Chapter 1: Further Theory of Modules

In this Chapter, we assume that R is a ring with identity 1 6= 0 and M

is a unita right R-module.

§5. Injective Modules.

5.1. Definitions and Examples.

If we reverse all arrows in the homomorphism diagram that defines a

projective module, we obtain the dual notion, an injective module. In the
diagram below, the top row is exact, that is, f is a monomorphism.
f
0 −→ N → M

g↓ .h

5.1.1. Definition. The right R-module E is injective if given g : N −→ E, and

f : N −→ M monomorphism, there exists a homomorphism h : M −→ E (not
necessarily unique) such that g = hf, i.e., the above diagram is commutative.
We sometimes say that we have “lifted” g to h, or h is an “extension” of g.
5.1.2. Examples. 1. Let consider ZZ as an ZZ-module. Then f : 2ZZ −→ ZZ
defined by f (2n) = n, ∀n ∈ ZZ clearly cannot be extended to a homomorphism
f 0 : ZZ −→ ZZ. So ZZ-module ZZ isn’t injective.

 of 2× 2 upper triangular matrices over a ring k 6= 0.

2. Let R be the ring
0 k
Consider the ideal I = and f : I −→ R defined by
0 0
   
0 a 0 0
f( )= .
0 0 0 a

This is easily checked to be a right

 R-homomorphism.
 If f can be extended to
x y
R, there would exist a matrix such that
0 z
       
0 a x y 0 a 0 xa
f( )= . = (∀a ∈ k),
0 0 0 z 0 0 0 0

Typeset by AMS-TEX

1
2

which is clear impossible. This so that RR isn’t injective.

The above examples are noninjective cases, we will present the injective
cases below.

5.2. CHARACTERIZATIONS AND PROPERTIES.

As with projectives, there are several equivalent ways to characterize an

injective module.
5.2.1. Theorem. The following conditions on the right R-module E
are equivalent.
(1) E is injective.
(2) If the sequence of right R-modules and right R-module homomor-
f g
phisms 0 −→ N −→ M −→ L −→ 0 is exact, then the following sequence of
ZZ-modules is exact
Hom(g,1) HomR (f,1)
0 −→ HomR (L, E) −→ HomR (M, E) −→ HomR (N, E) −→ 0.

(3) Every exact sequence 0 −→ E −→ M −→ N −→ 0 splits.

Proof. (1) is equivalent to (2). In view of the exactness of the sec-
ond sequence, by Proposition 3.4.2, we only need to prove the exactness of it
at HomR (N, E), i.e., it says that we are to show that if f : N −→ M is a
monomorphism, then HomR (f, 1) : HomR (M, E) −→ HomR (N, E) is an epi-
morphism. But HomR (f, 1)(h) = hf, so given g : N −→ E, we must produce
h : M −→ E such that g = hf. This is precisely the definition of injectivity.
(2) implies (3). Let 0 −→ E −→ M −→ N −→ 0 be a short exact se-
quence, with f : E −→ M (necessarily monomorphism). Since E is an injective
module, we have the following diagram:
f
0 −→ E → M

1E ↓ .g

Thus there exists g : M −→ E such that gf = 1E , which means that the

exact sequence splits.
(3) implies (1). Given g : N −→ E, and f : N −→ M monomorphic, we
form the following diagram
 3

f
N → M

g↓ ↓ g0
f0
E → Q
in which Q = (E ⊕ M)/U, U = {(g(n), −f (n))|n ∈ N } and

f0 : E −→ Q
x 7−→ (x, 0) + U

g0 : M −→ Q
m 7−→ (0, m) + U
Then it can easily prove that U ≤ Q, f 0 , g 0 are homomorphisms and the square
is commutative.
Since f is monomorphic, so is f 0 . Indeed, let f 0 (x) = (x, 0) + U = 0.
Then (x, 0) ∈ U, so there exists n ∈ N such that x = g(n), −f (n) = 0. So n = 0
and hence x = 0.
Thus the sequence

f0
0 −→ E → Q → Q/Imf 0 −→ 0

is exact. By (3), there exists h : Q −→ E such that hf 0 = 1E . We now have

hg 0 : M −→ E with hg 0 f = hf 0 g = 1E g = g, proving that E is injective.
We proved in Section 4 that a direct sum of modules is projective iff each
component is projective. The dual result holds for injectives.
5.2.2. Proposition. A direct product
Y
Ej
i∈J

of right R-modules is injective iff each Ej , j ∈ J is injective. Consequently, a

finite direct sum is injective iff each summand is injective.
Proof. “ ⇒ ”. Fix i ∈ J . Let f : N −→ M Q be a monomorphism and
homomorphism g : N −→ Ei and let σηi g : N −→ Ei . Then there exists a
Q i∈J
homomorphism k : M −→ Ei such that kf = σηi g. Take h = πi k then h is
i∈J
a desired homomorphism.
4

Q
“ ⇐ ”. If f : N −→ M is injective and g : N −→ i Ei , let gi = πi g,
where πi is the projection of the direct product on Ei . Then finding h : M −→
Q
i Ei such that g = hf is equivalent to finding, for each i, a homomorphism
hi : M −→ Ei such that gi = hi f. [If pi g = hi f = pi hf for every i, then g = hf
by the uniqueness part of the universal mapping property for products.] The
last assertion holds because the direct sum of finitely many modules coincides
with the direct product.
In checking whether a right R-module E is injective, we are given g :
N −→ E, and f : N −→ M, with f monomorphism, and we must lift g to
h : M −→ E with g = hf. The next result drastically reduces the collection
of homomophisms f and g that must be examined. We may take M = R and
restrict N to a right ideal I of R, with f the inclusion homomorphism.
5.2.3. Theorem (Baer’s Criterion). The right R-module Q is injective if
and only if every R-homomorphism f : U −→ Q, where U is a right ideal of R,
can be extended to an R-homomorphism h : R −→ Q.
Proof. The “if” part is clear.
The “only if” part follows in two steps.
Step 1: Let α : A −→ B be a monomorphism and let ϕ ∈ HomR (A, Q). Let
C < B, C 6= B with Imα ≤ C and let γ : C −→ Q with ϕ(a) = γα(a), ∀a ∈ A.
We will prove that there is a submodule C1 ≤ B, C < C1 , C 6= C1 and a
homomorphism γ1 : C1 −→ Q such that γ1 |C = γ (hence also ϕ(a) = γ1 α(a).)
Indeed, let b ∈ B, b 6∈ C. Put C1 = C + bR. We divide it in two cases:
Case 1: C ∩ bR = 0 then immediately we could extend γ trivially to C1 .
Case 2: C ∩ bR 6= 0. Let

U = {u ∈ R|bu ∈ C}.

Then it is clear that U is a right ideal of R and

ψ: U −→ C
u 7−→ bu

is an R-homomorphism. Let ρ = γψ : U −→ Q. By assumption there is a

homomorphism τ : R −→ Q such that ρ = τ ι :
ι
U −→ R
ψ↓
C .τ
γ↓
Q
 5

We now define

γ1 : C1 −→ Q
.
c + br 7−→ γ(c) + τ (r)

We need to prove that γ1 is a mapping. Indeed, let c + br = c1 + br1 , c, c1 ∈

C, r, r1 ∈ R, then c − c1 = b(r1 − r) ∈ C ∩ bR or r − r1 ∈ U ⇒ γψ(r −
r1 ) = τ (r − r1 ) ⇒ γ(c − c1 ) = γ(b(r1 − r)) = γψ(r1 − r) = τ (r1 − r). Hence
γ(c) + τ (r) = γ(c1 ) + τ (r1 ).
Since γ and τ are R-homomorphisms, γ1 is also an R-homomorphism
and γ1 |C = γ by the definitions of γ1 .
Step 2: Let C0 = Imα and let α0 be the isomorphism of the module A onto C0
induced by α. In addition let γ0 = ϕα−1 . Then we have ϕ(a) = γα(a), ∀a ∈ A.
The homomorphism γ0 is now extended to the whose of B by Step 1 and Zorn’s
Lemma. For this let Γ be the set of all pairs (C, γ) with Imα = C0 ≤ C ≤ B
and γ : C −→ Q with γ|C0 = γ0 . This set is not empty as (C0 , γ0 ) ∈ Γ. An
ordering is defined in Γ by:

(Cγ) ≤ (C1 , γ1 ) ⇔ C ≤ C1 and γ1 |C = γ

Now let Λ be a non-empty totally ordered subset of Γ and let

[
D= C
(C,γ)∈Λ

then we have C0 ≤ D ≤ B. Further let

δ: D Q
,
d 7−→ γ(d)

for d ∈ C with (c, γ) ∈ Λ. Then by the above condition δ is a homomorphism

and δ|C0 = γ0 . Consequently (D, δ) is an upper bound of Λ in Γ.
Hence by Zorn’s Lemma, there exists a maximal element in Γ, which,
from Step 1, must be equal to a (B, ν) with ϕ = να. This completes the proof.
Since free modules are projective, we can immediately produce many
examples of projective modules. The primary source of injective modules lies a
bit below the surface.
5.2.4. Definitions. Let R be an integral domain. The R-module M is divisible
if each y ∈ M can be divided by any nonzero element r ∈ R, that is, there exists
x ∈ M such that xr = y.
6

5.2.5. Examples. 1. The additive group of rational numbers is a divisible

abelian group, as is Q/Z
I Z, the rationals mod ZZ.
2. The quotient field of any integral domain (regarded as an abelian
group) is divisible, for example Q/Z
I Z.
3. A cyclic group of finite order n > 1 can never be divisible, since it is
not possible to divide by n. For example, ZZ4 isn’t a divisble ZZ-module.
4. The group of integers ZZ is not divisible since the only possible divisors
of an arbitrary integer are ±1. It follows that a nontrivial finitely generated
abelian group, a direct sum of cyclic groups, is not divisible.
5.2.6. Proposition. Homomorphic image of a divisible module is divisible,
hence a quotient or a direct summand of a divisible module is divisible. Also, a
direct sum of modules is divisible iff each component is divisible.
Proof. It follows from the definition.
5.2.7. Proposition. If R is any integral domain, then an injective R-module
is divisible. If R is a PID, then an R-module is injective if and only if it is
divisible.
Proof. Assume E is injective, and let y ∈ E, r ∈ R, r 6= 0. Let I be the ideal
rR, and define an R-homomorphism f : I −→ E by f (rt) = yt. If rt = 0, then
since R is an integral domain, t = 0 and f is well defined. By Theorem 5.2.3, f
has an extension to an R-homomorphism h : R −→ E. Thus

y = f (r) = h(r) = h(1r) = h(1)r

so division by r is possible and E is divisible.

Conversely, assume that R is a PID and E is divisible. Let f : I −→ E
be a homomorphism, where I is an ideal of R. Since R is a PID, I = rR for
some r ∈ R. We have no trouble extending the zero mapping, so assume r 6= 0.
Since E is divisible, there exists x ∈ E such that xr = f (r). Define h : R −→ E
by h(t) = xt, ∀t ∈ R. If for any v ∈ R, then

h(rv) = x(rv) = (xr)v = f (r)v = f (rv)

so h extends f , proving E injective.

5.2.8. Examples. 1. Q/Z
I Z is an injective ZZ-module.
2. The quotient field of any integral domain is an injective ZZ-module.
3. A cyclic group of finite order n > 1 can never be injective. For
example ZZ4 isn’t an injective ZZ-module. But however, note that we can use
 7

Baer’s criterion to prove that ZZ4 is an injective ZZ4 -module and it is leaved for
readers.
4. The group of integers ZZ is not injective ZZ-module. A nontrivial
finitely generated abelian group, a direct sum of cyclic groups, is not an injective
ZZ-module.