MATH RI: ADVANCED ALGEBRA I (RINGS AND IDEALS)

HEESUNG YANG

Abstract. This notes covers ring theory of the first half of Dalhousie’s algebra comprehen-
sive exam syllabus which is not covered in Advanced Algebra I (MATH 5045). This notes
will cover Chapters VII.1–VII.6 (Basic ring theory), and the first half of Chapter VIII.2
(Principal ideal domains) of Dummit & Foote. Some propositions and lemmas are from the
past comprehensive exams; the proofs of those lemmas and propositions are included in this
notes as well.

1. Chapter VII.1: Introduction to Rings

Definition 1.1. A ring R is a set with two binary operations called addition (+) and
multiplication (·) such that
(1) hR, +i is an abelian group
(2) · is associative (i.e., (a · b) · c = a · (b · c) for all a, b, c ∈ R)
(3) · and + are distributive over one another (i.e., a(b+c) = ab+ac and (a+b)c = ac+bc).
Definition 1.2. A ring R is commutative if ab = ba for all a, b ∈ R. Otherwise a ring R is
non-commutative. A ring R has a unity (or has an identity) if · has an identity, which we
call it 1 (i.e., 1 ∈ R and 1 · a = a for all a ∈ R). An element a ∈ R is a unit if there exist a
left multiplicative inverse a0 and a right multiplicative inverse a00 such that a0 a = aa00 = 1.
Example. Z, R, and Z[x] are examples of (commutative) rings. M2 (Z), the 2 × 2-matrix ring
over Z is a (non-commutative) ring.
Proposition 1.1. a0 = a00 . In other words, a left multiplicative inverse of a and a right
multiplicative inverse of a are the same.
Proof. a0 a = 1, so a0 aa00 = a00 . Thus a0 = a00 . 
Definition 1.3. A ring R with unity 1 6= 0 is a division ring if every non-zero element
a ∈ R has a multiplicative inverse, i.e., for any a ∈ R there is b ∈ R such that ab = ba = 1.
Therefore, a field is a commutative division ring.
Proposition 1.2 (Additional properties of a ring). Let R be a ring.
(1) 0a = a0 = 0 for all a ∈ R.
(2) (−a)b = a(−b) = −(ab) for all a, b ∈ R, where −a is the additive inverse of a.
(3) (−a)(−b) = ab for all a, b ∈ R.
(4) If R has unity 1, then the unity is unique and −a = (−1)a.
Definition 1.4. A non-zero element a ∈ R is a zero-divisor if there exists b 6= 0 ∈ R such
that ab = 0 or ba = 0. If R is commutative, has unity 1, and has no zero-divisors, then
R is an integral domain (or domain in short). A field is an integral domain in which every
non-zero element is a unit.
Date: 7 April 2019.
1
Example. Z is a commutative ring with unity 1 and units ±1. Z has no zero divisors. Thus Z
is an integral domain. On the other hand, Z/6Z has unity 1 and the units are 1, 5. However,
Z/6Z has three zero divisors, namely 2, 3, 4. Notice that 2 · 3 = 4 · 3 = 0. Therefore Z/6Z is
not an integral domain.
Example. Z/pZ for p prime, Q, R, C, C(x) are examples of fields.
Proposition 1.3. Units cannot be zero divisors.
Proof. Let a be a unit, and suppose that there exists a non-zero b such that ab = 0. Since a
is a unit, there exists c such that ca = 1. Thus we have b = 1b = (ca)b = c(ab) = c0 = 0,
but this contradicts the fact that b is non-zero. Thus a cannot be a left zero-divisor. Now
suppose that there is b0 such that b0 a = 0. Since a is a unit, we have ac = 1. (We use
the same c thanks to Proposition 1.1.) Then b0 = b0 1 = b0 (ac) = (b0 a)c = 0c = 0, which
again contradicts the fact that b0 is non-zero. Therefore no unit can be a zero divisor as
required. 
Proposition 1.4 (Cancellation property of integral domains). Let R be a ring, and let
a, b, c ∈ R such that a is not a zero divisor in R. If ab = ac, then either a = 0 or b = c.
Therefore, if R is an integral domain, and ab = ac, then we have a = 0 or b = c.
Proof. If ab = ac, then ab − ac = a(b − c) = 0 per the distributive law. However, since a is
not a zero divisor, we have a = 0 or b − c = 0, as required. The second statement follows
upon noting that integral domains cannot have zero divisors. 
Corollary 1.1. Any finite integral domain is a field.
Proof. Let R be a finite integral domain, and let a ∈ R \ {0}. Clearly the map x 7→ ax is an
injection due to the cancellation property. The map is also surjective as R is finite. Thus
x 7→ ax is a bijection, so there must exist an inverse map – thus there exists b ∈ R so that
ba = 1. Therefore R is a field. 
Theorem 1.1 (Wedderburn’s little theorem). Any finite division ring is a field.
Definition 1.5. A subring S of the ring R is a subgroup of R that is closed under multipli-
cation.
Example. Z is a subring of Q; Q is a subring of R.
Example. nZ for any n ∈ Z is a subring of Z. However, Z/nZ is not a subring of Z for any
n ≥ 2.
Example. The ring of all differentiable functions from R to R is a subring of the ring all
continuous functions from R to itself. Both of these rings are subrings of the ring of all
functions from R to itself.
Example. Let K be a number field (i.e., any finite-dimensional field extension over Q), and
let OK be the set of elements in K whose minimal polynomial is monic. Then OK is a
subring of K; more specifically, OK is called the ring of integers of K. (To see where this
name comes from, consider the rational integer counterpart: note that x − a = 0 is the
minimal polynomial of a for any a ∈ Z; for any b = p/q ∈ Q \ Z where gcd(p, q) = 1, the
minimal polynomial is qx − p = 0 and cannot be monic.) For more information, please refer
to the PMATH 641 (Algebraic number theory) course notes and the MATH 5055 (Galois
theory) course notes.
2
Proposition 1.5 (Subring test). Suppose R is a ring, and that S is a non-empty subset of
R. If a − b, ab ∈ S for any a, b ∈ S, then S is a subring of R.
2. Chapter VII.2: Some examples of rings
2.1. Polynomial rings
Definition 2.1. Let
R[x] := {an xn + · · · + a1 x + a0 : n ≥ 0, an 6= 0, ai ∈ R},
where R is a commutative ring with unity 1 6= 0. Then an xn + · · · + a1 x + a0 is said to be a
polynomial in x with coefficients in R. Then R[x] is the ring of polynomials (or polynomial
ring) in the variable x with coefficients in R, where addition and multiplication are the same
as the standard addition and multiplication.
Remark. Clearly R is a subring of R[x], and any element in R is a constant polynomial.
Since we assume that R is a commutative ring with unity 1 6= 0, the polynomial ring R[x] is
also a commutative ring with unity 1, where the unity is the same 1 from R.
Proposition 2.1. Let R be an integral domain, and let p(x), q(x) be non-zero elements of
R[x]. Then
(1) R[x] is an integral domain;
(2) deg pq = deg p + deg q; and
(3) the units of R[x] are precisely the units of R.
Proof. Let deg p = n and deg q = m. Thus the leading term of p(x) is an xn ; the leading
term of q(x) is bm xm . Thus the leading term of p(x)q(x) is an bm xn+m . Since R is an
integral domain, an bm 6= 0. (Note that R has no zero divisors, and neither an nor bm can be
zero.) This proves both the first part and the second part. For the third part, suppose that
p(x)q(x) = 1. By the second part we see deg pq = deg p + deg q = 0; since deg p ≥ 0 for any
polynomial p(x) ∈ R[x], it follows that deg p(x) = deg q(x) = 0. Therefore p(x), q(x) ∈ R as
required. 
Remark. If R has zero divisors (e.g., Z/6Z), then R[x] also has zero divisors.
Proposition 2.2. Let R be a commutative ring with unity 1 6= 0, and R[x] the polynomial
ring of R. Let p(x) = an xn + · · · + a1 x + a0 . Then the following are equivalent:
(i) p(x) is a zero divisor in R[x].
(ii) there exists a non-zero c ∈ R such that cp(x) = 0.
Proof. ((i) ⇒ (ii)) Suppose that q(x) = bm xm + · · · + b1 x + b0 is a non-zero polynomial
of the least degree such that p(x)q(x) = 0. Then we have an bm = 0. Clearly bm 6= 0
since q(x) is of the least degree satisfying the desired property. Hence deg an q(x) < m,
and (an q(x))p(x) = an q(x)p(x) = 0. This forces an q(x) ≡ 0, and hence an bm−1 = 0. Note
that the coefficient of xm+n−1 of p(x)q(x) is an−1 bm + an bm−1 which is equal to 0. But
then an bm−1 = 0 so an−1 bm = 0. Hence deg an−1 q(x) < m, so an−1 q(x) ≡ 0. This implies
an−1 bm−1 = 0; and the coefficient of xn+m−2 of pq is an−1 bm−1 + an−2 bm + an bm−2 , which
must equal 0. Thus an−2 bm + an bm−2 = 0. But then an q(x) ≡ 0, so an bm−2 = 0. Hence
an−2 bm = 0. Repeating this argument, we see that ai bm = 0 for all 0 ≤ i ≤ n, so bm p(x) = 0
as desired.
((ii) ⇒ (i)) This direction is immediate. 
3
Remark. Suppose that S is a subring of R. Then the polynomial ring S[x] is also a subring
of the polynomial ring R[x].
2.2. Matrix rings
For any ring R and a positive integer n, we can define Mn (R), the set of all n × n matrices
with entries from R. With the usual matrix addition and multiplication operation, Mn (R)
has a ring structure. Such ring is called a matrix ring. For any non-trivial ring R, the matrix
ring Mn (R) is non-commutative for any n ≥ 2 (whether R is commutative or not) because
the matrix multiplication itself is not commutative. Also, Mn (R) cannot be an integral
domain either as it is possible to construct two non-zero square matrices with 1 or 0 for each
entry whose product is the zero matrix.
We examine some subrings of Mn (R). It is straightforward to verify that the set of
scalar matrices (a ∈ R on all diagonal entries, 0 everywhere else) is a subring of Mn (R).
Particularly, if a = 1, then that matrix serves as the unity of Mn (R). Thus Mn (R) has a
unity if and only if R has unity as well. Notice that the subring of scalar matrices is a subset
(and also a subring) of the set of upper triangular matrices, which is again not that hard to
verify that this is a subring of Mn (R). This example serves to demonstrate that the ”the
subring of” relation is transitive. The ring of scalar matrices is a subring of the ring of upper
triangular matrices; the ring of upper triangular matrices is a subring of Mn (R); and the
ring of scalar matrices is a subring of Mn (R).
Note that as long as det(A) 6= 0, A is invertible. Thus, the group of units of Mn (R) is
GLn (R) = {A ∈ Mn (R) : det(A) 6= 0}, also known as the general linear group of degree n
over R.
Finally, if S is a subring of R, Mn (S) is always a subring of Mn (R).
2.3. Group rings
Definition 2.2. Let R be a commutative ring with unity 1 6= 0, and suppose that G is a
finite group, say G = {g1 , g2 , . . . , gn } whose operation is written multiplicatively. Then the
group ring R[G] of G with coefficients in R is
R[G] := {a1 g1 + · · · + an gn : ai ∈ R},
with the addition and multiplication defined as follows:
(a1 g1 + · · · + an gn ) + (b1 g1 + · · · + bn gn ) = (a1 + b1 )g1 + · · · + (an + bn )gn
 
Xn X
(a1 g1 + · · · + an gn )(b1 g1 + · · · + bn gn ) =  ai bj  gk .
k=1 gi gj =gk

We shall write a1 g1 = a1 if g1 is the identity in G; similarly, we write 1gi = gi for any gi ∈ G

where 1 is the unity in R. Therefore, the 1 in R is the unity in R[G] also.
Remark. It is a routine exercise to check that these operations indeed form a ring. Also note
that the associativity of multiplication follows from the associativity of the group operation
in G.
Example. Suppose that G = D8 = hr, s : r4 = s2 = 1, rs = sr−1 i, and let R = Z. Then the
elements α = r2 + r − 2s and β = −3r2 + rs are in the group ring Z[D8 ]. Then we have
α + β = r − 2x2 − 2s + rs
4
 αβ = (r + r2 − 2s)(−3r2 + rs)
= −3r3 + r2 s − 3 + r3 s + 6r2 s − 2r3
= −3 − 5r3 + 7r2 s + r3 s,
so indeed α + β, αβ ∈ Z[D8 ].
Now we explore some properties of group rings.
Proposition 2.3. Let R be a commutative ring with unity 1 6= 0, and let G be a finite group.
Then the group ring R[G] is commutative if and only if G is an abelian group.
Proof. 
Proposition 2.4. G is a subgroup of the group of units of R[G].
Proof. Note that, for any g ∈ G we have 1R g = g. Thus the multiplication operation of R[G]
restricted to G is just the usual group operation defined in G. But then since G itself is a
group, every g has an inverse, which is a multiplicative inverse in R[G] as well. Thus g is in
the group of units of R[G], as required. 
Proposition 2.5. If #G > 1, then R[G] is not an integral domain.
Proof. We will prove that R[G] must have a zero-divisor if #G =: m > 1. Let g ∈ G. Then
note that 1 − g ∈ R[G] and 1 + g + g 2 + · · · + g m−1 ∈ R[G], and that
(1 − g)(1 + g + g 2 + · · · + g m−1 ) = 1 − g m .
But then g m = 1 since #G = m, so actually (1−g)(1+g +· · ·+g m−1 ) = 0. Since 1−g ∈ R[G]
is a zero-divisor, R[G] cannot be an integral domain. 
Proposition 2.6. Suppose that R is a commutative ring with unity 1 6= 0, and that S is a
subring of R. If G is a finite group, then S[G] is a subring of R[G]. If H is a subgroup of
G, then R[H] is a subring of R[G].

3. Chapter VII.3: Ring homomorphisms and quotient rings

Definition 3.1. Let R and S be rings. Then ϕ : R → S is a ring homomorphism if the
following axioms are satisfied for all a, b ∈ R:
(i) ϕ(a + b) = ϕ(a) + ϕ(b)
(ii) ϕ(ab) = ϕ(a)ϕ(b)
Additionally, if ϕ is:
(1) surjective, then ϕ is a ring epimorphism.
(2) injective, then ϕ is a ring monomorphism.
(3) a ring homomorphism to itself (i.e., R = S), then ϕ is a ring endomorphism.
(4) bijective, then ϕ is a ring isomorphism.
(5) both a ring isomorphism and a ring endomorphism, then ϕ is a ring automorphism.
Remark. Some textbooks stipulate that a ring homomorphism must preserve unity (i.e.,
ϕ(1R ) = 1S where each of 1R and 1S denotes the multiplicative identity of R and S, respec-
tively). Since Dummit & Foote does not have this stipulation, we shall assume that a unity
may not be preserved.
5
Example. Suppose ϕ : Z[x] → Z is defined as ϕ(p(x)) = p(0). This is a ring homomorphism.
Suppose that p(x) has constant term a and q(x) has constant term b (i.e., p(0) = a, q(0) = b).
Then (p + q)(0) = a + b = p(0) + q(0) and (pq)(0) = ab = p(0)q(0), preserving the ring
structure. Particularly, ker ϕ consists of all the polynomials whose constant term is 0. Note
that im ϕ = ϕ(Z[x]) = Z (i.e., ϕ is surjective), since for any a ∈ Z you can let p(x) = x + a.
Therefore ϕ is an example of a ring epimorphism.
Example. The following example shows that we must take caution to verify that both addition
and multiplication are preserved. Suppose that ϕn : Z → Z is defined as ϕn (x) = nx. ϕn (x)
is an additive group homomorphism since ϕn (x + y) = n(x + y) = nx + ny = ϕn (x) + ϕn (y),
for any n. However, ϕn does not preserve multiplication unless n = 0 or n = 1. We have
ϕn (xy) = nxy whereas ϕn (x)ϕn (y) = (nx)(ny) = n2 xy. If n = 0 or n = 1, then ϕn is either
the zero map (if n = 0) or the identity map (if n = 1) – so in these two cases, ϕn is a ring
homomorphism. Otherwise, n2 6= n so ϕn does not preserve multiplication. Therefore ϕn is
not a ring homomorphism if n 6= 0, 1.
Example. 2Z and 3Z are not isomorphic as rings. Suppose that ϕ : 2Z → 3Z is an isomor-
phism, and let ϕ(2) =: a. Then ϕ(4) = ϕ(2 + 2) = 2a whereas ϕ(4) = ϕ(2 · 2) = ϕ(2)2 = a2 .
Hence we must have a2 = 2a, and this is possible only when a = 0 or a = 2. Clearly 2 ∈
/ 3Z,
so a = 0. But then 2Z is generated by 2, and ϕ(2) = 0, so ϕ must be the zero map. But
this is impossible since neither 2Z nor 3Z are the zero sets. Hence 2Z and 3Z cannot be
isomorphic as rings.
Example. We want to find all the ring homomorphisms from Z to Z/20Z. Suppose that
f : Z → Z/20Z. Since Z = (1), the behaviour of f is completely determined by how f affects
1. First and foremost, any ring homomorphism must be an additive group homomorphism,
so hf (1)i ⊆ Z/20Z must be an additive subgroup. Also, note that f (1) = f (1 · 1) = f (1)2 ,
so we must have f (1)2 = f (1) also. Z/20Z is cyclic, so any of its subgroups is also cyclic. In
particular, h0i, h1i, h2i, h4i, h5i, h10i are the subgroups of Z/20Z.
(1) hf (1)i = {0}
In this case, f (1) = 0, so the ring homomorphism obtained from this case is
necessarily the zero map.
(2) hf (1)i = h1i = Z/20Z.
In this case f (1) = 1, so we have f (k) = f (1 + 1 + · · · + 1) = k; similarly,
0 = f (0) = f (k + (−k)) = f (k) + f (−k) = k + f (−k), so f (−k) = −k. Hence in
this case, f (z) = z (mod 20). Thus, if we are to take the convention that any ring
homomorphism must preserve the unity, this is the only ring homomorphism from Z
to Z/20Z.
(3) hf (1)i = h2i.
First we need f (1) such that f (1)2 = f (1). We need f (1)2 and f (1) to share the
same unit digit, so the only possible options are 6, 10 and 16. However 62 6≡ 6, 102 6≡
10 but 162 ≡ 16 (mod 20). f (1) = 16, so f (k) = 16k. Hence f (k) = f (1 · k) =
f (1)f (k) = 16f (k) = 256k ≡ 16k (mod 20), so 16 is the multiplicative identity.
(4) hf (1)i = h4i.
First we need f (1) such that f (1)2 = f (1). We need f (1)2 and f (1) to share
the same unit digit, so the only possible option is 16. So this case yields the same
isomorphism as the h2i case.
6
 (5) hf (1)i = h5i
In this case, only 5 and 15 have 52 ≡ 5, 152 ≡ 15, so f (1) may be either 5 or 15. Note
that f (1) must be the multiplicative identity of h5i since f (m) = f (1·m) = f (1)f (m)
for any m ∈ Z. But note that 15 · 5 = 75 ≡ 15 6= 5, so 15 cannot be the multiplicative
identity. Hence f (1) = 5. Indeed we see that f (m) = 5m, from which we have
f (m) = f (1 · m) = f (1)f (m) = 5f (m) = 25f (m) ≡ 5f (m) (mod 20) as required.
(6) hf (1)i = h10i = {0, 10}.
Since f (1) 6= 10 (because we already know f (1)2 = 100 ≡ 0 6≡ 10 (mod 20)),
f (1) = 0. Thus the zero map is the only possible choice, meaning that no ring
homomorphism can have h10i as its image.
So if fa : Z → Z/20Z is defined by fa (1) = a, then the following list gives the complete list
of ring homomorphisms: f0 , f1 , f5 , f16 .
Recall that from group theory that if ϕ : G → H is a group homomorphism, then ker ϕ is
a normal subgroup of G, and im ϕ = ϕ(G) is a subgroup of H (but not necessarily a normal
subgroup of H). Thus it is natural to wonder if an analogous result holds for rings. The
answer to this question is yes. Clearly, the ring counterpart of subgroups is subrings. What
about the ring counterpart of normal subgroups? This prompts us to introduce the following
definition.
Definition 3.2. Let R be a ring, let I ⊂ R, and let r ∈ R.
(1) rI := {ra : a ∈ I} and Ir := {ar : a ∈ I}.
(2) Suppose that I is not just a subset of R, but a subring of R. Then I is a left (resp.
right) ideal of R if I is closed under left (resp. right) multiplication by elements from
R. In other words, I is a left (resp. right) ideal if rI ⊆ I (resp. Ir ⊆ I) for all r ∈ R.
(3) If I is both a left ideal and a right ideal, then I is said to be an ideal (or a two-sided
ideal ) of R.
Remark. A corollary to the above definition is that if R is a commutative ring, then every
ideal is an ideal (i.e. two-sided) since left ideals and right ideals coincide.
Proposition 3.1 (Ideal test). Let R be a ring, and let I ⊂ R. Then I is an ideal if all of
the following conditions are satisfied:
(i) I 6= ∅
(ii) a − b ∈ I for all a, b ∈ I
(iii) ra ∈ I for all a ∈ I and r ∈ R (closed under multiplication by all the element of R,
not just by the elements of I).
If R has unity 1, then (ii) may be replaced with
(ii’) a + b ∈ I for all a, b ∈ I.
We shall introduce the following definition for the sake of giving an example of an ideal.
Definition 3.3. Let R be a commutative ring with unity. If x ∈ R such that xn = 0 for
some n ∈ Z, then x is said to be nilpotent. The nilradical of R η(R) is the set of nilpotent
elements of R, i.e., η(R) = {x ∈ R | xn = 0 for some n ∈ Z}.
Proposition 3.2. If R is a commutative ring with unity, then η(R) is an ideal of R.
Proof. Let a, b ∈ η(R) and r ∈ R. Suppose that n ∈ Z satisfies an = 0. Then rn an = (ra)n =
0, so ra ∈ η(R) as well. Therefore, it follows that any element of the form ak bl ∈ η(R). If
7
m ∈ Z is chosen such that bm = 0 (which also implies that (−b)m = 0), then we have
m+n
X m + n
m+n
(a − b) = am+n−k (−b)k .
k=0
k
But for any 0 ≤ k ≤ m, we have n ≤ m + n − k ≤ m + n, so the first m monomials of the
RHS above is 0. But the remaining monomials all have k > m, which implies that (−b)k = 0.
Hence (a − b)m+n = 0, so a − b ∈ η(R). 
Now we are ready to discuss the kernel and the image of a ring homomorphism.
Proposition 3.3. Let R and S be rings, and suppose that ϕ : R → S is a ring homomor-
phism.
(1) ϕ(R) = im ϕ is a subring of S (but not necessarily an ideal of S).
(2) ker ϕ is an ideal of R.
(3) If I is an ideal of S, then ϕ−1 (I) is an ideal of R.
Proof. (i) and (ii) are straightforward verification of axioms, so we shall prove (iii) only.
Since 0S ∈ I and f (0R ) = 0S , 0R ∈ ϕ−1 (I), thereby showing that f −1 (I) is non-empty.
That 0S ∈ I implies that f −1 (I) must contain ker ϕ also. Suppose that r1 , r2 ∈ ϕ−1 (I).
Then ϕ(r1 − r2 ) = ϕ(r1 ) − ϕ(r2 ) ∈ I as I is an ideal such that ϕ(r1 ), ϕ(r2 ) ∈ I. Hence
r1 − r2 ∈ ϕ−1 (I), so ϕ−1 (I) is closed under subtraction. Let r be any element in R. Then
ϕ(rr1 ) = ϕ(r)ϕ(r1 ) ∈ I since ϕ(r1 ) ∈ I, whence it follows rr1 ∈ ϕ−1 (I). One can use the
similar argument (just replace rr1 with r1 r) to show that ϕ−1 (I) is closed under both left
multiplication and right multiplication. Thus ϕ−1 (I) is an ideal in R containing ker ϕ. 
If G is a group and H a normal subgroup of G, then G/H is a quotient group, where
its group operation is induced by the group operation for G. Recall that H being a normal
subgroup of G is crucial in making the group operation well-defined. This is also a crucial fact
in deriving the first isomorphism theorem for groups. Furthermore, in group theory, we learnt
that H is a normal subgroup if and only if H is the kernel of some group homomorphism.
Does the similar claim hold for ideals? First, we shall introduce the notion of quotient rings,
and then proceed to verify whether the operations are well-defined. Also, since any ideal is
a subring, any ideal is automatically an additive normal subgroup. Thus R/I is always an
additive quotient group.
Definition 3.4. Let R be a ring, and I an ideal of R. Then the additive quotient group R/I
is called the quotient ring of R by I where the addition and the multiplication are defined
as follows:
(1) (r + I) + (s + I) = (r + s) + I for all r, s ∈ R, and
(2) (r + I) · (s + I) = rs + I for all r, s ∈ R.
We have yet to verify that the multiplication as defined above is well-defined and that such
multiplication satisfies all the necessary properties (such as distributive properties). However,
since the remaining properties such as distributive properties follow from the corresponding
axioms satisfied by R, it in fact suffices to prove that the multiplication is well-defined.
Suppose that α, β ∈ I. Then (r + α) + I = r + I and (s + β) + I = s + I. So we have
((r + α) + I)((s + β) + I) = (r + α)(s + β) + I = (rs + rβ + αs + αβ) + I.
8
Since I is an ideal of R, for any r, s ∈ R we have rβ, αs ∈ I. Therefore
(r + α)(s + β) + I = (rs + αβ) + I = rs + I,
since evidently αβ ∈ I. Hence the multiplication is well-defined.
Definition 3.5. Let R be a ring, and let I be an ideal of R. Then π : R → R/I defined by
r 7→ r + I is the natural projection of R onto R/I.
Theorem 3.1. Let R be a ring, and I an ideal of R. Then π is a ring epimorphism with
kernel I. Thus every ideal is the kernel of a ring homomorphism. Therefore, the set of ideals
and the set of kernels of ring homomorphisms coincide.
Proof. Since I is an ideal, R/I is a ring (hence an additive abelian group a fortiori). Hence
π is a group homomorphism with kernel I. So it remains to verify that π preserves multipli-
cation, which follows from the fact that
π(rs) = rs + I = (r + I)(s + I) = π(r)π(s).
That π is surjective follows from the definition of π. Finally, the last statement follows since
we previously proved that the kernel of a ring homomorphism is an ideal. 
We shall conclude this section with the isomorphism theorems for rings. Each isomorphism
theorem for rings has its counterpart for groups, as we shall see.
Theorem 3.2 (First isomorphism theorem for rings). If ϕ : R → S is a ring homomorphism,
then the kernel of ϕ is an ideal of R, the image of ϕ is a subring of S, and
R/ ker ϕ ∼
= ϕ(R) = im ϕ.
Proof. The first two statements are already proved in Proposition 3.3, so it suffices to prove
the last statement only. Let I = ker ϕ, and let ϕ : R/I → ϕ(R) be the homomorphism
induced by ϕ (i.e., ϕ : r + I 7→ ϕ(r)). Then ϕ−1 (im ϕ) is precisely the cosets of I. We
previously verified that ϕ is a well-defined group epimorphism, so it suffices to verify that ϕ
preserves multiplication and is injective. Indeed, we have
ϕ((r + I)(s + I)) = ϕ(rs + I) = ϕ(rs) = ϕ(r)ϕ(s) = ϕ(r + I)ϕ(s + I).
Suppose that ϕ(r) = 0. Then r ∈ ker ϕ so ϕ(r + I) = ϕ(r) = 0 if and only if r ∈ ker ϕ = I.
Therefore the kernel of ϕ is trivial. Hence R/ ker ϕ ∼
= ϕ(R) as rings, as desired. 
The proofs of the remaining three theorems are straightforward. First, use the correspond-
ing theorems for groups to get an additive group isomorphism or correspondence. Second,
verify that the same group isomorphism is a multiplicative map, thereby showing that it is
a ring isomorphism as well. Again, this is straightforward from the way multiplication is
defined in quotient rings.
Theorem 3.3 (Second isomorphism theorem for rings). Let A be a subring of R, and B an
ideal of R. Then
(i) A + B := {a + b : a ∈ A, b ∈ B} is a subring of R.
(ii) A ∩ B is an ideal of A.
(iii) (A + B)/B ∼= A/(A ∩ B).
Theorem 3.4 (Third isomorphism theorem for rings). Let I and J be ideals of R with I ⊆ J.
Then J/I is an ideal of R/I and (R/I)/(J/I) ∼
= R/J.
9
 The fourth isomorphism theorem for rings is called the lattice isomorphism theorem for
rings in some literature. The theorem, however, is more commonly known as the correspon-
dence theorem for rings, which is the name we will go by in the notes.
Theorem 3.5 (Correspondence theorem for rings). Let I be an ideal of R. Then there is
an inclusion-preserving bijection between the set of subrings A of R containing I and the set
of subrings of R/I. Furthermore, A is an ideal of R containing I if and only if A/I is an
ideal of R/I.
Example. We conclude with an application of the correspondence theorem. Suppose we want
to find all the ideals of Z[x]/(2, x2 + 1). First note that
Z[x]/(2, x2 + 1) ∼= F2 [x]/(x2 + 1),
and that (x2 + 1) = (x + 1)2 over F2 . So by the correspondence theorem, there is a bijection
between the set of ideals of F2 [x] containing (x2 + 1) corresponds and the set of ideals of
F2 [x]/(x2 + 1). F2 [x] is a PID, so every ideal can be written in the form (f (x)). Since
x2 + 1 = (x + 1)2 , it follows that f (x) must be one of 1, x + 1, and (x + 1)2 = (x2 + 1). Hence
the ideals of F2 [x]/(x2 + 1) are (0), (1), and (x + 1), so the ideals of Z[x]/(2, x2 + 1) are the
zero ideal, the entire ring, and (2, x + 1).

4. Chapter VII.4: Properties of ideals

Definition 4.1. Let I and J be ideals of R.
(1) I + J := {a + b | a ∈ I, b ∈ J} (the sum of I and J)
(2) IJ := {a1 b1 + · · · + an bn | ai ∈ I, bi ∈ J, n ∈ N} (the product of I and J)
(3) I n , called the nth power of I, is the set consisting of all finite sums of elements of the
form a1 a2 · · · an where ai ∈ I for all i.
Remark. Clearly, I, J ⊆ I + J. In fact, I + J is the smallest ideal of R containing both I
and J. From the definition, it is also clear that IJ ⊂ (I ∩ J); IJ may be strictly smaller
than I ∩ J. It should be noted that IJ must consist of all finite sums of the elements of the
form ab where a ∈ I, b ∈ J; note that the set {ab : a ∈ I, b ∈ J} need not be closed under
addition, so cannot be an ideal in general.
From now on, we shall assume that R is a ring with unity 1 6= 0.
Definition 4.2. Let A be any subset of the ring R.
(1) Let (A) be the smallest ideal of R generated by A. Then (A) is called the ideal
generated by A.
(2) The left ideal generated by A is RA := {r1 a1 + · · · + rn an | ri ∈ R, ai ∈ A, n ∈ N}.
Similarly, we define the right ideal generated by A to be AR := {a1 r1 + · · · + an rn |
ri ∈ R, ai ∈ A, n ∈ N}.
(3) The two-sided ideal generated by A is RAR := {r1 a1 r10 + · · · + rn an rn0 | ri , ri0 ∈ R, ai ∈
A, n ∈ N}.
(4) If #A is finite, then (A) is a finitely generated ideal. In this case, if A = {a1 , a2 , . . . , an },
then we omit the set brackets, and write (A) = (a1 , a2 , . . . , an ) instead.
(5) More specifically, if #A = 1, then (A) is a principal ideal. In this case, if A = {a},
then we omit the set brackets, and write (A) = (a).
10
 Recall that the subgroup of G generated by a subset B ⊂ G is
\
hBi = H
H⊇B
H subgroup

because the intersection of any non-empty collection of subgroups is a subgroup. We can do

the same things for ideals. In other words,
\
(A) = I.
I⊇A
I ideal

Now we will briefly explore why the uses of “the” are justified for (2) and (3) of Definition
4.2. Note that every left ideal generated by A must contain RA – since R contains 1, RA
necessarily contains A. Conversely, any left ideal containing A must contain all the elements
of the form r1 a1 + · · · + rn an for ri ∈ R, ai ∈ A for all i. Therefore such ideal must contain
RA. Hence, RA is precisely the left ideal generated by A. The same line of reasoning can
be used for AR and RAR as well.
Remark. If R is commutative, then evidently RA = AR = RAR = (A). If R is a commuta-
tive ring and I is a principal ideal (that is, I = (a) for some a ∈ R), then I = (a) = Ra =
{ra | r ∈ R}. If R is not commutative, then (a) = RaR = {r1 ar10 + · · · + rn arn0 | ri , ri0 ∈
R, n ∈ N}.
Example. If R = Z, then (n) = nZ is an ideal. In fact every ideal of Z is of the form (n) for
some n ∈ Z. Notice that if (n, m) is an ideal, then (d) = (n, m) where gcd(n, m) = d. Thus
Z is an example of a principal ideal domain (PID).
Example. Let R = Z[x] and I = (2, x). We show that I is an ideal but is not a principal ideal.
Suppose instead then that there exists p(x) ∈ R such that (p(x)) = (2, x). Every element in I
is of the form 2f (x)+xg(x), so it necessarily has an even constant term. Therefore 2 ∈ (p(x)),
so there is q(x) such that p(x)q(x) = 2. But then deg p(x)q(x) = 0 = deg p(x) + deg q(x),
so the generator p(x) must be of degree 0. Particularly, p(x) can only be ±1 or ±2. If
p(x) = ±1, then (p(x)) = Z[x](±1) = Z[x], but this contradicts the fact that polynomials
with an odd constant term are not in (p(x)). Therefore p(x) is 2 or −2. We also need
x ∈ (p(x)) = (2) = (−2). This means there is r(x) ∈ Z[x] so that 2r(x) = x. However this
is impossible since every coefficient of 2r(x) is even whereas x has coefficient 1. Therefore
(2, x) is not a principal ideal.
On the other hand, if Z[x] were replaced with Q[x], then (2, x) is in fact principal. Note
that 21 · 2 = 1 ∈ (2, x), so in fact (2, x) = (1) = Q[x] – this is because Q is a field.
The last example illustrates an important relationship between units, ideals, and fields,
which we shall state and prove below.
Proposition 4.1. Let I be an ideal of R.
(1) I = R if and only if I contains a unit.
(2) Assume that R is commutative. Then R is a field if and only if its only ideals are 0
and R.
Proof. ((1), ⇒) Suppose that I = R. Then I = (1), and clearly 1 is a unit.
11
 ((1), ⇐) Assume that I contains a unit u. Then there is v ∈ R such that vu = 1. Therefore
1 = vu ∈ I, so I = (1) = R as required.
((2), ⇒) Suppose that R is a field. Therefore every non-zero element of R has a multi-
plicative inverse. Thus once a non-zero element is in an ideal I, 1 is necessarily in I also. In
other words, there cannot be any ideals between 0 and R.
((2), ⇐) Suppose that R only has 0 and R as its only ideals. If u is any non-zero element
of R, then R = (u). Therefore 1 ∈ (u). This means that u has a multiplicative inverse. Our
choice of non-zero u is arbitrary, so it follows that every non-zero element of R is a unit.
This is precisely what it means for R to be a [Link] 
Corollary 4.1. Let F be a field. Then any ring homomorphism ϕ : F → R is either the
zero map or an injection.
Proof. Recall that ker ϕ is always an ideal of F . But since F is a field, ker ϕ can only be the
zero ideal or the entire field. If ker ϕ = (0), then ϕ is an injection. If ker ϕ = F , then ϕ is
the zero map. 
Next, we introduce two important classes of ideals: prime ideals and maximal ideals.
Prime ideals essentially generalizes the prime numbers from number theory. To illustrate
this point, we take a quick detour to elementary number theory.
Theorem 4.1. Suppose that a, b ∈ N and p is a prime. If p | ab then p | a or p | b. In other
words, at least one of a and b must be a multiple of p.
Proof. We will prove the contrapositive instead. Suppose that p - a and p - b. Then p cannot
show up in the prime factorization of a and that of b; therefore p cannot show up in the
prime factorization of ab either, so p - ab as required. 
Recall that if R = Z, then (p) = pZ. Therefore a ∈ (p) if and only if p | a. With this
insight, we can reformulate the above number-theoretic statement in terms of ideals.
Theorem 4.2. Let R = Z, p a prime, and a, b ∈ Z. If ab ∈ (p), then a ∈ (p) or b ∈ (p).
Prime ideals is the generalization of this idea to any general commutative ring.
Definition 4.3. Let R be a commutative ring. Then an ideal P is a prime ideal if:
(1) P is a proper ideal of R; and
(2) for any a, b ∈ R such that ab ∈ P , at least one of a and b lies in P .
Proposition 4.2. Let R be a commutative ring. Then R/P is an integral domain if and
only if P is a prime ideal.
Proof. (⇒) Suppose that R/P is an integral domain, and let r̄ := r + P, s̄ := s + P ∈ R/P
such that rs ∈ P . Then r̄s̄ = rs = 0̄ since rs ∈ P . But then R/P is an integral domain, so
at least one of r̄ = 0̄ and s̄ = 0̄ must hold. Therefore, at least one of r and s must lie in P ,
so P is a prime ideal.
(⇐) Suppose that P is a prime ideal, and suppose that rs ∈ P . This gives us
(r + P )(s + P ) = rs + P = 0 + P.
But then P is a prime ideal, so at least one of r and s must be in P . Therefore at least
one of r + P and s + P must be equal to 0 + P , so R/P cannot have any zero-divisor, as
required. 
12
Corollary 4.2. Let R be a commutative ring. Then R is an integral domain if and only if
(0) is a prime ideal.
Proof. (⇒) Suppose that R is an integral domain, and let ab = 0 ∈ (0). R has no zero-
divisor, so a = 0 or b = 0. Thus at least one of a and b is in (0). Thus (0) must be a prime
ideal.
(⇐) Suppose that (0) is a prime ideal, and that a, b ∈ R satisfy ab = 0 ∈ (0). But since
(0) is a prime ideal, we have a = 0 or b = 0. Hence R cannot have any zero divisor, so R is
an integral domain. 
We shall state two interesting results involving prime ideals before moving onto another
class of ideals.
Proposition 4.3. Let R be a commutative ring with unity. Suppose that I and J are ideals
of R, and suppose that P is a prime ideal of R. Then the following are true.
(i) If P contains no zero divisors, then R cannot contain any zero divisors either.
(ii) If I ∩ J ⊆ p, then I ⊆ p or J ⊆ p.
Proof. (i) Suppose a ∈ R is a zero divisors of R. Then there exists a non-zero b ∈ R such
that ab = 0. Since p is an ideal, necessarily 0 ∈ p. Hence ab ∈ p, so either a ∈ p or b ∈ p.
But since a and b are zero divisors of each other, it follows that p contains a zero divisor,
which is a contradiction. Hence R cannot contain any zero divisor either.
(ii) Suppose that I 6⊆ p and J 6⊆ p. Then there exist r ∈ I and s ∈ J such that r, s ∈ / p.
Since p is prime, it follows that rs ∈
/ p. But then rs ∈ I and rs ∈ J since I and J are ideals.
Hence rs ∈ I ∩ J. Thus rs is in I ∩ J but not in p. Therefore I ∩ J 6( p as required. 
Proposition 4.4. Suppose that R is a commutative ring with unity. Then η(R) is contained
in every prime ideal of R.
Proof. Suppose that N = η(R), and let P be a prime ideal of R. If x ∈ N , then xm ∈ P
for some m ∈ Z since x is in the nilradical of R. Let n be the smallest integer such that
xn ∈ P . Suppose that n > 1 so that x ∈ / P . Then xn = x · xn−1 ∈ P , so it follows that x ∈ P
n−1
or x ∈ P . But this contradicts the minimality of n, so it follows that x ∈ P . Since this
works for every prime ideal, it follows that
\
N⊆ P.
P prime

As for the reverse inclusion, we will show that if x ∈/ N , then one can find a prime ideal that
2
does not contain x. Let S = {1, x, x , . . . }, and consider the set of ideals of R that does not
intersect with S. Clearly this set is non-empty since the zero ideal does not intersect with
S. By Zorn’s lemma, there is a maximal element M in this set.
We want to show that M is a prime ideal. Suppose that there are some elements such
that a ∈ / M and B ∈ / M but ab ∈ M . If Aa := {z ∈ R : az ∈ M }, then Aa (and similarly,
Ab ) is an ideal of R such that M ( Aa . But recall that M is a maximal element of the
set of ideals not intersecting with S; therefore Aa must intersect with S, so there is some m
0
such that xm ∈ Aa . Similarly, there is some m0 such that xm ∈ Ab . Therefore we see that
0 0 0
xm+m = xm xm ∈ Aab . But since ab ∈ M , it follows that Aab = M , so xm+m ∈ M . This is a
contradiction, so M is a prime ideal that
T does not intersect with S. Thus we found a prime
ideal that does not contain x, so x ∈ / P , as required.
13
 
Another important class of ideals is the class of ideals that are not contained in any proper
ideal.
Definition 4.4. An ideal M of a ring R is a maximal ideal if M 6= R, and the only ideals
containing M are M and R.
Proposition 4.5. In a ring with unity, every proper ideal is contained in a maximal ideal.
Proposition 4.6. Let R be a commutative ring. Then R/M is a field if and only if M is a
maximal ideal.
Proof. (⇐) Thanks to the correspondence theorem for rings (Theorem 3.5), there is a bi-
jection between the set of ideals containing M and the set of ideals of R/M . But the only
ideal of R containing M is R since M is maximal, so the only ideals of R/M are R/M and
0. Therefore, R/M is a commutative ring that has no non-trivial ideals, so R/M is a field.
(⇒) Since R/M is a field, the only ideals of R/M are R/M and 0. So by the correspondence
theorem for rings, there is only one ideal containing M , and that ideal is R. Therefore M is
a maximal ideal of R. 
Proposition 4.7. Let R be a commutative ring with unity. An ideal M of R is maximal if
and only if for every r ∈ R \ M there exists x ∈ R such that 1 − rx ∈ M .
Proof. (⇐) Suppose that J is an ideal of R such that M ( J ⊂ R, and let r ∈ J \ M ⊆
R \ M . J is an ideal of R, so rx ∈ J where x satisfies the given assumption. But then
1R − rx ∈ M ( J, so (rx) + (1R − rx) = 1R ∈ J, which forces J = R. Therefore M is
maximal.
(⇒) Suppose that M is maximal. Since r ∈ R \ M , we have M ( (r) + M ⊂ R. Thus,
(r) + M = R by the maximality of R. (r) + M is an ideal, so 1R ∈ (r) + M . Thus there
exist x ∈ R and m ∈ M so that rx + m = 1R . Now it follows that 1R − rx = m ∈ M as
desired. 
Corollary 4.3. Every maximal ideal of a commutative ring is a prime ideal.
Proof. Let M be a maximal ideal of a commutative ring R. Then R/M is a field, so R/M
is an integral domain as well. Hence M is a prime ideal, as required. 
Example. The set of non-zero prime ideals of Z and the set of maximal ideals of Z coincide.
In fact, the only non-zero prime ideals are of the form (p) = pZ for some prime p. We already
observed that any ideals of (p) for p prime is a prime ideal. If n is not a prime (without loss
of generality, assume n ∈ N), then there exist 1 < a, b < n such that n = ab. Thus n ∈ (n)
but a, b ∈/ (n), so (n) cannot be a prime ideal. Indeed, if n is composite, then Z/(n) = Z/nZ,
and if ab = n then b is a zero-divisor of a. Recall that if p is a prime, then Z/(p) = Z/pZ,
and Z/pZ forms the finite field of p elements (Fp ). Thus (p) is not just a prime ideal, but
also a maximal ideal. Z is an integral domain (in fact, a principal ideal domain, so it is a
fortiori an integral domain), so (0) is a prime ideal. But (0) ( (p) ( Z for any prime p, so
(0) is not a maximal ideal. Indeed, Z is not a field.
Example. (x) is maximal and prime in Q[x] since Q[x]/(x) ∼ = Q, and Q is a field. However,
(x) in Z[x] is prime but not maximal in Z[x], since Z[x]/(x) ∼ = Z, which is an integral domain
but is not a field.
14
Example. (x2 +1) is both maximal and prime in R[x]. Recall from field theory that R[x]/(x2 +
1) ∼
= R(i) ∼
= C (see the MATH 5055 notes for more information on this), the field of complex
numbers.

5. Chapter VII.5: Rings of fractions

In this section, we aim to prove that every commutative ring R (for convenience, we shall
always assume that a ring has unity 1) is a subring of a larger ring Q such that every non-zero
element that is not a zero divisor is a unit. Our focus will be on constructing this larger
ring containing R. Before doing this for general case, it always helps to consider a familiar
specific example. So for the sake of a familiar example, let R = Z. We will explore why this
larger ring Q is Q that we are familiar with.
Recall that Q consists of elements of the form p/q where p, q ∈ Z. Thus every rational
number has multiple representations. Indeed, a/b = c/d if and only if ad = bc (clearly,
b, d 6= 0); using the equivalence relation notion, we conclude that (a, b) ∼ (c, d) if and only
if ad = bc. The addition and multiplications are the operations we are all familiar with.
Proving that the addition and multiplication are well-defined is a routine exercise, so we
shall skip the verification. For any non-zero p/q ∈ Q, the multiplicative inverse is q/p ∈ Q,
so every non-zero element is a unit. Hence, Q has a field (and hence a ring) structure, and
Z is a subring of Q since every element b ∈ Z can be written of the form b/1 ∈ Q. Let
N := Z \ {0}. Then Q = N −1 Z. The “ring of fractions” is thus the generalization of this
idea onto any commutative ring with unity 1.
Definition 5.1. We say that a set S is multiplicatively closed if S is a non-empty set such
that xy ∈ S whenever x, y ∈ S.
Definition 5.2. Suppose R is a commutative ring with unity 1, and that D is a non-empty
multiplicatively closed subset of R that contains 1 but does not contain 0 and any zero
divisors. Then the ring of fractions of D with respect to R is Q = D−1 R.
Theorem 5.1 (Existence and uniqueness of the ring of fractions). Let R be a commutative
ring with unity 1. Let D be any non-empty multiplicatively closed subset of R that contains
1 but not 0 and any of the zero divisors of R. Then there is a commutative ring Q with 1
such that R is a subring of Q, and every element in D is a unit in Q. Additionally,
(1) Q = D−1 R. That is, every element of Q is of the form r/d where r ∈ R and d ∈ D.
In particular if D = R \ {0}, then Q is a field.
(2) Q is the smallest ring containing R in which all elements of D become units, in the
following sense. Let S be any commutative ring with unity, and let ϕ : R → S be any
injective ring homomorphism such that ϕ(d) is a unit in S for every d ∈ D. Then
there is an injective homomorphism Φ : Q → S such that Φ|R = ϕ. Thus, any ring
containing an isomorphic copy of R in which all the elements of D become units must
also contain an isomorphic copy of Q.
Proof. See p261-263 of Dummit & Foote. 
Corollary 5.1. If R is an integral domain and D = R \ {0}, then Q is a field.
If R is an integral domain, there is a special term to refer to its ring of fractions Q.
15
Definition 5.3. Suppose that R is an integral domain and D = R \ {0}, then Q is called
the field of fractions of R.
Corollary 5.2. Let R be an integral domain, and let Q be the field of fractions of R. If
a field F contains a subring r0 isomorphic to R then the subfield of F generated by R0 is
isomorphic to Q. Therefore, the smallest field containing an integral domain R is its field of
fractions.
Proof. Suppose ϕ : R → R0 ⊆ F be a ring isomorphism from R to R0 . In particular,
ϕ : R → F is an injective homomorphism from R into the field F . Let Φ : Q → F be the
extension of ϕ. Recall that Φ is injective, so Φ(Q) is an isomorphic copy of Q in F ; indeed,
ϕ(R) = R0 , so ϕ(R) ⊆ Φ(Q). For any r1 , r2 ∈ R, we have ϕ(r1 )ϕ(r2 )−1 = ϕ(r1 r2−1 ), so every
subfield containing ϕ(R) = R0 has elements of this form. But then every element in Q is
precisely of the form r1 r2−1 for some r1 , r2 ∈ R. Therefore every subfield of F containing R0
must contain Φ(Q). Therefore Φ(Q) is the subfield of F generated by R0 . 

We conclude with a universal property associated with rings of fractions.

Proposition 5.1 (Universal property of localization). Let S be a multiplicative subset of a
commutative ring R. Then ϕs : R → S −1 R defined by ϕs (r) = rs/s is well-defined. If T
is another commutative ring with identity, and ψ : R → T is a ring homomorphism such
that ψ(S) is a subset of the set of units in T , then there is a unique ring homomorphism
ψ : S −1 R → T such that ψϕs = ψ.
Proposition 5.2. Suppose that R is an integral domain, and S a multiplicatively closed
subset of R. Then S −1 R is isomorphic to a subring of the quotient field of R.
Proof. Consider the map ϕ : R → Frac(R) defined by ϕ(r) = r/1, and ψs : R → S −1 R
be ψs (r) = rs/s. We claim that ϕ(S) consists of units of Frac(R). Suppose otherwise. If
ϕ(s) = s/1 = 0, then st = 0 where t ∈ R \ {0}. But since R is an integral domain, it follows
s = 0, but this contradicts the fact that 0 ∈/ S. Thus ϕ(S) contains non-zero elements, so
every element in ϕ(S) is a unit in Frac(R). Thus by the universal property of localization
there is a unique ring homomorphism η : S −1 R → Frac(R) such that η ◦ ψs = ϕ. To prove
the claim, it suffices to show that η must be injective. Suppose that η(r/s) = 0 where r ∈ R
and s ∈ S. Then (r/s)/1 = r/s = 0. Multiply by both sides by st (where s, t ∈ S hence
st ∈ S; note that since both s, t 6= 0, it follows st 6= 0) to get rt = 0. But then R is an
integral domain, so it follows r = 0. Hence η is injective. 

6. Chapter VII.6: The Chinese remainder theorem

Recall Bézout’s identity from the number theory, which states that if d = gcd(n, m), then
there exist infinitely many solutions (x, y) to the equation nx + my = k where k is a multiple
of d. Thus, if n and m are coprime, then there exist x and y so that nx + my = 1. Therefore,
if we view nZ and mZ as ideals of Z we have nZ + mZ = gcd(m, n)Z. We can generalize
this notion to other rings. We will restrict our attention to commutative rings with unity 1.

Definition 6.1. We say that ideals A, B of a ring R is comaximal if A + B = R.

16
Theorem 6.1 (Chinese remainder theorem). Let A1 , A2 , . . . , Ak be pairwise comaximal ideals
in R. Then the map
k
Y
ϕ : R 7→ R/Ai
i=1
defined by
ϕ(r) := (r + A1 , . . . , r + Ak )
is a ring epimorphism with kernel A1 ∩ · · · ∩ Ak , which is equal to A1 A2 A3 · · · Ak . Therefore,
we have the isomorphism
k k k
Ai ∼
Y \ Y
R/ Ai = R/ = R/Ai .
i=1 i=1 i=1

Proof. We will suppose that the base case (k = 2) holds, and suppose that the claim holds
for k − 1 ideals. Now consider A = A1 and B = A2 A3 · · · Ak . If we prove that A1 and
A2 A3 · · · Ak are comaximal, then we can use the base case to finish the inductive step. Since
A1 , A2 , . . . , Ak are mutually comaximal, for any 2 ≤ i ≤ k there exists yi ∈ Ai and xi ∈ A1
so that xi + yi = 1. Hence (x2 + y2 ) · · · (xk + yk ) = 1. Note that (x2 + y2 ) · · · (xk + yk )
can be written in the form X + y2 y3 . . . yk . Since each individual term in X is a mutliple
of xi for some 2 ≤ i ≤ k it follows that X ∈ A1 . Thus 1 ∈ A1 + (A2 A3 · · · Ak ), proving
the comaximality of A1 and A2 A3 · · · Ak . But then since we assume that the claim holds for
k = 2, the inductive step is complete.
Now it remains to prove the base case. Suppose k = 2 and ϕ(r) = (r + A1 , r + A2 ). Since
A1 and A2 are comaximal, there exist a1 ∈ A1 and a2 ∈ A2 so that a1 +a2 = 1. Hence for any
(p+A1 , q+A2 ), we have ϕ(pa2 +qa1 ) = (pa2 +A1 , qa1 +A2 ) = (pa2 +pa1 +A1 , qa1 +qa2 +A2 ) =
(p(a2 + a1 ) + A1 , q(a1 + a2 ) + A2 ) = (p + A1 , q + A2 ). Hence ϕ is surjective. Clearly ϕ is a
homomorphism since ϕ is a projection map. Finally, suppose ϕ(r) = (0 + A1 , 0 + A2 ). This
implies that r ∈ A1 and r ∈ A2 , so ker ϕ = A1 ∩ A2 . So by the first isomorphism theorem
for rings, we have
R/(A1 ∩ A2 ) ∼
= R/A1 × R/A2 ,
as desired. Thus it only remains to show that A1 ∩ A2 = A1 A2 . Recall that A1 A2 consists
of finite sums of elements of the form ab where a ∈ A1 and b ∈ A2 . Therefore every element
in A1 A2 clearly belongs to both A1 and A2 , per definition of ideals. Hence A1 A2 ⊆ A1 ∩ A2 .
Suppose that a ∈ A1 ∩ A2 , and suppose a1 + a2 = 1 where a1 ∈ A1 and a2 ∈ A2 . Then
a = a1 = a(a1 + a2 ) = aa1 + aa2 = a1 a + aa2 ∈ A1 A2 , so the equality follows as required. 

7. Chapter VIII.2 (first half only): Principal ideal domains

Definition 7.1. Let R be a commutative ring with unity 1 such that for every ideal I, there
exists a ∈ R such that I = (a). Then R is a principal ideal domain (PID).
Definition 7.2. Suppose that R is an integral domain, and a, b, d ∈ R. We say that d is a
divisor of a (written d | a) if there exists x ∈ R such that dx = a. If d | a and d | b, then d is
a common divisor of a and b. If D is a common divisor of a and b such that d | D for any
common divisor d of a and b, then D is a greatest common divisor of a and b.
Definition 7.3. We say that a, b ∈ R are associates if there exists a unit element u ∈ R
such that a = ub.
17
Remark. It should be remarked that there may be more than one GCDs of a and b, provided
one exists. A GCD is not always defined for any two or more arbitrary elements in an integral
domain. For any two or more elements in a commutative ring with unity to have a GCD,
we need the unique factorization condition. Thus, we can discuss GCD for any two or more
elements if a ring is a unique factorization domain (UFD). A fortiori every PID is a UFD.
From the definition of a divisor, we see that d | a if and only if (a) ⊆ (d). Indeed, since
there is x ∈ R such that dx = a, we have a = dx ∈ (d), so (a) ⊆ (d). This gives us an
alternative way of defining divisibility in the context of ring theory. Also, if d is a common
divisor of a and b, and D a greatest common divisor, then necessarily d | D, so we have
(D) ⊆ (d). Hence, D is a greatest common divisor of a and b if and only if (D) is the
smallest principal ideal containing both a and b. This gives rise to the following proposition.

Proposition 7.1. Let R be a commutative ring. If a and b are non-zero elements of R such
that (a, b) = (d), then d is a GCD of a and b.
Observe that the definition wrote a greatest common divisor rather than the greatest
common divisor. It is known that the greatest common divisor is unique up to a sign for the
integers (hence unique in N), but we don’t know if such uniqueness can be generalized in
arbitrary principal ideal domains. The ensuing discussion will answer affirmatively to this
question.
Proposition 7.2. Let R be an integral domain, and d, d0 ∈ R such that (d) = (d0 ). Then d0
and d are associates, i.e., there exists a unit u ∈ R such that d0 = ud. Therefore, a greatest
common divisor of any two elements is unique up to multiplication by a unit.
Proof. If either d or d0 is zero, then the claim follows trivially, so assume that both d and d0
are non-zeros. Since d ∈ (d0 ) there is x ∈ R such that d = xd0 . Similarly, d0 ∈ (d) so d0 = yd
for some y ∈ R. Putting them together, we have d = xd0 = xyd, so d(1 − xy) = 0. But then
d is non-zero, and R is an integral domain; thus, xy = 1, i.e., both x and y are units. The
second part is immediate from the fact that any two greatest common divisors of a and b
generate the same ideal, namely the smallest principal ideal containing both a and b. 
Proposition 7.3. Let R be a PID, and let a and b be non-zero elements of R. Suppose that
d is a generator for the principal ideal generated by a and b. Then the following are true.
(1) d is a GCD of a and b
(2) d is unique up to multiplication by a unit of R. In other words, greatest common
divisors of a and b are associates of each other.
(3) (Generalized Bézout’s identity) d can be written as a R-linear combination of a and
b: that is, there exist x, y ∈ R such that d = ax + by.
Proof. The first two claims follow immediately from Propositions 7.1 and 7.2. Since (d) =
(a, b), we have d ∈ (a, b). Therefore there exist x, y ∈ R such that ax+by = d, as required. 
A PID also satisfies a few other important conditions. Specifically, we will focus on two
properties. First, every prime ideal is a maximal ideal in a PID. (Recall that in general,
every maximal ideal is a prime ideal, whereas the converse is not true.) Second, every PID
satisfies the ascending chain condition, whose definition we shall introduce now.
18
Definition 7.4. Suppose that R is a commutative ring, and that
I1 ⊆ I2 ⊆ I3 ⊆ · · ·
is an arbitrary ascending chain of ideals of R. If there exists N such that In = In+1 for any
n ≥ N , then we say that R satisfies an ascending chain condition. A ring satisfying the
ascending chain condition is called a Noetherian ring.
Proposition 7.4. Every PID is Noetherian.
S
Proof. Let I1 ⊆ I2 ⊆ · · · be an ascending chain of ideals of R, a PID. If S
I := In , then I is
indeed an ideal. But R is a PID, so there is a ∈ R so that I = (a). a ∈ In , so there exists
some N such that a ∈ IN , meaning (a) ⊆ IN ⊆ I. Thus IN = I = (a), so Im = I for any
m ≥ N , as desired. 
Remark. In fact, a (commutative) ring R is Noetherian if and only if every ideal of R is
finitely generated. So Proposition 7.4 can be derived as a corollary to this theorem. For the
proof of this more general theorem and more information on Noetherian rings, refer to the
PMATH 646 (Commutative algebra) notes.
Proposition 7.5. Let R be a principal ideal domain, and let (p) be a non-zero prime ideal
of R. Then (p) is a maximal ideal. Thus, every non-zero prime ideal in a PID is also a
maximal ideal.
Proof. Suppose that I = (m) is an ideal of R containing (p). (p) ⊆ (m), so there exists
r ∈ R such that p = rm. Since rm ∈ (p), at least one of r and m must lie in (p). If m ∈ (p)
then (m) ⊆ (p), so (p) = (m) = I. Suppose instead that r ∈ (p). Then there is s ∈ R such
that r = ps. Hence p = rm = psm, or p(1 − sm) = 0. (p) is a non-zero prime ideal, so p 6= 0,
which forces 1 − sm = 0. Therefore sm = 1, so m is in fact a unit of R. Thus in this case,
I = (m) = R, as required. 
Corollary 7.1. Suppose that R is a commutative ring such that the polynomial ring R[x] is
a PID. Then R is a field.
Proof. Suppose that R[x] is a principal ideal domain. R is a subring of R[x], and R[x] has
unity 1 if and only if R does. These facts lead us to conclude that R must at least be an
integral domain. Recall that R[x]/(x) ∼= R; since R is an integral domain, (x) must be a
prime ideal. But R[x] is a PID, so (x) is a maximal ideal also. R[x]/(x) is thus in fact a
field, so R is also a field. 
The last proposition of this section provides the converse of Propositions 7.3 and 7.4.
Proposition 7.6. Suppose that R is an integral domain that satisfies the following two
conditions.
(i) (Generalized Bézout’s identity) Any two non-zero elements x, y ∈ R have a GCD
which can be written in the form rx + sy for some r, s ∈ R.
(ii) (Ascending chain condition on principal ideals) If x1 , x2 , x3 , . . . are non-zero elements
of R such that xi+1 | xi for all i ≥ 1, then there is a positive integer N such that xn
is a unit times xN for all n ≥ N .
Then R is a PID.
19
Proof. Let I be any finitely generated ideal of R, i.e., I = (x1 , . . . , xn ). By the first condition,
we have (x, y) = (d) where d = gcd(x, y). Indeed, d = rx + sy ∈ (x, y), so (d) ⊆ (x, y).
Conversely, since d | x and d | y, it follows that x, y ∈ (d), or (x, y) ⊆ (d). Hence (d) = (x, y).
Thus repeatedly applying this argument to I (pick any of the two generators, apply this
reasoning, and repeat this procedure) eventually gives us some D ∈ R such that I = (D).
Suppose that J = (x1 , x2 , . . . ) is an ideal of R that is infinitely generated. Then the chain
of ideals (x1 ) ( (x1 , x2 ) ( (x1 , x2 , x3 ) ( · · · is an ascending chain of ideals that does not
stop. But from the first point, we know that there is yi such that (x1 , x2 , . . . , xi ) = (yi ), so
our chain of ideals becomes in fact (y1 ) ( (y2 ) ( (y3 ) ( · · · . Therefore we have yi+1 | yi for
any i ≥ 1. But the second condition implies that for some sufficiently large N , we must have
yn = un yN for all n ≥ N (for some appropriate unit un ), which implies that (yn ) = (yN )
for all n ≥ N . This contradicts the fact that the chain (y1 ) ( (y2 ) ( · · · does not stabilize.
Hence every ideal of R must be finitely generated, so every ideal is generated by a single
element. Hence R is a PID. 

Department of Mathematics and Statistics, Dalhousie University, 6316 Coburg Rd, Hal-
ifax, NS, Canada B3H 4R2
E-mail address: hsyang@[Link]

20