AN INTRODUCTION TO THE ZARISKI TOPOLOGY

OSCAR MICHEL

Abstract. We give an introduction to the spectrum of a ring and its Zariski

topology, a fundamental tool in algebraic geometry. In addition, we cover the
ring theory and topology necessary for defining and proving basic properties
of the Zariski topology. Finally, we give examples of various ring spectra.

Contents
1. Introduction 1
2. Rings and Ideals 2
3. Topology 8
4. The Spectrum of a Ring 10
Acknowledgments 16
References 16

1. Introduction
Algebraic geometry is the branch of math that studies problems in geometry
that can be solved with algebra, and vice versa. Modern algebraic geometry unfor-
tunately has a reputation for being very difficult and inaccessible to learn. Many
standard algebraic geometry textbooks are written at a graduate level or higher.
The idea that one needs advanced techniques from category theory and commuta-
tive algebra to gain an appreciation for algebraic geometry is far from the truth,
however. It is still possible for the undergraduate student to engage with aspects of
the theory, and it is the goal of this paper to introduce an essential tool of modern
algebraic geometry using only undergraduate ring theory and topology.
In this paper we will study the spectrum of a ring, which gives a way to define a
topological space that can be created from any ring. This topological space, called
the Zariski topology, gives a geometric way to interpret the algebra of a ring using
the language of topology. A quick Google search of “the Zariski topology” is enough
to see its relevance in the theory of modern algebraic geometry, but many sources
will still be saturated with graduate level material. The good news is that there is
still a lot one can learn about the spectrum of a ring without having to know what
a sheaf or a scheme is. We have tried to combine the material that only relies on
basic ring theory and topology into a single source.
This paper should be accessible to second or third year undergraduate math
majors. The paper is divided into three main sections so that readers familiar with
ring theory or topology may skip ahead. Readers who have had a first course in
group theory should have no trouble reading this paper. We will begin with an
overview of ideals in rings, so readers who are unfamiliar with the definition of a
1
2 OSCAR MICHEL

ring, a subring, or a product of rings may see [1]. We will assume that all rings are
commutative and with unity. In addition, we assume ring homomorphisms send 1
to 1. No topology background is necessary for reading this paper.

2. Rings and Ideals

We begin our study of ring theory with the definition of an ideal. An ideal is
very similar to a normal subgroup in group theory. This is because ideals allow one
to construct quotient rings similar to the way quotient groups are constructed using
a normal subgroup. As we will see, many theorems of quotient groups reappear in
the form of quotient rings.
Definition 2.1. Let R be a ring and I be a subgroup of R under addition. For
r ∈ R, consider the set rI = {ri | i ∈ I}. If rI ⊆ I for all r ∈ R, then I is said to
be an ideal of R.
Before we continue our study of ideals, we will pause to introduce some notation
for constructing ideals out of ring elements. If A is some subset of a ring R, then
(A) will be the smallest ideal of R containing A. Such an ideal always exists because
A is contained in the ideal R. We note that if A is finite, A = {a1 , a2 , . . . , an }, then
the ideal (A) is also the set of all R-linear combinations of the ai . Sometimes we
will write the ideal (0) simply as 0.
One way to understand the definition of an ideal is to consider the ring of integers
Z. For any n ∈ Z, the set nZ = {nx | x ∈ Z} is an ideal of Z. The ideal 2Z is
exactly the set of even integers. Adding two even numbers together will always give
an even number, and multiplying an even number by any integer will result in an
even number. The key difference between the additive and multiplicative structure
of an ideal is that multiplication by elements outside the ideal must always stay
inside the ideal. This is not true for addition. Going back to our example of the
even integers, an even number will not remain even if an odd number is added to it.
As we will see next, the multiplicative structure of an ideal allows for a well-defined
construction of a quotient ring.
Proposition 2.2. Let R be a ring and let I be an ideal of R. Then the additive
quotient group R/I is a ring under addition and multiplication defined by:
(r + I) + (s + I) = (r + s) + I
(r + I) × (s + I) = (rs) + I.
Proof. Since R is an abelian group and I is a normal subgroup, R/I is automatically
an abelian quotient group under addition. We will prove that multiplication in the
quotient group is well defined. The remaining ring axioms should be verified by the
reader. If we chose representatives r, s ∈ R and i, j ∈ I then,
(r + i)(s + j) = rs + rj + is + ij.
Since I is closed under multiplication by elements of R, each of rj, is, and ij is in
I. Furthermore, their sum rj + is + ij is in I. Writing this in terms of cosets we
have the desired result:
(r + I) × (s + I) = (rs) + I.

 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 3

In group theory one can understand the structure of a group through group
homomorphisms. The isomorphism theorems for groups establish a relationship
between groups, normal subgroups, quotient groups, and group homomorphisms.
This perspective is very useful for studying rings, too. The following theorems will
prepare us for proving two isomorphism theorems for rings.
Theorem 2.3. Let R and S be rings and let ϕ : R → S be a ring homomorphism.
Then, the image of ϕ is a subring of S, and ker ϕ is an ideal of R.
Proof. If s1 , s2 ∈ im(ϕ), then there are r1 , r2 ∈ R such that s1 = ϕ(r1 ) and
s2 = ϕ(r2 ). From the homomorphism property, we know s1 + s2 = ϕ(r1 ) + ϕ(r2 ) =
ϕ(r1 + r2 ) and s1 s2 = ϕ(r1 )ϕ(r2 ) = ϕ(r1 r2 ). Hence, s1 + s2 ∈ im(ϕ) and s1 s2 ∈
im(ϕ). Finally, 1 ∈ im(ϕ) because ϕ(1) = 1, and this proves that im(ϕ) is a subring
of S.
Next, suppose r1 , r2 ∈ ker ϕ. Since ϕ(r1 ) = ϕ(r2 ) = 0, it follows again from the
homomorphism property that ϕ(r1 + r2 ) = 0 which proves r1 + r2 ∈ ker ϕ. Now
let a be any element of R, and let r ∈ ker ϕ. Multiplying ϕ(a) and ϕ(r), we see
ϕ(ar) = ϕ(a)ϕ(r) = ϕ(a)0 = 0, and ar ∈ ker ϕ. 
Theorem 2.4. Let ϕ : R → S be a ring homomorphism. If J is an ideal of S, then
ϕ−1 (J) is an ideal of S.
Proof. Suppose r1 , r2 ∈ ϕ−1 (J). By definition, ϕ(r1 ), ϕ(r2 ) ∈ J, and ϕ(r1 ) −
ϕ(r2 ) = ϕ(r1 − r2 ) ∈ J because J is an ideal of S. It follows that r1 + r2 ∈ ϕ−1 (J)
which proves ϕ−1 (J) is closed under addition. Next, suppose a ∈ ϕ−1 (J) and
r ∈ R. Since J is an ideal of S, ϕ(a)ϕ(r) ∈ J. This implies ar ∈ ϕ−1 (J) because
ϕ(ar) = ϕ(a)ϕ(r) ∈ J. 
Remark 2.5. It is not true in general that if ϕ : R → S is a ring homomorphism,
then ϕ(J) is an ideal if J is an ideal. However, if ϕ is a surjective homomorphism,
then ϕ(J) is an ideal in S.
We are now ready to prove two ismomorphism theorems for rings. In total,
there are four standard isomorphism theorems for rings, but only two of them will
be presented here. The first isomrophism theorem is a useful tool to prove two
rings are ismorphic, and establishes a relationship between ring homomorphisms
and quotient rings.
Theorem 2.6 (The First Isomorphism Theorem for Rings). If ϕ : R → S is a ring
homomorphism, then R/ ker ϕ is isomorphic to the image of ϕ. In particular, if ϕ
is surjective, then R/ ker ϕ ∼
= S.
Proof. Let I = ker ϕ. First we note that R/I is a valid ring because ker ϕ is an ideal
by Theorem 2.3. Consider the following map π : R/I → im(ϕ) where r + I 7→ ϕ(r).
First we will prove that this map is well defined. We will use the notation r to denote
the coset r + I. Suppose for some r1 , r2 ∈ R, r1 = r2 . Then r1 − r2 ∈ I = ker ϕ,
which means
π(r1 ) = ϕ(r1 ) = ϕ(r1 + (r2 − r2 )) = ϕ(r1 − r2 ) + ϕ(r2 ) = 0 + ϕ(r2 ) = π(r2 ).
Next we will prove π is an isomorphism between rings R/ ker ϕ and im(ϕ). First
note that π is a homomorphism.
π(r1 r2 )) = π(r1 r2 ) = ϕ(r1 r2 ) = ϕ(r1 )ϕ(r2 ) = π(r1 )π(r2 )
4 OSCAR MICHEL

π(r1 + r2 ) = π(r1 + r2 ) = ϕ(r1 + r2 ) = ϕ(r1 ) + ϕ(r2 ) = π(r1 ) + π(r2 )

The map π is surjective. For every ϕ(r) ∈ im(π) we have π(r) = ϕ(r). Finally
suppose π(r1 ) = π(r2 ). Then π(r1 ) − π(r2 ) = 0 and we get π(r1 ) − π(r2 ) =
ϕ(r1 ) − ϕ(r2 ) = ϕ(r1 − r2 ) = 0. This means r1 − r2 ∈ ker ϕ = I, so r1 = r2 . This
proves the map is injective, and hence an isomorphism. 
If ψ : R → R/ ker ϕ is the projection map from R to the quotient ring R/ ker ϕ,
then the diagram below illustrates the proof of Theorem 2.6.
ϕ
R S

ψ π

R/ ker ϕ
To see how the first isomorphism is useful, we will give an example. Recall
the ring of Gaussian integers: Z[i] = {a + bi | a, b ∈ Z}. We will prove that
Z[x]/(x2 + 1) ∼
= Z[i]. Consider the homomorphism ϕ : Z[x] → Z[i] given by p(x) 7→
p(i). This map is surjective because every Gaussian integer a + bi is mapped to
by its corresponding linear polynomial a + bx. Furthermore, the kernel of ϕ is the
ideal (x2 + 1), thereby proving Z[x]/(x2 + 1) ∼
= Z[i] by Theorem 2.6.
When constructing quotient rings, the ideal structure of the original ring is
preserved. For example, in the ring of integers there are three ideals containing 4Z:
4Z ⊂ 2Z ⊂ Z. In the ring Z/4Z, there are three ideals total: (0), (2), and Z/4Z.
It in not a coincidence that the number of ideals in the quotient ring is the same
as the number of ideals that contain 4Z. This relationship will be made precise in
the next theorem.
Theorem 2.7 (Lattice Isomorphism Theorem). Let I be an ideal of a ring R.
There is an inclusion preserving bijection between the set of ideals of R containing
I and the set of ideals of R/I.
Proof. Let φ : R → R/I be the projection map from R to the quotient ring R/I.
For each ideal J ⊇ I, there is a correspond ideal φ(J) ⊆ R/I since φ is surjective.
Similarly, if K is an ideal of R/I, then φ−1 (K) is an ideal of R. To prove there
is a bijection, we will show φ−1 (φ(J)) = J for J ⊇ I. Written explicitly as a set,
φ−1 (φ(J)) = {a ∈ R | φ(a) ∈ φ(J)}. When written in this way, it is clear that
J ⊆ φ−1 (φ(J)). If φ(a) ∈ φ(J), then there is a b ∈ J such that
φ(a) = φ(b)
⇒ φ(a − b) = 0
⇒a−b∈I
⇒ a − b = c, for some c ∈ I
⇒a=b+c
Since b ∈ J and c ∈ I ⊆ J, we have shown a ∈ J, and thus φ−1 (φ(J)) ⊆ J. 
The lattice isomorphishm theorem is a fundamental result related to the ideal
structure of a ring. Oftentimes ideals can best be understood by looking at their
containment relationship to other ideals. Next, we will define two special classes of
ideals: prime ideals and maximal ideals.
 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 5

Definition 2.8. An ideal M of a ring R is a maximal ideal if M is a proper ideal

of R, and the only ideals of R containing M are R and M .
We will state the following theorem, called Krull’s Theorem, without proof.
Krull’s Theorem addresses the existence of maximal ideals in a ring, and it’s proof
uses Zorn’s Lemma which is equivalent to the axiom of choice. The reader can find
a proof of Krull’s Theorem in [1].
Theorem 2.9 (Krull’s Theorem). Every proper ideal of a ring is contained in some
maximal ideal
Maximal ideals are “maximal” in the sense that they are a maximum element
in a chain of ideal inclusions. It is important to note that a ring can contain many
maximal ideals. The ring of integers contains infinitely many of them. The maximal
ideals of Z are the ideals pZ for a prime number p. However, the question of finding
the maximal ideals for any general ring R is not always as straightforward as finding
the maximal ideals of Z. Fortunately for us, there is a way to determine whether
some ideal M is maximal by looking at the quotient ring R/M .
Recall that an element u ∈ R is called a unit if there exists an element v of R
such that uv = 1.
Lemma 2.10. If I is an ideal of a ring R then I = R if and only if I contains a
unit.
Proof. If I = R, then I contains the unit 1. Conversely, if u is a unit in I with
inverse v, then 1 = uv ∈ I. If I contains 1, then I = R because r = 1 · r ∈ I for
every r ∈ R. 
Lemma 2.11. A ring R is a field if and only if its only ideals are 0 and R.
Proof. If R is a field, then every nonzero ideal of R contains some unit, and so
the only nonzero ideal of R must be R itself by Lemma 2.10. Conversely, if R is
the only nonzero ideal of R, then for any nonzero u ∈ R we have (u) = R. Thus
1 ∈ (u), and so there must be some v ∈ R such that uv = 1. Since every nonzero
element of R is a unit, R is a field. 
Theorem 2.12. If M is an ideal of a ring R, then M is a maximal ideal if and
only if the ring R/M is a field.
Proof. By definition, M is maximal if and only if there are no proper ideals I that
contain M . By Theorem 2.7, the ideals of R that contain M correspond bijectively
with the ideals of R/M . If R/M is a field, then the only ideals of R/M are (0) and
R/M by Lemma 2.11. Hence, M is maximal exactly when R/M is a field. 
Corollary 2.13. A ring R is a field if and only if 0 is a maximal ideal of R.
The reader may be familiar with the fact that the ring Z/pZ is a field whenever
p is prime. This should be true intuitively because if n is composite, n = pq, then
the ring Z/nZ contains zero divisors with p · q = pq = 0. However, the statement
can be proven in a simpler way using Theorem 2.12. If we accept the fact that the
maximal ideals of of Z are pZ for prime p, then the claim immediately follows from
Theorem 2.12.
The second class of ideals we will be studying are the prime ideals. As one would
expect, the prime ideals of Z are pZ for a prime p. The set of maximal ideals and
6 OSCAR MICHEL

nonzero prime ideals in Z are the same, but this is not true for every ring. Prime
ideals are based on a certain notion of primality in the integers. That is, if p is
prime and p divides ab, then p must divide at least one of a or b. With this in
mind, we now present the definition of a prime ideal.
Definition 2.14. Let R be a ring and P be a proper ideal of R. The ideal P is
a prime ideal if for every a, b ∈ R, whenever the product ab is an element of P , at
least one of a or b is an element of P .
Just as in the case for maximal ideals, there is a way to test whether an ideal is
prime by analyzing its quotient ring. Once we have this result, it will follow that
every maximal ideal is a prime ideal. We remind the reader that an integral domain
is a nonzero ring where the product of any two nonzero elements is nonzero.
Theorem 2.15. An ideal P of a ring R is prime if and only if R/P is an integral
domain.
Proof. An element r ∈ R will be an element of P if and only if r = 0 in the quotient
ring R/P . Therefore, a product ab ∈ R will be in P if and only if ab = ab = 0 in
R/P , and this will be true whenever R/P is an integral domain. 

Corollary 2.16. Every maximal ideal of a ring is a prime ideal.

Proof. This follows directly from theorems 2.12 and 2.15 since every field is an
integral domain. 

We now turn to studying the behavior of prime ideals under ring homomor-
phisms. If one has a homomorphism ϕ : R → S between rings R and S, then the
prime ideal structure of S is preserved under the inverse image of ϕ. Later, we
will see that this fact is essential in order to show the existence of continuous maps
between ring spectra.
Theorem 2.17. Let ϕ : R → S be a ring homomorphism between rings R and S.
If P is a prime ideal of S, then ϕ−1 (P ) is a prime ideal of R.
Proof. Denote ϕ−1 (P ) by Q. We know by Theorem 2.4 that Q is an ideal of
R, so what is left to prove is that Q is a prime ideal. Consider the canonical
homomorphism ψ : S → S/P . Composing ψ with ϕ gives us the following ring
homomorphism from R to S/P :
ϕ ψ
R S S/P.

We will refer to this map as π = ψ ◦ ϕ. First we claim that ker π = Q. This is

because any element in ker π must map to P = ker ψ in S, and therefore must be
in ϕ−1 (P ) = Q. By The First Isomorphism Theorem, R/Q is isomorphic to im(π)
which is a subring of S/P . Since P is a prime ideal, S/P is an integral domain.
Any subring of an integral domain is also an integral domain, so Q is a prime ideal
since R/Q is an integral domain. 

It should be noted that Theorem 2.17 can be proven in a more straightforward

manner using the standard definition of a prime ideal. However, we prefer our
proof because it emphasizes the viewpoint of using quotient rings and ring homo-
morphisms to solve problems in ring theory.
 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 7

We now present two important constructions between ideals of a ring. These

constructions will appear later in our discussion of the Zariski topology. Let R be
a ring, and let I and J be ideals of R.
Definition 2.18. The sum of I and J is the ideal I + J = {i + j | i ∈ I, j ∈ J}. If
I + J = R, then I and J are said to be comaximal.
Definition 2.19. The product of I and J is the ideal consisting of all finite sums
of products of the form ij where i ∈ I and j ∈ J.
Theorem 2.20. Let I and J be ideals of a ring R. Then the following are true:
(1) I + J is the smallest ideal of R containing both I and J
(2) The ideal IJ is contained in I ∩ J. Furthermore, if I + J = R, then
IJ = I ∩ J.
Proof.
(1) We want to show that if K is some ideal containing both I and J then
I + J ⊆ K. Take some x ∈ I + J and write x = i + j for i ∈ I and j ∈ J.
Since K is an ideal that contains both I and J, we know i, j ∈ K and
i + j ∈ K. Since every element of I + JPis in K, we have I + J ⊆ K.
n
(2) Suppose for some x ∈ IJ we have x = k=1 ik jk . Each term of the sum is
in I, and similarly each term is in J. Therefore each term is in I ∩ J, and
so their sum x must also be in I ∩ J. If I + J = R, then 1 ∈ I + J. This
means that 1 = i + j for some i ∈ I and j ∈ J. We have already proved
that IJ ⊆ I ∩ J, so what needs to be proven is that I ∩ J ⊆ IJ. Let x be
an element of I ∩ J. Note that x = x · 1 = (i + j)x = ix + jx. This proves
x ∈ IJ, because x is an element of both I and J, and we have written x as
a finite sum of elements of I multiplied by elements of J.

Theorem 2.21. Let I and J be ideals of a ring R. If P is some prime ideal of R
that contains IJ, the P contains either Ior J.
Proof. If I is contained in P we are done, so suppose I is not contained in P . This
means there is some i ∈ I such that i ∈ / P . For every j ∈ J, ij ∈ IJ ⊆ P . Since
P is a prime ideal, j must be in P , as i was assumed to not be in P . Since this is
true for all j ∈ J, we can conclude J ⊆ P . 
Definition
√ 2.22. If I is an ideal of a ring R, the radical of I, denoted rad(I) or
I, is defined to be the set
rad(I) = {r | rn ∈ I, n ∈ N}
Proposition 2.23. For any ideal I, rad(I) is an ideal of R.
Proof. If r ∈ rad(I) and a ∈ R, then (ar)n = an rn ∈ I. Hence ar ∈ rad(I) by
definition. If a, b ∈ rad(I), then there is some n large enough such that an ∈ I and
bn ∈ I. By the Binomial Theorem,
2n  
2n
X 2n 2n−k k
(a + b) = a b .
k
k=0
Since every term in the sum has a or b with degree at least n, each term is in I,
and a + b ∈ rad(I). 
8 OSCAR MICHEL

√
An element x ∈ R is said to be nilpotent if xn = 0 for some n. The ideal 0,
sometimes called the nilradical of R, is the set of all nilpotent elements in R.
√
Theorem 2.24. If P is a prime ideal of a ring R, then 0 ⊂ P .
√
Proof. If x ∈ 0, then xn = 0 for some n. Since 0 ∈ P , we have 0 = xn ∈ P . We
prove that x is in P by induction. If n = 2, then x2 = x · x ∈ P implies that x ∈ P
since P is prime. Now assume x ∈ P if xn−1 ∈ P . Since xn = xn−1 x ∈ P , it follows
that x ∈ P by the inductive hypothesis and the fact that P is prime. 
We now present a few theorems about ring products. The Chinese Remainder
Theorem gives a way to know when a ring is really a product of rings “in disguise”.
The proof of the Chinese Remainder Theorem is not difficult, but it is on the
lengthier side and is not important to know for our purposes. For these reasons,
we have omitted the proof, but the curious reader may find a full proof in [1].
Theorem 2.25 (Chinese Remainder Theorem). Let A1 , A2 , . . . , Ak be ideals of a
ring R. If for every i 6= j Ai and AJ are comaximal (Ai + Aj = R), then
R/(A1 A2 . . . Ak ) = R/(A1 ∩ A2 ∩ . . . ∩ Ak ) ∼
= R/A1 × R/A2 × . . . × R/Ak
As an example, let’s apply the Chinese Remainder Theorem to the ring Z/6Z
to prove that Z/6Z ∼ = Z/3Z × Z/2Z. Since 3 − 2 = 1, we have 1 ∈ 3Z + 2Z, and
3Z + 2Z = Z. The intersection of 3Z and 2Z will correspond to integers that are
divisible by 3 and 2 which is exactly the ideal 6Z. Applying the Chinese Remainder
Theorem proves Z/6Z ∼ = Z/3Z × Z/2Z.
Definition 2.26. An element e of a ring R is idempotent if e2 = e.
Remark 2.27. Every ring contains the trivial idempotents 1 and 0. If e is idempo-
tent, then so is 1 − e.
Theorem 2.28. If a ring R has nontrivial idempotents if and only if R is a product
of rings.
Proof. If R = S × T , then (1, 0)2 = (12 , 02 ) = (1, 0). Thus R contains the nontrivial
idempotent (1, 0). Conversely, suppose R contains a nontrivial idempotent e. We
will apply Theorem 2.25 to the ideals (e) and (1 − e). These ideals are comaximal
because e + (1 − e) = 1, i.e. (1) + (1 − e) = (1) = R. The fact that (e) · (1 − e) =
e − e2 = 0 proves the product of the ideals is (0). By the Chinese Remainder
Theorem, R ∼ = R/(0) ∼ = R/(e) × R/(1 − e). Since e is nontrivial, (e) and (1 − e)
will never be equal to (0) or (1), and so the ring product R/(e) × R/(1 − e) will be
nontrivial. 

3. Topology
In the section we give the background in topology necessary for defining the
spectrum of a ring. We will cover the definition of a topological space and other
basic notions in topology. This section is by no means a comprehensive introduction
to topology, and we only include the theory of topology that is necessary for defining
the Zariski topology. For a more complete introduction to topology, see [2]. With
that being said, this section is self-contained, and we do not assume any background
in topology. Readers who already have a basic understanding of topology should
feel free to skip ahead to the next section and refer back to this one as needed.
 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 9

Definition 3.1. A topological space hX, T i includes a nonempty set X and T ,

where T is a collection of subsets of X called open sets. Open sets must satisfy the
following properties:
(1) The empty set ∅ and X are open, i.e. ∅ ∈ T and X ∈ T .
(2) The union of any collectionS of open sets is open, i.e. if {Ai }i∈I is a collection
of open sets of T , then i∈I Ai ∈ T .
(3) The intersection of a finite collection of openTsets is an open set, i.e. if
n
{Ai }ni=1 is a collection of open set of T , then i=1 Ai ∈ T .
From now on, we may refer to a topological space hX, T i simply as X when it
is clear. It is important to understand, however, that in order to clearly define a
topological space both T and X need to be specified. Giving X alone is not enough,
as there can be many choices for different topological spaces over a given base set.
Definition 3.2. Let X be a topological space. A subset A of X is said to be a
closed set if its complement X \ A is open.
When it is convenient, we may refer to the complement of an open set as A
instead of X \ A. When using the notation A, we will always be taking the com-
plement of A relative to the base set X. Learning topology requires getting a good
grasp on the relationship between open sets versus closed sets. As a word of advice
to the reader, when learning a new idea in topology it is useful to reimagine the
idea in terms of both open and closed sets. The duality between open and closed
sets often times allows one to replace definitions and theorems stated in terms of
open sets with equivalent ones stated in terms of closed sets. Instead of specifying
a collection of open subsets when defining a topological space over a set X, it is
sometimes more convenient to start with a collection of closed subsets of X. Using
DeMorgan’s Law, one can derive an equivalent definition of a topological space
hX, T i by specifying a collection of closed sets instead of open sets. This definition
will be used in the next section when we define the Zariski topology.
Definition 3.3. An equivalent definition of a topological space hX, T i can be
defined where T is a collection of closed sets as follows:
(1) The empty set ∅ and X are closed.
(2) The union of a finite collection of closed sets is closed.
(3) The intersection of any collection of closed sets is closed.
Ring homomorphisms are useful in ring theory because they allow one to study
one ring in terms of another. There is a similar kind of map in topology called a
continuous map.
Definition 3.4. Let X and Y be topological spaces. A map f : X → Y is continous
if the preimage of every open set of Y is an open set of X.
Remark 3.5. The reader should verify that a map f : X → Y between topological
spaces X and Y is continuous if the preimage of every closed set of Y is a closed
set of X. We will use this definition in the next section.
The δ,  definition of continuity from calculus is a special case of the more general
Definition 3.4. Since open sets remain open under the inverse image of a continuous
map, the continuous map gives a way to measure similarity between topological
spaces. In fact, two topological spaces are “isomorphic” if there is a continuous
10 OSCAR MICHEL

way to travel back and forth between the two spaces. This is made precise in the
next definition.
Definition 3.6. Two topological spaces Xand Y are homeomorphic if there is a
there is a continuous bijection f : X → Y with a continuous inverse.
We will now present several examples of topological spaces.
Example 3.7. The indiscrete topology on a nonempty set X is a topological space
with open sets X and ∅. The indiscrete topology is sometimes called the “trivial
topology” because it is a relatively uninteresting topology that can be defined over
any set.
Example 3.8. The cofinite topology on a nonempty set X is a topological space
where open subsets of X include the empty set or sets with a finite complement.
In other words,
T = {A ⊆ X | A = ∅ or A is finite}
The cofinite topology is indeed a valid topological space because
(1) The empty set is an open set by definition and X = ∅.
(2) If {Xα }α∈Λ is a collection of open sets then
[ \
Xα = Xα .
α∈Λ α∈Λ

This is finite because for every Xα ,

\
Xα ⊆ Xα ,
α∈Λ

and every Xα is assumed to be finite.

(3) If X1 and X2 are open sets then X1 ∩ X2 = X1 ∪ X2 . Since X1 and X2 are
both finite, then their union will be finite as well.
Example 3.9. Let hX, TX i be a topological space and Y be a subset of X. The
subspace topology of X on Y , hY, TY i, consists of open sets
TY = {Y ∩ A | A ∈ TX }
Finally we will give two more definitions which describe properties of a topolog-
ical space. These will be used later, and will give us a way to study a ring in a
geometric way.
Definition 3.10. A topological space is disconnected if it is the union of two
nonempty disjoint open sets. A topological space is said to be connected if it is not
disconnected.
Definition 3.11. A topological space X is irreducible if it cannot be written as
the union of two nonempty proper closed subsets of X.

4. The Spectrum of a Ring

At this point, the reader should be familiar with the ring theory and topology
presented in the previous two sections. With these tools, we are now at a point
where we can define a topological space from a ring. First we will begin by defining
a point set. This will serve as our underlying base set for the Zariski topology.
 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 11

Definition 4.1. Let R be a commutative ring with 1. The spectrum of R, denoted

Spec(R), is the set of all prime ideals of R. We will additionally define mSpec(R)
to be the set of all maximal ideals of R . Since every maximal ideal is a prime ideal,
it is always true that mSpec(R) ⊂ Spec(R).
Example 4.2. The points of Spec(Z) are the ideals (0) and pZ where p is a prime
number. The points of mSpec(Z) are almost the exact same, with one important
exception: mSpec(Z) does not include (0).
At this point we will pause to make some important distinctions before defining
the Zariski topology attached to Spec(R). Previously, we thought of ideals as a
set which includes certain ring elements. However, now we will be constructing a
topological space on the set of prime ideals of R. The points of this topological
space are the ideals themselves, rather than the ring elements which make up the
ideal. In other words, the ideal should be viewed as a kind of discrete object or point
which encodes information about the ring. The set Spec(R) should be thought of
as the collection of these points. On its own, Spec(R) is nothing but a set, and at
this point we have yet to turn it into a topological space. To do this, we need to
specify the closed sets of Spec(R).
Definition 4.3. If R is a ring, and I is an ideal of R, then define the set
V (I) = {p ∈ Spec(R) | p ⊃ I}
as the set of all prime ideals containing I. In the case of mSpec(R), we will use the
notation Vm (I) to denote the set of all maximal ideals containing I.
Lemma 4.4. If R is a ring and I is an ideal of R, then V (I) = ∅ if and only if
I = R.
Proof. Since every prime ideal is proper, no prime ideal can contain R, hence
V (R) = ∅. Conversely, suppose for some ideal I we have V (I) = ∅. If I were
a proper ideal, then by Theorem 2.9 it must be contained in some maximal ideal
M of R. We know that every maximal ideal is prime, and this means M ∈ V (I).
By hypothesis V (I) = ∅, so I cannot be a proper ideal of R, and thus I = R. 
Proposition 4.5. Closed sets of the form V(I) for ideals I of a ring R form a
topology on Spec(R). This topology is called the Zariski topology.
Proof. Since every ideal contains (0), V (0) = Spec(R), and by Lemma 4.4 we have
V (R) = ∅. Therefore, Spec(R) and ∅ are closed sets of the Zariski topology. Next,
we claim that if {V (Iα )}α∈Λ is any collection of closed sets, then
!
\ X
V (Iα ) = V Iα
α∈Λ α∈Λ
P P
P containing every Iα in R. If p ∈ V ( Iα ),
RecallPthat α∈Λ Iα is the smallest ideal
then Iα ⊆ p. For all α ∈ Λ, Iα ⊆ Iα ⊆ p. This is equivalent to saying
\
Iα ⊆ p
α∈Λ
or, !
X \
V Iα ⊆ V (Iα ).
α∈Λ α∈Λ
12 OSCAR MICHEL

\
To prove the other direction, suppose p ∈ V (Iα ). Then for any α, the ideal Iα
P
is contained in p. However, since Iα is the smallest ideal of R containing every
Iα , it must be true that X
Iα ⊆ Iα ⊆ p.
α∈Λ
Thus, !
X
p∈V Iα .
α∈Λ
Finally, we will prove that V (I)∪V (J) = V (IJ). Suppose p ∈ V (IJ). By definition,
the prime ideal p contains IJ. By Theorem 2.21, p contains at least one of I or J,
which is the same as saying p ∈ V (I) ∩ V (J). This proves V (IJ) ⊆ V (I) ∩ V (J).
If p ∈ V (I) ∩ V (J), then I ⊆ p or J ⊆ p. Without loss of generality assume
I ⊆ p. From Theorem 2.20, it follows that IJ ⊆ I ∩ J ⊆ I ⊆ p. This proves that
p ∈ V (IJ), and V (I) ∩ V (J) ⊆ V (IJ). 
Proposition 4.5 defines a topology on Spec(R). The difference between Spec(R)
and “the Zariski topology” is that Spec(R) is nothing but a set, while the Zariski
topology is the topological space whose base set is Spec(R). We should note that
this distinction is not always carefully made. Oftentimes Spec(R) is used inter-
changeably with “the Zariski topology,” but usually the meaning can be inferred
from context.
Proposition 4.6. If R is a ring, then the closed points of Spec(R) correspond to
V (M ) = {M } where M is a maximal ideal of R.
Proof. Since every maximal ideal is prime, M ∈ V (M ) because M ⊆ M . The only
other ideal that contains M is R. However, R is not a prime ideal because it is not
a proper ideal. Hence V (M ) = {M }. If P is a prime ideal that is not maximal,
then {P } cannot be closed. Any closed set V (I) that contains P will also contain
some maximal ideal that contains P . 
√
Theorem 4.7. If I is an ideal, then V (I) = V ( I).
√ √
Proof. Since I ⊆ I, it follows that V (I) ⊆ I. To prove the √ other direction, we
want to show that if p is a prime ideal such that I ⊆ p, then I ⊆ p. For an ideal
J ⊇ I, we will use the notation R/J to denote the ideal that is the image of the
projection map from R to R/I.
√ If p is prime, then the √
ideal R/p is prime in R/I.
By Theorem 2.24, R/p ⊇ 0. Note that the ideal √ R/ I is the√nilradical of the
quotient ring R/I. Therefore we have R/p ⊇ R/ I, and so p ⊇ I by the Lattice
Isomorphism Theorem. 
√ √
Corollary 4.8. If V (I) = V (J), then I = J.
Readers who are familiar with topological notion of compactness may find the
following remark interesting.
Remark 4.9. If R is a ring, then Spec(R) is compact.
Proof. If {Iα }α∈Λ is a collection of ideals of a ring R where
\
V (Iα ) = ∅ = V (R)
α∈Λ
 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 13

then
X
Iα = R
α∈Λ
P
Since 1 ∈ R, then 1 ∈ α∈Λ Iα so 1 = iα1 + iα2 + . . . + iαk for iαj ∈ Iαj . From this
Pk
we can conclude 1 ∈ j=1 Iαj . Finally we have
 
Xk k
\
V Iαj  = V (Iαj ) = V (R) = ∅
j=1 j=1

which concludes the proof. 

Compactness is usually a very important property of a topological space, but

the fact that Spec(R) is compact is not very useful for studying the spectrum of a
ring. For this reason, we chose not to cover compactness in the topology section.
However, the next theorem we are about to prove is very important because it will
connect several ideas in ring theory to topology. Specifically, we will be able to use
ring homomorphisms to generate continuous maps between ring spectra.

Theorem 4.10. Let φ : R → S be a ring homomorphism for rings R and S. Then

the map φ? : Spec(S) → Spec(R) defined by φ? (P ) = φ−1 (P ) is continuous.

Proof. To prove φ? is continuous, we must show that if V (I) is some closed set of
Spec(R), then φ?−1 (V (I)) = V (J) for some ideal J of Spec(S). We will prove that
J = φ(I).

φ?−1 (V (I)) = {p ∈ Spec(S) | φ? (p) ∈ V (I)}

= {p ∈ Spec(S) | I ⊆ φ? (p)}
= {p ∈ Spec(S) | φ(I) ⊆ p}
= V ((φ(I)).

Notice that the direction of the maps in Theorem 4.10 reverse. The original ring
homomorphism goes from R to S, but the continuous map goes from Spec(S) to
Spec(R). Next, we will look at what happens when Theorem 4.10 is applied to the
case of a quotient ring.

Theorem 4.11. For a ring R and an ideal I, let φ : R → R/I be the projection
map sending r 7→ r + I. If φ∗ is the continuous map from Spec(R/I) to Spec(R)
given by φ, then the following are true:
(1) The image of φ∗ is V (I).
(2) The map is injective.
(3) The topology on Spec(R/I) is homeomorphic to the topology on V (I).

Proof.
14 OSCAR MICHEL

(1) Since the prime ideals containing I are the inverse images of the prime
ideals of R/I, we have
im(φ∗ ) = {φ∗ (p) | p ∈ Spec(R/I)}
= {φ−1 (p) | p ∈ Spec(R/I)}
= {p | p ⊇ I, p is a prime ideal}
= V (I).
(2) If ϕ(p1 ) = ϕ(p2 ), then φ(p1 ) = φ(p2 ). Since there is a bijection between
the ideals of R/I and the ideals of R containing I, p1 = p2 .
(3) The prime ideals of R/I are the ideals p/I for prime ideals p ∈ V (I). This
gives a continuous bijection from V (I) to Spec(R/I).


The consequence of this theorem is that we can view the closed set V (I) as a
topological space itself by looking at Spec(R/I). Next, we will present two more
theorems that will connect topological properties of the ring spectrum to the alge-
braic properties of a ring.
√
Theorem 4.12. If R is a ring, then Spec(R) = Spec(R/ 0).
√
Proof.
√ For every p ∈ Spec(R), we have 0 ⊆ p by Theorem 2.24. In other
√ words,
V ( 0) = Spec(R), and Theorem 4.11 proves that Spec(R) = Spec(R/ 0). 

Theorem 4.13. Let R1 and R2 be rings, and t denote the disjoint union of sets.
Then Spec(R1 × R2 ) = Spec(R1 ) t Spec(R2 ).
Proof. The reader should verify that the an ideal of R1 × R2 is of the form I1 × I2 ,
where I1 is an ideal of R1 and I2 is an ideal of R2 . Let P1 × P2 be a prime ideal of
R1 × R2 . Then the quotient ring R1 /P1 × R2 /P1 must be an integral domain, but
the product of two integral domains is never an integral domain. Therefore one of
P1 or P2 is a prime ideal, and the other is equal to its corresponding ring. 

The converse of Theorem 4.13 is also true. Namely, if Spec(R) is disconnected,

then R is a product of rings. We will prove this fact in two parts. First we will
prove the theorem is true for a ring with no nilpotents, and then we will prove the
general case.
Theorem 4.14. Suppose for a ring R, Spec(R) is disconnected, i.e. Spec(R) =
X tY for closed sets X and Y . If R contains no nilpotent elements, then R ∼
= S ×T
where X = Spec(S) and Y = Spec(T ).
Proof. Since Spec(R) = X t Y , we have X ∩ Y = ∅ and X ∪ Y = Spec(R). With
X and Y being closed, we write X = V (I) and Y = V (J) for ideals I and J of R.
Therefore

(4.15) X ∩ Y = V (I) ∩ V (J) = V (I + J) = ∅ = V (R)

and
(4.16) X ∪ Y = V (I) ∪ V (J) = V (IJ) = Spec(R) = V (0)
 AN INTRODUCTION TO THE ZARISKI TOPOLOGY 15

We can conclude from Equation 4.15 and Lemma 4.4 that I + J = R. This means
I and J are comaximal. Now we can apply the Chinese Remainder Theorem to see
R/(IJ) ∼= R/I × R/J
√
If R has no nilpotent elements, then 0 = (0), and hence IJ = 0 by Corollary 4.8.
Substituting (0) in for IJ we are left with
R/(0) =∼R= ∼ R/I × R/J
From Theorem 4.11, we know that V (I) = Spec(R/I) and V (J) = Spec(R/J). 
Theorem 4.17. If Spec(R) is disconnected, then R is a product of rings.
√
Proof. If Spec(R) is disconnected,
√ then Spec(R/ 0) √ is disconnected √by Theorem
4.12. Since every nilpotent in R maps to 0√ in R/ 0, the√ring R/ 0 contains
no nilpotents. Applying√Theorem 4.14 to R/ 0, we get R/ 0 = S × T for some
rings S and T . Since R/ 0 is a product of rings, it contains nontrivial idempotents
√
by Theorem 2.28. In general, R contains nontrivial idempotents if R/ 0 contains
nontrivial idempotents, so R = S 0 × T 0 is also a product of rings. The proof of
the fact that R contains nontrivial idempotents is intricate, and for this reason we
direct the reader to [10] instead of proving it here. 
The final theorem we present in this paper will not be proven, but the reader
may refer to [11] as a reference. However, we will see an application of this theorem
in the following discussion.
√
Theorem 4.18. The ideal 0 ⊂ R is prime if and only if Spec(R) is irreducible.
Corollary 4.19. If R is an integral domain, then Spec(R) is irreducible.
We have presented several theorems that allow us to study the algebraic prop-
erties of a ring through the topology of its spectrum. Now we will provide some
examples to see these theorems in action. We will not be completely rigorous in our
discussion of these examples. There will be some hand-waving, and this is because
we want to emphasize the geometric intuition, rather than proving everything in
detail.
The ring we will be using for the following examples is the polynomial ring
C[x, y]. Since C is an algebraically closed field, the maximal ideals of C[x, y] are
(x − a, y − b) for a, b ∈ C. The reason why this is true is because of a theorem called
Hilbert’s Nullstellensatz which this fact is a special case of. For more information
on Hilbert’s Nullstellensatz, see [6]. This means that we can identify mSpec(C[x, y])
with C × C because every maximal ideal is uniquely determined by two complex
numbers. Additionally, a polynomial f (x, y) is contained in the ideal (x − a, y − b) if
and only if f (a, b) = 0. In other words, Vm (f ) contains the zeros of the polynomial
f . This allows use to visualize mSpec(C[x, y]/(f )) as the curve f = 0 in C × C.
As an example, if f (x, y) = y − x2 , then mSpec(C[x, y]/(y − x2 )) looks like the
parabola y = x2 . Of course, some suspension of disbelief is required here because
of the extra dimensions in C × C, but hopefully the idea is clear. What about
Spec(C[x, y]/(y − x2 ))? It will contain all the individual points of C × C where
y − x2 = 0, and it will contain the irreducible component (y − x2 ) layered on top
of these points.
On the other hand, consider the polynomial x(x−1) ∈ C[x, y]. This polynomial is
reducible because it has factors x and (x − 1). This means that Spec(C[x, y]/(x(x −
16 OSCAR MICHEL

1))) is not irreducible, and moreover C[x, y]/(x(x − 1)) is not an integral domain.
In fact, Spec(C[x, y]/(x(x − 1))) is disconnected. This makes sense when we think
of the zeros of x(x − 1) as two separate spaces in the complex plane. On one hand
we have the zeroes along the curve x = 0, and then there is the separate curve of
x = 1. Since Spec(C[x, y]/(x(x − 1))) is disconnected, it is isomorphic to a product
of rings by Theorem 4.18. In particular Spec(C[x, y]/(x(x − 1))) ∼ = C[x] × C[y].
Acknowledgments
Firstly, I would like to thank my mentor Karl Schaefer for his time and effort
helping me write and learn the material for this paper. He is a fantastic teacher, and
he really made me exited to learn about math. I also thank Peter May for organizing
the REU, and I am very grateful to have had the opportunity to participate in such
an amazing program. Finally, I would like to thank the other graduate students
who spent their time running problem sessions and mentoring other students.

References
[1] David S. Dummit, Richard M. Foote. Abstract Algebra. John Wiley & Sons, Inc. 2004.
[2] James R. Munkres. Topology. Prentice Hall, Inc. 2000.
[3] https://math.stackexchange.com/questions/423623/show-that-a-ring-with-disconnected-
spectrum-is-a-product-of-two-subrings
[4] https://math.stackexchange.com/questions/1216203/prime-ideals-in-a-finite-direct-product-
of-rings
[5] Bianca Viray. https://sites.math.washington.edu/ bviray/teaching RingHomomorphismsAn-
dIsomorphisms.pdf
[6] Akhil Mathew. The CRing Project. http://math.uchicago.edu/~amathew/CRing.pdf
[7] John Terilla. Zariski Topology. https://math.mit.edu/~jhirsh/notes2_zariski.pdf
[8] John Hunter, Bruno Nachtergaele. Applied Analysis. https://www.math.ucdavis.edu/
~hunter/book/ch4.pdf
[9] Ravi Vakil. https://math.stanford.edu/~vakil/0708-216/216class06.pdf
[10] https://math.stackexchange.com/questions/326452/if-mathop-mathrmspeca-is-not-
connected-then-there-is-a-nontrivial-idempot
[11] https://math.stackexchange.com/questions/541943/spectrum-of-a-ring-is-irreducible-if-and-
only-if-nilradical-is-prime-atiyah-mac