Lecture Notes on Module Theory

by

Prof Shiv Datt Kumar

Department of Mathematics

Motilal Nehru National Institute of Technology

Allahabad (UP), India, Pin-211004

E-mail: sdt@mnnit.ac.in
Shiv Datt Kumar Modules

2
Chapter 1

Module Theory

The notion of an ideal is important in number theory and Algebraic Geometry.

Since operations of linear algebra such as formation of quotients, products and

tensor products are closed for modules but not for ideals, it is useful to study

ideals from a module theoretic set up.

Projective modules play the role of a vector space while studying linear al-

gebra over a general commutative ring. Flat modules are more general than

projective modules. Some times, geometric objects defined over a field behave

differently over a bigger field. Such questions are best studied by scalar ex-

tension of rings using tensor products. Some properties are preserved under

flat extension but faithfully flat extensions are useful as they have nice descent

properties.

1.1 Module

The notion of vector space over a ring is known as module over a ring.

Definition 1.1.1. (Module)

Let M be a non empty set and R be a ring with identity. A left R-module

M is an abelian group (M, +) together with a map · : R × M → M satisfying

the following conditions:

3
Shiv Datt Kumar Modules

1. α.(x + y) = α.x + α.y,

2. (α + β).x = α.x + β.x,

3. α.(β.x) = (αβ).x,

4. 1.x = x, ∀α, β ∈ R, ∀x, y ∈ M and 1 ∈ R.

Remark: In modules α.x = 0 ; α = 0 or x = 0 contrary to vector spaces.

Example 1.1.2. 1. M = Vector space V over R = field k

2. M = Abelian group G, over R = Z

3. M = C[0, 1] = {f : [0, 1] → R | f is continuous }, over R = R.

4. If I is an ideal of ring R, then R, I and R/I are R-module.

5. Let S be a multiplicative subset of R. Then the localization S −1 R is an

R- module.

6. Let R be a ring. Then Mn (R) is a left R-module if we define for r ∈ R

and A ∈ Mn (R)  
r 0 0 ··· 0
 
0 r 0
 ··· 0
 
rA = 0 0 r ··· 0 A
 
 
. . . ..
 .. .. ..

 . 

 
0 0 0 r

Clearly, Mn (R) can also be made a right R-module.

Definition 1.1.3. A module M is said to be finitely generated over R if there

exists a finite set S ⊂ M which generates M .

Definition 1.1.4. A subset N ⊂ M is called a submodule, if N is a subgroup

of the abelian group M and ax ∈ N for all a ∈ R and x ∈ N .

Example 1.1.5. 1. Any subspace W of a vector space V is a submodule.

4
Shiv Datt Kumar Modules

2. All polynomials of degree atmost n is a submodule of the R-module R[X].

3. The module 0 and M are submodules of M called improper submodules.

1.2 Free Modules

Simplest type of modules are free modules. Every vector space always has a

basis but this is not true in the case of module. Let R be a ring and I be

an ideal. Then R/I is a R-module but it has no basis as R/I has no linearly

independent subset. A module with basis is called a free module.

Definition 1.2.1. An R-module M is called Free module if ∃ a subset S (called

a basis for M ) such that

1. S is linearly independent

2. S generates M .

Note that if M is a free module with basis consisting of n elements, then

M∼
= Rn and n is called a rank of M .

Example 1.2.2. 1. Zn = {(x1 , x2 , . . . , xn ) | xi ∈ Z} is free over Z with

basis {e1 , e2 , . . . , en }.

2. M = {0̄, 2̄, 4̄} is a module over Z6 , which is not free.

In vector space theory we have two following Theorems:

Theorem 1.2.1. Every vector space is a free module.

Theorem 1.2.2. In a vector space V over a division ring D

1. every linearly independent subset can be extended to a basis.

2. every subset which spans V contains a basis.

It is a natural question to what extent these results are true for

1. modules over arbitrary rings.

5
Shiv Datt Kumar Modules

2. free modules over a commutative ring.

3. free modules over a non-commutative ring.

Observe the following table

S.No. Vector space Free Module over

arbitrary commutative
ring ring
1 Any linearly independent subset can be Not true Not true
extended to basis
2 Any spanning subset contains a basis Not true Not true
3 Any subspace of Vector space is again Not true Not true
Vector space
4 Any subspace of finitely generated Vec- Not true Not true
tor space is again finitely generated
5 any two basis have same cardinality Not true true

• Example of a free module in which a linearly independent subset can not

be extended to a basis. Take R = Z, M = Z = [{1}]. Note that {2} is

linearly independent over Z but can not be extended to a basis for any

b 6= 0, for {2, b} is linearly dependent, because 2.b − b.2 = 0 but 2 6= 0,

b 6= 0.

• Example of a free module for which a subset X spans M but X does not

contain a basis. R = Z, M = Z. Then X = {2, 3} or {m, n} such that

(m, n) = 1 generates M over R but not linearly independent.

• Example of a finitely generated free module M having a submodule which

is neither free nor finitely generated. Take R = K[x1 , x2 , . . . , xn , . . .] a

polynomial ring and M = R. Then M is free with basis {1}. Now take

N = (x1 , x2 , . . . , xn , . . .) which is neither finitely generated nor free R-

module.

• Example of free module M which has basis having different cardinality.

Let V be a vector space of countably infinite dimensional over a division

ring D and R = EndD (V ). Then R is free over R with {1}. Bn =

6
Shiv Datt Kumar Modules

{f1 , f2 , . . . , fn } for having n-elements, let B = {ek }∞

k=1 be a basis of V .

Define f1 , f2 , . . . , fn .

f1 f2 f3 ... fn
e1 e1 0 0 ... 0
e2 0 e1 0 ... 0
.. .. .. .. .. ..
. . . . . .
en 0 0 0 ... e1
en+1 e2 0 0 ... 0
en+2 0 e2 0 ... 0
.. .. .. .. .. ..
. . . . . .
e2n 0 0 0 ... e2
.. .. .. .. ..
. . . . ... .
ekn+1 ek+1 0 0 ... 0
ekn+2 0 ek+1 0 ... 0
.. .. .. .. .. ..
. . . . . .
e(k+1)n 0 0 0 ... ek+1

P
αi fi = 0, αi ∈ R. Evaluating on successive blocks of n even terms.

αi fi (ek+1 ) = 0, ∀k. This implies that αi = 0, ∀i and ∀f ∈ R, f =

Pn
i=1 αi fi , where αi ∈ R are defined by their values on B.

Proposition 1.2.3. Let M be a left R-module. Then

1. r0M = 0M for all r ∈ R

2. 0R m = 0M for all m ∈ M .

3. r(−m) = (−r)m = −rm for all m ∈ M and r ∈ R

Definition 1.2.4. Let M and N be R modules. A map f : M −→ N is called

a homomorphism of R modules if

(i) f (x + y) = f (x) + f (y), x, y ∈ M .

(ii) f (ax) = af (x), a ∈ R, x ∈ M .

Example 1.2.5. (i) For any fixed a ∈ R, the map f : M −→ M , given by

f (x) = ax, is a homomorphism.

7
Shiv Datt Kumar Modules

(ii) For any submodule N of M , the inclusion map i : N −→ M is a homo-

morphism.

Let N be a submodule of M . Consider the quotient abelian group M/N =

{x+N |x ∈ M } with the scalar multiplication given by a.(x+N ) = ax+N, a ∈ R,

x ∈ M . Then M/N acquires the structure of an R-module called the quotient

module M/N . The map p : M −→ M/N defined by p(x) = x + N , x ∈ M is a

homomorphism of modules called the projection.

Definition 1.2.6. : A homomorphism of modules which is both injective and

surjective is called an isomorphism.

Proposition 1.2.3. Let f : M −→ N be a homomorphism of M onto N . Then

kernel of f = {x ∈ M |f (x) = 0} is a submodule K of M and the quotient

module M/K is isomorphic to N .

Proof. Clearly K satisfies the conditions for a submodule. The map f : M/K −→

N given by f (x + K) = f (x), for all x ∈ M , is well defined and is an isomor-

phism.

Corollary 1.2.4. Let M be an R-module, N a submodule of M and K a sub-

M/K
module of N . Then M/N ' N/K .

M/K
Proof. The composition of the projections M −→ M/K −→ N/K is surjective
M/K
and has kernel N . Hence M/N ' N/K .

We now consider some operations on modules. If N and K are submodules

of M , then N ∩K is a submodule of M but in general, N ∪K is not a submodule

of M . The smallest submodule of M containing N ∪ K is called the submodule

generated by N and K.

Proposition 1.2.5. The submodule S generated by N and K is

N + K = {x + y|x ∈ N, y ∈ K}.

8
Shiv Datt Kumar Modules

Proof. Clearly N + K is a submodule of M . Also N ⊂ N + K and K ⊂ N + K

so that S ⊂ N + K. Conversely for any x ∈ M, y ∈ K, We have x, y ∈ S so that

x + y ∈ S. Thus N + K ⊂ S and S = N + K.

Corollary 1.2.6. If N1 , N2 , . . . , Nk are submodules of M , then submodule gen-

Pk
erated by N1 , N2 , . . . , Nk is equal to i=1 Ni = N1 + N2 + . . . + Nk .
P
Let A be a subset of M and I an ideal in R. Then the set {x = ai xi |ai ∈

I, xi ∈ A} is a submodule of M and is denoted by IA. In particular if I = R

and A = {x}, IA is denoted by Rx.

Definition 1.2.7. An R-module M is called cyclic if M = Rx for some x ∈ M .

1.3 Cyclic Module

Proposition 1.3.1. An R-module M is cyclic if and only if M ∼

= R/I, for some
ideal I in R.

Proof. If M is cyclic, then M = Rx for some x ∈ M . The natural map θ : R →

M , given by θ(x) = ax is a surjective homomorphism. The annihilator of an

R-module M is defined as Ann(M ) = {a ∈ R|aM = 0}. Clearly Ann(M ) is an

ideal of R. If I = kerθ, then R/I ' M . Conversely if M ' R/I, then M is

cyclic, as R/I is cyclic being generated by 1 = 1 + I.

Definition 1.3.1. M is called a faithful R-module if Ann(M ) = 0.

Definition 1.3.2. Module M is called a finitely generated module over R if

M = M1 + M2 + . . . + Mn , where each Mi is cyclic. If Mi = Rxi , then

x1 , x2 , . . . , xn is called a generating set for M .

Example 1.3.3. The module of polynomials over R of degree atmost n is

generated by 1, X, X 2 , . . . , X n . Clearly this generating set is not unique, as

1, 1 + X, X 2 , . . . , X n is also a generating set for the same module.

9
Shiv Datt Kumar Modules

Definition 1.3.4. : The module M is called direct sum of submodules M1 , M2 , . . . , Mn ,

if every x ∈ M can be uniquely expressed as x = x1 + x2 + . . . + xn , xi ∈ Mi ,

1 ≤ i ≤ n. The direct sum is denoted by M = M1 ⊕ M2 ⊕ . . . ⊕ Mn .

Proposition 1.3.2. An R-module M = M1 ⊕ M2 ⊕ . . . . ⊕ Mn if and only if

(i) M = M1 + M2 + . . . + Mn and

(ii) Mi ∩ (M1 + M2 + Mi−1 + Mi+1 + . . . + Mn ) = 0 for all i, 1 ≤ i ≤ n.

Proof. Suppose M = M1 ⊕ . . . ⊕ Mn . Then clearly (i) is true. To prove (ii)

suppose x ∈ Mi ∩ (M1 + M2 + Mi−1 + Mi+1 + . . . + Mn , then x ∈ Mi and

x = y1 + y2 + . . . + yi−1 + yi+1 + . . . yn , yj ∈ Mj , j 6= i. Since x = 0 + . . . +

0 + x + 0 + . . . + 0, with x in the i-th place, by uniqueness, we have x = 0.

Conversely assume conditions (i) and (ii). By (i), each x ∈ M can be expressed

as x = x1 + x2 + . . . + xn , xi ∈ Mi . Suppose x = y1 + y2 + . . . + yn , yi ∈ Mi .

Then 0 = (x1 − y1 ) + . . . + (xi − yi ) + . . . + (xn − yn ), so that (xi − yi ) ∈ Mi

and xi − yi = −[(x1 − y1 ) + . . . + (xi−1 − yi−1 ) + (xi+1 − yi+1 ) + . . . (xn − yn )] ∈

(M1 + . . . + Mi−1 + Mi+1 + . . . + Mn ). Hence by (ii), xi − yi = 0, i.e. xi = yi ,

1 ≤ i ≤ n. Thus M = M1 ⊕ M2 ⊕ . . . ⊕ Mn .

Remarks : (1) Definition of direct sum can be extended to any collection

of modules. An R-module M is a direct sum of a collection of submodules Mα ,

α ∈ I if each x ∈ M can be expressed uniquely as x = xα1 +xα2 +. . .+xαk , xαi ∈

P
Mαi , α1 , α2 , . . . , αn ∈ I. We denote this by M = ⊕ Mα .

(2) If M = M1 ⊕ M2 , x ∈ M , then x can be uniquely expressed as x = x1 + x2 ,

x1 ∈ M1 , x2 ∈ M2 . The mapping π1 : M −→ M1 , π2 : M −→ M2 defined by

π1 (x) = x1 , π2 (x) = x2 are called projections.

(3) A cyclic R-module M = Rx is called free if Ann(x) = 0. An R-module M

P
is called free if it can be expressed as a direct sum M = ⊕ Mα where each

Mα is a free cyclic R-module. If Mα = Rxα , then the collection {xα } is called

a basis of the free module M .

10
Shiv Datt Kumar Modules

Example 1.3.5. : (i) Rn = {(a1 , . . . , an )|ai ∈ R} is a free R-module with basis

e1 = (1, 0, . . . , 0), e2 = (0, 1, . . . 0), . . . en = (0, 0, . . . 0, 1).

(ii) Zn the group of integers modulo n is not free Z-module as each x ∈ Zn

has a non-zero annihilator. Clearly, basis of a free module is not unique.

Proposition 1.3.3. Any two bases of a free module have the same cardinality.

Proof. Let M be a free module with basis xα ∈ I. Choose a maximal ideal m

of R and let R/m = k. Then V = M/M m is annihilated by m and hence it is a

k-vector space. If xα = xα + M m, then it is easy to verify that xα is a basis of

V over k. Since any two bases of a vector space have the same cardinality, the

result follows.

Corollary 1.3.4. If a free module F has a basis with n elements, then any other

basis of F also has n elements. and n is called the rank of F .

Example 1.3.6. : (i) A cyclic module M = Rx with Ann(x) = 0 is free module

of rank one.

(ii) The R-module Rn is a free module of rank n.

(iii) The R-module R[X] is free module with basis 1, X, X 2 , . . . , X n , . . . and

has countable rank.

Remark 1.3.1. 1. If R is a commutative ring, then every left R-module can

also considered as a right R-module by defining x.a = a.x.

2. Every left R-module M gives rise to a ring homomorphism f from R to

End(M, +) defined by f (a)m = a.m. Conversely given any ring homo-

morphism f : R → End(M, +), End(M, +) is an R-module defined by

a.m = f (a)m.

11
Shiv Datt Kumar Modules

1.4 Noetherian modules

Ascending chain condition(ACC): A module M over R is said to satisfy

ACC if given any chain N1 ⊆ N2 ⊆ . . . ⊆ Nr ⊆ Nr+1 ⊆ . . . of submodules of

M , ∃ n0 ∈ N such that Nr = Nn0 , ∀r ≥ n0 .

Maximal Condition: An R-module M is said to satisfy maximal condition

if given any non-empty family {Mα | α ∈ ∧} of submodules, it has a maximal

element.

Zorn’s Lemma: A partially ordered non-empty set in which every chain is

bounded above has a maximal element.

Theorem 1.4.1. Let M be an R-module. Then the following are equivalent:

1. M satisfies ACC.

2. every submodule of M is finitely generated.

3. M satisfies maximal condition.

(1) ⇒ (2) Let L ⊆ M be a submodule. If L = (0), nothing to do. If not,

let x1 ∈ L. and x1 6= 0. Then {x1 } may be L. If not, ∃ x2 ∈ L such that

< x1 >(< x1 , x2 >. Then L may be < x1 , x2 >, if not proceed, because of

ACC, we get L =< x1 , . . . , xn >.

(2) ⇒ (1) Let M1 ⊆ M2 ⊆ . . . Mn ⊆ . . . be a ascending chain of submodules and

let L = ∪Mn , n ∈ N. Then L is a submodule of M , which is finitely generated

by assumption. Let L =< x1 , . . . , xn >. Suppose < x1 , . . . xn >⊆ MN , for some

N . Hence ∪Mn = L ⊆ MN i.e. Mn ⊆ MN , ∀n. But MN ⊆ Mn ∀n ≥ N .

Hence Mn = MN , ∀n ≥ N . This means that M satisfies ACC.

(1) ⇒ (3) Let X = {Mα |α ∈ ∧} be a non empty family of submodules of M . Let

Mα1 ∈ X. It may be maximal, if not ∃ Mα2 ∈ X such that Mα1 ⊂ Mα2 . Now

Mα2 may be maximal, if not proceed. Because of ACC, this process terminates

after finitely many terms giving us a maximal member of the family.

(3) ⇒ (1) Let M1 ⊆ M2 ⊆ . . . Mn ⊆ . . . be a chain of ascending submodules.

Then X = {Mn |n ∈ N} is a non empty family of submodules. Since M satisfies

12
Shiv Datt Kumar Modules

maximal condition, ∃ Mn0 ∈ X, such that Mn ⊆ Mn0 , ∀n. Since Mn0 ⊆ Mn ,

∀n > n0 . Then Mn0 = Mn , ∀ n > n0 .

Definition 1.4.1. A module M over a ring R is called Noetherian module if

every submodule of M is finitely generated. A ring R is called Noetherian ring

if every ideal of R is finitely generated.

Example 1.4.2. • Fields, Finite rings, PID’s are Noetherian rings.

• The ring of polynomials in finitely-many variables over the integers or a

field.

Proposition 1.4.2. Let L be a submodule of a module M and L, M/L are

Noetherian. Then M is Noetherian.

U +L
Proof. Let U be a submodule of M . Then is a submodule of M/L,
L
U ∼ U +L
which is finitely generated and = . Since U ∩ L is a submodule
U ∩L L
U
L, U ∩ L is finitely generated. Suppose = {x1 + U ∩ L, . . . , xr + U ∩ L}.
U ∩L
Let U ∩ L = {y1 , . . . , ys }. Then U = [{x1 , . . . , xr , y1 . . . ys }]. Let x ∈ U .
U
Then x + U ∩ L ∈ . Thus x = α1 (x1 + U ∩ L) + . . . + αr (xr + U ∩ L).
U ∩L
P P
This implies x − (α1 x1 + . . . αr xr ) ∈ U ∩ L ⇒ x − αi xi = βj yj . Hence
P P
x = αi xi + βj yj .

1.5 Hilbert Basis Theorem

Theorem 1.5.1. Let R be a commutative ring with identity. Then R is Noethe-

rian iff R[X] is Noetherian.

Proof. If R[X] is Noetherian, then the map α : R[X] −→ R defined by α(f (x)) =

f (0) is a surjective homomorphism. Since homomorphic image of a Noetherian

module is Noetherian, R is Noetherian. Conversely, suppose R is Noetherian

and let I be an ideal in R[X], which is not finitely generated. Let f1 ∈ I be a

polynomial of least degree. Suppose fk , (k ≥ 1) has already been chosen. Let

fk+1 be a polynomial of least degree in I − (f1 , f2 , . . . , fk ). Let nk be the degree

13
Shiv Datt Kumar Modules

of fk and ak be the leading coefficient of fk , for (k = 1, 2, . . .). Then by choice of

the fk , we have n1 ≤ n2 ≤ . . . and (a1 ) ⊆ (a1 , a2 ) ⊆ . . . is a chain of ideals that

is not stationary. For suppose (a1 , a2 , . . . , ak ) = (a1 , a2 , . . . , ak , ak+1 ). Then we

Pk Pk
have an equation ak+1 = i=1 bi ai . Then g := fk+1 − i=1 bi X nk+1 −ni fi is

a polynomial of lower degree than fk+1 . Note that g ∈ / (f1 , f2 , . . . , fk ) for if

Pk
g ∈ (f1 , f2 , . . . , fn ), then fk+1 ∈ (f1 , f2 , . . . , fn ), because i=1 bi X nk+1 −ni fi ∈

(f1 , f2 , . . . , fn ). Contradicting the choice of fk+1 .

1.6 Exercises

1. If N and K are submodules of M , show that

N + K/K ' N/N ∩ K.

2. Show that if M has a finitely generated submodule N such that the quo-

tient M/N is finitely generated, then M is finitely generated.

3. Show by an example that an increasing union of finitely generated sub-

modules of M need not be finitely generated.

0
4. A submodule N of M is called a direct summand of M if M = N ⊕ N
0
for some submodule N of M . If K ⊂ N ⊂ M are submodules, show the

following.

(a) If N is a direct summand of M , N/K is a direct summand of M/K.

(b) If K is a direct summand of N and N is a direct summand of M ,

then K is a direct summand of M .

(c) If K is a direct summand of M , then K is a direct summand of

N . If further N/K is a direct summand of M/K, then N is a direct

summand of M .

P
5. If M = ⊕ Mα , show that M is finitely generated if and only if Mα = 0

for all but finite α and each Mα is finitely generated.

14
Shiv Datt Kumar Modules

6. Let N and K be submodules of M with I = Ann(N ) and J = Ann(K).

Show that Ann(N ∩ K) ⊃ I + J and give an example to show that the

inclusion can be strict inclusion.

7. let N and K be submodules of M . Define quotient (N : K) = {a ∈

R|aK ⊂ N }. Show that

(i) (N : K) = Ann( N N
+K
)

(ii) (N : Rx)x = N ∩ Rx, x ∈ M .

8. Show that any R-module M can be expressed as the quotient of a free

R-module F . If M is finitely generated, show that F can be chosen to be

free of finite rank.

9. Let F be a free module with basis e1 , e2 , . . . , en . Show that a necessary

P
condition for a ∈ F with a = ai ei to be a part of another basis of F
P
is that a is unimodular, i.e. bi ai = 1, for some bi ∈ R. Deduce that a

necessary condition for a vector (a1 , a2 , . . . , an ) ∈ Rn to be completed to

a non-singular matrix over R is that it is unimodular.

10. Every sub module of a Noetherian module is a Noetherian.

11. Homomorphic image of a Noetherian module is a Noetherian.

12. Let M1 , M2 , . . . , Mr be modules over R. Then M = M1 × M2 × . . . × Mr

is Noetherian iff Mi is Noetherian.

13. Let R be a Noetherian ring. Then every finitely generated module over R

is Noetherian.

15
Shiv Datt Kumar Projective Modules

16
Chapter 2

Projective Modules

2.1 Localization

Projective modules are free modules over local rings. Local rings are obtained

quite often by the process of localization.

Definition 2.1.1. Let R be a ring with 1. A subset S ⊂ R is called a multi-

plicative closed in R if

1. 0 ∈
/S

2. 1 ∈ S

3. ab ∈ S, ∀a, b ∈ S

Example 2.1.2. • Let R be an Integral domain. Then S = R − {0}.

• If p is prime ideal, then S = R − {p}

S
• If {pi } is a set of prime ideals of R, then S = R − pi .

Consider the set M ×S. Define ∼ on M ×S by (x, s) ∼ (y, t) ⇔ s1 (tx−sy) =

0, for some s1 ∈ S. Note that ∼ is an equivalence relation. The set of equivalence

x
classes is denoted by MS = { | x ∈ M, s ∈ S} = S −1 M If S = R − p, then
s
MS ≡ Mp . Define addition and multiplication in MS as follows: xs + yt =
xt + ys x y xy
st and s.t = st . Then Ms is an Rs -module. Also

17
Shiv Datt Kumar Projective Modules

1. Rs ⊗R M ∼
= Ms .

2. Let f : R → Rs be a ring homomorphism and the natural map f is defined

x
by f (x) =
1

2.2 Exact Sequences

Definition 2.2.1. A sequence of R-modules and R-homomorphisms of the type

f1 f2 fi fi+1 fi+2 fn fn+1

M0 −→ M1 −→ M2 . . . −→ Mi −−−→ Mi+1 −−−→ . . . −→ Mn −−−→ . . .

is called an exact sequence if Imagefi = Kerfi+1 , 0 ≤ i ≤ n.

An exact sequence of the type

f g
0 → M1 −
→ M2 −
→ M3 → 0

is called a short exact sequence, if

(i) f is injective

(ii) g is surjective

(iii) Image (f ) = Ker (g) i.e. gof = 0.

Example 2.2.2. : (i) The sequence

i p
0 → 2Z →
− Z−
→ Z2 → 0

is an exact sequence, where i is the inclusion map and p is the projection of

Z onto Z2 = Z/2Z.

(ii) For any two R-modules M and N , the sequence

i p
0→M →
− M ⊕N −
→ N → 0,

is exact, where i is the inclusion map given by i(x) = (x, 0) and p is the projec-

tion given by p(x, y) = y, x ∈ M , y ∈ N .

0 f g 00
Definition 2.2.3. An exact sequence 0 → M −
→M −
→ M −→ 0 of R-module
00
splits, if there exits an R-homomorphism t : M −→ M such that gt = IM 00 ,
00
the identity map on M .

18
Shiv Datt Kumar Projective Modules

Example 2.2.4. : Sequence

i p
0→M →
− M ⊕N −
→ N → 0,

is a split exact sequence with the splitting map t : N −→ M ⊕ N given by

t(y) = (0, y).

0 f g 00
Proposition 2.2.1. If 0 −→ M −
→M −
→M −→ 0 is a split exact sequence,
0 00
then M ' M ⊕ M .

00
Proof. Let t : M −→ M be a splitting so that gt = IM 00 . This implies that
00
t is injective for if t(x) = 0, x ∈ M , then x = gt(x) = 0. If x ∈ M , then
00
x = tg(x) + x − tg(x) ∈ t(M ) + Kerg, as g(x − tg(x)) = 0.
00 00
Moreover if y ∈ f (M ) ∩ Kerg, then y = t(z), then y = t(z) , z ∈ M and
00
z = gt(z) = g(y) = 0. Hence y = t(z) = 0. Hence M = t(M ) ⊕ Kerg =
00 00 0
t(M ) ⊕ Imf ' M ⊕ M as f and t are injective maps.

0 f g 00
Corollary 2.2.2. Let 0 −→ M −
→ M −
→ M −→ 0 be an exact sequence of
0
R-modules which splits. Then there exists an R-homomorphism s : M −→ M

such that sf = IM 0 .

00 0 00
Proof. Let t : M −→ M be a splitting, then we have M = f (M ) ⊕ t(M ).
0
Take π1 to be the projection of M onto f (M ), and let s = f −1 π −1 . Then

sf = IM 0 .

2.3 Module HomR (M, N )

Let M and N be R modules. Then the set HomR (M, N ) of all R-module homo-

morphisms from M to N has a natural R-module structure for the operations,

(f + g)(x) = f (x) + g(x), x ∈ M, f, g ∈ S.

(a.f )(x) = af (x), a ∈ R, x ∈ M , f ∈ S.

We want to see how this module changes when either of the two modules M

or N is fixed and the other module changes.

19
Shiv Datt Kumar Projective Modules

0
Let M be a fixed R-module. Any homomorphism f : N −→ N of R-
0
modules, induces a homomorphism f∗ : HomR (M, N ) −→ HomR (M, N ) given
0 00
by f∗ (α) = f oα, α ∈ HomR (M, N ). Then (gf )∗ = g∗ f∗ , g ∈ HomR (N, N )

and I∗ : Hom(M, N ) → HomR (M, N ), the identity map. Similarly, for any
0
fixed module M and a homomorphism h : N −→ N , there exists an R-module
0
homomorphism h∗ : HomR (N, M ) −→ HomR (N , M ) given by h∗ (β) = βoh,
00
β ∈ HomR (N, M ). Then (gh)∗ = h∗ g ∗ , where g ∈ HomR (N, N ) and I ∗ = Id,

the identity map.

Proposition 2.3.1. For any given R-module M and an exact sequence

0 f g 00
0 −→ N −
→N −
→ N −→ 0,

the induced sequence

0 f∗ g∗ 00
0 −→ HomR (M, N ) −→ HomR (M, N ) −→ HomR (M, N )

is exact.

0
Proof. Clearly f∗ is injective, for if f∗ (α) = 0, α ∈ HomR (M, N ), then f α = 0.

This implies α = 0, for f is injective. Moreover g∗ f∗ = (gf )∗ = 0∗ as gf = 0.

Thus Im(f∗ ) ⊂ Ker(g∗ ). Conversely, let β ∈ Kerg∗ so that g∗ (β) = gβ = 0.

For any x ∈ M , gβ(x) = 0, then β(x) ∈ Ker(g) = image(f ). Thus β(x) = f (y)
0 0
for some unique y ∈ N , for f is injective. This defines a map α : M −→ N by

setting α(x) = y where β(x) = f (y). Clearly α is a homomorphism and f oα = β

i.e. f∗ (α) = β. Thus β ∈ Image(f∗ ). Hence Image (f∗ ) = Ker (g∗ ).

fm ν
Example 2.3.1. Sequence 0 → Z −−→ Z −
→ Zm → 0 is exact but
ν∗ fm ∗
0 → Hom(Zm , Z) −→ Hom(Z, Z) −−→ Hom(Z, Z) → 0 not exact, for fm ∗ is

not surjective for m > 1, for fm ∗ (φ) = (φofm ). Then φofm (x) = φ(m.x) =

mφ(x) 6= 1, if m > 1.

Example 2.3.2. It is not in general true that f ∗ is surjective. Consider

f g
the exact sequence 0 −→ Z −
→ Q −
→ Q/Z −→ 0 and take N = Z. Then
f∗
HomZ (Q, Z) −→ HomZ (Z, Z) is not surjective as the identity map IZ has no

pre-image, for Z is not a direct summand of Q.

20
Shiv Datt Kumar Projective Modules

Proposition 2.3.2. Let N be an R-module and

0 f g 00
0 −→ M −
→M −
→ M −→ 0,

be an exact sequence of R-modules. Then the induced sequence

00 g∗ f∗ 0
0 −→ HomR (M , N ) −→ HomR (M, N ) −→ HomR (M , N )

is exact.

00
Proof. Let α ∈ HomR (M , N ) be such that g ∗ (α) = 0, Then αog = 0. Since

g is surjective (due to exactness), α = 0. Thus g ∗ is injective. Now gof = 0

(given). Then f ∗ g ∗ = (gof )∗ = 0∗ = 0. Therefore Image (g ∗ ) ⊆ Ker (f ∗ ).

Let α ∈ Ker (f ∗ ). Then f ∗ (α) = 0. This implies αof = 0. So ker(α) ⊇

image(f ) = Ker (g). Then by fundamental theorem of homomorphism ∃

M
unique homomorphism α : −→ N such that αoν = α. Also since g is
Ker (g)
M 00
surjective, we have an isomorphism g : −→ M such that goν = g.
Ker g
Then g ∗ (αog −1 ) = αog −1 og = αoν = α. Thus α ∈ Image (g ∗ ). Hence

Ker (f ∗ ) = Image (g ∗ ).

Remark 2.3.1. Hom is not right exact functor.

Example 2.3.3. In general g∗ is not surjective. Consider the exact sequence

f g
0 −→ 2Z −
→ Z −
→ Z2 −→ 0 and take M = Z2 . Then g∗ : HomZ (Z2 , Z) −→

HomZ (Z2 , Z2 ) is not surjective, for HomZ (Z2 , Z) = 0.

Example 2.3.4. Sequence 0 → Z → Z → Zm → 0 is exact but

0 → Hom(Zm , Z) → Hom(Zm , Z) → Hom(Zm , Zm ) → 0

is not exact for Hom(Zm , Z) = 0 and Hom(Zm , Zm ) ∼

= Zm .

21
 Shiv Datt Kumar Projective Modules

Lemma 2.3.3. (Five Lemma) Consider the commutative diagram

2.
1. Proof. Suppose f3 (m) = 0. Then f4 α3 (m) = β3 f3 (m) = β3 (0) = 0. Since f4 is

injective, α3 (m) = 0 ⇒ m ∈ Ker(α3 ) = Image(α2 ). Thus ∃m2 ∈ M2 . Then

α2 (m2 ) = m. Further 0 = f3 (m) = f3 (α2 (m2 )) = β2 (f2 (m)) by commutativity

of the diagram. This implies f2 (m2 ) ∈ ker(β2 ) = image (β1 ) (Exactness). Then

∃n1 ∈ N1 such that β1 (n1 ) = f2 (m2 ). Since f1 is surjective, ∃ m1 ∈ M1 such

that f1 (m1 ) = n1 . Now f2 (α1 (m1 )) = β1 (f1 (m1 )) (commutativity of diagram).

Thus f2 (α1 (m1 )) = β1 (n1 ) = f2 (m2 ). Since f2 is injective α1 (m1 ) = m2 .

But we already have α2 (m2 ) = m. Hence m = α2 (α1 (m1 )) = 0 for image

α1 = kerα2 . Therefore f3 is injective.

Let n ∈ N3 , then β3 (n) ∈ N4 . Since f4 is onto, ∃m4 ∈ M4 such that f4 (m4 ) =

β3 (n). Now f5 (α4 (m4 )) = β4 (f4 (m4 )) = β4 (β3 (n)) = 0. Since f5 is injective

α4 (m4 ) = 0, then m4 ∈ ker(α4 ) = Image(α3 ). Hence ∃m3 ∈ M3 such that

α3 (m3 ) = m4 . Since β3 (f3 (m3 )) = f4 (α3 (m3 )) = f4 (m4 ) = β3 (n). Thus

β3 (n − f3 (m3 )) = 0. Thus n − f3 (m3 ) ∈ ker(β3 ) = Image(β2 ), then ∃ n2 ∈ N2

such that β2 (n2 ) = n − f3 (m3 ). Since f2 is surjective ∃m2 ∈ M2 such that

f2 (m2 ) = n2 . Now n − f3 (m3 ) = β2 (n2 ) = β2 (f2 (m2 )) = f3 (α2 (m2 )) i.e n =

f3 (m3 + α2 (m)). Take m = m3 + α2 (m2 ). Then f3 (m) = n i.e f3 is surjective.

α β
Proposition 2.3.4. A short exact sequence 0 → M1 −
→ M2 −
→ M3 → 0 is split

exact iff ∃ s : M2 → M1 such that soα = IM1 .

Proof. Suppose t : M3 → M2 such that β o t = IM3 . Let x ∈ M2 . Consider

β(x − tβ(x)) = β(x) − β(x) = 0. Then x − tβ(x) ∈ ker(β) = Image(α). Since α

is injective, ∃ unique element s(x) (say) in M1 such that α(s(x)) = x−tβ(x) is a

homomorphism, then there exists y ∈ M1 such that α(y) = x. Also α(s(α(y)) =

α(y) − tβ(α(y)) = α(y), for βα = 0. Then soα = IM1 .

α β
Proposition 2.3.5. Let 0 → M1 −
→ M2 −
→ M3 → 0 be an split exact sequence.

22
Shiv Datt Kumar Projective Modules

Then ∃ an isomorphism f2 : M2 → M1 ⊕ M3 such that the diagram

α β
0 −−−−→ M1 −−−−→ M2 −−−−→ M3 −−−−→ 0
    
 f1  f2  f3  
y y y y y
i p2
0 −−−−→ M1 −−−−→ M1 ⊕ M3 −−−−→ M3 −−−−→ 0
is commutative.

2.4 Universal problem

Let R be a commutative ring with identity, S a non empty set. Does ∃ a pair

(M, i) such that for any pair (N, j) ∃ a map unique homomorphism φ : M → N

such that the diagram

S
i /M
j
φ

N
is commutative ?

Solution of this problem exists and is unique.

Solution: Consider F (S) = {f : S → R|f (x) = 0 for all but finitely many

elements }.

Define (f + g)(x) = f (x) + g(x) and αf (x) = α(f (x)), α ∈ R. Then F (S) is an

R-module. Define i : S → F (S) by


 0
 if x 6= y
i(x)(y) =

 1 if x = y,

J = {x1 . . . xn } finite subset such that f (S − J) = 0 and φoi(x) = j(x).

Then (F (S), i) is a solution of the problem (verify).

Solution of the universal problem F (S) is called free module on S.

Proposition 2.4.1. Every module is isomorphic to quotient of a free module.

Proof. Let M be an R-modules and F (M ) be a free module on M . Then

consider IM : M → M identity map. Hence by the universal property of

(F (M ), i), ∃ unique φ : F (M ) → M such that

23
Shiv Datt Kumar Projective Modules

M
i / F (M )
IM
φ
" 
M
commutes i.e. φoi = I. Hence φ is a surjective homomorphism from F (M ) to
F (M ) ∼
M . Then by fundamental theorem of homomorphism = M.
Ker φ

2.5 Projective modules

Certain types of modules P have the property that for any surjective homo-
00
morphism g : M −→ M , the induced homomorphism g ∗ : HomR (P, M ) −→
00
HomR (P, M ) is surjective. These are the projective modules defined as follows.

Definition 2.5.1. Let M , N be A-modules. Then A-module P is called pro-

jective if given a surjective A-linear map f : M −→ N and a A-linear map

g : P −→ N there is an A-linear map h : P −→ M such that the following

diagram

P
h
g
~ 
M
f
/N /0

commutes.

Example 2.5.2. 1. Let A be a commutative Noetherian ring and R = A × A.

Then P = {0} × A is a projective R-module, which is not free.

2. Every Projective module over principal ideal domain is a free module. If

R is a principal ideal domain, then any projective R-module is free, as any

submodule of a free R-module is free. In particular, any projective module

over K[X], K field, is free. In general every finitely generated projective

module over K[X1 , . . . , Xn ], where K is a P.I.D., is free.

3. Let R = Z6 , P = {0, 2, 4} and Q = {0, 3}, ideal of R. Then R = P ⊕ Q,

so that P is R-projective. Now P is not R-free as any free module of rank

n over R = Z6 is isomorphic to Rn , which has 6n elements.

24
Shiv Datt Kumar Projective Modules

Following is a classical example of a projective module which is not free.

Example 2.5.3. Let A = R[X1 , X2 , X3 ]/(X1 2 + X2 2 + X3 2 − 1) be the algebraic

coordinate ring of sphere S 2 . Let xi denote the images of Xi under the natural

map, i = 1, 2, 3. Consider the A- homomorphism f : A3 −→ A defined by

f ((u, v, w)) = ux1 + vx2 + wx3 . Since X1 2 + X2 2 + X3 2 = 1, f is surjective.

Ker(f ) = {(u, v, w) ∈ A3 |x1 u + x2 v + x3 w = 0}. Clearly 0 −→ Ker(f ) −→

A3 −→ A −→ 0, is a split exact and Ker(f ) ⊕ A = A3 . Therefore Ker(f ) is

a projective A-module. This module can be considered as module of sections of

tangent bundle T (S 2 ) = {(x1 , x2 , x3 , u, v, w) ∈ S 2 × A3 |x1 u + x2 v + x3 w = 0}

over S 2 . This is not free for otherwise T (S 2 ) would become trivial vector bundle.

It is well known that T (S n ) is trivial if and only if n = 1, 3 or 7.

Axiom of Choice: Let {Xα : α ∈ I} be a family of non empty sets. Then

there exists a map C : I → ∪α∈I Xα such that C(α) ∈ Xα , for all α.

Proposition 2.5.1. Every free module is projective.

Proof. Let (F (S), i) be a free module on S. Consider the diagram

S
i / F (S)
j
f
" 
M
β
/N /0

Since β is onto, β −1 (f oi)(x) 6= φ, ∀x ∈ S. Hence by axiom of choice there

exists a map C : S → M such that C(x) = β −1 (f oi(x)) ⇒ βoC = f oi. Since

(F (S), i) is a free module on S, ∃ unique homomorphic φ : F (S) −→ M such

that φoi = C. Now βo(φoi) = βoC = f oi. Since (F (S), i) is a free module and

i is injective Therefore βoφ = f . Therefore F (S) is a projective module.

Theorem 2.5.2. Let M , N and P be A-modules. Then the following conditions

are equivalent :

1. Module P is projective.

25
Shiv Datt Kumar Projective Modules

2. Every exact sequence

0 −→ N −→ M −→ P −→ 0, of A-linear maps splits.

3. P is a direct summand of a free module i.e. there is an A-module Q such

that P ⊕ Q is free.

α β
Proof. (1) ⇒ (2) Suppose P is projective and 0 → M −
→N −
→ P → 0 exact.

Since IP : P → P is a homomorphism and P is projective, therefore ∃ t : P → N

such that βot = IP i.e t is a splitting.

(2) ⇒ (3) Suppose every exact sequence of the above type splits. Then

0 → Ker(β) → F (P ) → P → 0 splits. Then by using Five lemma ∃ isomor-

phism f : F (P ) → ker(β) ⊕ P such that the diagram

0 −−−−→ ker(β) −−−−→ F (P ) −−−−→ P −−−−→ 0

    
 Id  f Id  
y y y y y
0 −−−−→ ker(β) −−−−→ ker(β) ⊕ P −−−−→ P −−−−→ 0

is commutative and f is an isomorphism. Thus that F (P ) ∼

= Ker (β) ⊕ P
i.e. P is direct summand of a free module.

(3) ⇒ (1) Suppose F = P ⊕ Q. Let p1 : F −→ P defined by p1 (x, y) = x

β
and i1 : P −→ F as i1 (x) = (x, 0). Let M −
→ N → 0 be exact and f :

P → N be a homomorphism. Since F (being free) is projective. There exists a

homomorphism φ : F → M such that diagram

F =P ⊕Q /P

φ f
 
M
β
/N /0

commutes. Then βo(φoi1 ) = f . Thus P is projective.

Q
Proposition 2.5.3. Module P = ⊕ Pα is projective iff each Pα is projective.

26
Shiv Datt Kumar Projective Modules

Example 2.5.4. Submodule of a projective module need not be projective. For

example, A = {0, 2} is a submodule of Z4 but A is not projective module over

Z4 .

Example 2.5.5. Quotient of a projective module need not be projective.

Proposition 2.5.4. An R-module P is projective if and only if for any surjective

0
homomorphism g : M −→ M , the induced homomorphism g ∗ : HomR (P, M ) −→
0
HomR (P, M ) is surjective.

Proof. First suppose P is free, with basis {ei |i ∈ I(indexing set)}. Given α ∈
0
HomR (P, M ), let α(ei ) = xi . Since g is surjective choose yi ∈ M with g(yi ) =

xi . Then the map β : P −→ M , defined by β(ei ) = yi , i ∈ I can be extended

to an R-linear map β by defining β( ai ei ) = ai yi . Clearly g ∗ (β) = α. This

P P

shows that g ∗ is surjective. If P is projective, there exist an R-module Q with

0
P ⊕Q = F , free. Let π : F −→ P the projection map. Given α ∈ HomR (P, M ),
0
consider απ ∈ HomR (F, M ). There exists β ∈ HomR (F, M ) with g ∗ (β) = απ.

Then β1 = β|P : P −→ M has the property that g ∗ (β1 ) = α. Hence g ∗ is

surjective.

Conversely, suppose P has property stated in the proposition. Express P as

a quotient of a free module F . Hence there is a surjective map g : F −→ P .

Then the induced map g ∗ : HomR (P, F ) −→ HomR (P, P ) is surjective. In

particular, there exists β ∈ HomR (P, F ) with g ∗ (β) = gβ = IP . Therefore the

g
exact sequence 0 −→ Kerg −→ F −
→ P −→ 0 splits. Then P is isomorphic to

a direct summand of F . Hence P is projective.

Proposition 2.5.5. An R-module P is projective if and only if every exact se-

0 f g
quence 0 −→ M −
→M −
→ P −→ 0 splits.

Proof. If P is projective, the identity map IP : P −→ P can be lifted to a map

φ : P −→ M such that gφ = IP , i.e. the exact sequence

0
0 −→ M −→ M −→ P −→ 0 splits. To prove the converse, express

P as the quotient of a free module F with Kernel K so that the sequence

27
Shiv Datt Kumar Projective Modules

0 −→ K −→ F −→ P −→ 0 is exact. By assumption, this sequence splits so

that P is a direct summand of F and hence projective.

Corollary 2.5.6. Projective module P is finitely generated iff P is a direct sum-

mand of a free module of finite rank.

Remark 2.5.1. Now let A be any ring and M a finite projective A-module. For

each p ∈ Spec(A), the localization Mp is a free module over Ap and we write

n(p) for its rank. Then n is a function from Spec(A) to Z, and is constant on

every connected component (since n(p) = n(q) if p ⊃ q). This function n is

called the rank of M . If the rank is constant r over the whole Spec(A), then

we say that M is a projective module of rank r.

Note that rank is a continuous function (with the discrete topology on Z and

Zariski topology on Spec(A)). Moreover, rank is a constant function for every

finitely generated projective A-module P if A has no non-trivial idempotent

elements.

2.6 Tensor Products

Let M , N , K be R-modules. A map θ : M × N −→ K is R-bilinear, if

1. θ(a1 x1 + a2 x2 , y) = a1 θ(x1 , y) + a2 θ(x2 , y), for all x1 , x2 ∈ M , a1 , a2 ∈ R.

2. θ(x, b1 y1 + b2 y2 ) = b1 θ(x, y1 ) + b2 θ(x, y2 ), for all y1 , y2 ∈ N , b1 , b2 ∈ R i.e.

θ is R-linear in x, for each fixed y ∈ N and R-linear in y for each fixed

x ∈ M.

The study of bilinear maps from M × N to K can be reduced to the study

of linear maps on a suitably defined module (tensor product of M and N ) with

values in K.

Definition 2.6.1. Let M and N be R-modules. Then tensor product of M and N

is a pair (T, θ) where T is an R-module and θ : M × N −→ T an R-bilinear map

with the property that for any R-module K and R-bilinear map f : M ×N −→ K,

there exists a unique R-linear map fe : T −→ K with feθ = f .

28
Shiv Datt Kumar Projective Modules

Theorem 2.6.1. : The tensor product of modules M and N exists and it is

unique up to isomorphism.

Proof. Uniqueness: Let (T, θ) and (T 0 , θ0 ) be two pairs which satisfy the con-

dition of tensor product. Consider the following diagrams:

θ /T θ0 / T0
M ×N and M ×N
φ ψ
θ0
# 0 # 
θ
T T
By definition of the tensor product there exist unique R-linear maps φ and ψ

such that φθ = θ0 and ψθ0 = θ. Now consider the following diagrams:

M ×N
θ /T and M ×N
θ /T

ψφ IT
#  # 
θ θ
T T
Since (ψφ)θ = ψ(φθ) = ψθ0 = θ and IT θ = θ, we have the ψφ = IT . Similarly

φψ = IT 0 . Hence T ∼
= T 0.
Existence: Consider free left R-module (F (M × N ), i) on M × N , where i :

M × N −→ F (M × N ) is a map. Let L be a submodule of F (M × N ) generated

by the elements of the types

(i) i(x + y, u) − i(x, y) − i(y, u)

(ii) i(x, u + v) − i(x, u) − i(x, v)

(iii) i(xa, u) − i(x, au).

Let T = F (M × N )/L and θ = νoi. The image θ(m, n) = i(m, n) + A

is denoted m ⊗ n. The abelian group T denoted by M ⊗R N is called tensor

product of M and N over R.

Properties of tensor products:

Proposition 2.6.2. Let R be commutative ring with identity and M , N , L be

R-modules. Then

1. M ⊗R N ∼
= N ⊗R M , m ⊗ n 7→ n ⊗ m.

2. R ⊗R M ∼
= M , a ⊗ m 7→ am.

29
Shiv Datt Kumar Projective Modules

3. M ⊗R (N ⊗ L) = (M ⊗R N ) ⊗R L.

4. M ⊗R (N + L) ∼
= M ⊗R N + M ⊗R L.

5. Tensor is a right exact functor.

Example 2.6.3. 1. Z3 ⊗ Z5 = {0}.

2. Zm ⊗ Zn ∼ ¯ where d = (a, b).

= Zd , (ā, b̄) 7→ ab mod d = ab,

3. Q ⊗Z Z ∼
= Q.

2.7 Exercises

1. If M is an R-module. Then HomR (R, M ) ∼

= M for f → f (1) is an
isomorphism.

2. Define projective module. By an example show that submodule of a pro-

jective module need not be projective.

α β
3. Let M1 −→ M2 −→ M3 −→ 0 be an exact sequence of R - modules and
β∗ α∗
N be a R - module. Then 0 −→ HomR (M3 , N ) −→ HomR (M2 , N ) −→

HomR (M1 , N ) is exact.

α β
4. Let 0 −→ M1 −→ M2 −→ M3 be an exact sequence of R - modules and
β∗ ∗ α
N be a R - module. Then 0 −→ HomR (N, M1 ) −→ HomR (N, M2 ) −→

HomR (N, M3 ) is exact.

5. An R- module is projective iff every short exact sequence 0 −→ M −→

N −→ P −→ 0 splits.

6. Direct summand of a projective module is projective.

7. Let φ : M −→ F be a surjective homomorphism of a finitely generated

module M onto a free module F . Show that Ker(φ) is finitely generated.

8. Let n be a positive integer with n = d1 d2 . Consider the sequence given by

β α
0 → d2 Zn −
→ Zn −
→ d1 Zn → 0. Where β is the inclusion map and (α) is

30
Shiv Datt Kumar Projective Modules

the multiplication by d1 . Show that the sequence is exact. Show further

that the sequence splits if and only if d1 and d2 are relatively prime.

9. For any R-module M , show that

(a) HomR (R, M ) ' M .

P P
(b) HomR (⊕ Mi , N ) ' ⊕ HomR (Mi , N ).
P P
(c) HomR (M, ⊕ Ni ) ' ⊕ HomR (M, Ni ).

10. Show that if P is a projective module, then there exists a free module F

with P ⊕ F free.

11. Let 0 → M1 → M2 → M3 → 0 be an exact sequence of R-modules and N

be an R-module. Then

0 → Hom(M3 , N ) → HomR (M2 , N ) → HomR (M, N ) and

0 → Hom(N, M1 ) → HomR (N, M2 ) → Hom(N, M3 ) are exact.

12. An element e ∈ R is called idempotent if e2 = e. Show that if every

projective R-module is free, then the only idempotents of R are 0 and 1.

13. Let M be an R-module. An element x ∈ M is called unimodular if the

cyclic submodule Rx is a free direct summand of M . Show that x ∈ M is

unimodular if and only if there exists an R homomorphism φ : M −→ R

with φ(x) = 1. In particular if M is free with basis ei , 1 ≤ i ≤ n,

P
x= ai ei is unimodular if and only if (a1 , . . . , an ) is a unimodular row.

14. A finitely generated projective R-module P is said to be stably free if

P ⊕ Rm ' Rn for some m, n. Show that every stably free R-module is

free iff every unimodular row over R can be completed to a non-singular

matrix.

[If R = k[X1 , . . . , Xn ], k field, every finitely generated projective R-module

is free. See proof of Serre’s conjecture by D. Quillen and A. A. Suslin in

the Lecture Notes in Mathematics, No. 635, Springer Verlag]

31