4.<br />

3.7.7. Proposition. Let F, F ′ be flat modules. Then F ⊗R F ′ is flat.<br />

Pro<strong>of</strong>. Let M → N be injective. Then by 2.6.10 M ⊗R (F ⊗R F ′ ) → N ⊗R<br />

(F ⊗R F ′ ) is (M ⊗R F ) ⊗R F ′ → (N ⊗R F ) ⊗R F ′ being injective by using the<br />

definition twice.<br />

3.7.8. Proposition. Let R → S be a ring homomorphism and F a flat R-module.<br />

The change <strong>of</strong> ring module F ⊗R S is a flat S-module.<br />

Pro<strong>of</strong>. Let M → N be an injective homomorphism <strong>of</strong> S-modules. Then by 2.7.5<br />

M ⊗S (F ⊗R S ′ ) → N ⊗S (F ⊗R S) is M ⊗R F → N ⊗R F being injective since<br />

F is a flat R-module.<br />

3.7.9. Proposition. Let R be a ring and F a module. The following are equivalent.<br />

(1) F is a flat module.<br />

(2) HomR(F, E) is an injective module for any injective module E.<br />

Pro<strong>of</strong>. Let M → N be injective. By 3.6.13 M ⊗R F → N ⊗R F is injective if and<br />

only if HomR(N ⊗R F, E) → HomR(M ⊗R F, E) is surjective for any injective<br />

module E. By 2.6.13 this is HomR(N, HomR(F, E)) → HomR(M, HomR(F, E)).<br />

3.7.10. Corollary. Given a short exact sequence<br />

0<br />

<br />

M<br />

<br />

N<br />

<br />

F<br />

where F is a flat module. Then M is flat if and only if N is flat.<br />

Pro<strong>of</strong>. Let E be an injective module. By 3.7.9 and 3.6.3 the sequence<br />

0<br />

<br />

HomR(F, E)<br />

<br />

HomR(N, E)<br />

<br />

0<br />

<br />

HomR(M, E)<br />

is split exact. By 3.6.5 and 3.6.6 HomR(N, E) is injective if and only if HomR(M, E)<br />

is so. Conclusion by 3.7.9.<br />

3.7.11. Corollary. Given a short exact sequence<br />

0<br />

<br />

M<br />

where F is a flat module. For any module L there is a short exact sequence<br />

0<br />

<br />

L ⊗R M<br />

<br />

N<br />

<br />

L ⊗R N<br />

<br />

F<br />

<br />

0<br />

<br />

L ⊗R F<br />

<br />

0<br />

<br />

0


56 3. EXACT SEQUENCES OF MODULES<br />

Pro<strong>of</strong>. Let E be an injective module. By 3.7.9 and 3.6.3 the sequence<br />

0<br />

<br />

HomR(F, E)<br />

is split exact. So also the sequence<br />

0<br />

<br />

HomR(L, HomR(F, E))<br />

<br />

HomR(N, E)<br />

<br />

HomR(M, E)<br />

<br />

HomR(L, HomR(N, E))<br />

<br />

HomR(L, HomR(M, E))<br />

is split exact. By 2.6.13 this is natural isomorphic to the sequence<br />

0<br />

Conclusion by 3.6.13.<br />

<br />

HomR(L ⊗R F, E)<br />

<br />

HomR(L ⊗R N, E)<br />

<br />

HomR(L ⊗R M, E)<br />

3.7.12. Proposition. A module F is flat, if for any ideal I ⊂ R<br />

is exact.<br />

0 → I ⊗R F → R ⊗R F<br />

Pro<strong>of</strong>. By 3.7.9 it suffices to see that HomR(F, E) is injective for any injective E.<br />

By 3.6.7 this amounts to HomR(R, HomR(F, E)) → HomR(I, HomR(F, E)) being<br />

surjective. By 2.6.13 this homomorphism is HomR(R⊗RF, E) → HomR(I⊗R<br />

F, E) which is surjective since E is injective.<br />

3.7.13. Exercise. (1) Show that Z/(n) is not a flat Z-module for n = 0, 1.<br />

(2) Show that Z/(2) is a flat Z/(6)-module.<br />

(3) Show that Z/(2) is not a flat Z/(4)-module.<br />

<br />

0<br />

<br />

0<br />

<br />

0


4<br />

Fraction constructions<br />

4.1. Rings <strong>of</strong> fractions<br />

4.1.1. Lemma. Let R be a ring and U ⊂ R such that 1 ∈ U and for u, v ∈ U the<br />

product uv ∈ U. On U × R is defined a relation<br />

(u, a) ∼ (u ′ , a ′ ) ⇔ there is v ∈ U such that vu ′ a = vua ′<br />

(1) The relation is an equivalence relation.<br />

(2) If (u, a) ∼ (u ′ , a ′ ) and (v, b) ∼ (v ′ , b ′ ) then (uv, va + ub) ∼ (u ′ v ′ , v ′ a ′ +<br />

u ′ b ′ ).<br />

(3) If (u, a) ∼ (u ′ , a ′ ) and (v, b) ∼ (v ′ , b ′ ) then (uv, ab) ∼ (u ′ v ′ , a ′ b ′ ).<br />

Pro<strong>of</strong>. The claims are proved by simple calculations. (1) Symmetry is clear. Reflexive<br />

follows as 1 ∈ U. Transitive: if (u, a) ∼ (u ′ , a ′ ), (u ′ , a ′ ) ∼ (u ′′ , a ′′ ) then<br />

vu ′ a = vua ′ , v ′ u ′′ a ′ = v ′ u ′ a ′′ . Since multiplication is commutative (v ′ vu ′ )u ′′ a =<br />

v ′ u ′′ vua ′ = (v ′ vu ′ )ua ′′ give (u, a) ∼ (u ′′ , a ′′ ). (2) From wu ′ a = wua ′ , w ′ v ′ b =<br />

w ′ vb ′ follow that w ′ vv ′ wu ′ a = w ′ vv ′ wua ′ , wuu ′ w ′ v ′ b = wuu ′ w ′ vb ′ . So now<br />

ww ′ (u ′ v ′ )(va + ub) = ww ′ (uv)(v ′ a ′ + u ′ b ′ ) as needed. (3) Is similar.<br />

4.1.2. Definition. Let R be a ring. U ⊂ R is a multiplicative subset if 1 ∈ U<br />

and for u, v ∈ U the product uv ∈ U. The ring <strong>of</strong> fractions U −1R is given by<br />

on U × R under the relation 4.1.1<br />

equivalence classes a<br />

u<br />

(u, a) ∼ (u ′ , a ′ ) ⇔ there is v ∈ U such that vu ′ a = vua ′<br />

The addition is<br />

a b<br />

+<br />

u v<br />

and the multiplication is<br />

a b<br />

·<br />

u v<br />

The canonical ring homomorphism is<br />

= va + ub<br />

uv<br />

= ab<br />

uv<br />

ι : R → U −1 R, a ↦→ a<br />

1<br />

4.1.3. Proposition. Let φ : R → S be a ring homomorphism and U ⊂ R a<br />

multiplicative subset. If all elements in φ(U) ⊂ S are units, then there exists a<br />

unique ring homomorphism φ ′ : U −1 R → S such that φ = φ ′ ◦ ι.<br />

φ<br />

R <br />

<br />

ι <br />

<br />

U −1R Pro<strong>of</strong>. φ ′ ( a<br />

u ) = φ(a)φ(u)−1 is the well defined unique ring homomorphism. Observe<br />

that the elements <strong>of</strong> form bv −1 satisfy the rules for fractions.<br />

57<br />

φ ′<br />

<br />

<br />

S


58 4. FRACTION CONSTRUCTIONS<br />

4.1.4. Proposition. Let U ⊂ R be a multiplicative subset.<br />

(1) a<br />

1 = 0 in U −1R if and only if Ann(a) ∩ U = ∅.<br />

(2) The canonical ring homomorphism R → U −1R is injective if and only if U<br />

consists only <strong>of</strong> nonzero divisors.<br />

Pro<strong>of</strong>. a<br />

1 = 0 if and only if va = 0 for some v ∈ U.<br />

4.1.5. Corollary. If R is a domain and U is the nonzero elements then K = U −1 R<br />

is a field and the canonical homomorphism identifies R as a subring.<br />

4.1.6. Definition. The field K in 4.1.5 is the fraction field <strong>of</strong> R.<br />

4.1.7. Corollary. If R is a domain with fraction field K and φ : R → L is an<br />

injective ring homomorphism into a field, then there is a unique homomorphism<br />

K → L extending φ.<br />

4.1.8. Definition. The total ring <strong>of</strong> fractions <strong>of</strong> R is U −1 R, where U is the set <strong>of</strong><br />

nonzero divisors <strong>of</strong> R.<br />

4.1.9. Corollary. A ring is a subring <strong>of</strong> its total ring <strong>of</strong> fractions.<br />

4.1.10. Proposition. Let U = {u n } the powers <strong>of</strong> an element u ∈ R. There is an<br />

isomorphism<br />

R[X]/(uX − 1) → U −1 R, X ↦→ 1<br />

u<br />

The ring <strong>of</strong> fractions with only one denominator is <strong>of</strong> finite type.<br />

Pro<strong>of</strong>. A pair <strong>of</strong> inverse homomorphism are constructed by 1.6.7 and 4.1.3.<br />

4.1.11. Exercise. (1) Show that if U contains a nilpotent element, then U −1 R = 0.<br />

(2) Show that<br />

Ker R → U −1 R = {a ∈ R|ua = 0, for some u ∈ U}<br />

(3) Let U = {u1, . . . , um} and u = u1 · · · um. Then show that<br />

U −1 R = {u n } −1 R<br />

(4) Let R1, R2 be domains with fraction fields K1, K2. Show that the total ring <strong>of</strong> fractions<br />

<strong>of</strong> R1 × R2 is K1 × K2.<br />

(5) Show that the ring<br />

{ a<br />

∈ Q|a ∈ Z, n ∈ N}<br />

2n is <strong>of</strong> finite type over Z.<br />

4.2. Modules <strong>of</strong> fractions<br />

4.2.1. Lemma. Let R be a ring and U ⊂ R multiplicative. On U × M is defined<br />

a relation<br />

(u, x) ∼ (u ′ , x ′ ) ⇔ there is v ∈ U such that vu ′ x = vux ′<br />

(1) The relation is an equivalence relation.<br />

(2) If (u, x) ∼ (u ′ , x ′ ) and (v, y) ∼ (v ′ , y ′ ) then (uv, vx + uy) ∼ (u ′ v ′ , v ′ x ′ +<br />

u ′ y ′ ).<br />

(3) If (u, a) ∼ (u ′ , a ′ ) in U × R 4.1.1 and (v, x) ∼ (v ′ , x ′ ) then (uv, ax) ∼<br />

(u ′ v ′ , a ′ x ′ ).<br />

Pro<strong>of</strong>. The claims are proved by simple calculations. See the pro<strong>of</strong> 4.1.1.


4.2. MODULES OF FRACTIONS 59<br />

4.2.2. Definition. Let R be a ring and U a multiplicative subset. The module <strong>of</strong><br />

fractions U −1M is given by equivalence classes x<br />

u on U × M under the relation<br />

4.2.1<br />

(u, x) ∼ (u ′ , x ′ ) ⇔ there is v ∈ U such that vu ′ x = vux ′<br />

The addition is<br />

x y<br />

+<br />

u v<br />

and the U −1R-scalar multiplication is<br />

a y<br />

·<br />

u v<br />

The canonical homomorphism is<br />

= vx + uy<br />

uv<br />

= ay<br />

uv<br />

M → U −1 M, x ↦→ x<br />

1<br />

4.2.3. Lemma. Let R be a ring and f : M → N a homomorphism <strong>of</strong> R-modules.<br />

Then there is a homomorphism U −1f : U −1M → U −1N, x f(x)<br />

u ↦→ u <strong>of</strong> U −1R modules.<br />

Pro<strong>of</strong>. The claims are proved by simple calculations. If (u, x) ∼ (u ′ , x ′ ) then<br />

vu ′ x = vux ′ and therefore vu ′ f(x) = vuf(x ′ ) showing (u, f(x)) ∼ (u ′ , f(x ′ )).<br />

So the map is well defined. The rest is similar.<br />

4.2.4. Proposition. The construction 4.2.2<br />

and 4.2.3<br />

M ↦→ U −1 M<br />

f : M → N ↦→ U −1 f : U −1 M → U −1 N<br />

is a functor from R-modules to U −1 R-modules.<br />

Pro<strong>of</strong>. Follows from the definitions by simple calculations as in the pro<strong>of</strong> <strong>of</strong> 4.2.3.<br />

For example, U −1 (f + g) = U −1f + U −1g, follows from f(x)+g(x)<br />

u<br />

g(x)<br />

u .<br />

= f(x)<br />

u +<br />

4.2.5. Remark. The induced homomorphism relates to the canonical homomorphism<br />

such that the diagram is commutative.<br />

M<br />

f<br />

<br />

N<br />

<br />

<br />

U −1M U −1f <br />

U −1N That is, the canonical homomorphism is a natural homomorphism.<br />

4.2.6. Proposition. Let U ⊂ R be a multiplicative subset and M a module.<br />

(1) x<br />

1 = 0 in U −1M if and only if Ann(x) ∩ U = ∅.<br />

(2) The canonical homomorphism M → U −1M is injective if and only if U<br />

consists only <strong>of</strong> nonzero divisors on M.<br />

Pro<strong>of</strong>. x<br />

1 = 0 if and only if vx = 0 for some v ∈ U.


60 4. FRACTION CONSTRUCTIONS<br />

4.2.7. Proposition. If Mα is a family <strong>of</strong> modules, then the homomorphism<br />

U −1 ( <br />

Mα) → <br />

is a natural isomorphism <strong>of</strong> U −1 R-modules.<br />

α<br />

α<br />

U −1 Mα<br />

Pro<strong>of</strong>. This is the method <strong>of</strong> common denominators in a finite sum.<br />

xi<br />

=<br />

ui<br />

1 <br />

(Πj=iuj)xi<br />

Πiui<br />

i<br />

4.2.8. Exercise. (1) Show that if U contains a nilpotent element, then U −1 M = 0.<br />

(2) Show that<br />

Ker M → U −1 M = {x ∈ M|ux = 0, for some u ∈ U}<br />

(3) Let U = {u1, . . . , um} and u = u1 · · · um. Then show that<br />

U −1 M = {u n } −1 M<br />

(4) Show that U −1 M = 0 if and only if U ∩ Ann(x) = ∅ for all x ∈ M.<br />

(5) Show that the fraction homomorphism <strong>of</strong> a composition is the composition <strong>of</strong> the<br />

respective fraction homomorphisms.<br />

(6) Let M be a free R-module. Show that U −1 M is a free U −1 R-module<br />

4.3. Exactness <strong>of</strong> fractions<br />

4.3.1. Proposition. Let R be a ring and U a multiplicative subset. Given an exact<br />

sequence <strong>of</strong> R-modules<br />

M f<br />

Then the following sequence is exact<br />

U −1 M<br />

i<br />

g<br />

<br />

N<br />

<br />

U −1N <br />

L<br />

<br />

U −1L Pro<strong>of</strong>. If y<br />

u ∈ U −1N maps to g(y)<br />

u = 0 then there is v ∈ U such that 0 = vg(y) =<br />

g(vy). Choose x ∈ M such that f(x) = vy. Then x<br />

exactness.<br />

4.3.2. Corollary. Given a short exact sequence<br />

0<br />

f<br />

<br />

M<br />

Then the following sequence is exact<br />

0<br />

<br />

U −1M g<br />

<br />

N<br />

<br />

U −1N <br />

L<br />

vu<br />

maps to f(x)<br />

vu<br />

<br />

0<br />

<br />

U −1L If the first sequence is split exact, also the second sequence is split exact.<br />

<br />

0<br />

= y<br />

u proving<br />

4.3.3. Corollary. For a homomorphism f : M → N there are natural isomorphisms<br />

<strong>of</strong> U −1 R-modules.<br />

(1) U −1 Ker f Ker U −1 f.<br />

(2) U −1 Im f Im U −1 f.<br />

(3) U −1 Cok f Cok U −1 f.


4.3. EXACTNESS OF FRACTIONS 61<br />

Pro<strong>of</strong>. Represent the statements using short exact sequences. (1) The kernel is<br />

determined by the exact sequence 0 → Ker f → M → N, 3.1.4. (3) The cokernel<br />

is determined by the exact sequence M → N → Cok f → 0, 3.1.5. (2) The<br />

image is determined by the exact sequence 0 → Ker f → M → Im f → 0, 2.3.5,<br />

3.1.4.<br />

4.3.4. Corollary. For submodules N, L ⊂ M there are natural identifications <strong>of</strong><br />

U −1 R-submodules and factor modules.<br />

(1) U −1 (M/N) = U −1 M/U −1 N.<br />

(2) U −1 (N + L) = U −1 N + U −1 L.<br />

(3) U −1 (N ∩ L) = U −1 N ∩ U −1 L.<br />

Pro<strong>of</strong>. Represent the statements using short exact sequences. (1) 0 → N → M →<br />

M/N → 0 is short exact giving 0 → U −1 N → U −1 M → U −1 (M/N) → 0. The<br />

wanted isomorphism follows form 3.1.5. (2) N + L is the image <strong>of</strong> N ⊕ L → M<br />

so conclude by 4.3.3. (3) N ∩ L is the kernel <strong>of</strong> N ⊕ L → M so conclude by<br />

4.3.3.<br />

4.3.5. Corollary. For ideals I, J ⊂ R there are natural identifications in U −1 R.<br />

(1) U −1 (R/I) = U −1 R/U −1 I.<br />

(2) U −1 (I + J) = U −1 I + U −1 J.<br />

(3) U −1 (I ∩ J) = U −1 I ∩ U −1 J.<br />

(4) U −1 (IJ) = U −1 IU −1 J.<br />

Pro<strong>of</strong>. (1) (2) (3) These are special cases <strong>of</strong> 4.3.4. (4) Both sides have the same<br />

, a ∈ I, b ∈ J, u ∈ U.<br />

generators ab<br />

u<br />

4.3.6. Proposition. Let R → U −1 R be the canonical homomorphism.<br />

(1) For an ideal I ⊂ R the extended ideal<br />

IU −1 R = U −1 I<br />

(2) For an ideal J ⊂ U −1 R the extended contracted ideal<br />

U −1 (J ∩ R) = J<br />

(3) For an ideal I ⊂ R the contracted extended ideal<br />

I ⊂ IU −1 R ∩ R<br />

Pro<strong>of</strong>. (1) This is clear. (2) U −1 (J ∩ R) ⊂ J is true for any ring homomorphism<br />

1.2.6. If b<br />

b<br />

1 b<br />

u ∈ J then 1 ∈ J giving b ∈ J ∩ R and ub = u ∈ U −1 (J ∩ R). (3) This<br />

is true for any ring homomorphism 1.2.6.<br />

4.3.7. Proposition. Let R → U −1 R be the canonical homomorphism. For an<br />

ideal I ⊂ R the contracted extended ideal<br />

I = IU −1 R ∩ R<br />

if and only if each u ∈ U is a nonzero divisor on R/I.<br />

Pro<strong>of</strong>. Apply the snake lemma 3.2.4 to<br />

0<br />

0<br />

<br />

I<br />

<br />

<br />

0<br />

<br />

R<br />

<br />

<br />

<br />

U −1R/U −1I U −1R/U −1I <br />

0<br />

<br />

R/I<br />

<br />

0


62 4. FRACTION CONSTRUCTIONS<br />

and get the exact sequence<br />

Conclusion from 4.2.6.<br />

0 → I → IU −1 R ∩ R → Ker(R/I → U −1 (R/I)) → 0<br />

4.3.8. Corollary. Let P ⊂ R be a prime ideal. Then U −1 P ⊂ U −1 R is either a<br />

prime ideal or the whole ring.<br />

Pro<strong>of</strong>. By 4.1.5 and 4.3.5 U −1 R/U −1 P = U −1 (R/P ) is either 0 or a domain.<br />

4.3.9. Corollary. Let R be a principal ideal domain. Then U −1 R is a principal<br />

ideal domain.<br />

Pro<strong>of</strong>. A restricted ideal is principal by hypothesis and the extension <strong>of</strong> a principal<br />

ideal is principal. Conclude by 4.3.6.<br />

4.3.10. Proposition. Let R be a unique factorization domain. Then U −1 R is a<br />

unique factorization domain.<br />

Pro<strong>of</strong>. By 4.3.7 the extension <strong>of</strong> an irreducible element is either a unit or an irreducible<br />

element. Since a principal ideal ( a a<br />

u ) = ( 1 ), a factorization into irreducibles<br />

in R gives a factorization in U −1R. Now if (a) = (p1) . . . (pn) is a factorization<br />

in R and ( a<br />

1 ) is irreducible in U −1R. Then all but one pi<br />

1 is a unit in U −1R so<br />

( a pi<br />

1 ) = ( 1 ) is a prime ideal 4.3.9. The conditions 1.5.3 are satisfied.<br />

4.3.11. Exercise. (1) Let U ⊂ R be multiplicative. Show that<br />

U −1 (R[X]) = (U −1 R)[X]<br />

(2) Let φ : R → S be a ring homomorphism and U ⊂ R a multiplicative subset. Show<br />

that<br />

U −1 S = φ(U) −1 S<br />

4.4. Tensor modules <strong>of</strong> fractions<br />

4.4.1. Proposition. Let R be a ring and U a multiplicative subset. For any module<br />

M, the homomorphism<br />

M ⊗R U −1 R → U −1 M, x ⊗ a ax<br />

↦→<br />

u u<br />

is a natural isomorphism <strong>of</strong> U −1R-modules. Pro<strong>of</strong>. By 2.6.3 there is an R-module homomorphism x ⊗ a<br />

u<br />

b a<br />

ba bax b ax<br />

(x ⊗ ) = x ⊗ ↦→ =<br />

v<br />

u<br />

vu<br />

vu<br />

map U −1 M → M ⊗R U −1 R, x<br />

u<br />

v<br />

ax → u . By definition<br />

u , so the this is a U −1 R-homomorphism. The<br />

↦→ x ⊗ 1<br />

u<br />

is an inverse.<br />

4.4.2. Remark. The two constructions, module change <strong>of</strong> ring to a fraction ring<br />

and fraction module are natural isomorphic functors from modules to modules over<br />

the fraction ring.<br />

4.4.3. Corollary. Let R be a ring and U a multiplicative subset. Then U −1 R is a<br />

flat R-module.<br />

Pro<strong>of</strong>. This follows from 4.4.1 and 4.3.1.<br />

4.4.4. Corollary. Let R be a ring and U a multiplicative subset and M, N modules.<br />

Then there is a natural isomorphism<br />

U −1 (M ⊗R N) U −1 M ⊗ U −1 R U −1 N


4.5. HOMOMORPHISM MODULES OF FRACTIONS 63<br />

Pro<strong>of</strong>. This follows from 2.7.5 and 4.4.1.<br />

4.4.5. Corollary. Let I ⊂ R be an ideal and U a multiplicative subset For a<br />

module M module<br />

U −1 (IM) U −1 IU −1 M<br />

Pro<strong>of</strong>. Use that IM = Im(I ⊗R M → M) and 4.4.4.<br />

4.4.6. Exercise. (1) Let M be a flat R-module. Show that U −1 M is a flat U −1 Rmodule.<br />

(2) Let M be a projective R-module. Show that U −1 M is a projective U −1 R-module.<br />

4.5. Homomorphism modules <strong>of</strong> fractions<br />

4.5.1. Proposition. Let R be a ring and U a multiplicative subset. For any modules<br />

M, N there is a natural homomorphism<br />

<strong>of</strong> U −1 R-modules.<br />

U −1 HomR(M, N) → Hom U −1 R(U −1 M, U −1 N)<br />

Pro<strong>of</strong>. Given f : M → N and u ∈ U the setting x<br />

homomorphism U −1 M → U −1 N.<br />

v ↦→ f(x)<br />

uv is a U −1R 4.5.2. Proposition. Let R be a ring and U a multiplicative subset. For any Rmodule<br />

M and any U −1 R-modules N there is a natural isomorphism<br />

<strong>of</strong> U −1 R-modules.<br />

HomR(M, N) → Hom U −1 R(U −1 M, N)<br />

Pro<strong>of</strong>. This is the change <strong>of</strong> rings isomorphism 2.7.6 interpreted according to<br />

4.4.2.<br />

4.5.3. Example. Let R be a ring and U a multiplicative subset. R → U −1 R the<br />

canonical homomorphism. The induced module functor maps an R-module M to<br />

the U −1 R-module HomR(U −1 R, M). The natural isomorphism 2.7.10 is<br />

for any U −1 R-module N.<br />

HomR(N, M) Hom U −1 R(N, HomR(U −1 R, M))<br />

4.5.4. Example. The homomorphism 4.5.1 is in general neither injective nor surjective.<br />

(1) Not surjective:<br />

(2) Not injective:<br />

0 = HomZ(Q, Z) → HomQ(Q, Q) = Q<br />

0 = HomZ(Q, Q/Z) ⊗Z Q → HomQ(Q, Q/Z ⊗Z Q) = 0<br />

4.5.5. Exercise. (1) Let M be a Z-module and N a Q-module. Show that there is an<br />

isomorphism HomZ(M, N) HomQ(M ⊗Z Q, N).


64 4. FRACTION CONSTRUCTIONS<br />

4.6. The polynomial ring is factorial<br />

4.6.1. Definition. Let R be a unique factorization domain and let f = anX n +<br />

· · · + a0 be a polynomial over R. Then the content <strong>of</strong> polynomial f, c(f), is the<br />

greatest common divisor <strong>of</strong> the coefficients a0, . . . , an.<br />

4.6.2. Proposition (Gauss’ lemma). Let R be a unique factorization domain. For<br />

polynomials f, g ∈ R[X]<br />

c(fg) = c(f)c(g)<br />

Pro<strong>of</strong>. Assume by cancellation that c(f), c(g) are units in R. For any irreducible<br />

p ∈ R the projections <strong>of</strong> f, g in R/(p)[X] are nonzero. Since R has unique factorization<br />

the ideal (p) is a prime ideal. It follows that the projection <strong>of</strong> the product<br />

fg in R/(p)[X] is also nonzero and therefore p is not a common divisor <strong>of</strong> the<br />

coefficients <strong>of</strong> the product fg.<br />

4.6.3. Proposition. Let R be a unique factorization domain. Then the ring <strong>of</strong><br />

polynomials R[X] is a unique factorization domain.<br />

Pro<strong>of</strong>. Let K be the fraction field <strong>of</strong> R, then the polynomial ring K[X] is a principal<br />

ideal domain. Let f ∈ R[X] and use unique factorization in K[X] to get<br />

0 = a ∈ R and p1, . . . , pn ∈ R[X], irreducible in K[X], such that<br />

af = p1 . . . pn<br />

Assume by 4.6.2 that a = 1 and c(p1), . . . , c(pn) are units in R. Apply 4.6.2 and<br />

1.6.5 to see that p1, . . . , pn are irreducible in R[X]. An irreducible p ∈ R generates<br />

a prime ideal (p) ⊂ R[X]. A non constant irreducible p ∈ R[X] generates a prime<br />

ideal (p) ⊂ K[X] and therefore also a prime ideal (p) ⊂ R[X]. So conditions<br />

1.5.3 are satisfied.<br />

4.6.4. Theorem. Let K be a field. Then the polynomial ring K[X1, . . . , Xn] is a<br />

unique factorization domain.<br />

Pro<strong>of</strong>. Follows by induction from 4.6.3.<br />

4.6.5. Exercise. (1) Let f ∈ Z[X] be monic and assume f = gh where g, h ∈ Q[X]<br />

are monic. Show that g, h ∈ Z[X].<br />

(2) Let f ∈ Z[X] be monic and irreducible in Z/(n)[X]. Show that f is irreducible<br />

Q[X].<br />

(3) Let K be a field. Show that the polynomial ring K[X1, X2, . . . ] in countable many<br />

variables is a unique factorization domain


5<br />

Localization<br />

5.1. Prime ideals<br />

5.1.1. Theorem (Krull). A nonzero ring contains a maximal ideal.<br />

Pro<strong>of</strong>. The nonempty set <strong>of</strong> ideals different from R is ordered by inclusion. Given<br />

an increasing chain Iα then ∪Iα is an ideal different from R which is a maximum<br />

for the chain. Conclusion by Zorn’s lemma.<br />

5.1.2. Corollary. Any proper ideal in a ring is contained in a maximal ideal.<br />

Pro<strong>of</strong>. The factor ring is nonzero and contains a maximal ideal. Conclusion by<br />

1.2.10.<br />

5.1.3. Proposition. Let P1, . . . , Pn ⊂ R be ideals with at most 2 not being prime<br />

ideals. If an ideal<br />

I ⊂ P1 ∪ · · · ∪ Pn<br />

then I ⊆ Pi for some i.<br />

Pro<strong>of</strong>. Assume n > 1 and I is not contained in any sub union. Moreover assume<br />

the numbering such that P3, . . . , Pn are prime ideals. Then for each i there is<br />

The element<br />

ai ∈ (I ∩ Pi)\ ∪j=i Pj<br />

an + a1 . . . an−1<br />

is in I but not in any Pi, giving a contradiction. So n = 1.<br />

5.1.4. Proposition. Let P1, . . . , Pn ⊂ R be prime ideals and I any ideal. If for<br />

some a ∈ R<br />

a + I ⊂ P1 ∪ · · · ∪ Pn<br />

then I ⊆ Pi for some i.<br />

Pro<strong>of</strong>. If a ∈ ∩iPi then conclusion by 5.1.3. On the contrary after renumbering<br />

there exists j with 1 ≤ j < n such that<br />

a ∈ P1 ∩ · · · ∩ Pj\Pj+1 ∪ · · · ∪ Pn<br />

Assume no inclusions between the prime ideals and choose by 5.1.3<br />

b ∈ I ∩ Pj+1 ∩ · · · ∩ Pn\P1 ∪ · · · ∪ Pj<br />

Then a + b /∈ ∪iPi contradicts the hypothesis.<br />

5.1.5. Proposition. Let R → U −1 R be the canonical homomorphism. Extension<br />

and contraction gives a bijective correspondence between prime ideals in R disjoint<br />

from U and all prime ideals in U −1 R.<br />

(1) For a prime ideal P ⊂ R\U the extended ideal P U −1 R is a prime ideal in<br />

U −1 R and the contracted P U −1 R ∩ R = P .<br />

65


66 5. LOCALIZATION<br />

(2) For a prime ideal Q ⊂ U −1 R the contracted ideal Q ∩ R is a prime ideal<br />

and the extended (Q ∩ R)U −1 R = Q<br />

Pro<strong>of</strong>. Conclusion by 4.3.6, 4.3.7, 4.3.8.<br />

5.1.6. Corollary. Let R be a ring and U a multiplicative subset.<br />

(1) An ideal maximal among the ideals disjoint from U is a prime ideal.<br />

(2) Any ideal disjoint from U is contained in a prime ideal disjoint form U.<br />

Pro<strong>of</strong>. The prime ideals disjoint from U are the prime ideals in U −1 R.<br />

5.1.7. Proposition. The nilradical <strong>of</strong> a ring R is the intersection <strong>of</strong> all prime ideals<br />

P .<br />

√ <br />

0 = P<br />

Pro<strong>of</strong>. By 1.3.8 the nilradical is contained in any prime ideal. Suppose u ∈ R is<br />

not nilpotent. Then {u n } −1 R is nonzero. Then contraction <strong>of</strong> a maximal ideal,<br />

5.1.1, is a prime ideal in R not containing u.<br />

5.1.8. Corollary. Let R be a ring.<br />

(1) The radical <strong>of</strong> an ideal I is the intersection <strong>of</strong> all prime idealsP containing I<br />

√ <br />

I = P<br />

P<br />

I⊂P<br />

(2) For ideals I, J ⊂ R, √ I ∩ J = √ I ∩ √ J.<br />

(3) If U is a multiplicative subset, then U −1√ 0 = √ 0 in U −1 R.<br />

If R is reduced, then U −1 R is reduced.<br />

Pro<strong>of</strong>. (1) Use 5.1.7 on the factor ring R/I. (2) This follows from (1). (3) Use the<br />

correspondence 5.1.5.<br />

5.1.9. Definition. A prime ideal minimal for inclusion among prime ideals is a<br />

minimal prime ideal.<br />

5.1.10. Proposition. Any prime ideal <strong>of</strong> Q ⊂ R contains a minimal prime ideal<br />

P ⊂ Q.<br />

Pro<strong>of</strong>. The set <strong>of</strong> prime ideals in R is ordered by inclusion. Given a decreasing<br />

chain Pα then ∩Pα is a prime ideal. Conclusion by Zorn’s lemma.<br />

5.1.11. Corollary. Let R ⊂ S be a subring and P ⊂ R a minimal prime ideal.<br />

Then there is a minimal prime ideal Q ⊂ S contracting to P = Q ∩ R.<br />

5.1.12. Exercise. (1) Let K ⊂ R be an infinite subfield and I, P1, . . . , Pn any ideals.<br />

Show that if I ⊂ P1 ∪ · · · ∪ Pn then I ⊂ Pi for some i.<br />

(2) Let P, P1, P2 be proper ideals. Show that if P is a maximal ideal and P n ⊂ P1 ∪ P2<br />

then P = P1 <strong>of</strong> P = P2.


5.2. LOCALIZATION OF RINGS 67<br />

5.2. Localization <strong>of</strong> rings<br />

5.2.1. Definition. A ring R which contains precisely one maximal ideal P is a<br />

local ring and denoted (R, P ). The residue field <strong>of</strong> R is R/P denoted by k(P ).<br />

A ring homomorphism φ : R → S <strong>of</strong> local rings (R, P ), (S, Q) is a local ring<br />

homomorphism if φ(P ) ⊂ Q.<br />

5.2.2. Lemma. Let R be a ring and Q = ∅ a subset. Then R is a local ring with<br />

maximal ideal Q if and only if R\Q is the set <strong>of</strong> units in R.<br />

Pro<strong>of</strong>. Use that an ideal I = R if and only if it contains a unit.<br />

5.2.3. Lemma. A ring homomorphism φ : R → S <strong>of</strong> local rings (R, P ), (S, Q) is<br />

a local ring homomorphism if the extended ideal P S ⊂ Q or the contracted ideal<br />

Q ∩ R = P . The residue homomorphism k(P ) → k(Q) is a field extension.<br />

Pro<strong>of</strong>. The contraction Q ∩ R is a prime ideal containing P . The rest is clear.<br />

5.2.4. Lemma. Let R be a ring and P a prime ideal. Then U = R\P is a multiplicative<br />

subset. The ring <strong>of</strong> fractions U −1 R is a local ring. The maximal ideal is<br />

the extended ideal P U −1 R. The residue field is U −1 R/P U −1 R which is canonical<br />

isomorphic to the fraction field <strong>of</strong> R/P .<br />

5.2.5. Definition. Let R be a ring, P a prime ideal and U = R\P . The localized<br />

ring at P is the local ring RP = U −1 R. The residue field is denoted k(P ) =<br />

RP /P RP .<br />

5.2.6. Example. Let the ring be Z.<br />

(1) The local ring at (0) is the fraction field Q = Z (0).<br />

(2) The local ring Z (p) for a prime number p is identified with a subring <strong>of</strong> Q<br />

Z (p) = { m<br />

|p not dividing n}<br />

n<br />

The residue field Fp = Z (p)/(p).<br />

(3) Any nonzero ideal in Z (p) is principal <strong>of</strong> the form (pn ) for some n.<br />

5.2.7. Proposition. Let (R, P ) be a local ring. One <strong>of</strong> the following conditions is<br />

satisfied:<br />

(1) The characteristic char(R) = 0. P ∩ Z = (0) and Q ⊂ R is a subfield.<br />

Q → k(P ) is a field extension.<br />

(2) The characteristic char(R) = 0. P ∩ Z = (p), p a prime number. Z (p) ⊂ R<br />

is a local subring. Fp → k(P ) is a field extension.<br />

(3) The characteristic char(R) = p n , a power <strong>of</strong> a prime number. Z/(p n ) ⊂ R<br />

is a local subring. Fp → k(P ) is a field extension.<br />

Pro<strong>of</strong>. (1) (2) are clear by 5.2.3 and 5.2.6. (3) If the characteristic is nonzero then<br />

any prime ideal contracts Q ∩ Z = (p). So a prime number q = p gives a unit in<br />

R. There is a local homomorphism Z (p) → R, 4.1.3. The nontrivial kernel is (p n )<br />

by 5.2.6.<br />

5.2.8. Example. A field K is a local ring with maximal ideal (0). The power series<br />

ring K[[X]] is a local ring with maximal ideal (X) and residue field K.<br />

5.2.9. Remark. Let R be a domain.<br />

(1) The local ring at (0) is the fraction field K = R (0).


68 5. LOCALIZATION<br />

(2) Any local ring RP is identified with a subring <strong>of</strong> the fraction field K.<br />

(3) The intersection<br />

R = <br />

RP , P a maximal ideal<br />

P<br />

5.2.10. Remark. Let R × S be a product <strong>of</strong> rings.<br />

(1) A prime ideal is <strong>of</strong> the form P × S or R × Q for uniquely determined prime<br />

ideals P ⊂ R or Q ⊂ S.<br />

(2) The local ring at P ×S is identified with RP through the projection R ×S →<br />

R.<br />

(3) The local ring at R × Q is identified with SQ through the projection R × S →<br />

S.<br />

5.2.11. Proposition. Let P be a prime ideal and R → RP the canonical homomorphism.<br />

Extension and contraction gives a bijective correspondence between<br />

prime ideals in R contained in P and all prime ideals in RP .<br />

(1) For a prime ideal Q ⊂ P the extended ideal QRP is a prime ideal in RP and<br />

the contracted QRP ∩ R = Q.<br />

(2) For a prime ideal Q ′ ⊂ RP the contracted ideal Q ′ ∩ R ⊂ P is a prime ideal<br />

and the extended (Q ′ ∩ R)RP = Q ′<br />

Pro<strong>of</strong>. This is a special case <strong>of</strong> 5.1.5.<br />

5.2.12. Lemma. Let Q ⊂ P ⊂ R be prime ideals. Then QRP is a prime ideal in<br />

RP and canonically<br />

Pro<strong>of</strong>. By fraction rules a w av<br />

u / v = uw .<br />

RQ = (RP )QRP<br />

5.2.13. Definition. The intersection <strong>of</strong> all maximal ideals in a ring is the Jacobson<br />

radical.<br />

5.2.14. Remark. The Jacobson radical contains the nilradical. In a local ring the<br />

Jacobson radical is the maximal ideal.<br />

5.2.15. Exercise. (1) Show that a local ring is never a product <strong>of</strong> two nonzero rings.<br />

(2) Show that a ∈ R is in the Jacobson radical if and only if 1+ab is a unit for all b ∈ R.<br />

(3) Let p be a prime number. Describe the prime ideals in the ring Z (p).<br />

(4) Let P be a prime ideal. Show that k(P ) is the fraction field <strong>of</strong> R/P .<br />

(5) Let (R, P ) be a local ring. Show that (R[[X]], (P, X) is a local ring and the canonical<br />

homomorphism R → R[[X]] is a local homomorphism.<br />

5.3. Localization <strong>of</strong> modules<br />

5.3.1. Definition. Let R be a ring, P a prime ideal and U = R\P . For a module<br />

M, the localized module at P is the module MP = U −1 M over the local ring RP .<br />

For a homomorphism f : M → N the localized homomorphism is fP : MP →<br />

NP and the residue homomorphism is f(P ) : M ⊗R k(P ) → N ⊗R k(P ).<br />

The constructions are functors, 2.7.4, 4.2.4.<br />

5.3.2. Lemma. Let Q ⊂ P ⊂ R be prime ideals and M an R-module Then QRP<br />

is a prime ideal in RP and canonically<br />

Pro<strong>of</strong>. See pro<strong>of</strong> <strong>of</strong> 5.2.12.<br />

MQ = (MP )QRP


5.3. LOCALIZATION OF MODULES 69<br />

5.3.3. Proposition. Let R be a ring and P a prime ideal. If Mα is a family <strong>of</strong><br />

modules, then the homomorphism<br />

( <br />

Mα)P → <br />

(Mα)P<br />

is an isomorphism <strong>of</strong> RP -modules.<br />

Pro<strong>of</strong>. See 4.2.7.<br />

5.3.4. Corollary. For a homomorphism f : M → N<br />

(1) (Ker f)P Ker fP .<br />

(2) (Im f)P Im fP .<br />

(3) (Cok f)P Cok fP .<br />

Pro<strong>of</strong>. See 4.3.3.<br />

α<br />

5.3.5. Corollary. Let R be a ring and P a prime ideal. For submodules N, L ⊂ M<br />

(1) (M/N)P MP /NP .<br />

(2) (N + L)P NP + LP .<br />

(3) (N ∩ L)P NP ∩ LP .<br />

Pro<strong>of</strong>. See 4.3.4.<br />

5.3.6. Proposition. Let R be a ring, P a prime ideal and M a module.<br />

(1) MP M ⊗R RP .<br />

(2) MP /P RP MP M ⊗R k(P ).<br />

Pro<strong>of</strong>. See 4.4.1.<br />

5.3.7. Proposition. Let R be a ring, P a prime ideal.<br />

(1) For an R module M and an RP -module N there is a natural isomorphism<br />

M ⊗R RP ⊗RP N M ⊗R N<br />

(2) For an R module M, L there is a natural isomorphism<br />

Pro<strong>of</strong>. See 4.2.7 and 2.7.4.<br />

(M ⊗R L)P MP ⊗RP LP<br />

5.3.8. Definition. Let R be a ring. F is a locally free module is FP is a free<br />

RP -module for all prime ideals P .<br />

5.3.9. Lemma. Let R be a ring. F is a locally free module if FQ is a free RQmodule<br />

for all maximal ideals Q.<br />

Pro<strong>of</strong>. A prime ideal P ⊂ Q is contained in a maximal ideal. By 5.3.6 FP <br />

(FQ)PQ is free.<br />

5.3.10. Example. A free module is a locally free module.<br />

5.3.11. Exercise. (1) Let P ⊂ R be a prime ideal and R → S a ring homomorphism.<br />

Show that RP → SP is a ring homomorphism.<br />

(2) Let Q ⊂ S be a prime ideal and R → S a ring homomorphism. Show that RQ∩R →<br />

SQ is a local ring homomorphism.<br />

(3) Let R = K × L be a product <strong>of</strong> fields. Show that ideal K × {0} is locally free but<br />

not free.<br />

α


70 5. LOCALIZATION<br />

5.4. Exactness and localization<br />

5.4.1. Proposition. Let R be a ring and M a module. The following conditions<br />

are equivalent.<br />

(1) M = 0.<br />

(2) MP = 0 for all prime ideals P .<br />

(3) MP = 0 for all maximal ideals P .<br />

Pro<strong>of</strong>. (1) ⇒ (2) ⇒ (3) is clear. (3) ⇒ (1): Let 0 = x ∈ M be given. Then<br />

Ann(x) ⊂ P is contained in a maximal ideal, 5.1.2. Clearly 0 = x<br />

1 ∈ MP<br />

contradicts (3).<br />

5.4.2. Corollary. Let R be a ring and f : M → N a homomorphism. The following<br />

conditions are equivalent.<br />

(1) f is injective.<br />

(2) fP is injective for all prime ideals P .<br />

(3) fP is injective for all maximal ideals P .<br />

Pro<strong>of</strong>. Use 5.4.1 on Ker f.<br />

5.4.3. Corollary. Let R be a ring and f : M → N a homomorphism. The following<br />

conditions are equivalent.<br />

(1) f is surjective.<br />

(2) fP is surjective for all prime ideals P .<br />

(3) fP is surjective for all maximal ideals P .<br />

Pro<strong>of</strong>. Use 5.4.1 on Cok f.<br />

5.4.4. Corollary. Let R be a ring and f : M → N a homomorphism. The following<br />

conditions are equivalent.<br />

(1) f is an isomorphism.<br />

(2) fP is an isomorphism for all prime ideals P .<br />

(3) fP is an isomorphism for all maximal ideals P .<br />

5.4.5. Corollary. Let R be a ring and<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

a sequence <strong>of</strong> homomorphisms. The following conditions are equivalent.<br />

(1) The sequence is short exact.<br />

(2) The sequence<br />

0<br />

<br />

MP<br />

fP <br />

NP<br />

is short exact for all prime ideals P .<br />

(3) The sequence<br />

0<br />

<br />

MP<br />

fP <br />

NP<br />

is short exact for all maximal ideals P .<br />

<br />

L<br />

gP <br />

LP<br />

gP <br />

LP<br />

5.4.6. Corollary. Let R be a ring and F a module. The following conditions are<br />

equivalent.<br />

(1) F is flat.<br />

(2) FP is flat for all prime ideals P .<br />

<br />

0<br />

<br />

0<br />

<br />

0


(3) FP is flat for all maximal ideals P .<br />

5.5. FLAT RING HOMOMORPHISMS 71<br />

Pro<strong>of</strong>. Let 0 → M → N. Use 5.3.7 and 5.4.2 on M ⊗R F → N ⊗R F .<br />

5.4.7. Proposition. Let R be a ring and M a module. Then there is an exact<br />

sequence<br />

0 → M → <br />

P maximal<br />

Pro<strong>of</strong>. Let 0 = x ∈ M be given. Then Ann(x) ⊂ P is contained in a maximal<br />

ideal, 5.1.2. Clearly 0 = x<br />

1 ∈ MP .<br />

5.4.8. Corollary. Let R be a ring. Then there is an injective ring homomorphism<br />

R → <br />

P maximal<br />

5.4.9. Corollary. Let R be a ring. The following conditions are equivalent.<br />

(1) R is reduced.<br />

(2) RP is reduced for all prime ideals P .<br />

(3) RP is reduced for all maximal ideals P .<br />

Pro<strong>of</strong>. Use 5.1.8 and 5.4.10.<br />

5.4.10. Exercise. (1) Let R be a ring and<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

a split exact sequence. Show that the localized sequence is split exact for all prime<br />

ideals P .<br />

RP<br />

MP<br />

<br />

L<br />

5.5. Flat ring homomorphisms<br />

5.5.1. Definition. A flat R-module F is faithfully flat if for any M = 0 the tensor<br />

product M ⊗R F = 0.<br />

5.5.2. Example. A nonzero free module is faithfully flat.<br />

5.5.3. Lemma. Let F be a faithfully flat R-module and M → N a homomorphism.<br />

If M ⊗R F → N ⊗R F is injective, then M → N is injective.<br />

Pro<strong>of</strong>. Let f : M → N. Then 0 → Ker f ⊗R F → M ⊗R F → N ⊗R F is exact.<br />

It follows, that Ker f = 0.<br />

5.5.4. Definition. A ring homomorphism R → S is a flat ring homomorphism if<br />

S is a flat R-module and a faithfully flat ring homomorphism if S is faithfully flat.<br />

5.5.5. Proposition. Let R → S be a faithfully flat ring homomorphism. for any<br />

ideal I ⊂ R the extended contracted returns I, i.e.<br />

I = IS ∩ R<br />

Pro<strong>of</strong>. Tensor the homomorphism R/I → R/IS∩R with S. The induced R/I⊗R<br />

S → R/IS ∩ R ⊗R S is canonically isomorphic to the identity S/IS → S/IS.<br />

By 5.5.3 R/I → R/IS ∩ R is injective giving I = IS ∩ R.<br />

5.5.6. Corollary. A faithfully flat ring homomorphism R → S is injective.<br />

5.5.7. Proposition. Let φ : R → S be a ring homomorphism. The following<br />

conditions are equivalent<br />

<br />

0


72 5. LOCALIZATION<br />

(1) φ is flat.<br />

(2) φP : RP → SP is flat for all prime ideals P ⊂ R.<br />

(3) φP : RP → SP is flat for all maximal ideals P ⊂ R.<br />

Pro<strong>of</strong>. Use 5.4.6.<br />

5.5.8. Proposition. A flat local homomorphism (R, P ) → (S, Q) is faithfully flat.<br />

Pro<strong>of</strong>. Let 0 = x ∈ M be an R-module. R/ Ann(x) Rx and I = Ann(x) ⊂ P ,<br />

so IS ⊂ Q. Then R/I ⊗R S S/IS = 0 and therefore M ⊗R S = 0 giving the<br />

claim.<br />

5.5.9. Proposition (going-down). Let φ : R → S be a flat ring homomorphism<br />

and Q ⊂ S a prime ideal. For any prime ideal P ′ ⊂ P = Q ∩ R there is a prime<br />

ideal Q ′ ⊂ Q contracting to P ′ = Q ′ ∩ R.<br />

Pro<strong>of</strong>. The local homomorphism RP → SQ is faithfully flat. The ring k(P ′ ) ⊗R<br />

SQ is nonzero and therefor contains a maximal ideal Q ′′ . The contraction to S,<br />

Q ′ = Q ′′ ∩ S contracts to P ′ = R ∩ Q ′ .<br />

5.5.10. Proposition. Let R → S be a flat homomorphism. The following conditions<br />

are equivalent.<br />

(1) R → S is faithfully flat.<br />

(2) Any prime ideal P ⊂ R is the contraction P = Q ∩ R <strong>of</strong> a prime ideal<br />

Q ⊂ S.<br />

Pro<strong>of</strong>. Assume (2). Let M = 0 be an R-module. Then MP = 0 for some P . Let<br />

P = Q ∩ R, then (M ⊗R S)Q MP ⊗RP SQ = 0 as RP → SQ is faithfully flat.<br />

So (1) is true.<br />

5.5.11. Proposition. The inclusion R → R[X1, . . . , Xn] is a faithfully flat homomorphism<br />

Pro<strong>of</strong>. The R-module R[X1, . . . , Xn] is free.<br />

5.5.12. Exercise. (1) Show that a free module is faithfully flat.<br />

(2) Let R → S and S → T be flat homomorphisms. Show that the composite R → S is<br />

flat.<br />

(3) Show that Q is a flat but not faithfully flat Z-module.<br />

(4) Let R be a ring and I = √ 0 the nilradical. Show that IR[X] is the nilradical <strong>of</strong><br />

R[X].


6<br />

Finite modules<br />

6.1. Finite Modules<br />

6.1.1. Definition. Let R be a ring. A finite module is generated by finitely many<br />

elements. The finite free module with standard basis e1, . . . , en is denoted R n .<br />

6.1.2. Lemma. Let R be a ring and M a module. The following are equivalent.<br />

(1) M is generated by n elements x1, . . . , xn.<br />

(2) There is a surjective homomorphism R n → M → 0, ei ↦→ xi.<br />

Pro<strong>of</strong>. See 2.4.12.<br />

6.1.3. Proposition. Let R → S be a ring homomorphism. If an R-module M is<br />

generated by n elements x1, . . . , xn. Then the change <strong>of</strong> rings S-module M ⊗R S<br />

is generated by x1 ⊗ 1, . . . , xn ⊗ 1 over S.<br />

Pro<strong>of</strong>. Follows from 6.1.2 and 3.4.1<br />

6.1.4. Corollary. Let R be a ring and U a multiplicative subset. If M is a finite<br />

R-module, then U −1 M is a finite U −1 R-module.<br />

6.1.5. Proposition. For a short exact sequence<br />

0<br />

f<br />

<br />

M<br />

the following h<strong>old</strong><br />

(1) If N is finite, then L is finite.<br />

(2) If M, L are finite, then N is finite.<br />

g<br />

<br />

N<br />

Pro<strong>of</strong>. (1) If y1, . . . , yn generates N, then g(y1), . . . , g(yn) generates L. (2) Choose<br />

u : R n → M → 0 and v : R m → L → 0 exact. By 3.5.4 there is w : R m → N<br />

such that g ◦ w = v. There is a diagram<br />

0<br />

0<br />

<br />

Rn <br />

u<br />

<br />

<br />

M<br />

Conclusion by the snake lemma 3.2.4.<br />

6.1.6. Corollary. Let<br />

0<br />

f<br />

f<br />

<br />

M<br />

<br />

L<br />

Rn ⊕ Rm <br />

f◦u+w<br />

<br />

g<br />

<br />

N<br />

g<br />

<br />

N<br />

<br />

0<br />

Rm <br />

v<br />

<br />

<br />

L<br />

be a split exact sequence. Then N is finite if and only if M, L are finite.<br />

Pro<strong>of</strong>. Let u be a retraction <strong>of</strong> f. By 6.1.5 Im u = M is finite. The rest is contained<br />

in 6.1.5.<br />

73<br />

<br />

L<br />

<br />

0<br />

0<br />

<br />

0


74 6. FINITE MODULES<br />

6.1.7. Corollary. Let f : M → N be a homomorphism.<br />

(1) If M is finite, then Im f is finite.<br />

(2) If Ker f, Im f are finite, then M is finite.<br />

(3) If N is finite, then Cok f is finite.<br />

(4) If Im f, Cok f are finite, then N is finite.<br />

Pro<strong>of</strong>. Use 6.1.5 on the exact sequences 3.1.8.<br />

6.1.8. Corollary. Let M, N be modules. Then M ⊕ N is finite if and only if M<br />

and N are finite.<br />

Pro<strong>of</strong>. Use 6.1.6.<br />

6.1.9. Proposition. Let R be a ring and M, N finite modules. Then M ⊗R N is<br />

finite.<br />

Pro<strong>of</strong>. Use 6.1.2 and 6.1.5. Let R m → M and R n → N be surjective. Then<br />

R m ⊗R R n → M ⊗R N is surjective, 3.4.1.<br />

6.1.10. Proposition. Let R be a ring and M a module. The following are equivalent.<br />

(1) M is finite and projective.<br />

(2) M is a direct summand in a finite free module.<br />

Pro<strong>of</strong>. Use 6.1.2 and 6.1.8.<br />

6.1.11. Proposition. Let R be a ring and M, N finite modules.<br />

(1) If M is projective, then HomR(M, N) is finite.<br />

(2) If M, N are projective, then HomR(M, N) is projective.<br />

Pro<strong>of</strong>. Use 6.1.8.<br />

6.1.12. Proposition. Let F be a finite projective module and E an injective module.<br />

(1) F ⊗R E is injective.<br />

(2) HomR(F, E) is injective.<br />

Pro<strong>of</strong>. (1) (2) Both modules become summands in injective modules.<br />

6.1.13. Proposition. Let R be a ring and U a multiplicative subset. For a finite<br />

module M the following h<strong>old</strong>:<br />

(1) U −1 M = 0 if and only if there is a u ∈ U such that uM = 0.<br />

(2) Ann(U −1 M) = U −1 Ann(M) in U −1 R.<br />

Pro<strong>of</strong>. Let x1, . . . , xn generate M. (1) U −1 M = 0 if and only if u1x1 = · · · =<br />

unxn. Put u = u1 . . . un. (2) Ann(M) = Ann(x1) ∩ · · · ∩ Ann(xn). Now use (1)<br />

and 4.3.4.<br />

6.1.14. Proposition. Let R be a ring and Mα a family <strong>of</strong> modules. For any finite<br />

module N there is a natural isomorphism<br />

HomR(N, <br />

Mα) <br />

HomR(N, Mα)<br />

α<br />

Pro<strong>of</strong>. A homomorphism f : N → Mα has an image in a finite sum.<br />

6.1.15. Exercise. (1) Show that <br />

n Z/(n) is not a finite Z-module.<br />

α


6.2. FREE MODULES 75<br />

(2) Let K be a field and R = K[X1, X2, . . . ] the polynomial ring in countable many<br />

variables. Show that R is a finite module, but the ideal (X1, X2, . . . ) is not a finite<br />

module.<br />

6.2. Free Modules<br />

6.2.1. Definition. Let R be a ring and let R n be the free module with standard<br />

basis e1, . . . , en.<br />

(1) Let A = (aij) be a m × n-matrix with m rows and n columns, where the<br />

entry aij ∈ R. Identify Rn with n-columns. Then matrix multiplication<br />

x = (xj) ↦→ y = Ax, yi = <br />

j<br />

aijxj<br />

defines a homomorphism R n → R m .<br />

(2) Let f : R n → R m be a homomorphism. Then define a m × n-matrix A =<br />

(aij) by<br />

f(ej) = <br />

i<br />

aijei<br />

6.2.2. Proposition. (1) The dictionary defined in 6.2.1 gives a canonical isomorphism<br />

between the module <strong>of</strong> m × n-matrices and HomR(R n , R m ).<br />

(2) Matrix multiplication corresponds to composition <strong>of</strong> homomorphisms and the<br />

identity matrix corresponds to the identity homomorphism.<br />

(3) Invertible matrices correspond to isomorphisms.<br />

Pro<strong>of</strong>. Do linear <strong>algebra</strong> homework.<br />

6.2.3. Definition. Let R be a ring.<br />

(1) Let A = (aij) be a m × n-matrix. The (m − 1) × (n − 1) matrix derived<br />

from A be deleting i-row and j-column is Aij.<br />

(2) For a square matrix A the determinant is defined by row expansion and induction:<br />

det(a11) = a11<br />

det A = <br />

i<br />

(−1) 1+j a1j det A1j<br />

(3) The determinant <strong>of</strong> a k × k-matrix derived from A by choosing entries from<br />

k rows and columns is a k-minor <strong>of</strong> A.<br />

(4) If A is a n × n-matrix, then the c<strong>of</strong>actor matrix A ′ = (a ′ ij ) has entries<br />

given by (n − 1)-minors.<br />

a ′ ij = (−1) i+j det Aji<br />

6.2.4. Proposition. (1) The determinant <strong>of</strong> the identity matrix is 1.<br />

(2) The determinant is calculated by any expansion<br />

det A = <br />

(−1) i+j aij det Aij = <br />

i<br />

(3) If A, B are n × n-matrices then the product rule h<strong>old</strong>s<br />

det AB = det A det B<br />

j<br />

(−1) i+j aij det Aij


76 6. FINITE MODULES<br />

(4) Let A be an n × n-matrix with c<strong>of</strong>actor matrix A ′ . Then the matrix Cramer’s<br />

rule h<strong>old</strong>s<br />

AA ′ = A ′ A = det A(1n)<br />

Where (1n) is the n × n identity matrix.<br />

(5) A square matrix A is invertible if and only if det A is a unit in R.<br />

Pro<strong>of</strong>. More linear <strong>algebra</strong> homework.<br />

6.2.5. Proposition. Let f : R n → R n be a homomorphism represented by an<br />

n × n-matrix A.<br />

(1) f is a surjective if and only if det A is a unit.<br />

(2) f is injective if and only if det A is a nonzero divisor.<br />

Pro<strong>of</strong>. (1) If f is surjective, then a section is represented by a matrix B such that<br />

AB = (1n). Then det A is a unit. Conversely by 6.2.4. (2) If det A is a nonzero<br />

divisor, then by 6.2.4 A ′ A = det A(1n) gives an injective homomorphism, so f is<br />

injective. If det A is a zero divisor, then there is number k < n such that Ann(1 −<br />

minors) = · · · = Ann(k − minors) = 0 and 0 = b ∈ Ann((k + 1) − minors).<br />

Assume that the k−minor from first k rows and columns ck+1 has ck+1b = 0. Let<br />

cj(−1) k+1+j be the k-minor from first k rows and first k + 1 columns excluding<br />

number j and put cj = 0, j > k + 1. Then A(cj) is a column with entries being<br />

(k + 1) − minors so f((bcj)) = A(bcj) = 0 and f is not injective.<br />

6.2.6. Proposition. Let R be a ring and f : R n → R m homomorphism.<br />

(1) If f is surjective, then n ≥ m.<br />

(2) If f is injective, then n ≤ m.<br />

Pro<strong>of</strong>. (1) If n < m let p : R m → R n be the projection onto first n coordinates.<br />

Then f ◦p is surjective and represented by an m×m-matrix A with m-column zero.<br />

A section to f ◦ p is represented by an m × m-matrix B such that BA = (1m). But<br />

the product AB must have a zero m-column, so the contradiction gives n ≥ m.<br />

(2) If n > m let i : R m → R n be injection onto first m coordinates. Then i ◦ f is<br />

injective and represented by an n × n-matrix A with n-row zero. Then det A = 0<br />

contradicting 6.2.5. So n ≤ m.<br />

6.2.7. Proposition. (1) A finite free module has a finite basis.<br />

(2) The number <strong>of</strong> elements in a basis for a finite free module is independent <strong>of</strong><br />

the basis.<br />

Pro<strong>of</strong>. Let F be finite free generated by n elements. If y1, . . . , ym is part <strong>of</strong> a<br />

basis, then by projection F → R m there is a surjective homomorphism R n → R m .<br />

Conclusion by 6.2.6.<br />

6.2.8. Definition. The number <strong>of</strong> elements 6.2.7 in a basis for a finite free module<br />

F is the rank, rankR F .<br />

6.2.9. Proposition. If x1, . . . , xn generates a free module F <strong>of</strong> rank n, then they<br />

constitutes a basis.<br />

Pro<strong>of</strong>. Choose a basis and an isomorphism f : R n → F . The homomorphism<br />

g : R n → F, ei ↦→ xi is surjective. The composite f −1 ◦ g : R n → R n is<br />

surjective and therefore by 6.2.5 an isomorphism. Then g is an isomorphism and<br />

xi a basis.


6.3. CAYLEY-HAMILTON’S THEOREM 77<br />

6.2.10. Proposition. Let F, F ′ be finite free modules. Then<br />

(1) F ⊕ F ′ is free and rankR F ⊕ F ′ = rankR F + rankR F ′ .<br />

(2) F ⊗R F ′ is free and rankR F ⊗R F ′ = rankR F · rankR F ′ .<br />

(3) HomR(F, F ′ ) is free and rankR HomR(F, F ′ ) = rankR F · rankR F ′ .<br />

6.2.11. Exercise. (1) Let R n → R n be a surjective homomorphism. Show that it is an<br />

isomorphism.<br />

6.3. Cayley-Hamilton’s theorem<br />

6.3.1. Remark. Let R be a ring and f : M → M a homomorphism. By 2.1.13<br />

view M as an R[X]-module, where Xx = f(x) for x ∈ M. The homomorphism<br />

1.6.7, 2.6.9, R[X] → HomR(M, M), a ↦→ aM, X ↦→ f is a ring homomorphism.<br />

The image is R[f] the smallest subring containing 1M, f. M is naturally a R[f]module<br />

and the R[X]-module above is the restriction <strong>of</strong> scalars.<br />

6.3.2. Proposition. Let A be an n × n-matrix and I the ideal generated by the<br />

entries aij. The polynomial<br />

det(X1n − A) = a0 + a1X + . . . an−1X n−1 + X n<br />

has a0, . . . , an−1 ∈ I and gives the relation<br />

as n × n-matrix.<br />

a0(1n) + a1A + . . . an−1A n−1 + A n = 0<br />

Pro<strong>of</strong>. View R n as a module over the ring R[X], 6.3.1, with scalar multiplication<br />

Xx = Ax, x ∈ R n<br />

Let the n × n-matrix U = X(1n) − A over R[X] have c<strong>of</strong>actor matrix U ′ . The<br />

relations above give U ′ Uej = 0, written out by 6.2.4<br />

det U ej = 0<br />

for all j. That is det U M = 0. By calculation<br />

in R[X], with ai ∈ I.<br />

det U = a0 + a1X + · · · + an−1X n−1 + X n<br />

6.3.3. Proposition. Let I ⊂ R be an ideal and f : M → M a homomorphism<br />

on a finite module generated by n elements. Suppose Im f ⊂ IM, then there exist<br />

a0, . . . , an−1 ∈ I such that<br />

in HomR(M, M).<br />

a01M + a1f + . . . an−1f n−1 + f n = 0<br />

Pro<strong>of</strong>. Let x1, . . . , xn generate M and write<br />

f(xj) = <br />

i<br />

aijxi<br />

for an n × n-matrix A with entries aij ∈ I. View this over the ring R[X], 6.3.1.<br />

Then<br />

Xxj = <br />

i<br />

aijxi


78 6. FINITE MODULES<br />

Let the n × n-matrix U = X(1n) − A over R[X] have c<strong>of</strong>actor matrix U ′ . The<br />

relations above give U ′ Uxj = 0, written out by 6.2.4<br />

det U xj = 0<br />

for all j. That is det U M = 0. By calculation<br />

in R[X], with ai ∈ I.<br />

det U = a0 + a1X + · · · + an−1X n−1 + X n<br />

6.3.4. Proposition. Let I ⊂ R be an ideal and M a finite module. If IM = M<br />

then I + Ann(M) = R. That is there is a ∈ I such that (1 + a)M = 0.<br />

Pro<strong>of</strong>. By 6.3.3<br />

Put a = a0 + · · · + an−1.<br />

a01M + · · · + 1M = (a0 + · · · + an−1)1M + 1M = 0<br />

6.3.5. Corollary. Let I ⊂ R be an ideal and M a finite module. If IM = M and<br />

all elements 1 + I are nonzero divisors on M, then M = 0.<br />

6.3.6. Corollary. Let I ⊂ R be an ideal and N ⊂ M a submodule. Suppose M/N<br />

is a finite module and M = N + IM. Then I + (N : M) = R.<br />

6.3.7. Proposition. Let R be a ring and M a finite module. If a homomorphism<br />

f : M → M is surjective, then it is an isomorphism.<br />

Pro<strong>of</strong>. Regard M, f as a module over R[X] 6.3.1. Then (X)M = M, so by 6.3.4<br />

there is p ∈ R[X] such that 1 + pX ∈ Ann(M). For any u ∈ Ker f, calculate<br />

u = u + p(f) ◦ f(u) = (1 + pX)u = 0. So f is an isomorphism.<br />

6.3.8. Exercise. (1) Let √ 0M = M. Show that M = 0.<br />

6.4. Nakayama’s Lemma<br />

6.4.1. Proposition. Let (R, P ) be a local ring and M a finite module. The following<br />

conditions are equivalent.<br />

(1) M = 0.<br />

(2) P M = M.<br />

(3) M ⊗R k(P ) = 0.<br />

Pro<strong>of</strong>. (1) ⇒ (2) ⇔ (3) is clear. (2) ⇒ (1): Elements in 1 + P are units in R, so<br />

by 6.3.5 M = 0.<br />

6.4.2. Corollary. Let (R, P ) be a local ring and N ⊂ M a submodule. Suppose<br />

M/N is a finite module and M = N + P M. Then N = M.<br />

6.4.3. Corollary. Let (R, P ) be a local ring and M, N finite modules. If M ⊗R<br />

N = 0, then M = 0 or N = 0.<br />

Pro<strong>of</strong>. If M, N = 0 then M ⊗R k(P ), N ⊗R k(P ) = 0 are vector spaces over<br />

k(P ). Now M ⊗R N ⊗R k(P ) M ⊗R k(P )⊗ k(P ), N ⊗R k(P ) = 0, giving the<br />

statement.<br />

6.4.4. Corollary. Let R be a ring and M a finite module. The following conditions<br />

are equivalent.<br />

(1) M = 0.


(2) P MP = MP for all prime ideals P .<br />

(3) P MP = MP for all maximal ideals P .<br />

(4) M ⊗R k(P ) = 0 for all prime ideals P .<br />

(5) M ⊗R k(P ) = 0 for all maximal ideals P .<br />

Pro<strong>of</strong>. Combine 6.4.1 with 5.4.1.<br />

6.4. NAKAYAMA’S LEMMA 79<br />

6.4.5. Corollary. Let (R, P ) be a local ring and f : M → N a homomorphism.<br />

Assume N is finite. The following conditions are equivalent.<br />

(1) f is surjective.<br />

(2) f(P ) is surjective.<br />

Pro<strong>of</strong>. f is surjective if and only if Cok f = 0. Cok f is finite, so it is zero if and<br />

only if Cok f ⊗R k(P ) = Cok(f(P )) = 0, 6.4.1.<br />

6.4.6. Corollary. Let R be a ring and f : M → N a homomorphism. Assume N<br />

is finite. The following conditions are equivalent.<br />

(1) f is surjective.<br />

(2) f(P ) is surjective for all prime ideals P .<br />

(3) f(P ) is surjective for all maximal ideals P .<br />

6.4.7. Corollary. Let (R, P ) be a local ring and M a finite module. Let x1, . . . , xn ∈<br />

M. The following conditions are equivalent.<br />

(1) x1, . . . , xn generate M.<br />

(2) x1, . . . , xn generate M ⊗R k(P ).<br />

6.4.8. Corollary. Let R be a ring and M a finite module. Let x1, . . . , xn ∈ M.<br />

The following conditions are equivalent.<br />

(1) x1, . . . , xn generate M.<br />

(2) x1, . . . , xn generate M ⊗R k(P ) for all prime ideals P .<br />

(3) x1, . . . , xn generate M ⊗R k(P ) for all maximal ideals P .<br />

6.4.9. Proposition. Let f : R n → R m be a homomorphism represented by an<br />

m × n-matrix A. The following are equivalent.<br />

(1) f is surjective.<br />

(2) n ≥ m and the ideal (m − minors) = R.<br />

Pro<strong>of</strong>. (1) ⇒ (2): If f is surjective, then for any maximal ideal f(P ) is surjective<br />

linear map. So n ≥ m and some m-minor is nonzero in k(P ). Therefore (m −<br />

minors) is not contained in P , so (m − minors) = R. (2) ⇒ (1): f(P ) is<br />

surjective for all maximal ideals, so f surjective by 6.4.6.<br />

6.4.10. Proposition. Let f : R n → R m be a homomorphism represented by an<br />

m × n-matrix A. The following are equivalent.<br />

(1) f is injective.<br />

(2) n ≤ m and the ideal Ann(n − minors) = 0.<br />

Pro<strong>of</strong>. See the pro<strong>of</strong> 6.2.5 (2).<br />

6.4.11. Exercise. (1) Let R be a domain and f : R n → R m a homomorphism represented<br />

by an m × n-matrix. Show that f is injective if and only if n ≤ m and some<br />

n − minor = 0.


80 6. FINITE MODULES<br />

6.5. Finite Presented Modules<br />

6.5.1. Definition. Let R be a ring. A finite presented module is a module M<br />

having an exact sequence<br />

R n → R m → M → 0<br />

or equivalently there is a short exact sequence<br />

with N finite.<br />

0 → N → R m → M → 0<br />

6.5.2. Example. A finite projective module is finite presented, 6.1.10.<br />

6.5.3. Lemma. For a short exact sequence<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

the following h<strong>old</strong>:<br />

(1) If M, L are finite presented, then N is finite presented.<br />

(2) If L is finite presented and N is finite , then M is finite.<br />

(3) If N is finite presented and M is finite , then L is finite presented.<br />

Pro<strong>of</strong>. (1) Choose u : R n → M → 0 and v : R m → L → 0 exact with finite<br />

kernels. By 3.5.4 there is w : R m → N such that g ◦ w = v. There is a diagram<br />

0<br />

0<br />

<br />

Rn <br />

u<br />

<br />

<br />

M<br />

f<br />

<br />

L<br />

Rn ⊕ Rm <br />

f◦u+w<br />

<br />

g<br />

<br />

N<br />

<br />

0<br />

Rm <br />

v<br />

<br />

<br />

L<br />

By the snake lemma 3.2.4 the sequence 0 → Ker u → Ker f ◦u+w → Ker v → 0<br />

is exact. By 6.1.5 Ker f ◦ u + w is finite. (2) Choose v : R m → L → 0 exact with<br />

finite kernel and w : R m → N such that g ◦ w = v. There is a diagram<br />

0<br />

0<br />

<br />

0<br />

<br />

f<br />

<br />

M<br />

<br />

R m<br />

w<br />

<br />

<br />

N<br />

g<br />

Rm <br />

v<br />

<br />

<br />

L<br />

By the snake lemma 3.2.4 the sequence 0 → Ker w → Ker v → M → Cok w →<br />

0 is exact. By 6.1.5 M is finite. (3) Choose w : R m → N → 0 exact with finite<br />

kernel. Then v = g ◦ w : R m → L → 0 is exact and there is a diagram<br />

0<br />

0<br />

<br />

0<br />

<br />

f<br />

<br />

M<br />

<br />

R m<br />

w<br />

<br />

<br />

N<br />

g<br />

Rm <br />

v<br />

<br />

<br />

L<br />

By the snake lemma 3.2.4 the sequence 0 → Ker w → Ker v → M → 0 is exact.<br />

By 6.1.5 Ker v is finite and L is finite presented.<br />

6.5.4. Corollary. Let<br />

0<br />

f<br />

<br />

M<br />

g<br />

<br />

N<br />

be a split exact sequence. Then N is finite presented if and only if M, L are finite<br />

presented.<br />

<br />

L<br />

<br />

0<br />

0<br />

<br />

0<br />

0<br />

<br />

0<br />

0<br />

<br />

0


Pro<strong>of</strong>. By 3.1.13 there is a split exact sequence<br />

so the statement follows from 6.5.3.<br />

0<br />

6.5. FINITE PRESENTED MODULES 81<br />

<br />

v<br />

L <br />

u<br />

N <br />

M<br />

6.5.5. Corollary. Let f : M → N be a homomorphism.<br />

(1) If M is finite and Im f finite presented, then Ker f is finite.<br />

(2) If Ker f, Im f are finite presented, then M is finite presented.<br />

(3) If N is finite presented and Im f finite, then Cok f is finite presented.<br />

(4) If Im f, Cok f are finite presented, then N is finite presented.<br />

Pro<strong>of</strong>. Use the sequences 3.1.8.<br />

6.5.6. Proposition. Let R be a ring and M, N finite presented modules.<br />

(1) M ⊕ N is finite presented.<br />

(2) M ⊗R N is finite presented.<br />

Pro<strong>of</strong>. (1) This is clear from 6.5.4. (2) If M = R n then M ⊗R N is finite presented<br />

by (1). In general chose u : R n → M → 0 exact with finite kernel. The sequence<br />

Ker u ⊗R N → R n ⊗R N → M ⊗R N → 0 is exact. So Ker u ⊗ 1N is finite.<br />

Conclusion by 6.5.5.<br />

6.5.7. Proposition. Given submodules N, L ⊂ M. Then<br />

(1) If M/N, M/L are finite and M/N + L is finite presented, then M/N ∩ L is<br />

finite.<br />

(2) If M/N, M/L are finite presented and M/N ∩ L is finite, then M/N + L is<br />

finite presented.<br />

Pro<strong>of</strong>. Use the sequence 3.2.7.<br />

6.5.8. Proposition. Let R be a ring and U a multiplicative subset.<br />

(1) For a finite module M and any module N the natural homomorphism<br />

<br />

0<br />

U −1 HomR(M, N) → Hom U −1 R(U −1 M, U −1 N)<br />

is injective.<br />

(2) For a finite presented module M and any module N the homomorphism<br />

U −1 HomR(M, N) → Hom U −1 R(U −1 M, U −1 N)<br />

is a natural isomorphism.<br />

Pro<strong>of</strong>. (1) If M = R this is an isomorphism. Then this is also an isomorphism<br />

for M = R n since both functors respect finite direct sums. In general choose<br />

0 → K → R n → M → 0 exact. There is a diagram<br />

0<br />

0<br />

<br />

U −1 HomR(M, N)<br />

<br />

<br />

HomU −1R(U −1M, U −1N) <br />

U −1 HomR(R n , N)<br />

<br />

<br />

HomU −1R(U −1Rn , U −1N) giving injectivity. (2) In (1) K is finite, so the last vertical map is injective. Conclusion<br />

by the five lemma 3.2.8.<br />

<br />

U −1 HomR(K, N)<br />

<br />

<br />

HomU −1R(U −1K, U −1N)


82 6. FINITE MODULES<br />

6.5.9. Corollary. Let P ⊂ R be a prime ideal and M a finite presented module.<br />

For any module N there is a natural isomorphism<br />

HomR(M, N)P HomRP (MP , NP )<br />

6.5.10. Proposition. Let R be a ring and F a module. The following conditions<br />

are equivalent.<br />

(1) F is flat.<br />

(2) For any module N and a relation 0 = <br />

i yi ⊗ xi ∈ N ⊗R F , there exist<br />

zj ∈ F and aij ∈ R such that 0 = <br />

i aijyi ∈ N and xi = <br />

j aijzj ∈ F .<br />

(3) For any relation 0 = <br />

i bixi ∈ F , there exist zj ∈ F and aij ∈ R such that<br />

0 = <br />

i aijbi ∈ R and xi = <br />

j aijzj ∈ F .<br />

Pro<strong>of</strong>. (1) ⇒ (2): Let f : Rn → N, ei ↦→ yj, then 0 → Ker f ⊗R F → F n →<br />

N ⊗R F is exact. By assumption (xi) ∈ Ker f ⊗R F , so (xi) = <br />

j aij ⊗ zj<br />

<br />

with (aij) ∈ Ker f. (2) ⇒ (3) is clear. (3) ⇒ (1): Let I ⊂ R be an ideal and<br />

bi ⊗ xi ∈ Ker(I ⊗R F → F ). Then 0 = <br />

i aijbi and xi = <br />

j aijzj. Now<br />

calculate bi ⊗ xi = <br />

j i aijbi ⊗ xi = 0. By 3.7.12 F is flat.<br />

6.5.11. Proposition. Let (R, P ) be a local ring and F a finite presented module.<br />

The following conditions are equivalent.<br />

(1) F is free.<br />

(2) F is projective.<br />

(3) F is flat.<br />

(4) P ⊗R F → F is injective.<br />

Pro<strong>of</strong>. (1) ⇒ (2) ⇒ (3) ⇒ (4) are clear. (4) ⇒ (1): Choose xi ∈ F such that<br />

xi ⊗ 1 give a basis for F ⊗R k(P ). The homomorphism f : R n → F, ei ↦→ xi is<br />

surjective by 6.4.5. P ⊗R Ker f → P n → P ⊗R F → 0 is exact, so P Ker f =<br />

Ker(P n → P ⊗R F ) = Ker(P n → P ⊗R F → F ) = Ker f. by the hypothesis.<br />

Ker f is finite 6.5.5 and therefore Ker f = 0 by 6.4.1, so F is free.<br />

6.5.12. Corollary. Let (R, P ) be a local ring and f : F → F ′ a homomorphism<br />

<strong>of</strong> finite free modules. The following conditions are equivalent.<br />

(1) f has a retraction u : F ′ → F .<br />

(2) f is injective and Cok f is free.<br />

(3) f(P ) is injective.<br />

Pro<strong>of</strong>. (1) ⇒ (2): Cok f is projective, so free by 6.5.11. (2) ⇒ (3) is clear. (3) ⇒<br />

(1): Let f ∨ : F ′∨ → F ∨ be the dual homomorphism. f ∨ (P ) is surjective, so f ∨ is<br />

surjective, 6.4.5. A section v <strong>of</strong> f ∨ gives a retraction u = v ∨ .<br />

6.5.13. Corollary. Let R be a ring and F a finite presented module. The following<br />

conditions are equivalent.<br />

(1) F is projective.<br />

(2) F is flat.<br />

(3) F is locally free.<br />

(4) FP is free for all maximal ideals P .<br />

Pro<strong>of</strong>. (1) ⇒ (2) ⇒ (3) ⇒ (4) are clear by 6.5.11. (4) ⇒ (1): Let N → L → 0<br />

be exact. By hypothesis HomRP (FP , NP ) → HomRP (FP , LP ) → 0 is exact for<br />

all maximal ideals. By 6.5.9 HomR(F, N)P → HomR(F, L)P → 0 is exact for


6.6. FINITE RING HOMOMORPHISMS 83<br />

all maximal ideals. By 5.4.3 HomR(F, N) → HomR(F, L) → 0 is exact and F is<br />

projective.<br />

6.5.14. Exercise. (1) Let I ⊂ R be an ideal. Show that R/I is a finite presented<br />

R-module if and only if I is a finite ideal.<br />

(2) Show that Q is a flat, but not projective Z-module.<br />

6.6. Finite ring homomorphisms<br />

6.6.1. Definition. A ring homomorphism φ : R → S is a finite ring homomorphism<br />

if S is a finite R-module. If R ⊂ S is a subring, then a finite ring homomorphism<br />

is a finite ring extension.<br />

6.6.2. Proposition. Let R be a ring.<br />

(1) Let f ∈ R[X] be a monic polynomial. Then the homomorphism R →<br />

R[X]/(f) is finite.<br />

(2) Let f : M → M be a homomorphism <strong>of</strong> a finite R-module. Then the homomorphism<br />

6.3.1, R → R[f] is finite.<br />

Pro<strong>of</strong>. (2) Follows from (1) and 6.3.3.<br />

6.6.3. Lemma. Let φ : R → S be a finite ring homomorphism. If N is a finite<br />

S-module, then by restriction along φ the R-module N is finite.<br />

6.6.4. Proposition. Let R ⊂ S be a finite ring extension <strong>of</strong> domains. Then R is a<br />

field if and only if S is a field.<br />

Pro<strong>of</strong>. Let R be a field, a minimal equation 6.3.3 for scalar multiplication by a<br />

nonzero b ∈ S, bS as R-module homomorphism<br />

gives<br />

b n + · · · + a0 = 0<br />

b −1 = −a −1<br />

0 (an−1b n−2 + · · · + a1) ∈ S<br />

Let S be a field and 0 = a ∈ R. An equation 6.3.3 for scalar multiplication a −1<br />

S as<br />

R-homomorphism<br />

a −n + · · · + a0 = 0<br />

gives<br />

a −1 = −(a0a n−1 + · · · + an−1) ∈ R<br />

6.6.5. Corollary. Let R → S be a finite ring homomorphism. A prime ideal Q ⊂ S<br />

is maximal if and only if the contraction Q ∩ R is maximal.<br />

6.6.6. Proposition (going-up). Let R ⊂ S be a finite ring extension and P ⊂ R a<br />

prime ideal. Then there is a prime ideal Q ⊂ S contracting<br />

.<br />

P = Q ∩ R<br />

Pro<strong>of</strong>. RP ⊂ SP is a finite ring extension. Since SP = 0 there is a maximal<br />

ideal in SP contracting to P RP by 6.6.5. The corresponding prime ideal Q ⊂ S<br />

contracts to P .<br />

6.6.7. Proposition. Let R ⊂ S be a finite ring extension and E an R-module. The<br />

following are equivalent.


84 6. FINITE MODULES<br />

(1) E is an injective R-module.<br />

(2) HomR(S, E) is an injective S-module.<br />

Pro<strong>of</strong>. (1) ⇒ (2): This is 3.6.10. (2) ⇒ (1): Let E ⊂ E ′ be an injective envelope,<br />

3.6.15. HomR(S, E) → HomR(S, E ′ ) is injective and identifies HomR(S, E)<br />

as an essential submodule, since S is a finite R-module. E ′ is injective, so also<br />

HomR(S, E ′ ). It follows that<br />

HomR(S, E)<br />

<br />

HomR(R, E)<br />

E<br />

HomR(S, E ′ )<br />

<br />

<br />

HomR(R, E ′ )<br />

<br />

E ′<br />

with the right down map being surjective. So E = E ′ .<br />

6.6.8. Exercise. (1) Show that Z → Z[ √ −5] is finite.<br />

(2) Let p be a prime number. Show that Z → Z (p) is not finite.


7<br />

Modules <strong>of</strong> finite length<br />

7.1. Simple Modules<br />

7.1.1. Definition. A nonzero module M is a simple module if 0 and M are the<br />

only submodules.<br />

7.1.2. Proposition. Let f : M → M ′ be a homomorphism<br />

(1) If M is simple, then f is either zero or injective.<br />

(2) If M ′ is simple, then f is either zero or surjective.<br />

(3) If M, M ′ are both simple, then f is either zero or an isomorphism.<br />

Pro<strong>of</strong>. This follows from 2.3.3.<br />

7.1.3. Proposition. A simple module is isomorphic to a factor module R/P for<br />

some uniquely determined maximal ideal P . A simple module is finite, in fact<br />

generated by one element.<br />

Pro<strong>of</strong>. A nonzero f : R → M must be surjective and R/P is simple exactly when<br />

P is maximal. Clearly P = Ann(M).<br />

7.1.4. Corollary. Any nonzero finite module has a simple factor module.<br />

Pro<strong>of</strong>. If M = 0 then MP = 0 for some maximal ideal. By 6.4.1 M ⊗R k(P ) =<br />

M/P M = 0. Choose a nonzero linear projection M/P M → k(P ), giving M →<br />

k(P ) → 0 exact.<br />

7.1.5. Example. (1) If K is a field, a one dimensional vector space is simple.<br />

(2) If R is a principal ideal domain, then R/(p) is a simple module for all irreducible<br />

(p).<br />

(3) Z/(p) are simple for all prime numbers p.<br />

(4) Let K be a field. K[X]/(X − a) is a simple K[X]-module.<br />

7.1.6. Lemma. Let I ⊂ R be an ideal. A simple R/I-module is a simple Rmodule.<br />

7.1.7. Exercise. (1) Show that Z does not contain any simple modules.<br />

(2) Show that Q does not have any simple factor Z-module.<br />

(3) Let L, L ′ ⊂ M be simple submodules. Show that either L ∩ L ′ = 0 <strong>of</strong> L = L ′ .<br />

(4) Let L = L ′ ⊂ M be simple submodules. Show that L ⊕ L ′ = L + L ′ .<br />

7.2. The Length<br />

7.2.1. Definition. Let R be a ring. A module M is <strong>of</strong> finite length if it admits a<br />

composition series by submodules<br />

0 = M0 ⊂ M1 ⊂ · · · ⊂ Mn−1 ⊂ Mn = M<br />

such that each factor Mi/Mi−1 is a simple module.<br />

85


86 7. MODULES OF FINITE LENGTH<br />

7.2.2. Lemma. Any two finite composition series have the same number <strong>of</strong> submodules.<br />

Pro<strong>of</strong>. Let l(M) be the least length <strong>of</strong> a composition series. If M is simple then<br />

l(M) = 1. Let 0 = M0 ⊂ M1 ⊂ · · · ⊂ Mn = M be any composition series. For a<br />

submodule N ⊂ M there is a filtration 0 = N ∩M0 ⊂ N ∩M1 ⊂ · · · ⊂ N ∩Mn =<br />

N with factors N ∩Mi/N ∩Mi−1 ⊂ Mi/Mi−1 being either simple or 0. It follows<br />

that l(N) ≤ l(M). If l(N) = l(M) then each factor is nonzero and therefore<br />

N = M. Applying this to the composition series <strong>of</strong> M gives l(M) ≤ n ≤ l(M)<br />

as wanted.<br />

7.2.3. Definition. The number <strong>of</strong> nontrivial submodules in a filtration as above<br />

will be denoted ℓR(M) and is the length <strong>of</strong> M.<br />

7.2.4. Proposition. Given an exact sequence<br />

0 → N → M → L → 0<br />

Then M has a finite length if and only if both N and L have finite length. In that<br />

case<br />

ℓR(M) = ℓR(N) + ℓR(L)<br />

Pro<strong>of</strong>. A finite filtration with simple quotients in M induces filtrations in both N<br />

and L, and vice versa.<br />

The filtrations in N and L give a filtration in M with the same quotients, so the<br />

length formula follows from 7.2.2.<br />

7.2.5. Corollary. Let f : M → N be a homomorphism.<br />

(1) If M has finite length if and only if Ker f, Im f have finite length. If finite<br />

length<br />

ℓR(M) = ℓR(Ker f) + ℓR(Im f)<br />

(2) If N has finite length if and only if Im f, Cok f have finite length. If finite<br />

length<br />

ℓR(N) = ℓR(Im f) + ℓR(Cok f)<br />

Pro<strong>of</strong>. Use the sequences 3.1.8.<br />

7.2.6. Corollary. Let f : M → M be a homomorphism on a module <strong>of</strong> finite<br />

length. Then<br />

ℓR(Ker f) = ℓR(Cok f)<br />

and the following are equivalent:<br />

(1) f is injective.<br />

(2) f is surjective.<br />

(3) f is an isomorphism.<br />

Pro<strong>of</strong>. Use the sequences 3.1.8.<br />

7.2.7. Corollary. Let N ⊂ M be a submodule and suppose M has finite length.<br />

(1) ℓR(N) ≤ ℓR(M).<br />

(2) ℓR(N) = ℓR(M) if and only if N = M.<br />

7.2.8. Corollary. Let N, L ⊂ M be a submodules and suppose N, L has finite<br />

length. Then N + L, N ∩ L have finite length and<br />

ℓR(N + L) + ℓR(N ∩ L) = ℓR(N) + ℓR(L)


7.3. ARTINIAN RINGS 87<br />

7.2.9. Proposition. Given submodules N, L ⊂ M. Then the following are equivalent.<br />

(1) M/N, M/L have finite length.<br />

(2) M/N ∩ L has finite length.<br />

If finite length<br />

ℓR(M/N + L) + ℓR(M/N ∩ L) = ℓR(M/N) + ℓR(M/L)<br />

Pro<strong>of</strong>. Use the sequences 3.2.8.<br />

7.2.10. Proposition. A R-module M <strong>of</strong> finite length is finite and generated by<br />

ℓR(M) or less elements.<br />

Pro<strong>of</strong>. There is a sequence 0 → N → M → L → 0 with L simple. Conclusion<br />

by induction.<br />

7.2.11. Proposition. Let I ⊂ R be an ideal. Suppose an R/I-module M has finite<br />

length. Then M has finite length as R-module and<br />

ℓ R/I(M) = ℓR(M)<br />

Pro<strong>of</strong>. This follows from 7.1.6 and 7.2.2.<br />

7.2.12. Example. Let K be a field. A module M is <strong>of</strong> finite length if it is a finite<br />

dimensional vector space. Then<br />

7.2.13. Exercise. (1) Compute<br />

(2) Compute<br />

ℓZ(Z/(p n1<br />

1<br />

ℓK(M) = rankK M<br />

. . . pnk<br />

k )) = n1 + · · · + nk<br />

ℓ K[X](K[X]/((X − a1) n1 . . . (X − ak) nk )) = n1 + · · · + nk<br />

7.3. Artinian Rings<br />

7.3.1. Lemma. Let M be a module. The following conditions are equivalent.<br />

(1) Any decreasing sequence Mi <strong>of</strong> submodules is stationary, Mn = Mn+1 for<br />

n >> 0.<br />

(2) Any nonempty subset <strong>of</strong> submodules <strong>of</strong> M contains a minimal element.<br />

Pro<strong>of</strong>. (1) ⇒ (2): Suppose a nonempty subset <strong>of</strong> submodules do not contain a<br />

minimal element. Then choose a non stationary sequence. (2) ⇒ (1): A descending<br />

sequence containing a minimal element is stationary.<br />

7.3.2. Definition. A module M which satisfies the conditions <strong>of</strong> 7.3.1 is an artinian<br />

module.<br />

7.3.3. Definition. A ring R is an artinian ring if R is an artinian module over<br />

itself.<br />

7.3.4. Proposition. Let 0 → N → M → L → 0 be an exact sequence <strong>of</strong> modules<br />

over the ring R. Then M is artinian if and only if N and L are artinian.


88 7. MODULES OF FINITE LENGTH<br />

Pro<strong>of</strong>. Let M be artinian. A chain in N gives a chain in M, so N is artinian. A<br />

chain in L gives a chain in M, which becomes stationary, so also the original chain,<br />

so L is artinian. Conversely if N and L are artinian and Mi a chain in M, then<br />

the induced chains in N and L become stationary. By the snake lemma 3.2.4 the<br />

original chain is stationary.<br />

7.3.5. Corollary. Let f : M → N be a homomorphism.<br />

(1) M is artinian if and only if Ker f, Im f are artinian.<br />

(2) N is artinian if and only if Im f, Cok f are artinian<br />

Pro<strong>of</strong>. Use the sequences 3.1.8.<br />

7.3.6. Proposition. Let f : M → M be a homomorphism on an artinian module.<br />

Then the following are equivalent<br />

(1) f is injective<br />

(2) f is an isomorphism<br />

Pro<strong>of</strong>. There is a number n such that Im f ◦n = Im f ◦2n . For x ∈ M there is y<br />

such that f ◦n (x) = f ◦2n (y). Then f ◦n (x − f ◦n (y)) = 0 so x = f ◦n (y) since f is<br />

injective. It follows that f ◦n is surjective and so is f.<br />

7.3.7. Proposition. Given submodules N, L ⊂ M. Then the following are equivalent.<br />

(1) M/N, M/L are artinian<br />

(2) M/N ∩ L is artinian<br />

Pro<strong>of</strong>. use the sequences 3.2.8.<br />

7.3.8. Corollary. Given ideals I, J ⊂ R. Then the following are equivalent.<br />

(1) R/I, R/J are artinian<br />

(2) R/I ∩ J is artinian<br />

7.3.9. Corollary. Let R be an artinian ring and I an ideal. Then the factor ring<br />

R/I is artinian.<br />

The product <strong>of</strong> two artinian rings is artinian.<br />

7.3.10. Proposition. An artinian domain is a field.<br />

Pro<strong>of</strong>. By 7.3.6 scalar multiplication with a nonzero element is an isomorphism.<br />

7.3.11. Proposition. Let R be an artinian ring. Then all prime ideals are maximal<br />

and there are only finitely many such.<br />

Pro<strong>of</strong>. By 7.3.10 primes are maximal. If Pi are different maximal ideals, then<br />

P1 · · · Pn+1 ⊂ P1 · · · Pn is a strictly decreasing chain. So there are only finitely<br />

many maximal ideals.<br />

7.3.12. Proposition. Let R be an artinian ring.<br />

(1) The factor ring R/ √ 0 is a finite product <strong>of</strong> fields.<br />

(2) The nilradical √ 0 is nilpotent, √ 0 k = 0 for some k.


7.3. ARTINIAN RINGS 89<br />

Pro<strong>of</strong>. (1) Let P1, . . . , Pn be the prime and maximal ideals 7.3.11. The nilradical<br />

√ 0 = P1 ∩ · · · ∩ Pn by 5.1.7. Conclude by Chinese remainders 1.4.3. (2) √ 0 k =<br />

√ 0 k+1 for some k. If √ 0 k = 0 let (a) be minimal among ideals I such that<br />

I √ 0 k = 0. By minimality (a) = (a) √ 0 k , so a = ab for some b ∈ √ 0 k . But b is<br />

nilpotent 1.3.8, so a = 0 gives a contradiction. It follows, that √ 0 k = 0.<br />

7.3.13. Proposition. A ring R is artinian if and only if it has finite length.<br />

Pro<strong>of</strong>. The factor module √ 0 i / √ 0 i+1 is an artinian module over R/ √ 0 which is<br />

a product <strong>of</strong> fields, so it has finite length.<br />

7.3.14. Corollary. Let R be an artinian ring and M a module. The following are<br />

equivalent.<br />

(1) M is finite.<br />

(2) M has finite length.<br />

(3) M is finite presented.<br />

7.3.15. Corollary. Let R be artinian and R → S a finite ring homomorphism.<br />

(1) S is artinian.<br />

(2) A finite length S-module N is by restriction <strong>of</strong> scalars a finite length Rmodule.<br />

(3) A finite length R-module M gives by change <strong>of</strong> rings M ⊗R S as finite length<br />

S-module.<br />

7.3.16. Proposition. Let M be a R-module <strong>of</strong> finite length.<br />

(1) The ring R/ Ann(M) is artinian.<br />

(2) There are only finitely many prime ideals Ann(M) ⊂ P .<br />

(3) Any prime ideal Ann(M) ⊂ P is maximal.<br />

(4) M is finite presented.<br />

Pro<strong>of</strong>. (1) Let x ∈ M, then R/ Ann(x) Rx is artinian. If x1, . . . , xn generate<br />

M, then Ann(M) = Ann(x1)∩· · ·∩Ann(Xn). By 7.3.8 R/ Ann(M) is artinian.<br />

7.3.17. Example. Let K be a field. The ring R = <br />

artinian.<br />

(1) The maximal ideals in R are<br />

(2) The simple types are<br />

N<br />

Pi = {a : N → K|a(i) = 0}<br />

R/Pi Ji = {a : N → K|a(j) = 0, i = j}<br />

(3) Ji are the only simple ideals and the sum is an ideal<br />

<br />

Ji = J = R<br />

(4) J is an ideal which has no complement in R.<br />

i<br />

K = {a : N → K} is not<br />

7.3.18. Exercise. (1) Show that a vector space is artinian if and only if it is finite dimensional.<br />

(2) Show that Z is not artinian.<br />

(3) Show that R[X] is not artinian for a nonzero ring R.<br />

(4) Show that Q[X]/(X 2 − X) is artinian.


90 7. MODULES OF FINITE LENGTH<br />

(5) Show that a ring with finitely many ideals is artinian.<br />

(6) Let R be a principal ideal domain and a = 0. Show that R/(a) is artinian.<br />

7.4. Localization<br />

7.4.1. Proposition. Any artinian ring is the product <strong>of</strong> finitely many artinian local<br />

rings. Let R be artinian with maximal ideals P1, . . . , Pk. Then there is n1, . . . , nk<br />

such that P n1<br />

1 . . . P nk<br />

k = 0 and<br />

R R/P n1<br />

nk<br />

1 × · · · × R/Pk Each R/P ni<br />

1 is a local artinian ring.<br />

Pro<strong>of</strong>. This follows from 7.3.11, 7.3.12 and Chinese remainders 1.4.2.<br />

7.4.2. Example. A reduced artinian ring is a finite product <strong>of</strong> fields.<br />

7.4.3. Corollary. Let R be an artinian ring and U ⊂ R a multiplicative subset.<br />

The ring <strong>of</strong> fractions U −1 R is an artinian ring.<br />

7.4.4. Example. Let P ⊂ R be a maximal ideal and M a module. Assume s /∈ P<br />

and n ∈ N.<br />

(1) Rs + P n = A.<br />

(2) Scalar multiplication by s : M/P n M → M/P n M, x → sx is an isomorphism.<br />

(3) The canonical map M/P n M → (M/P n M)P is an isomorphism.<br />

(4) If M is finite then M/P n M has finite length.<br />

7.4.5. Proposition. Let R be artinian with maximal ideals P1, . . . , Pk. A finite<br />

module M has a decomposition<br />

M <br />

The length is<br />

Pro<strong>of</strong>. Let R R/P n1<br />

ℓR(M) = <br />

R/P nk<br />

k and MPi M ⊗R R/P ni<br />

i .<br />

1<br />

i<br />

i<br />

MPi<br />

ℓRP i (MPi )<br />

nk<br />

× · · · × R/Pk . Then M M ⊗R R/P n1<br />

1 × · · · × M ⊗R<br />

7.4.6. Proposition. Let M be an R-module <strong>of</strong> finite length and Ann(M) ⊂ P1, . . . , Pk<br />

the maximal ideals. Then MPi is a finite length RPi-module and<br />

ℓR(M) = <br />

ℓRP (MPi i )<br />

Pro<strong>of</strong>. The ring R/ Ann(M) R/P n1<br />

1<br />

i<br />

× · · · × R/P nk<br />

k<br />

by 7.3.16 and 7.4.1.<br />

7.4.7. Proposition. Let (R, P ) → (S, Q) be a local ring homomorphism and assume<br />

that k(P ) → k(Q) is a finite extension. If N is a finite length S-module, then<br />

N is a finite length R-module and<br />

Pro<strong>of</strong>. Reduce to N = k(Q).<br />

ℓR(N) = rank k(P )(k(Q)) · ℓS(N)<br />

7.4.8. Proposition. Let (R, P ) → (S, Q) be a local ring homomorphism and assume<br />

that S/P S is a finite length S-module. Let M be a finite length R-module.


7.5. LOCAL ARTINIAN RING 91<br />

(1) M ⊗R S is a finite length S-module.<br />

(2) In general<br />

ℓS(M ⊗R S) ≤ ℓS(S/P S) · ℓR(M)<br />

(3) If R → S is flat then<br />

ℓS(M ⊗R S) = ℓS(S/P S) · ℓR(M)<br />

Pro<strong>of</strong>. The case M = k(P ) is clear. Conclude by induction.<br />

7.4.9. Exercise. (1) Let K ⊂ L be a finite field extension and W a finite vector space<br />

over L. Show that<br />

rankK(W ) = rankK(L) · rankL(W )<br />

(2) Let K ⊂ L be a finite field extension and V a finite vector space over K. Show that<br />

rankL(V ⊗K L) = rankK(V )<br />

7.5. Local artinian ring<br />

7.5.1. Lemma. Let (R, P ) be a local artinian ring and M a finite module. The<br />

following are equivalent.<br />

(1) M = 0.<br />

(2) HomR(k(P ), M) = 0.<br />

(3) HomR(M, k(P )) = 0.<br />

Pro<strong>of</strong>. Clear by a filtration.<br />

7.5.2. Proposition. Let (R, P ) be a local artinian ring and k(P ) ⊂ E an injective<br />

envelope.<br />

(1) The are isomorphisms<br />

k(P ) HomR(k(P ), k(P )) HomR(k(P ), E)<br />

(2) For a finite module M, the module HomR(M, E) has finite length and<br />

ℓR(HomR(M, E)) = ℓR(M)<br />

Pro<strong>of</strong>. (1) A nonzero homomorphism f : k(P ) → E has Im f = k(P ) since the<br />

extension is essential. (2) This follows from (1) by use <strong>of</strong> a filtration.<br />

7.5.3. Corollary. Let (R, P ) be a local artinian ring and k(P ) ⊂ E an injective<br />

envelope. Then E has finite length<br />

ℓR(E) = ℓR(R)<br />

7.5.4. Proposition. Let (R, P ) be a local artinian ring and k(P ) ⊂ E an injective<br />

envelope. There is a natural isomorphism for any finite module M<br />

x ↦→ evx : M → HomR(HomR(M, E), E)<br />

Pro<strong>of</strong>. The case M = k(P ) is clear by 7.5.2. Let 0 → N → M → L → 0 be a<br />

short exact sequence. Then the following diagram has exact rows.<br />

0<br />

0<br />

<br />

N<br />

<br />

<br />

HomR(HomR(N, E), E)<br />

<br />

M<br />

<br />

<br />

HomR(HomR(M, E), E)<br />

Conclusion by the five lemma 3.2.8 and induction on the length.<br />

<br />

L<br />

<br />

<br />

HomR(HomR(L, E), E)<br />

<br />

0<br />

<br />

0


92 7. MODULES OF FINITE LENGTH<br />

7.5.5. Corollary. Let (R, P ) be a local artinian ring and k(P ) ⊂ E an injective<br />

envelope. There is an isomorphism<br />

11E : R → HomR(E, E)<br />

7.5.6. Proposition. Let (R, P ) be a local artinian ring and k(P ) ⊂ E an injective<br />

envelope. Let M → E ′ be an injective envelope <strong>of</strong> a finite module. Then<br />

(1) HomR(k(P ), M) HomR(k(P ), E ′ )<br />

(2) E ′ E n , where n = ℓR(HomR(k(P ), M).<br />

(3) ℓR(E ′ ) = ℓR(HomR(k(P ), M) ℓR(E).<br />

Pro<strong>of</strong>. (1) A homomorphism f : k(P ) → E ′ has Im f ⊂ M since the extension<br />

is essential. (2) By induction on length <strong>of</strong> M follows that E ′ E n , where n is<br />

determined by (1) and 7.5.2. (3) This follows from (2).<br />

7.5.7. Proposition. Let (R, P ) be a local artinian ring and k(P ) ⊂ E an injective<br />

envelope. The following are equivalent.<br />

(1) R is injective.<br />

(2) R E.<br />

(3) HomR(k(P ), R) HomR(k(P ), k(P )) k(P ).<br />

Pro<strong>of</strong>. (1) ⇒ (2): By 7.5.6 R E n and by 7.5.3 n = 1. (2) ⇒ (3): Immediate<br />

from 7.5.6. (3) ⇒ (1): Let R ⊂ E ′ be an injective envelope. By 7.5.6 R = E ′ .<br />

7.5.8. Definition. A ring satisfying the conditions 7.5.7 is a local artinian Gorenstein<br />

ring.<br />

7.5.9. Example. Let R be a principal ideal domain and p ∈ R an irreducible<br />

element. Then R/(p n ) is a local artinian Gorenstein ring.<br />

7.5.10. Exercise. (1) Let p be a prime number. Show that Z/(p k ) is a local artinian<br />

Gorenstein ring.


8<br />

Noetherian modules<br />

8.1. Modules and submodules<br />

8.1.1. Lemma. Let M be a module. The following conditions are equivalent.<br />

(1) Any increasing sequence Mn <strong>of</strong> submodules is stationary, Mn = Mn+1 for<br />

n >> 0.<br />

(2) Any nonempty subset <strong>of</strong> submodules <strong>of</strong> M contains a maximal element.<br />

(3) Any submodule <strong>of</strong> M is finite.<br />

Pro<strong>of</strong>. (1) ⇔ (2): See the pro<strong>of</strong> <strong>of</strong> 7.3.1. (2) ⇒ (3): Let N be a submodule <strong>of</strong><br />

M and choose a maximal element N ′ in the set <strong>of</strong> finite submodules <strong>of</strong> N. For<br />

y ∈ N, the module N ′ + Ry is finite, so N ′ = N ′ + Ry gives N = N ′ finite.<br />

(3) ⇒ (1): The union ∪Mn is a submodule, generated by x1, . . . , xm ∈ Mk, so<br />

Mn = ∪Mn, n > k.<br />

8.1.2. Definition. A module M which satisfies the conditions <strong>of</strong> 8.1.1 is a noetherian<br />

module.<br />

8.1.3. Proposition. Let 0 → N → M → L → 0 be an exact sequence <strong>of</strong> modules<br />

over the ring R. Then M is noetherian if and only if N and L are noetherian.<br />

Pro<strong>of</strong>. See the pro<strong>of</strong> <strong>of</strong> 7.3.4.<br />

8.1.4. Corollary. Let f : M → N be a homomorphism.<br />

(1) M is noetherian if and only if Ker f, Im f are noetherian.<br />

(2) N is noetherian if and only if Im f, Cok f are noetherian.<br />

Pro<strong>of</strong>. Use the sequences 3.1.8.<br />

8.1.5. Proposition. Let f : M → M be a homomorphism on a noetherian module.<br />

Then the following are equivalent<br />

(1) f is surjective<br />

(2) f is an isomorphism<br />

Pro<strong>of</strong>. Analog <strong>of</strong> the pro<strong>of</strong> <strong>of</strong> 7.3.4. There is a number n such that Ker f ◦2n =<br />

Ker f ◦n . For x ∈ Ker f ◦n there is y such that x = f ◦n (y). Then f ◦2n (y) =<br />

f ◦n (x) = 0, so y ∈ Ker f ◦2n = Ker f ◦n . Then x = f ◦n (y) = 0 and f is<br />

injective.<br />

8.1.6. Proposition. A finite direct sum <strong>of</strong> noetherian modules is noetherian.<br />

8.1.7. Proposition. Given submodules N, L ⊂ M. Then the following are equivalent.<br />

(1) M/N, M/L are noetherian<br />

(2) M/N ∩ L is noetherian<br />

93


94 8. NOETHERIAN MODULES<br />

Pro<strong>of</strong>. Use the sequence 3.2.7.<br />

8.1.8. Example. A module <strong>of</strong> finite length is noetherian.<br />

8.1.9. Exercise. (1) Show that <br />

N Z is not noetherian.<br />

(2) Show that <br />

N Z is not noetherian.<br />

(3) Show that Q is not a noetherian Z-module.<br />

8.2. Noetherian rings<br />

8.2.1. Definition. A ring R is a noetherian ring if R is a noetherian module.<br />

8.2.2. Proposition. The following conditions are equivalent.<br />

(1) R is noetherian.<br />

(2) Any ideal is finite.<br />

(3) Any increasing sequence <strong>of</strong> ideals is stationary.<br />

(4) Any nonempty subset <strong>of</strong> ideals <strong>of</strong> contains an ideal maximal for inclusion.<br />

Pro<strong>of</strong>. This is follows from 8.1.1.<br />

8.2.3. Proposition. The nilradical in a noetherian ring is nilpotent, i.e. √ 0 n = 0<br />

for some n. Some power <strong>of</strong> the radical <strong>of</strong> an ideal is contained in the ideal, i.e.<br />

√ I n ⊂ I for some n<br />

Pro<strong>of</strong>. √ 0 = (b1, . . . , bn) such that b k i = 0. Then ( aibi) nk = 0.<br />

8.2.4. Proposition. Given ideals I, J ⊂ R. Then the following are equivalent.<br />

(1) R/I, R/J are noetherian<br />

(2) R/I ∩ J is noetherian<br />

Pro<strong>of</strong>. This is a special case <strong>of</strong> 8.1.7.<br />

8.2.5. Lemma. (1) Any factor ring <strong>of</strong> a noetherian ring is a noetherian ring.<br />

(2) A principal ideal domain is noetherian.<br />

(3) If M is a noetherian R-module, then R/ Ann(M) is a noetherian ring.<br />

Pro<strong>of</strong>. (1), (2) are clear. (3) Analog <strong>of</strong> the pro<strong>of</strong> <strong>of</strong> 7.3.16. Let x ∈ M, then<br />

R/ Ann(x) Rx is noetherian. If x1, . . . , xn generate M, then Ann(M) =<br />

Ann(x1) ∩ · · · ∩ Ann(Xn). By 8.2.4 R/ Ann(M) is artinian.<br />

8.2.6. Proposition. Let R be a noetherian ring. Any finite R-module is noetherian.<br />

Pro<strong>of</strong>. Let M be a finite R-module. There is a surjective homomorphism R n →<br />

M → 0, 6.1.2. Conclusion by 8.2.3.<br />

8.2.7. Corollary. Let R be a noetherian ring. Any finite R-module is finite presented.<br />

8.2.8. Lemma. Let R be a noetherian ring and M, N noetherian modules.<br />

(1) M ⊗R N is noetherian.<br />

(2) HomR(M, N) is noetherian.<br />

Pro<strong>of</strong>. Conclusion by 6.1.8, 6.1.9, 8.2.6.


8.3. FINITE TYPE RINGS 95<br />

8.2.9. Proposition. Let R be a noetherian ring and U a multiplicative subset. For<br />

a finite module M and any module N the homomorphism<br />

is an isomorphism.<br />

U −1 HomR(M, N) → Hom U −1 R(U −1 M, U −1 N)<br />

Pro<strong>of</strong>. Conclusion by 6.5.8, 8.2.7.<br />

8.2.10. Proposition. Let R be a noetherian ring. If E is an injective R-module,<br />

then U −1 E is an injective U −1 R-module.<br />

Pro<strong>of</strong>. Let I ⊂ R be an ideal. Then HomR(R, E) → HomR(I, E) → 0 is exact.<br />

By 8.2.9 Hom U −1 R(U −1 R, U −1 E) → Hom U −1 R(U −1 I, U −1 E) → 0 is exact.<br />

So U −1 E is injective 3.6.7, using that any ideal is extended 4.3.6.<br />

8.2.11. Proposition. For a ring R, the following are equivalent.<br />

(1) R is a noetherian ring.<br />

(2) For any family Eα <strong>of</strong> injective modules, the sum <br />

α Eα is injective.<br />

Pro<strong>of</strong>. (1) ⇒ (2): Let I ⊂ R be an ideal. A homomorphism f : I → Eα<br />

has Im f contained in a finite sum, which is injective 3.6.6. So f extends to R →<br />

Eα, giving injectivity. (2) ⇒ (1): Let In ⊂ R be an increasing chain <strong>of</strong> ideals.<br />

Put I = ∪In and choose an injective envelope I/In ⊂ En. The homomorphism<br />

f : I → En extends to f ′ : R → En. Since Im f is contained in a finite<br />

sum, I/In = 0 for n >> 0.<br />

8.2.12. Proposition. If R is a ring such that every prime ideal is finite, then it is<br />

noetherian.<br />

Pro<strong>of</strong>. If R is not noetherian, then the set <strong>of</strong> not finite ideals is nonempty and by<br />

Zorn’s lemma it has a maximal element I. Since I is not prime there is a, b /∈ I<br />

and ab ∈ I. The ideals I + (a) and I : (a) are both greater that I and therefore<br />

finite. If I + (a) = (a1, . . . , am, a), ai ∈ I and I : (a) = (b1, . . . , bn), then<br />

I = (a1, . . . , am, ab1, . . . , abn) is finite. It follows that R must be noetherian.<br />

8.2.13. Exercise. (1) Show that a principal ideal domain is noetherian.<br />

(2) Let I ⊂ R be an ideal in a noetherian ring. Show that R/I is noetherian.<br />

(3) Let K be a field and R = K[X1, X2, . . . ] the polynomial ring in countable many<br />

variables. Show that R is not noetherian.<br />

(4) Let K be a field. Show that the ring R = <br />

N K is not noetherian.<br />

8.3. Finite type rings<br />

8.3.1. Proposition (Hilbert’s basis theorem). Let R be a noetherian ring. Then the<br />

ring <strong>of</strong> polynomials R[X] is noetherian.<br />

Pro<strong>of</strong>. Assume I ⊂ R[X] to be a not finite ideal. Choose a sequence f1, f2, · · · ∈<br />

I such that<br />

fi = aiX di + terms <strong>of</strong> lower degree , ai = 0<br />

and fi+1 has lowest degree in I − (f1, . . . , fi). The ideal <strong>of</strong> leading coefficients is<br />

finitely generated by a1, . . . , an. Then an+1 = b1a1 + · · · + bnan and d1 ≤ · · · ≤<br />

dn+1 = d gives<br />

fn+1 − b1X d−d1 f1 − · · · − bnX d−dn fn<br />

in I − (f1 . . . , fn) <strong>of</strong> degree less than d. By contradiction the ideal I is finite.


96 8. NOETHERIAN MODULES<br />

8.3.2. Corollary. Let R be a noetherian ring and R → S a finite type ring over R.<br />

Then S is noetherian.<br />

8.3.3. Corollary. Let R be a noetherian ring and M a finite module. Then the ring<br />

R ⊕ M, 2.1.14, is noetherian.<br />

8.3.4. Example. Let I ⊂ R be an ideal in a noetherian ring. Then there are noetherian<br />

rings.<br />

(1) GIR = ⊕n≥0I n /I n+1 .<br />

(2) BIR = ⊕n≥0I n .<br />

8.3.5. Theorem (Krull’s intersection theorem). Let I be an ideal in a noetherian<br />

ring R. Then there is a ∈ I such that<br />

1 + a ∈ Ann( <br />

I n )<br />

Pro<strong>of</strong>. Let I = (u1, . . . , um). If b ∈ I n then b = fn(u1, . . . , um) where fn ∈<br />

R[X1, . . . , Xm] are homogeneous <strong>of</strong> degree n. By Hilbert’s basis theorem, 8.3.1<br />

there is N such that<br />

fN+1 = f1g1 + · · · + fNgN<br />

where gn is homogeneous <strong>of</strong> degree N − n + 1 > 0. By substitution b ∈ bI and<br />

I ∩ I n = ∩I n . Conclusion by 6.3.4.<br />

8.3.6. Corollary. Let I be an ideal in a noetherian ring R such that the elements<br />

1 + a, a ∈ I are nonzero divisors. Then<br />

<br />

I n = 0<br />

n<br />

8.3.7. Corollary. Let I be an ideal in a noetherian ring R and M a finite module.<br />

Then there is a ∈ I such that<br />

1 + a ∈ Ann( <br />

I n M)<br />

Pro<strong>of</strong>. Use 8.3.5 on the ring R ⊕ M and the ideal I ⊕ M. Clearly (I + M) n =<br />

I n + I n−1 M.<br />

8.3.8. Corollary. Let I be an ideal in a noetherian ring R and M a finite module<br />

such that the elements 1 + a, a ∈ I are nonzero divisors on M. Then<br />

<br />

I n M = 0<br />

n<br />

8.3.9. Proposition. If R ⊂ S be a finite extension. Then R is noetherian if and<br />

only if S is noetherian.<br />

Pro<strong>of</strong>. Assume S is noetherian. Let Eα be a family <strong>of</strong> injective R-modules. Then<br />

HomR(S, Eα) is an injective S-module. Since S is finite over R, HomR(S, Eα) <br />

HomR(S, Eα) 6.1.14. By 6.6.7 Eα is injective and by 8.2.11 R is noetherian.<br />

8.3.10. Example. Let K be a field and R = K[X1, X2, . . . ]/(X1 − x2X3, x2 −<br />

x3X4, . . . ). Put P = (X1, X2, . . . ).<br />

(1) P is maximal<br />

(2) P 2 = P .<br />

(3) P (∩nP n ) = ∩nP n .<br />

n<br />

n


8.4. POWER SERIES RINGS 97<br />

8.3.11. Exercise. (1) Show that if R[X] is noetherian, then R is noetherian.<br />

(2) Show that the subring Z[2X, 2X 2 , . . . ] ⊂ Z[X] is not noetherian. Conclude that the<br />

extension is not finite.<br />

8.4. Power series rings<br />

8.4.1. Proposition. Let R be a noetherian ring. Then the power series ring R[[X]]<br />

is noetherian.<br />

Pro<strong>of</strong>. Let P ⊂ R[[X]] be a prime ideal. Then P +(X)/(X) = (a1, . . . , an) ⊂ R<br />

is a finite ideal. If X ∈ P then P = (a1, . . . , an, X) is finite. Suppose X /∈ P<br />

and choose fi = ai + terms <strong>of</strong> positive degree ∈ P . If g ∈ P then Xg1 =<br />

g − b11f1 + · · · + bn1fn ∈ P ∩ (X) for some bi1 ∈ R. Since P is prime, g1 ∈ P .<br />

Now Xg2 = g1 − b12f1 + · · · + bn2fn ∈ P ∩ (X) and so on. Put hi = bikX k ,<br />

then g = hifi. P is finite and R[[X]] is noetherian by 8.2.12.<br />

8.4.2. Proposition. Let R be a principal ideal domain. Then the power series ring<br />

R[[X]] is a unique factorization domain.<br />

Pro<strong>of</strong>. Let P ⊂ R[[X]] be a minimal nonzero prime ideal. Then P + (X)/(X) =<br />

(a) ⊂ R Suppose P = (X) and choose f = a + terms <strong>of</strong> positive degree ∈ P .<br />

If g ∈ P then Xg1 = g − b1f ∈ P ∩ (X) for some b1 ∈ R. Since P is prime,<br />

g1 ∈ P . Now Xg2 = g1 − b2f ∈ P ∩ (X) and so on. Put h = bkX k , then<br />

g = hf. P = (f) is principal. The conditions <strong>of</strong> 1.5.3 are satisfied since R[[X]] is<br />

noetherian 8.4.1.<br />

8.4.3. Proposition. Let I ⊂ R be an ideal in a noetherian ring. Then there is a<br />

canonical isomorphism<br />

R[[X]]/IR[[X]] R/I[[X]]<br />

Pro<strong>of</strong>. The projection R[[X]] → R[[X]]/IR[[X]] factors over R/I[[X]] since I is<br />

finitely generated. Then there is an inverse to the homomorphism 1.9.8.<br />

8.4.4. Corollary. If P ⊂ R is a prime ideal in a noetherian ring, then P R[[X]] ⊂<br />

R[[X]] is a prime ideal.<br />

8.4.5. Proposition. Let R be a noetherian ring.<br />

(1) The inclusion R[X] ⊂ R[[X]] is a flat homomorphism.<br />

(2) The inclusion R ⊂ R[[X]] is a faithfully flat homomorphism.<br />

Pro<strong>of</strong>. (1) Let I ⊂ R[X] be an ideal and let ai ⊗ fi ∈ K = Ker(I ⊗R[X] R[[X]] → R[[X]]. Then aifi = 0. Write fi = gi + Xnhi and get ai ⊗ fi =<br />

Xn (ai ⊗ hi). It follows that K ⊂ <br />

n Xn (I ⊗R[X] R[[X]]). I ⊗R[X] R[[X]]<br />

is a finite R[[X]]-module and 1 + aX is a unit, so conclusion by 8.3.8. (2) The<br />

homomorphism is flat by (1). For any maximal ideal P ⊂ R the homomorphism<br />

RP → RP [[X]] is local and flat.<br />

8.4.6. Exercise. (1) Show that if R[[X]] is noetherian, then R is noetherian.


98 8. NOETHERIAN MODULES<br />

8.5. Fractions and localization<br />

8.5.1. Proposition. Let R be a noetherian ring and U a multiplicative subset. Then<br />

U −1 R is a noetherian ring.<br />

Pro<strong>of</strong>. Any ideal is extended 4.3.6.<br />

8.5.2. Corollary. Let R be a noetherian ring and U a multiplicative subset. If M<br />

is a noetherian R-module, then U −1 M is a noetherian U −1 R-module.<br />

Pro<strong>of</strong>. By 6.1.3 change <strong>of</strong> ring <strong>of</strong> a finite module is finite.<br />

8.5.3. Corollary. Let R be a noetherian ring and M a finite module. Let P be a<br />

prime ideal. Then RP is a noetherian ring and MP is a finite RP -module.<br />

8.5.4. Proposition (Krull’s intersection theorem). Let (R, P ) be a noetherian local<br />

ring and M a finite module. Then<br />

<br />

P n M = 0<br />

Pro<strong>of</strong>. This follows from 8.3.7.<br />

n<br />

8.5.5. Corollary. Let (R, P ) be a noetherian local ring and I ⊂ P an ideal. Then<br />

<br />

I n = 0<br />

n<br />

8.5.6. Proposition. Let (R, P ) be a noetherian local ring and F a finite module.<br />

The following conditions are equivalent.<br />

(1) F is free.<br />

(2) F is projective.<br />

(3) F is flat.<br />

(4) P ⊗R F → F is injective.<br />

Pro<strong>of</strong>. This follows from 6.5.13.<br />

8.5.7. Exercise. (1) Is it true that if U −1 R is noetherian, then R is noetherian?<br />

8.6. Prime filtrations <strong>of</strong> modules<br />

8.6.1. Proposition. Let R be a ring and M = 0 a nonzero module. An ideal<br />

Ann(x) maximal in the set <strong>of</strong> ideals {Ann(y)|0 = y ∈ M} is a prime ideal.<br />

Pro<strong>of</strong>. Let Ann(x) be a maximal annihilator. Suppose a, b ∈ R such that ab ∈<br />

Ann(x) and b /∈ Ann(x). Then<br />

Ann(x) ⊆ Ann(bx) = R<br />

Consequently Ann(x) = Ann(bx) in particular a ∈ Ann(x).<br />

8.6.2. Corollary. Let R be a noetherian ring and M = 0 a nonzero module. Then<br />

there is x ∈ M such that Ann(x) is a prime ideal.<br />

8.6.3. Theorem. Let R be a noetherian ring and M = 0 a finite R-module. Then<br />

there exists a finite filtration <strong>of</strong> M by submodules<br />

0 = M0 ⊂ M1 ⊂ · · · ⊂ Mr−1 ⊂ Mr = M<br />

such that Mi/Mi−1, i = 1, . . . , r is isomorphic to an R-module <strong>of</strong> the form R/Pi<br />

where Pi is a prime ideal in R.


8.6. PRIME FILTRATIONS OF MODULES 99<br />

Pro<strong>of</strong>. The set <strong>of</strong> submodules <strong>of</strong> M for which the theorem is true is nonempty by<br />

8.2.6. Let N ⊂ M be maximal in this set. Suppose N = M. By 8.2.6 applied to<br />

M/N there is a chain N ⊂ N ′ ⊂ M such that N ′ /N is isomorphic to an R-module<br />

<strong>of</strong> the form R/P ′ where P ′ is a prime ideal. This contradicts the maximality <strong>of</strong> N.<br />

So N = M.<br />

8.6.4. Corollary. Let R be a nonzero noetherian ring. Then there exists a finite<br />

filtration <strong>of</strong> ideals<br />

0 = I0 ⊂ I1 ⊂ · · · ⊂ Ir−1 ⊂ Ir = R<br />

such that Ii/Ii−1, i = 1, . . . , r is isomorphic to an R-module <strong>of</strong> the form R/Pi<br />

where Pi is a prime ideal in R.<br />

8.6.5. Example. In Z there is a filtration<br />

0 =⊂ (p n ) ⊂ (p n−1 ) ⊂ · · · ⊂ (p) ⊂ Z<br />

<strong>of</strong> any length with factors Z/(p) for any prime number p.<br />

8.6.6. Proposition. A ring R is artinian if and only if it is noetherian and all prime<br />

ideals are maximal.<br />

Pro<strong>of</strong>. By 7.3.13 an artinian ring has finite length and therefore noetherian. Primes<br />

are maximal by 7.3.11. Conversely by 8.6.4 there is a finite composition series.<br />

8.6.7. Corollary. Let R be a ring and M a module. The following are equivalent<br />

(1) M has finite length.<br />

(2) R/ Ann(M) is artinian and M is finite.<br />

8.6.8. Proposition. If R ⊂ S be a finite extension. Then R is artinian if and only<br />

if S is artinian.<br />

Pro<strong>of</strong>. Suppose S is artinian. By 8.3.9 R is noetherian. By 6.6.5 any prime ideal<br />

in R is maximal. Conclusion by 8.6.6.<br />

8.6.9. Proposition. Let R be a noetherian ring. The number <strong>of</strong> minimal prime<br />

ideals is finite.<br />

Pro<strong>of</strong>. Choose a filtration 8.6.4, 0 = I0 ⊂ · · · ⊂ Ir = R with Ii/Ii−1 R/Pi,<br />

where Pi is a prime ideal. Let P be a minimal prime ideal in R. Then (Ii/Ii−1)P <br />

(R/Pi)P = 0 if and only if Pi ⊂ P . Thus P = Pi for some i since RP = 0.<br />

8.6.10. Proposition. Let R be a noetherian ring such that the local rings RP are<br />

domains for all maximal ideals P . Then R is a finite product <strong>of</strong> domains.<br />

Pro<strong>of</strong>. Let P1, . . . , Pn be the minimal primes 8.6.9. The intersection is 0 5.4.9 and<br />

they are comaximal since a domain has a unique minimal prime. Conclusion by<br />

Chinese remainders 1.4.2.<br />

8.6.11. Exercise. (1) Compute a filtration 8.6.3 <strong>of</strong> the Z-module Z/(36).


9<br />

Primary decomposition<br />

9.1. Support <strong>of</strong> modules<br />

9.1.1. Definition. Let R be a ring and M a module.<br />

(1) The set <strong>of</strong> prime ideals is the spectrum and denoted Spec(R).<br />

(2) For a ring homomorphism φ : R → S restriction defines a map<br />

(3) For a subset B ⊂ R<br />

is a subset <strong>of</strong> the spectrum.<br />

(4) The support <strong>of</strong> M is<br />

φ ∗ : Spec(S) → Spec(R)<br />

Q ↦→ φ −1 (Q)<br />

V (B) = {P ∈ Spec(R)|MP = 0}<br />

Supp(M) = {P ∈ Spec(R)|MP = 0}<br />

(5) A minimal prime ideal in Supp(M) is a minimal prime <strong>of</strong> M.<br />

9.1.2. Proposition. Let R be a ring.<br />

(1) Let I ⊂ R be an ideal. Then<br />

and is identified with Spec(R/I).<br />

(2) Let U be a multiplicative subset.<br />

Supp(R/I) = V (I)<br />

Supp(U −1 R) = {P ∈ Spec(R)|P ∩ U = ∅}<br />

and is identified with Spec(U −1 R).<br />

Pro<strong>of</strong>. This is a restatement <strong>of</strong> 1.3.5, 5.1.5.<br />

9.1.3. Proposition. Let 0 → N → M → L → 0 be a short exact sequence <strong>of</strong><br />

modules. Then<br />

Supp(M) = Supp(N) ∪ Supp(L)<br />

Pro<strong>of</strong>. This follows from 5.4.5.<br />

9.1.4. Corollary. (1) Let N ⊂ M be a submodule. Then<br />

Supp(M) = Supp(N) ∪ Supp(M/N)<br />

(2) Given submodules N, L ⊂ M. Then<br />

Supp(M/N ∩ L) ∪ Supp(M/N + L) = Supp(M/N) ∪ Supp(M/L)<br />

and<br />

Supp(M/N + L) ⊂ Supp(M/N) ∩ Supp(M/L)<br />

Pro<strong>of</strong>. (2) Use the sequence 3.2.7.<br />

101


102 9. PRIMARY DECOMPOSITION<br />

9.1.5. Proposition. Let R be a ring and M a module.<br />

(1) M = 0 if and only if Supp(M) = ∅.<br />

(2) For any module<br />

Supp(M) ⊂ V (Ann(M))<br />

.<br />

(3) If M is finite, then<br />

Supp(M) = V (Ann(M))<br />

Pro<strong>of</strong>. (1) See 5.4.1. (2) If MP = 0 then for u /∈ P there is x ∈ M such that<br />

ux = 0. So Ann(M) ⊂ P . (3) Let x1, . . . , xn generate M. If Ann(M) =<br />

∩ Ann(xi) ⊂ P then some Ann(xi) ⊂ P , so xi<br />

1 = 0 in MP .<br />

9.1.6. Proposition. Let N1, . . . , Nk ⊂ M be submodules such that Supp(M/Ni)∩<br />

Supp(M/Nj) = ∅, i = j. Then the homomorphism<br />

M/ ∩i Ni → <br />

M/Ni<br />

is an isomorphism.<br />

Pro<strong>of</strong>. By induction on k it is enough to treat the case k = 2. By 9.1.4 the support<br />

<strong>of</strong> the cokernel is empty.<br />

9.1.7. Proposition. Let R be a ring and M, N modules.<br />

(1)<br />

Supp(M ⊗R N) ⊂ Supp(M) ∩ Supp(N)<br />

(2) If M, N are finite, then<br />

Supp(M ⊗R N) = Supp(M) ∩ Supp(N)<br />

Pro<strong>of</strong>. There is an isomorphism (M ⊗R N) MP ⊗RP NP . (1) This is clear. (2)<br />

This follows from 6.4.3.<br />

9.1.8. Corollary. Let φ : R → S be a ring homomorphism and M an R-module.<br />

(1) For the change <strong>of</strong> rings module<br />

(2) If M is finite, then<br />

Supp(M ⊗R S) ⊂ φ ∗−1 (Supp(M))<br />

Supp(M ⊗R S) = φ ∗−1 (Supp(M))<br />

9.1.9. Corollary. Let R be a ring, I an ideal in R and M a finite R-module. Then<br />

Supp(M/IM) = Supp(M) ∩ V (I)<br />

9.1.10. Corollary. Let R be a ring, U a multiplicative subset and M a finite Rmodule.<br />

Then<br />

Supp(U −1 M) = Supp(M) ∩ Spec(U −1 R)<br />

9.1.11. Proposition. Let (R, P ) → (S, Q) be a local homomorphism and M a<br />

finite R-module. If Supp(M) = ∅ then Supp(M ⊗R S) = ∅.<br />

Pro<strong>of</strong>. The homomorphism R → S is faithfully flat 5.5.8.<br />

9.1.12. Proposition. Let M be a finite R-module and P ∈ Supp(M). Then there<br />

is a nonzero homomorphism M → R/P .<br />

i


9.2. ASS OF MODULES 103<br />

Pro<strong>of</strong>. The module HomR(M, R/P )P HomRP (MP , k(P )) is nonzero 7.1.4.<br />

9.1.13. Proposition. Let R be a noetherian ring and M a finite module. M has<br />

finite length if and only if Supp(M) consists only <strong>of</strong> maximal ideals.<br />

Pro<strong>of</strong>. This is 8.6.7.<br />

9.1.14. Proposition. Let R be a noetherian ring and F a finite module. If R has<br />

only finitely many maximal ideals and FP is free <strong>of</strong> rank n, then F is free <strong>of</strong> rank<br />

n.<br />

Pro<strong>of</strong>. If R is artinian, this is clear from 7.4.5. If P1, . . . , Pk are the maximal<br />

ideals, then choose x1, . . . , xn ∈ F giving a basis for F/P1 · · · PnF . The homomorphism<br />

R n → F, ei ↦→ xi is an isomorphism 6.4.6 and 8.1.5.<br />

9.1.15. Exercise. (1)<br />

9.2. Ass <strong>of</strong> modules<br />

9.2.1. Definition. Let M be an R-module. A prime ideal P ⊂ R is an associated<br />

prime ideal <strong>of</strong> M if P = Ann(x) for some x ∈ M. The set <strong>of</strong> prime ideals<br />

associated to M is Ass(M).<br />

9.2.2. Proposition. Let P ⊂ R be a prime ideal.<br />

(1) P ∈ Ass(M) if and only if there is an injective homomorphism R/P → M.<br />

(2)<br />

Ass(M) ⊂ Supp(M)<br />

(3) Let 0 = N ⊂ R/P be a nonzero submodule, then<br />

Pro<strong>of</strong>. This is clear from the definition.<br />

Ass(N) = {P }<br />

9.2.3. Proposition. Let 0 → N → M → L → 0 be a short exact sequence <strong>of</strong><br />

modules. Then<br />

Ass(N) ⊂ Ass(M) ⊂ Ass(N) ∪ Ass(L)<br />

Pro<strong>of</strong>. The left inclusion is trivial. Next let P ∈ Ass(M) such that P /∈ Ass(N).<br />

Choose a submodule L <strong>of</strong> M such that L R/P . Then Ass(L ∩ N) ⊂ Ass(L) ∩<br />

Ass(N). It follows that L ∩ N = 0, and therefore M/N contains a submodule<br />

isomorphic to R/P .<br />

9.2.4. Corollary. Let 0 → N → M → L → 0 be a split exact sequence <strong>of</strong><br />

modules.<br />

Ass(M) = Ass(N) ∪ Ass(L)<br />

9.2.5. Corollary. (1) Let N ⊂ M be a submodule. Then<br />

(2) Let M, N be modules. Then<br />

Ass(N) ⊂ Ass(M) ⊂ Ass(N) ∪ Ass(M/N)<br />

Ass(M ⊕ N) = Ass(M) ∪ Ass(N)<br />

(3) Given submodules N, L ⊂ M. Then<br />

Ass(M/N ∩ L) ⊂ Ass(M/N) ∪ Ass(M/L) ⊂ Ass(M/N ∩ L) ∪ Ass(M/N + L)<br />

Pro<strong>of</strong>. (3) Use the sequence 3.2.7.


104 9. PRIMARY DECOMPOSITION<br />

9.2.6. Proposition. Let M be a module and P ⊂ Ass(M). Then there is a submodule<br />

N ⊂ M such that<br />

Ass(N) = P , Ass(M/N) = Ass(M)\P<br />

Pro<strong>of</strong>. Choose by Zorn’s lemma N maximal in the set <strong>of</strong> submodules N ′ ⊂ M<br />

for which Ass(N ′ ) ⊂ P. Let Q ∈ Ass(M/N) and choose N ⊂ L ⊂ M with<br />

L/N R/Q. Then Ass(L) ⊂ P ∪ {Q}. By maximality <strong>of</strong> N follows that<br />

Ass(L) ⊂ P. So Q /∈ P and Q ∈ Ass(L) ⊂ Ass(M).<br />

9.2.7. Proposition. Let R be a noetherian ring and M a module. Then M = 0 if<br />

and only if Ass(M) = ∅.<br />

Pro<strong>of</strong>. If M = 0, then by 8.6.2 Ass(M) = ∅.<br />

9.2.8. Corollary. Let R be a noetherian ring and f : M → N a homomorphism.<br />

The following conditions are equivalent.<br />

(1) f is injective.<br />

(2) fP is injective for all prime ideals P ∈ Ass(M).<br />

Pro<strong>of</strong>. Use 9.2.7 on Ker f.<br />

9.2.9. Corollary. Let R be a noetherian ring and M a module. Then a ∈ R is a<br />

nonzero divisor on M if and only if<br />

a /∈ ∪ P ∈Ass(M)P<br />

The set <strong>of</strong> zero divisors on M is ∪ P ∈Ass(M)P .<br />

Pro<strong>of</strong>. aM is injective if and only if aMP<br />

This happens when a /∈ ∪ P ∈Ass(M)P .<br />

is injective for all P ∈ Ass(M), 8.2.8.<br />

9.2.10. Proposition. Let U ⊂ R be a multiplicative subset and M a module.<br />

(1) In general<br />

(2) If R is noetherian<br />

Ass(M) ∩ Spec(U −1 R) ⊂ Ass U −1 R(U −1 M)<br />

Ass(M) ∩ Spec(U −1 R) = Ass U −1 R(U −1 M)<br />

Pro<strong>of</strong>. Let P be a prime ideal. For any homomorphism R/P → M there is a<br />

commutative diagram<br />

R/P <br />

M<br />

<br />

U −1R/P <br />

<br />

U −1M (1) If P ∈ Ass(M) and P ∩ U = ∅ then U −1 P ∈ Ass(U −1 M). (2) If U −1 P ∈<br />

Ass(U −1 M), then P ∩ U = ∅. By 8.2.9 there is a diagram as above. R/P → M<br />

is injective by 9.2.8. So P ∈ Ass(M).<br />

9.2.11. Proposition. Let R be a noetherian ring and M a module. Then any minimal<br />

prime P ∈ Supp(M) is contained in Ass(M).<br />

Pro<strong>of</strong>. Assume P ∈ Supp(M) is minimal. Then the RP -module MP has support<br />

exactly in the maximal ideal, so {P RP } = Ass(MP ). Conclusion by 9.2.10.


9.2. ASS OF MODULES 105<br />

9.2.12. Definition. A non minimal prime ideal in Ass(M) is an embedded prime<br />

<strong>of</strong> M.<br />

9.2.13. Proposition. Let R be a noetherian ring and M a finite module. Then<br />

Ass(M) is a finite set.<br />

Pro<strong>of</strong>. Follows immediately from 9.2.4 and 8.1.5.<br />

9.2.14. Corollary. Let R be a noetherian ring and M a finite module. The following<br />

are equivalent<br />

(1) M has finite length.<br />

(2) Supp(M) consists <strong>of</strong> maximal ideals.<br />

(3) Ass(M) consists <strong>of</strong> maximal ideals.<br />

Pro<strong>of</strong>. (3) ⇒ (2): The minimal ideals in the support are maximal.<br />

9.2.15. Lemma. Let (R, P ) be a local ring M a module. Then P ∈ Ass(M) if<br />

and only if HomR(k(P ), M) = 0.<br />

9.2.16. Proposition. Let R be a noetherian ring and M a finite module. For any<br />

module N<br />

Ass(HomR(M, N)) = Supp(M) ∩ Ass(N)<br />

Pro<strong>of</strong>. By 8.2.9 HomR(M, N)P HomRP (MP , NP ). So reduce to the case<br />

where (R,P) is local. Now<br />

HomR(k(P ), HomR(M, N)) = HomR(M, HomR(k(P ), N))<br />

Conclusion by Nakayama’s lemma 6.4.1 and 9.2.15.<br />

= Hom k(P )(M ⊗R k(P ), HomR(k(P ), N))<br />

9.2.17. Proposition. Let R be a noetherian ring and F a finite module. Assume<br />

rank F ⊗R k(P ) = n for all primes P . Then F is locally free (projective) if and<br />

only if FP is free for all P ∈ Ass(R).<br />

Pro<strong>of</strong>. Let Q be a maximal ideal and 0 → K → R n Q → FQ → 0 exact. Ass(K) ⊂<br />

Ass(R), so KP = 0 for all P ∈ Ass(K). By 9.2.7 K = 0.<br />

9.2.18. Proposition. Let (R, P ) be a noetherian local ring. If there is a nonzero<br />

finite injective module E, then R is artinian.<br />

Pro<strong>of</strong>. Let Q be a prime and f : R/Q → E. If a ∈ P \Q then a R/Q is injective,<br />

so there is f ′ : R/Q → E such that f = f ′ ◦ a R/Q. That is P HomR(R/Q, E) =<br />

HomR(R/Q, E), so by Nakayama’s lemma 6.4.1 HomR(R/Q, E) = 0 if Q = P .<br />

By 9.2.7 0 = HomR(R/P, E) ⊂ HomR(R/Q, E), so Q = P . R is artinian by<br />

8.6.6.<br />

9.2.19. Exercise. (1) Show that<br />

(2) Show that<br />

Ass(Z/(n)) = {(p)|p prime dividing n}<br />

Ass(K[X, Y ]/(X) ∩ (X 2 , Y 2 )) = {(X), (X, Y )}<br />

and point out an embedded prime.<br />

(3) Let I ⊂ R be an ideal such that √ I = I. Show that R/I has no embedded prime<br />

ideals.


106 9. PRIMARY DECOMPOSITION<br />

(4) Let I, J ⊂ R be a ideals such that JRP ⊂ IRP for all P ∈ Ass(R/I). Show that<br />

J ⊂ I.<br />

9.3. Primary modules<br />

9.3.1. Definition. A submodule N ⊂ M is a primary submodule or more precisely<br />

P -primary if Ass(M/N) = {P }.<br />

9.3.2. Proposition. A prime ideal P ⊂ R is a P -primary submodule.<br />

Pro<strong>of</strong>. This is 9.2.2.<br />

9.3.3. Proposition. Let R be a noetherian ring and M a finite module. For a<br />

submodule N ⊂ M the following are equivalent.<br />

(1) N ⊂ M is primary for some prime<br />

(2) The set <strong>of</strong> zero divisors on M/N is contained in the radical Ann(M/N).<br />

Pro<strong>of</strong>. (1) ⇒ (2): Ann(M/N) = P the set <strong>of</strong> zero divisors by 9.2.9. (2) ⇒ (1):<br />

If P1, P2 ∈ Ass(M/N) then P1 ∪ P2 ⊂ Ann(M/N) ⊂ P1 ∩ P2, so P1 = P2.<br />

9.3.4. Corollary. Let R be a noetherian ring and I ⊂ R a proper ideal. The<br />

following are equivalent.<br />

(1) I ⊂ R is primary for some prime.<br />

(2) Any zero divisor in R/I is nilpotent.<br />

9.3.5. Corollary. Let R be a noetherian ring.<br />

(1) If an ideal I ⊂ R is P -primary then √ I = P .<br />

(2) If the radical √ I = P is a maximal ideal, then I ⊂ R is a P -primary<br />

submodule.<br />

(3) A finite power P n ⊂ R <strong>of</strong> a maximal ideal is a P -primary submodule.<br />

9.3.6. Proposition. Let R be a noetherian ring and M a finite module.<br />

(1) If N ⊂ M is P -primary, then Ann(M/N) ⊂ R is P -primary.<br />

(2) If N, N ′ ⊂ M are P -primary, then N ∩ N ′ is P -primary.<br />

Pro<strong>of</strong>. (1) This follows from 9.3.3. (2) This follows from 9.2.5.<br />

9.3.7. Proposition. Let R be a noetherian ring and M a finite module. Suppose<br />

N ⊂ M is P -primary and U ⊂ R is multiplicative subset.<br />

(1) If U ∩ P = ∅, then U −1 N ⊂ U −1 M is P U −1 R-primary.<br />

(2) If U ∩ P = ∅, then U −1 N = U −1 M.<br />

Pro<strong>of</strong>. This follows from 9.2.10.<br />

9.3.8. Exercise. (1) Let K be a field. Show that (X 2 , Y ) ⊂ K[X, Y ] is (X, Y )primary.<br />

(2) Let p be a prime number. Show that (p k ) ⊂ Z is a primary ideal.<br />

9.4. Decomposition <strong>of</strong> modules<br />

9.4.1. Definition. A submodule L ⊂ M has a primary decomposition if there<br />

exist a family Ni ⊂ M <strong>of</strong> Pi-primary submodules, such that<br />

L = N1 ∩ · · · ∩ Nn<br />

A primary decomposition is a reduced primary decomposition if Pi = Pj for i = j<br />

and no Ni can be excluded.


9.4. DECOMPOSITION OF MODULES 107<br />

9.4.2. Lemma. Let R be a noetherian ring and M a finite module. For each Pi ∈<br />

Ass(M) there is a submodule Ni ⊆ M such that Ass(Ni) = Ass(M) − {Pi} and<br />

Ass(M/Ni)) = {Pi}. M injects<br />

0 → M → <br />

M/Ni<br />

Pro<strong>of</strong>. The submodule Ni is given by 9.2.6. Ass(∩Ni) = ∅, so conclusion by<br />

9.2.2.<br />

9.4.3. Proposition. Let R be a noetherian ring and M a finite module. A proper<br />

submodule L ⊂ M has a reduced primary decomposition<br />

and for any such<br />

and<br />

is exact.<br />

Pro<strong>of</strong>. Apply 9.4.2 to M/L.<br />

L = N1 ∩ · · · ∩ Nn<br />

Ass(M/L) = {P1, . . . , Pn}<br />

0 → M/L → <br />

M/Ni<br />

9.4.4. Proposition. Let R be a noetherian ring and M a finite module. If<br />

L = N1 ∩ · · · ∩ Nn<br />

is a reduced primary decomposition <strong>of</strong> L ⊂ M and Pi is minimal in Ass(M/L),<br />

then<br />

and therefore uniquely determined.<br />

Ni = M ∩ LPi<br />

Pro<strong>of</strong>. Clearly Ni ⊂ M ∩ LPi . By localization Ass(M ∩ LPi /Ni) = ∅. So<br />

equality.<br />

9.4.5. Proposition. Let R be a noetherian ring and M a finite module. Let L ⊂ M<br />

such that M/L = 0 has finite length. If Ass(M/L) = {P1, . . . , Pn}, then there is<br />

a reduced primary decomposition<br />

where<br />

and an isomorphism<br />

L = N1 ∩ · · · ∩ Nn<br />

Ni = M ∩ LPi<br />

M/L <br />

M/Ni<br />

Pro<strong>of</strong>. This follows from 9.4.3, 9.4.4 and 9.1.6.<br />

9.4.6. Proposition. Let R be a noetherian ring and M a finite length module. If<br />

Ass(M) = {P1, . . . , Pn}, then there is a reduced primary decomposition<br />

where<br />

0 = N1 ∩ · · · ∩ Nn<br />

Ni = Pi ni M, M/Ni MPi<br />

i<br />

i<br />

i


108 9. PRIMARY DECOMPOSITION<br />

and isomorphisms<br />

M <br />

Pro<strong>of</strong>. This follows from 7.4.5.<br />

i<br />

MPi<br />

<br />

M/Pi niM 9.4.7. Proposition. Let R be a noetherian ring and M a finite module. Let L ⊂ M<br />

have a reduced primary decomposition<br />

L = N1 ∩ · · · ∩ Nn<br />

where Ni is Pi-primary. Assume U to be a multiplicative subset disjoint from<br />

exactly P1, . . . Pk. Then<br />

is a reduced primary decomposition.<br />

U −1 L = U −1 N1 ∩ · · · ∩ U −1 Nk<br />

Pro<strong>of</strong>. This follows from 9.3.7 and 9.4.3.<br />

9.4.8. Exercise. (1) Describe the primary decomposition over a field.<br />

9.5. Decomposition <strong>of</strong> ideals<br />

9.5.1. Proposition. A proper ideal I ⊂ R has a reduced primary decomposition<br />

and for any such<br />

and<br />

is exact.<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 9.4.3.<br />

I = Q1 ∩ · · · ∩ Qn<br />

Ass(R/I) = {P1, . . . , Pn}<br />

0 → R/I → <br />

R/Qi<br />

9.5.2. Proposition. If<br />

I = Q1 ∩ · · · ∩ Qn<br />

is a reduced primary decomposition and Pi is minimal in Ass(R/I), then<br />

and therefore uniquely determined.<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 9.4.4.<br />

Qi = R ∩ IRPi<br />

9.5.3. Definition. Let P be a prime ideal. The symbolic power <strong>of</strong> P is<br />

P (n) = R ∩ P n RP<br />

9.5.4. Proposition. Let<br />

P n = Q1 ∩ · · · ∩ Qn<br />

be a reduced primary decomposition <strong>of</strong> a power <strong>of</strong> a prime ideal P . If Ass(R/Q1) =<br />

{P }, then<br />

Q1 = P (n)<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 9.4.4.<br />

i<br />

i


9.5. DECOMPOSITION OF IDEALS 109<br />

9.5.5. Proposition. Let I ⊂ R be a proper ideal in a noetherian ring such that R/I<br />

has finite length. If Ass(R/I) = {P1, . . . , Pn}, then there is a reduced primary<br />

decomposition<br />

I = Q1 ∩ · · · ∩ Qn<br />

where<br />

and an isomorphism<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 9.4.3 and 9.5.4.<br />

Qi = R ∩ QPi<br />

R/I <br />

R/Qi<br />

9.5.6. Proposition. Let R be an artinian ring. If Ass(M) = {P1, . . . , Pn}, then<br />

there is a reduced primary decomposition<br />

where<br />

and isomorphisms<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 9.4.6.<br />

0 = Q1 ∩ · · · ∩ Qn<br />

Qi = Pi ni , R/Qi RPi<br />

R <br />

i<br />

RPi<br />

i<br />

<br />

R/Pi ni<br />

9.5.7. Proposition. Let a proper ideal I ⊂ R in a noetherian ring have a reduced<br />

primary decomposition<br />

I = Q1 ∩ · · · ∩ Qn<br />

where Qi is Pi-primary. Assume U to be a multiplicative subset disjoint from<br />

exactly P1, . . . Pk. Then<br />

is a reduced primary decomposition.<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 9.4.7.<br />

U −1 I = U −1 Q1 ∩ · · · ∩ U −1 Qk<br />

9.5.8. Example. Let R be a unique factorization domain. A factorization into<br />

powers <strong>of</strong> different irreducible primes is a reduced primary decomposition <strong>of</strong> a<br />

principal ideal.<br />

9.5.9. Proposition. Let R be a noetherian ring. The following are equivalent<br />

(1) R is reduced.<br />

(2) RP is a field for all P ∈ Ass(R).<br />

(3) RP is a domain for all P ∈ Ass(R).<br />

Pro<strong>of</strong>. (1) ⇒ (2): The maximal ideal P RP = 0. (3) ⇒ (1): This follows from<br />

√ 0 = ∩P ∈Ass(R)P .<br />

9.5.10. Corollary. Let R be a reduced noetherian ring. Then all elements in<br />

Ass(R) are minimal primes. That is, there are no embedded primes.<br />

9.5.11. Exercise. (1) Let I ⊂ R be an ideal. Show that if P = √ I is a maximal ideal,<br />

then I is a P -primary ideal.<br />

(2) Let I ⊂ R be an ideal. Show that if I contains a power <strong>of</strong> a maximal ideal P , then I<br />

is a P -primary ideal.<br />

i


110 9. PRIMARY DECOMPOSITION<br />

(3) Let K be a field and I = (X 2 , XY ) ⊂ K[X, Y ]. Show that √ I = (X), but I is not<br />

(X)-primary.


10<br />

Dedekind rings<br />

10.1. Principal ideal domains<br />

10.1.1. Lemma. Let R be a domain. The set <strong>of</strong> elements x ∈ M in a module with<br />

Ann(x) = 0 is a submodule.<br />

Pro<strong>of</strong>. The product <strong>of</strong> two nonzero elements is nonzero.<br />

10.1.2. Definition. Let R be a domain. An element x ∈ M in a module is a torsion<br />

element if Ann(x) = 0. By 10.1.1 the set <strong>of</strong> torsion elements is a submodule<br />

T (M) ⊂ M, the torsion submodule. If T (M) = 0 then M is a torsion free<br />

module.<br />

10.1.3. Lemma. Let R be a domain.<br />

(1) A flat module is torsion free.<br />

(2) M/T (M) is torsion free.<br />

(3) Let K be the fraction field. Then T (M) = Ker(M → M ⊗R K).<br />

(4) If M → N → L is exact, then T (M) → T (N) → T (L) is exact.<br />

(5) If U ⊂ R is multiplicative, then U −1 T (M) = T (U −1 M).<br />

Pro<strong>of</strong>. (1) If 0 = a ∈ R and F flat, then aF is injective. (2), (4), (5) These are<br />

clear. (3) This follows from 4.4.1.<br />

10.1.4. Corollary. Let R be a domain and M a module. The following are equivalent.<br />

(1) M is torsion free.<br />

(2) MP is torsion free for all prime ideals P .<br />

(3) MP is torsion free for all maximal ideals P .<br />

10.1.5. Lemma. Let R be a principal ideal domain. A submodule <strong>of</strong> a finite free<br />

module is free.<br />

Pro<strong>of</strong>. Let F ⊂ R n be a submodule and p : R n → R the last projection. Then<br />

p(F ) is a principal ideal and free. By induction F ∩ Ker p is free, so F F ∩<br />

Ker p ⊕ p(F ) is free.<br />

10.1.6. Proposition. Let R be a principal ideal domain.<br />

(1) A torsion free module is flat.<br />

(2) A finite torsion free module is free.<br />

Pro<strong>of</strong>. (1) For a nonzero ideal (a) ⊂ R the composite R (a) → R is aM.<br />

For a torsion free F the homomorphism aF : F (a) ⊗R F → F is injective.<br />

So F is flat by 3.7.12. (2) Let K be the fraction field. Let F ⊂ F ⊗R K be a<br />

torsion free submodule and suppose x1, . . . , xn ∈ F give a basis for F ⊗R K.<br />

Then F ′ = Rx1 ⊕ · · · ⊕ Rxn ⊂ F is a free submodule such that F/F ′ is a finite<br />

111


112 10. DEDEKIND RINGS<br />

torsion module. There is 0 = a ∈ Ann(F/F ′ ), so F aF ⊂ F ′ . Conclusion by<br />

10.1.5.<br />

10.1.7. Proposition. Let R be a principal ideal domain. A finite module M decomposes<br />

M = T (M) ⊕ F<br />

as a direct sum <strong>of</strong> the torsion submodule and a finite free submodule F .<br />

Pro<strong>of</strong>. The sequence 0 → T (M) → M → M/T (M) → 0 is split exact.<br />

10.1.8. Proposition. Let R be a principal ideal domain. A finite torsion module<br />

M has a primary decomposition<br />

Where<br />

M = ⊕ (p)Mp<br />

Mp = {x ∈ M|p n x = 0, for some n}<br />

Pro<strong>of</strong>. This is the primary decomposition 9.4.6.<br />

10.1.9. Proposition. Let R be a principal ideal domain and (p) an irreducible<br />

principal ideal. A finite torsion module M such that M = Mp has decomposition<br />

Where n1 ≥ · · · ≥ nk.<br />

M = R/(p n1 ) ⊕ · · · ⊕ R/(p nk )<br />

Pro<strong>of</strong>. Let n1 = n be such that p n ∈ Ann(M), but p n−1 x = 0 for some x ∈ M.<br />

The short exact sequence 0 → Rx → M → M/Rx → 0 <strong>of</strong> R/(p n )-modules<br />

is split exact, since R/(p n ) Rx is an injective module 7.5.9. Conclusion by<br />

induction on ℓR(M).<br />

10.1.10. Exercise. (1) Show that a nonzero finite Z-submodule <strong>of</strong> Q is a free module<br />

<strong>of</strong> rank 1.<br />

(2) Show that a finite torsion Z-module is a finite group.<br />

(3) Let K be a field. Show that a finite torsion K[X]-module is a finite K vector space.<br />

(4) Let R be a noetherian domain such that every nonzero prime ideal is maximal. Let<br />

M be a finite module. Show that the torsion submodule T (M) has finite length.<br />

10.2. Discrete valuation rings<br />

10.2.1. Definition. A local principal ideal domain, which is not a field, is a discrete<br />

valuation ring. A generator <strong>of</strong> the maximal ideal is a local parameter or a<br />

uniformizing parameter.<br />

10.2.2. Proposition. Let (R, (p)) be a discrete valuation ring. Then<br />

Pro<strong>of</strong>. See 8.5.5.<br />

∩n(p n ) = 0<br />

10.2.3. Proposition. Let (R, (p)) be a discrete valuation ring. Any nonzero ideal<br />

is <strong>of</strong> the form (p n ) for a unique n = 0, 1, 2, . . . .<br />

Pro<strong>of</strong>. Let 0 = x R, by 10.2.2 there is n such that x ∈ (p n ) − (p n+1 ). Since any<br />

ideal is finitely generated it follows that a nonzero ideal is <strong>of</strong> the form (p n ). n is<br />

unique by unique factorization.


10.3. DEDEKIND DOMAINS 113<br />

10.2.4. Corollary. Let (R, (p)) be a discrete valuation ring. Any nonzero element<br />

in the fraction field K <strong>of</strong> R has a unique representation up n where u is a unit and<br />

n ∈ Z.<br />

10.2.5. Definition. Let K be a field. A surjective map v : K\{0} → Z satisfying<br />

(1) v(xy) = v(x) + v(x).<br />

(2) If x + y = 0, v(x + y) ≥ min(v(x), v(y).<br />

is a valuation on K.<br />

10.2.6. Proposition. Let v be a valuation on a field K. Then<br />

R = {x|v(x) ≥ 0}<br />

is the discrete valuation ring <strong>of</strong> v. A local parameter is any element p, v(p) = 1.<br />

Pro<strong>of</strong>. Clear from the definition.<br />

10.2.7. Proposition. Let (R, (p)) be a discrete valuation ring with fraction field<br />

K. The map<br />

v : K\{0} → Z, up n ↦→ n<br />

is a valuation and R is the discrete valuation ring <strong>of</strong> v.<br />

Pro<strong>of</strong>. Clear from the definitions.<br />

10.2.8. Proposition. Let R, (p) be a discrete valuation ring. A finite module M<br />

has decomposition<br />

Where n1 ≥ · · · ≥ nk > 0.<br />

M = R/(p n1 ) ⊕ · · · ⊕ R/(p nk ) ⊕ R n<br />

Pro<strong>of</strong>. This is a case <strong>of</strong> 10.1.7 and 10.1.9.<br />

10.2.9. Exercise. (1) Let K be a field. Show that the subring K[[X 2 , X 3 ]] ⊂ K[[X]]<br />

is not a discrete valuation ring.<br />

10.3. Dedekind domains<br />

10.3.1. Definition. A noetherian domain R, which is not a field, is a Dedekind<br />

domain if all local rings RP at nonzero prime ideals are discrete valuation rings.<br />

10.3.2. Proposition. Let R be a Dedekind domain.<br />

(1) Any nonzero prime ideal is maximal.<br />

(2) If U ⊂ R is multiplicative, then U −1 R is a field or a Dedekind domain.<br />

Pro<strong>of</strong>. (1) (0), P are the only prime ideals in a prime ideal P . (2) This is clear<br />

from 5.2.11.<br />

10.3.3. Proposition. Let R be a domain which is not a field. The following conditions<br />

are equivalent.<br />

(1) R is a Dedekind domain.<br />

(2) Every nonzero proper ideal in R is a product <strong>of</strong> finitely many maximal ideals.<br />

Pro<strong>of</strong>. The primary ideals P (n) = P n . Conclusion by 9.5.4 and Chinese remainders<br />

1.4.2.<br />

10.3.4. Proposition. Let R be Dedekind domain.<br />

(1) If R is a unique factorization domain then it is a principal ideal domain.


114 10. DEDEKIND RINGS<br />

(2) If R has only finitely many maximal ideals then it is a principal ideal domain.<br />

Pro<strong>of</strong>. (1) Any nonzero prime ideal is principal. Conclusion by 10.3.3. (2) Let<br />

P, P2 . . . , Pn be the finitely many maximal ideals. Choose a ∈ P \P 2 ∪P2 · · ·∪Pn,<br />

5.1.3. Then (a) is P primary. By 10.3.3 (a) = P k , so (a) = P . As all maximal<br />

ideals are principal, conclusion by 10.3.3.<br />

10.3.5. Proposition. Let R be Dedekind domain. An ideal I is generated by at<br />

most two elements.<br />

Pro<strong>of</strong>. Let Ass(R/I) = {P1, . . . , Pn} and U = R\P1 ∪ · · · ∪ Pn, then by 10.3.4<br />

U −1 R is a principal ideal domain. By 10.1.6 U −1 R U −1 I, so choose by 8.2.9 a<br />

homomorphism f : R → I such that U −1 f is an isomorphism. Then f is injective<br />

9.2.8 and the ideal f(R) = (a) ⊂ R satisfies: Pi /∈ Ass(R/(a)) for any i. Let<br />

Q1, . . . , Qm ∈ Ass(I/(a)) and choose b ∈ I\Q1 ∪ · · · ∪ Qm. b is a nonzero<br />

divisor on I/(a) and therefore b I/(a) is an isomorphism as I/(a) has finite length.<br />

It follows that I = (a, b).<br />

10.3.6. Theorem. Let R be Dedekind domain.<br />

(1) A torsion free module is flat.<br />

(2) A finite torsion free module is projective.<br />

(3) Any ideal is projective.<br />

(4) Let F be a finite torsion free module. Then there is a number n and an ideal<br />

I such that<br />

F R n ⊕ I<br />

Pro<strong>of</strong>. (1), (2), (3) These follow from 10.1.6, 6.5.13. (4) Let R have fraction field<br />

K and assume rankK F ⊗RK = n+1. Choose a nonzero homomorphism F → R<br />

and get by induction on n, F I1⊕· · ·⊕In+1 for nonzero ideals Ij in R. It suffices<br />

to treat the case n = 1. Let Ass(R/I1) = {P1, . . . , Pm} and U = R\P1∪· · ·∪Pm,<br />

then by 10.3.4 U −1 R is a principal ideal domain. By 10.1.6 U −1 I2 U −1 R, so<br />

choose by 8.2.9 a homomorphism f : I2 → R such that U −1 f is an isomorphism.<br />

Then f is injective 9.2.8 and the ideal f(I2) ⊂ R satisfies: Pi /∈ Ass(R/f(I2)) for<br />

any i. It follows that I1 + f(I2) = R. Conclusion by a surjection I1 ⊕ I2 → R.<br />

10.3.7. Proposition. Let R be a Dedekind domain. A finite module M decomposes<br />

M = T (M) ⊕ F<br />

as a direct sum <strong>of</strong> the torsion submodule and a finite torsion free submodule F .<br />

Pro<strong>of</strong>. By 10.3.6 the projection M → M/T (M) splits.<br />

10.3.8. Proposition. Let R be a Dedekind domain. A finite torsion module M has<br />

decomposition<br />

M = R/P n1<br />

nk<br />

1 ⊕ · · · ⊕ R/Pk Where P1, . . . , Pk are not necessarily distinct maximal ideals.<br />

Pro<strong>of</strong>. This follows from 9.5.1 and 10.2.9.<br />

10.3.9. Corollary. Let R be a Dedekind domain and M a finite module. Then M<br />

has decomposition<br />

M = R/P n1<br />

1<br />

⊕ · · · ⊕ R/P nk<br />

k ⊕ Rn ⊕ Q1 · · · Ql<br />

Where P1, . . . , Pk, Q1, . . . , Ql are not necessarily distinct maximal ideals.


10.3. DEDEKIND DOMAINS 115<br />

10.3.10. Exercise. (1) Show that the ring Z[ √ −5] is a Dedekind domain.<br />

(2) Show that the ring Z[ √ 5] is not a Dedekind domain.


Bibliography<br />

A. Altman and S. Kleiman, Introduction to Grothendieck duality theory, Springer-Verlag 1970.<br />

M. Atiyah and I. Macdonald, An introduction to commutative <strong>algebra</strong>, Addison-Wesley 1969.<br />

N. Bourbaki, Algébre commutative, Hermann-Masson 1961-.<br />

M. Brodmann and R. Sharp, Local cohomology: an <strong>algebra</strong>ic introduction with geometric applications,<br />

Cambridge University Press 1997.<br />

W. Bruns and J. Herzog, Cohen-Macaulay rings, Cambridge University Press 1993.<br />

H. Cartan and S. Eilenberg, Homological <strong>algebra</strong>, Princeton 1956.<br />

J. Dieudonné, Cours de géométrie algébrique, Presses Universitaires de France 1974.<br />

D. Eisenbud, <strong>Commutative</strong> <strong>algebra</strong> with a view toward <strong>algebra</strong>ic geometry, Springer-Verlag 1996.<br />

W. Fulton, Algebraic curves, Benjamin inc. 1969.<br />

W. Fulton, Intercection theory, Springer-Verlag 1984.<br />

A. Grothendieck, Sur quelques points d’algégre homologique, Tôhoku Math, Journ., 9, 1957.<br />

A. Grothendieck, Eléments de géométrie algébrique, IHES 1960-67.<br />

A. Grothendieck, Séminaire de géométrie algébrique, IHES 1960-67.<br />

R. Hartshorne, Residues and duality, Springer-Verlag 1966.<br />

R. Hartshorne, Algebraic geometry, Springer-Verlag 1977.<br />

B. Iversen, Generic local structure in commutative <strong>algebra</strong>, Springer-Verlag 1974.<br />

B. Iversen, Cohomology <strong>of</strong> sheaves, Springer-Verlag 1986.<br />

B. Iversen, Local rings, Aarhus 1974-80.<br />

I.Kaplansky, <strong>Commutative</strong> rings, Allyn and Bacon 1970<br />

E. Kunz, Introduction to commutative <strong>algebra</strong>, Birkhäuser 1980.<br />

S. Lang, Algebra, Addison-Wesley 1965.<br />

H. Matsumura, <strong>Commutative</strong> <strong>algebra</strong>, Benjamin inc. 1970.<br />

J. Milne, http://www.jmilne.org.<br />

D. Mumford, Introduction to <strong>algebra</strong>ic geometry, Red and 1 inch thick.<br />

D. Mumford, Algebraic geometry I, complex projective varieties, Springer-Verlag 1976.<br />

M. Nagata, Local rings, Interscience Publ. 1962.<br />

C. Peskine, An <strong>algebra</strong>ic introduction to complex projective geometry, Cambridge University Press<br />

1997.<br />

M. Reid, Undergraduate commutative <strong>algebra</strong>, Cambridge University Press 1995.<br />

J-P. Serre, Algébre locale - multiplicités, Springer-Verlag 1965.<br />

J-P. Serre, Faisceaux algébrique cohérents, Ann. <strong>of</strong> Math., 61, 1955.<br />

I. Shafarevich, Basic <strong>algebra</strong>ic geometry, Springer-Verlag 1974.<br />

R. Sharp, Steps in commutative <strong>algebra</strong>, Cambridge University Press 1990.<br />

O. Zariski and P. Samuel, <strong>Commutative</strong> <strong>algebra</strong>, van Nostran 1958-60.<br />

117


118 BIBLIOGRAPHY


A/B, 9<br />

P -primary, 106<br />

0-sequence, 39<br />

abelian group, 9<br />

addition, 9<br />

annihilator, 27<br />

artinian module, 87<br />

artinian ring, 87<br />

artinian rings, 87<br />

ass <strong>of</strong> modules, 103<br />

associated prime ideal, 103<br />

associative, 9, 21<br />

basis, 29<br />

bilinear, 21<br />

binomial formula, 10<br />

canonical homomorphism, 59<br />

canonical ring homomorphism, 57<br />

Cayley-Hamilton’s theorem, 77<br />

change <strong>of</strong> ring, 36<br />

change <strong>of</strong> rings, 36<br />

characteristic, 12<br />

Chinese remainder theorem, 14<br />

Chinese remainders, 14<br />

coefficient, 16, 20<br />

c<strong>of</strong>actor matrix, 75<br />

cokernel, 25<br />

colon ideal, 11, 27<br />

comaximal ideals, 14<br />

commutative, 9<br />

composition series, 85<br />

constants, 16<br />

content <strong>of</strong> polynomial, 64<br />

contracted ideal, 11<br />

contravariant, 31<br />

cosets, 9<br />

Cramer’s rule, 76<br />

decomposable, 30<br />

decomposition <strong>of</strong> ideals, 108<br />

decomposition <strong>of</strong> modules, 106<br />

Dedekind domain, 113<br />

Dedekind domains, 113<br />

Index<br />

119<br />

degree, 16<br />

derivative, 18<br />

determinant, 75<br />

direct product, 28<br />

direct sum, 28, 29<br />

discrete valuation ring, 112<br />

discrete valuation ring <strong>of</strong> v, 113<br />

discrete valuation rings, 112<br />

distributive, 9<br />

divisible module, 53<br />

domain, 10<br />

dual homomorphism, 33<br />

dual module, 33<br />

embedded prime, 105<br />

essential extension, 54<br />

evaluation, 32<br />

evaluation map, 17<br />

exact sequence, 39<br />

exact sequences, 39<br />

exactness and localization, 70<br />

exactness <strong>of</strong> fractions, 60<br />

exactness <strong>of</strong> hom, 48<br />

exactness <strong>of</strong> tensor, 49<br />

extended ideal, 11<br />

factor group, 9<br />

factor module, 24<br />

factor ring, 11<br />

faithfully flat, 71<br />

faithfully flat ring homomorphism, 71<br />

field, 10<br />

field extension, 19<br />

fields, 19<br />

finite field extension, 19<br />

finite ideal, 11<br />

finite length, 85<br />

finite module, 73<br />

finite modules, 73<br />

finite presented module, 80<br />

finite presented modules, 80<br />

finite ring extension, 83<br />

finite ring homomorphism, 83<br />

finite ring homomorphisms, 83<br />

finite type ring, 17


120 INDEX<br />

finite type rings, 95<br />

finitely generated ring, 17<br />

five lemma, 46<br />

flat module, 54<br />

flat modules, 54<br />

flat ring homomorphism, 71<br />

flat ring homomorphisms, 71<br />

fraction field, 58<br />

fractions and localization, 98<br />

free module, 29<br />

free modules, 75<br />

Frobenius homomorphism, 12<br />

functor, 31<br />

Gauss’ lemma, 64<br />

generated, 23<br />

going-down, 72<br />

going-up, 83<br />

Gorenstein ring, 92<br />

greatest common divisor, 16<br />

Hilbert’s basis theorem, 95<br />

homomorphism, 9, 21<br />

homomorphism module, 30<br />

homomorphism modules, 30<br />

homomorphism modules <strong>of</strong> fractions, 63<br />

ideal, 11<br />

ideal generated by, 11<br />

ideal product, 11<br />

ideals, 11<br />

idempotent, 15<br />

identity, 9, 21<br />

identity isomorphism, 9, 21<br />

image, 25<br />

indecomposable, 30<br />

induced module, 37<br />

injections, 28<br />

injective envelope, 54<br />

injective module, 52<br />

injective modules, 52<br />

irreducible element, 15<br />

irreducible principal ideal, 15<br />

isomorphism, 9, 21<br />

Jacobson radical, 68<br />

kernel, 11, 25<br />

kernel and cokernel, 25<br />

Krull’s intersection theorem, 96, 98<br />

Krull’s theorem, 65<br />

leading coefficient, 16<br />

least common multiple, 16<br />

length, 86<br />

linear map, 22<br />

local artinian ring, 91<br />

local parameter, 112<br />

local ring, 67<br />

local ring homomorphism, 67<br />

localization, 90<br />

localization <strong>of</strong> modules, 68<br />

localization <strong>of</strong> rings, 67<br />

localized homomorphism, 68<br />

localized module, 68<br />

localized ring, 67<br />

locally free module, 69<br />

maximal ideal, 13<br />

minimal prime, 101<br />

minimal prime ideal, 66<br />

minor, 75<br />

module, 21<br />

module <strong>of</strong> fractions, 59<br />

modules and homomorphisms, 21<br />

modules and submodules, 93<br />

modules <strong>of</strong> fractions, 58<br />

monic polynomial, 16<br />

monomial, 16<br />

multiplication, 9<br />

multiplication <strong>of</strong> principal ideals, 15<br />

multiplicative subset, 57<br />

multiplicity, 18<br />

Nakayama’s lemma, 78<br />

natural homomorphism, 31<br />

natural isomorphism, 31<br />

negative, 9<br />

nilpotent, 14<br />

nilradical, 14<br />

noetherian module, 93<br />

noetherian ring, 94<br />

noetherian rings, 94<br />

noncommutative ring, 9<br />

nontrivial idempotent, 15<br />

nonzero divisor, 10, 22<br />

order, 20<br />

polynomial, 16<br />

polynomial ring, 16<br />

polynomials, 16<br />

power series, 20<br />

power series ring, 20<br />

power series rings, 97<br />

primary decomposition, 106<br />

primary modules, 106<br />

primary submodule, 106<br />

prime fields, 19<br />

prime filtrations <strong>of</strong> modules, 98<br />

prime ideal, 13<br />

prime ideals, 13, 65<br />

principal ideal, 11<br />

principal ideal domain, 15<br />

principal ideal domains, 111<br />

product ring, 10<br />

projection, 9, 24


projections, 28<br />

projective module, 50<br />

projective modules, 50<br />

proper ideal, 11<br />

radical, 14<br />

rank, 76<br />

reduced, 14<br />

reduced primary decomposition, 106<br />

reflexive module, 33<br />

residue field, 67<br />

residue homomorphism, 68<br />

restriction <strong>of</strong> scalars, 22<br />

retraction, 41<br />

ring, 9<br />

ring extension, 9<br />

ring generated, 17<br />

ring <strong>of</strong> fractions, 57<br />

rings, 9<br />

rings <strong>of</strong> fractions, 57<br />

root, 18<br />

roots, 18<br />

scalar multiplication, 21, 22<br />

section, 41<br />

short exact sequence, 40<br />

simple module, 85<br />

simple modules, 85<br />

simple root, 18<br />

snake homomorphism, 44<br />

snake lemma, 45<br />

spectrum, 101<br />

split exact sequence, 42<br />

standard basis, 29<br />

subfield, 19<br />

subgroup, 9<br />

submodule, 21<br />

submodule generated, 23<br />

submodules and factor modules, 23<br />

subring, 9<br />

sum and product, 28<br />

support, 101<br />

support <strong>of</strong> modules, 101<br />

symbolic power, 108<br />

tensor modules <strong>of</strong> fractions, 62<br />

tensor product, 33<br />

tensor product modules, 33<br />

tensor product ring, 38<br />

the length, 85<br />

the polynomial ring is factorial, 64<br />

the snake lemma, 43<br />

torsion element, 111<br />

torsion free module, 111<br />

torsion submodule, 111<br />

total ring <strong>of</strong> fractions, 58<br />

uniformizing parameter, 112<br />

INDEX 121<br />

unique factorization, 15<br />

unique factorization domain, 15<br />

unit, 10<br />

valuation, 113<br />

vector space, 22<br />

windmill lemma, 46<br />

zero, 9<br />

zero divisor, 10, 22<br />

zero ideal, 11<br />

zero module, 21<br />

zero submodule, 21

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!