Galois Theory of Cyclotomic Extensions

Winter School 2014, IISER Bhopal

Romie Banerjee, Prahlad Vaidyanathan

I. Introduction
1. Course Description
The goal of the course is to provide an introduction to algebraic number theory, which
is essentially concerned with understanding algebraic field extensions of the field of ra-
tional numbers, Q.

We will begin by reviewing Galois theory:

1.1. Rings and Ideals, Field Extensions

1.2. Galois Groups, Galois Correspondence
1.3. Cyclotomic extensions

We then discuss Ramification theory:

1.1. Dedekind Domains

1.2. Inertia groups
1.3. Ramification in Cyclotomic Extensions
1.4. Valuations

This will finally lead to a proof of the Kronecker-Weber Theorem, which states that If
Q ⊂ L is a finite Galois extension whose Galois group is abelian, then ∃n ∈ N such that
L ⊂ Q(ζn ), where ζn denotes a primitive nth root of unity

2. Pre-requisites
A first course in Galois theory. Some useful books are :

2.1. Ian Stewart, Galois Theory (3rd Ed.), Chapman & Hall (2004)
2.2. D.J.H. Garling, A Course in Galois Theory, Camridge University Press (1986)
2.3. D.S. Dummit, R.M. Foote, Abstract Algebra (2nd Ed.), John Wiley and Sons (2002)

2
3. Reference Material
3.1. M.J. Greenberg, An Elementary Proof of the Kronecker-Weber Theorem, The
American Mathematical Monthly, Vol. 81, No. 6 (Jun.-Jul. 1974), pp. 601-607.
3.2. S. Lang, Algebraic Number Theory, Addison-Wesley, Reading, Mass. (1970)
3.3. J. Neukrich, Algebraic Number Theory, Springer (1999)

4. Pre-requisites
4.1. Definition:
(i) Rings
(ii) Commutative Ring
(iii) Units in a ring
(iv) Field
4.2. Examples:
(i) Q, R, C
(ii) Fp := Z/pZ for p ∈ Z prime
(iii) Definition: Let k be a field
(a) A polynomial over k
(b) Polynomial ring k[x]
(c) Degree of a polynomial
4.3. (Euclidean Division): Let k be a field. If f, g ∈ k[x] with g 6= 0, ∃t, r ∈ k[x] such
that f = tg + r and deg(r) < deg(g)
4.4. k[x] is a principal ideal domain (PID).
4.5. Definition: Let k be a field.
(i) f | g in k[x]
(ii) GCD of f and g.
4.6. Theorem: Let k be a field. If f, g ∈ k[x], then the GCD d of f and g exists in k[x].
Furthermore, ∃s, t ∈ k[x] such that d = sf + tg. We write d = (f, g)
4.7. Definition:
(i) Ideal in a ring
(ii) Maximal ideal
(iii) Irreducible polynomial f ∈ k[x]
4.8. Theorem: For f ∈ k[x], the following are equivalent:
(i) f is irreducible in k[x]

3
 (ii) The ideal (f ) is a maximal ideal
(iii) k[x]/(f ) is a field
4.9. Examples:
(i) Polynomials of degree 1 are automatically irreducible
(ii) x2 − 2 is irreducible in Q[x], but not R[x]
(iii) Polynomial of degree 2 or 3 is irreducible in k[x] iff it does not contain a root
in k (Exercise)
4.10. (Gauss Lemma): Let f ∈ Z[x] be monic, then f is irreducible in Z[x] iff it is
irreducible in Q[x]
4.11. (Rational Root theorem): Let f (x) = a0 + a1 x + . . . + an xn ∈ Z[x] have a root
p/q ∈ Q where (p, q) = 1. Then
(i) p | a0 and q | an
(ii) In particular, if f is monic, then every rational root of f must be an integer.
4.12. (Eisenstein’s Criterion): Let f (x) = a0 + a1 x + . . . + an xn ∈ Z[x]. Suppose ∃p ∈ Z
prime such that
(i) p | ai for all i ∈ {0, 1, . . . , n − 1}
(ii) p - an
(iii) p2 - a0
Then f is irreducible in Q[x]
4.13. (Reduction (mod p)): Let f (x) = a0 + a1 x + . . . + an xn ∈ Z[x] and p ∈ Z prime
such that p - an . If f is irreducible in Zp [x], then f is irreducible in Q[x].
(The converse is not true: f (x) = x4 + 1 ∈ Z[x] is reducible in Z2 [x])
4.14. (Fundamental Theorem of Algebra): For any non-constant f ∈ C[x], ∃α ∈ C such
that f (α) = 0

4
II. Field Extensions
1. Simple Extensions
1.1. Remark: In this course, we will consider two kinds of fields :
(i) Finite fields, such as Fp = Z/pZ, p ∈ Z prime
(ii) Subfields of C (which are necessarily infinite), such as Q, R, C
We will begin by only considering subfields of C, and discuss finite fields later.
1.2. Definition:
(i) Field extension k ⊂ L
(ii) Smallest field k(X) generated by a field k ⊂ C and a set X ⊂ C.
(iii) A field extension k ⊂ L is called simple if ∃α ∈ L such that L = k(α). α is
called a primitive element of the field extension. Note: The primitive element
may not be unique (See Example 1.3(ii))
1.3. Examples:
(i) Q ⊂ R, Q ⊂ C are field extensions, but neither are simple. (Proof later)
(ii) R ⊂ C is a simple extension. C = R(i). Note that C = R(i + 1) as well, so
the primitive element may not be unique.
(iii) Every subfield k ⊂ C contains Q. So Q ⊂ k is a field extension. (Exercise)
√
(iv) Let F = {a + b 2 : a, b ∈ Q}, then
(a) F is a field
(b) Hence, Q ⊂ F is a field extension
√
(c) Note that F = Q( 2)
√ √ √ √
(v) Let K = Q( 2, 3), then K = Q( 2 + 3) and is hence a simple extension
(vi) Let k be a field and f ∈ k[x] be irreducible. Set L = k[x]/(f ), then k ⊂ L is
a field extension.
1.4. Definition/Remark: Let k ⊂ C be a field and α ∈ C
(i) α is said to be algebraic over k if ∃0 6= f ∈ k[x] such that f (α) = 0.
(ii) α is said to be transcendental over k if it is not algebraic.
1.5. Examples:
(i) If α ∈ k, then α is algebraic over k

5
 √
(ii) 2 is algebraic over Q
(iii) π is transcendental over Q (without proof)
(iv) π is algebraic over R
1.6. Theorem: Let k ⊂ L be a field extension and α ∈ L be algebraic over k. Then ∃
unique polynomial f ∈ k[x] such that
(i) f is monic and irreducible
(ii) f (α) = 0
Furthermore, if g ∈ k[x] is any polynomial, then g(α) = 0 iff f | g in k[x]. This is
called the minimal polynomial of α over k and is denoted by mα := mα,k .
Proof. Let
I = {g ∈ k[x] : g(α) = 0}
Then by hypothesis, I 6= {0}. Also, I is clearly an ideal in k[x]. Since k[x] is a
PID, ∃f ∈ I such that
I = (f )
By dividing by the leading coefficient, we may assume f is monic.
(i) f is irreducible: If f = gh, then 0 = f (α) = g(α)h(α). Since C is a field,
either g(α) = 0 or h(α) = 0. Assume WLOG that g(α) = 0, then g ∈ I to
f | g. However, g | f as well, so g = cf for some constant c ∈ k. Hence,
deg(h) = 0, which means that h ∈ k
(ii) Suppose g ∈ k[x] is irreducible, monic and satisfies g(α) = 0. Then, g ∈ I
and so f | g. But g is irreducible, so f = cg for some c ∈ k. They are both
monic, so c = 1. Hence we get uniqueness as well.
(iii) Since f is irreducible, I is a maximal ideal, and so k[x]/I is a field. Define

ϕ : k[x] → C by f 7→ f (α)

Then ϕ is a homomorphism and I = ker(ϕ). Thus, Image(ϕ) is a field.

(iv) Clearly, k ⊂ Image(ϕ) and α ∈ Image(ϕ). So

k(α) ⊂ Image(ϕ)

Conversely, if β ∈ Image(ϕ), then ∃f ∈ k[x] such that β = f (α). Thus,

β ∈ k(α). Hence,
Image(ϕ) = k(α)

1.7. Examples:
(i) If α ∈ k, then mα (x) = x − α
√
(ii) If k = Q, α = 2, then mα (x) = x2 − 2

6
 √ √
(iii) If k = R, α = 2, then mα (x) = x − 2
(iv) If k = Q, ζp = e2πi/p , where p is a prime, then

mζp (x) = Φp (x) = xp−1 + xp−2 + . . . + x + 1

1.8. Definition: Let k ⊂ L1 and k ⊂ L2 be field extensions

(i) A homomorphism of field extensions is a field homomorphism ϕ : L1 → L2
such that ϕ(α) = α for all α ∈ k.
(ii) An isomorphism of field extensions is a bijective homomorphism. If such an
isomorphism exists, we write
L1 ∼
=k L2

1.9. Theorem: Let k ⊂ C be a field and α ∈ C be algebraic over k. Then

k[x]/(mα ) ∼
=k k(α)

Proof. Example 1.3.(vi) and Theorem 1.6

1.10. Corollary: Let k ⊂ L be a field extension and α, β ∈ L be algebraic over k with
the same minimal polynomial. Then there is an isomorphism of field extensions
k(α) ∼
=k k(β) which sends α 7→ β.

Exercises
1.1. Let k be a field and let f ∈ k[x] be of degree 2 or 3. Prove that f is irreducible in
k[x] iff f does not have a root in k.
1.2. (i) Prove that f (x) is irreducible in Z[x] iff f (x + 1) is irreducible.
(ii) Let p ∈ Z be a prime, then prove that, for any 1 ≤ k ≤ p − 1,
 
p
p|
k

(iii) Use (i) and (ii) to prove that

p−1
xp − 1 X i
Φp (x) := = x
x−1 i=0

is irreducible in Z[x]
1.3. Let f (x) = x4 + 1 ∈ Z[x]. Use Part (i) of the previous problem to prove that f is
irreducible in Z[x]
1.4. Let k be a field. Define ϕ : Z → k by n 7→ n · 1.
(i) Prove that ker(ϕ) is either trivial or ∃p ∈ Z such that ker(ϕ) = pZ.

7
 (ii) If k ⊂ C is any field, then prove that Q ⊂ k
(iii) If k is a finite field, then prove that ∃p ∈ Z prime such that Zp ⊂ k.
Definition: We say a field has characteristic n ∈ Z iff ker(ϕ) = nZ. We write
char(k) = n

Note: By the proof of (i), n must either be 0 or a prime. Hence,

(i) If k ⊂ C, then char(k) = 0
(ii) If k is a finite field, then char(k) = p for some prime p ∈ Z
1.5. (i) If f (x) ∈ Z[x] is irreducible, then prove that f (−x) ∈ Z[x] is irreducible.
(ii) Determine the minimal polynomial of α = eπi/5 over Q.
1.6. Let ϕ : Q → C be a homomorphism of fields, then prove that ϕ(x) = x ∀x ∈ Q
Note: Since ϕ is a field homomorphism, ϕ(1) = 1 must hold by definition.
√
1.7. Determine all possible homomorphisms ϕ : Q( 2) → C
[Hint: Use the previous √
problem to prove that any such homomorphism √ is com-
pletely determined by ϕ( 2). Now determine the possible values of ϕ( 2)]
1.8. Let ω = e2πi/3 . Prove that there is an isomorphism
√ √
Q( 2) ∼
3 3
=Q Q(ω 2)

1.9. Is the set

L := {a + bπ : a, b ∈ Q}
a field?
1.10. (i) Let k ⊂ L be a field extension, then L is a k-vector space.
(ii) If k ⊂ L1 and k ⊂ L2 are two extensions, then a homomorphism ϕ : L1 → L2
of k−extensions is a k-linear map of vector spaces.
Definition: Let k ⊂ L be a field extension
(i) The degree of the extension, denoted by [L : k], is the dimension of the
k-vector space L.
(ii) The field extension is called finite if [L : k] < ∞.
1.11. Prove that
√
(i) Q ⊂ Q( 2) is a finite extension of degree 2.
(ii) Q ⊂ R is not a finite extension. [Hint: Q is countable, and R is not]
1.12. If k is a finite field, then prove that ∃p ∈ N prime and n ∈ N such that |k| = pn
1.13. Let Q ⊂ L be a field extension of degree 2 with Q-basis {1, α}.
(i) Prove that α satisfies a polynomial f ∈ Q[x] of degree 2.
√
(ii) Prove that ∃r ∈ Q such that L = Q( r)
[Hint: Use the quadratic formula]

8
2. Finite and Algebraic Extensions
2.1. Recall: If k ⊂ L is a field extension, then
(i) L is a k-vector space
(ii) The dimension of L over k is called the degree of the extension, and is denoted
by [L : k]
(iii) k ⊂ L is called a finite extension iff [L : k] < ∞
2.2. Theorem: Let k ⊂ C be a field and α ∈ C be algebraic over k. Let mα ∈ k[x] be
the minimal polynomial of α over k, and let n = deg(mα ). Then
(i) {1, α, α2 , . . . , αn−1 } is a basis for k(α) over k
(ii) In particular, [k(α) : k] = deg(mα ) < ∞
Proof. Let S = {1, α, α2 , . . . , αn−1 }.
(i) S is linearly independent: Suppose ∃a0 , a1 , . . . , an−1 ∈ k not all zero such that
n−1
X
ai α i = 0
i=0

Then α satisfies the non-zero polynomial f (x) = a0 + a1 x + . . . + an−1 xn−1 ∈

k[x]. This contradicts the minimality of n = deg(mα ).
(ii) S spans k(α): Suppose β ∈ k(α), then by Theorem 1.9, ∃f ∈ k[x] such that
β = f (α). By Euclidean division, ∃t, r ∈ k[x] such that

f = tmα + r, where deg(r) < n

Clearly, f (α) = r(α), so β = r(α) and deg(r) < n. Hence, β ∈ Span(S).

2.3. Examples:
√ √
(i) Q( 2) = {a + b 2 : a, b ∈ Q} (This agrees with Example 1.3(iv))
√
(ii) Q( 3 2) = {a + b21/3 + c22/3 : a, b, c ∈ Q}
(iii) Let p ∈ Z be a prime number and ζp := e2πi/p ∈ C, then

[Q(ζp ) : Q] = deg(Φp ) = p − 1

2.4. (Tower Law) If k ⊂ F and F ⊂ L are two field extensions, then

[L : k] = [L : F ][F : k]

9
 Proof. Suppose [F : k] is infinite, then [L : k] is infinite. Similarly, if [L : F ] is
infinite, so is [L : k]. Hence, we may assume that the RHS is finite. Suppose
S1 := {x1 , x2 , . . . , xn }
is a k-basis for F , and
S2 := {y1 , y2 , . . . , ym }
is an F -basis for L. We want to show that
S := {xi yj : 1 ≤ i ≤ n, 1 ≤ j ≤ m}
has nm elements and is a k-basis for L.
(i) S has nm elements : Suppose not, then xi1 yj1 = xi2 yj2 . If yj1 = yj2 , then
xi1 = xi2 and we are done. If not, then
xi1 yj1 − xi2 yj2 = 0
But {yj1 , yj2 } is F -linearly independent, so xi1 = xi2 = 0. This contradicts
the k-linear independence of {xi1 }.
(ii) S spans L over k: Suppose β ∈ L, then ∃a1 , a2 , . . . , am ∈ F such that
m
X
β= aj yj
i=1

But aj ∈ F implies that ∃αi,j ∈ k such that

n
X
aj = αi,j xi
i=1
P
Hence, β = i,j αi,j xi yj
(iii) S is k-linearly independent in L: Suppose ∃αi,j ∈ k such that
X
αi,j xi yj = 0
i,j
Pm
Collect the coefficient of yj into aj and write j=1 aj yj = 0 where aj ∈
F . However this would imply that aj = 0 for all j. Again, the k-linear
independence of {xi } implies that αi,j = 0 for all i, j.

√
2.5. Example: Let f (x) = x3 + 6x + 2 ∈ Q[x]. Then f is irreducible over Q( 4 2)
√
Proof. Suppose not, then by Exercise 1.1, f must have a root α ∈ Q( 4 2). Hence,
we have a tower √4
Q ⊂ Q(α) ⊂ Q( 2)
However, [Q(α) : Q] = deg(mα ). But f is irreducible over Q by Eisenstein, so it
must be the minimal polynomial of α. Hence, [Q(α) : Q] = 3, which, by the Tower
Law, would imply that 3 | 4. This is a contradiction.

10
2.6. Corollary: Let k ⊂ F1 and k ⊂ F2 be field extensions (all contained in C). Let L
denote the smallest field containing both F1 and F2 . Then
(i) [L : F2 ] ≤ [F1 : k]
(ii) [L : k] ≤ [F1 : k][F2 : k]
L is called the compositum of F1 and F2 and is denoted by F1 F2
Proof. (i) Again, assume [F1 : k] and [F2 : k] are finite. Suppose S := {x1 , x2 , . . . , xn }
is a k-basis for F2 which contains 1. Claim:
n
X
L = SpanF1 (S) = { ai x i : ai ∈ F 1 }
i=1

Let F3 = SpanF1 (S)

(a) F3 ⊂ L since F1 ⊂ L and S ⊂ L and L is a field.
(b) F1 ⊂ F3 since 1 ∈ S
(c) F2 ⊂ F3 since F2 = Spank (S) and k ⊂ F1
(d) F3 is a field:
A. F3 is clearly closed under addition since it is a vector space.
B. If x = ni=1 αi xi and y = nj=1 βj xj with αi , βj ∈ F1 , then
P P

X
xy = α i β j xi xj
i,j

Now, xi xj ∈ F2 = Spank (S). Since k ⊂ F1 it follows that,

xy ∈ SpanF1 (S)
Pn
C. If 0 6= x = i=1 αi xi ∈ F3 . Then consider the map

T : F3 → F3 given by y 7→ xy

Then T is clearly F1 -linear, and is injective (why?). Since [F3 : F1 ] <

∞, it follows that T is surjective. Hence, ∃y ∈ F3 such that xy = 1.
(e) Since F3 is a field and F1 and F2 are contained in F3 , it follows that
L ⊂ F3 . By part (a), we have that

L = F3 = SpanF1 (S)

Hence
[L : F2 ] ≤ |S| = [F1 : k]

(ii) Part (ii) follows from (i) and the fact that [L : k] = [L : F2 ][F2 : k]

11
 2.7. Definition: A field extension k ⊂ L is said to be algebraic if every α ∈ L is algebraic
over k.
2.8. Theorem:
(i) If k ⊂ L is finite extension, then it is algebraic.
(ii) If α ∈ L is algebraic over k, then k ⊂ k(α) is algebraic.
Proof. (i) Suppose k ⊂ L is a finite extension of degree n. Let α ∈ L, then the
set
S = {1, α, α2 , . . . , αn }
has n+1 elements and so it must be k-linearly dependent. Hence, ∃a0 , a1 , . . . , an ∈
k not all zero such that n
X
ai α i = 0
i=0
Pn
and so α satisfies the non-zero polynomial f (x) = i=0 ai xi ∈ k[x]
(ii) Follows from part (i) and Theorem 2.1

2.9. Definition: An extension k ⊂ L is said to be finitely generated if ∃α1 , α2 , . . . αn ∈ L

such that
L = k(α1 , α2 , . . . , αn )
In other words, L is the smallest field containing k and the set {α1 , α2 , . . . , αn }.
2.10. Theorem: k ⊂ L is a finite extension iff it is algebraic and finitely generated.
Proof. (i) Suppose k ⊂ L is a finite extension, then let S be a k-basis for L.
Then
L = k(S)
and so k ⊂ L is finitely generated. Also, by Theorem 3.2, k ⊂ L is algebraic.
(ii) Suppose k ⊂ L is finitely generated and algebraic, then ∃{α1 , α2 , . . . , αn } ⊂ C
such that
L = k(α1 , α2 , . . . , αn )
By hypothesis, αi is algebraic over k, and so

[k(αi ) : k] < ∞

By Theorem 2.6(ii) and induction, it follows that

n
Y
[L : k] ≤ [k(αi ) : k] < ∞
i=1

Hence, k ⊂ L is finite extension.

12
2.11. Example: Let k ⊂ C be a field and set

F = {α ∈ C : α is algebraic over k}

Then
(i) F is a field
(ii) k ⊂ F is an algebraic extension
(iii) If k = Q, then k ⊂ F is infinite
2.12. Theorem: Let k ⊂ C be any field and set F to be the field of numbers that are
algebraic over k. For every non-constant f ∈ F [x], ∃α ∈ F such that f (α) = 0.
Proof. Suppose f (x) = a0 + a1 x + . . . + an xn ∈ F [x] is a non-constant polynomial,
then ∃α ∈ C such that f (α) = 0. But then, α is algebraic over F . We want to
show that α is algebraic over k. So consider

F0 = k(a0 , a1 , . . . , an )

Each ai ∈ F , so ai is algebraic over k. By Theorem 2.10

[F0 : k] < ∞

Also, α is algebraic over F0 . So by Theorem 2.2,

[F0 (α) : F0 ] < ∞

So by the Tower Law

[F0 (α) : k] < ∞
which implies that α is algebraic over k by Theorem 2.1. Hence, α ∈ F by
definition.
2.13. Remark:
(i) We say that F is algebraically closed.
(ii) F is the smallest field that contains k and is algebraically closed (Why?)
(iii) F is called the algebraic closure of k, and is denoted by k
(iv) If k is any field (even a finite field), we can construct a field L such that
(a) k ⊂ L is algebraic
(b) L is algebraically closed
(c) If k ⊂ M is any field extension satisfying the above two properties, then
∃ an injective homomorphism σ : L → M .
In other words, L is unique upto isomorphism. This field L is called the
algebraic closure of k, and is denoted by k

13
Exercises
2.1. Prove that
√ √
(i) 3 ∈/ Q( 2)
√ √
(ii) [Q( 2, 3) : Q] = 4
2.2. If k ⊂ L is a finite extension such that [L : k] is prime, then
(i) If F is a field such that k ⊂ F ⊂ L, then prove that either F = k or F = L
(ii) Prove that ∃α ∈ L such that L = k(α) (ie. k ⊂ L is a simple extension)
2.3. Suppose k ⊂ F1 and k ⊂ F2 are two field extensions whose degrees are relatively
prime. Then prove that

[F1 F2 : k] = [F1 : k][F2 : k]

2.4. Let p ∈ Z be prime and let ζp := e2πi/p . Determine [Q(21/p , ζp ) : Q]

2.5. Let k ⊂ C be a field and α ∈ C be algebraic over k, and suppose ϕ : k(α) → C is
a field homomorphism such that ϕ |k = idk . Let

β := ϕ(α)

Then prove that

(i) For any f ∈ k[x],
ϕ(f (α)) = f (β)

(ii) β is algebraic over k

(iii) β and α have the same minimal polynomial over k.
2.6. Use the previous problem and Corollary 1.10 to answer the following question: If
α ∈ C is algebraic over Q, determine the number of homomorphisms

ϕ : Q(α) → C

2.7. For each n ∈ N, define

An := {f ∈ Q[x] : deg(f ) ≤ n}
[
Bn := {α ∈ C : f (α) = 0}
f ∈An

(i) Prove that An is countable.

(ii) Prove that Bn is countable.
(iii) Prove that Q is countable.
Conclude that there exist real numbers that are transcendental.

14
3. The Galois Group
3.1. Recall: Let k ⊂ C be a field, α ∈ C algebraic over k
(i) If k ⊂ k(α) is a finite extension, and
( n−1 )
X
k(α) = ai αi : ai ∈ k
i=0

where n = deg(mα )
(ii) If ϕ : k(α) → C is a homomorphism such that ϕ |k = idk , then

Xn−1 n−1
X
i
ϕ( ai α ) = ai ϕ(α)i
i=0 i=0

So, ϕ is completely determined by the complex number ϕ(α). By Exercise

2.5, ϕ(α) must be a root of the minimial polynomial mα of α. Hence, if

S = {k-homomorphisms ϕ : k(α) → C}, and

T = {roots of mα in C}

Then we have a map µ : S → T given by

ϕ 7→ ϕ(α)

We have shown above that this function µ is injective.

(iii) By Corollary 1.10, if β ∈ T , then ∃ an isomorphism k(α) ∼
=k k(β). In
particular, we get a k-homomorphism ϕ : k(α) → C such that ϕ(α) = β.
This proves that µ is surjective. Hence,

|{k-homomorphisms ϕ : k(α) → C}| = |{roots of mα in C}|

3.2. Examples: List all homomorphisms from k → C:

(i) k = Q: By Exercise 1.6, there is only one map, the inclusion
√
(ii) k = Q( 2): By Exercise 1.7, there are two maps given by
√ √
σ1 : 2 7→ 2
√ √
σ2 : 2 7→ − 2

(iii) k = Q(ω): There are two maps given by

σ1 : ω →
7 ω
σ2 : ω →7 ω2

15
 √
(iv) k = Q( 3 2): There are three maps given by
√
3
√
3
σ1 : 2 7→ 2
√
3
√3
σ2 : 2 7→ ω 2
√ √
σ3 : 2 7→ ω 2 2
3 3

√ √
(v) k = Q( 2, 3): There are atmost 4 maps given by

σ1 = id
√ √ √ √
σ2 = 2 7→ − 2 and 3 7→ 3
√ √ √ √
σ3 = 2 7→ 2 and 3 7→ − 3
√ √ √ √
σ4 = 2 7→ − 2 and 3 7→ − 3
√
(vi) k = Q( 3 2, ω): There are atmost 6 maps given by

σ1 = idL
√ √
σ2 = 2 7→ 2 and ω 7→ ω 2
3 3

√3
√3
σ3 = 2 7→ ω 2 and ω 7→ ω
√ √
σ4 = 2 7→ ω 2 and ω 7→ ω 2
3 3

√ √
σ5 = 2 7→ ω 2 2 and ω 7→ ω
3 3

√ √
σ6 = 2 7→ ω 2 2 and ω 7→ ω 2
3 3

3.3. Theorem: Let f ∈ k[x] be a monic irreducible polynomial, then all the roots of
f in C are distinct (ie. f does not have any multiple roots. We say that such a
polynomial is separable)
Proof. Suppose β ∈ C is a multiple root of f , then (x − β)2 | f in C[x]. Hence,
f 0 (β) = 0, whence f | f 0 in k[x] (since f must be the minimal polynomial of β over
k). However, deg(f 0 ) < deg(f ), so this is impossible.
3.4. (Primitive Element Theorem): Let k ⊂ L be a finite extension of subfields of C,
then it is a simple extension. ie. ∃θ ∈ L such that L = k(θ)
3.5. Corollary: Let k ⊂ L be a finite extension of subfields of C, then

the number of k-homomorphisms ϕ : L → C = [L : k]

3.6. Definition: A field extension k ⊂ L is said to be normal if every k-homomorphism

ϕ : L → C maps L into itself.
3.7. Lemma: If k ⊂ L is a finite normal extension, then, for any k-homomorphism
ϕ : L → C, ϕ : L → L is bijective.

16
 Proof. Since k ⊂ L is normal, ϕ(L) ⊂ L. Hence we may consider

ϕ:L→L

Note that ϕ is a k-linear transformation, and it is injective since it is a homomor-

phism of fields. Since [L : k] < ∞, it must be surjective.
3.8. Definition:
(i) A field extension that is both finite and normal is called a Galois extension.
(ii) If k ⊂ L is a Galois extension, then we define the Galois group of the extension
as
Galk (L) = {k-homomorphisms ϕ : L → C}

Note: Galk (L) is a group under composition.

3.9. Examples:
(i) Galk (k) = {idk }
√
(ii) GalQ (Q( 2)) ∼= Z2
(iii) GalQ (Q(ω)) ∼
= Z2
√ √
(iv) GalQ (Q( 2, 3)) ∼ = Z2 × Z2
√ √
Proof. Let L = Q( 2, 3). By Example 3.2, we have atmost 4 maps ϕ : L →
C. By Corollary 3.5, there √
are exactly √
4 maps, so these must
√ √be all of them.
For each of these maps, ϕ( 2) and ϕ( 3) belong to Q( 2, 3). Hence, by
Lemma 3.7, they must be bijective as well. Hence,

Galk (L) = {σ1 , σ2 , σ3 , σ4 }

Since it is a group of order 4, Galk (L) ∼

= Z4 or Z2 × Z2 . However,

σi2 = idL ∀i

Hence, Galk (L) ∼= Z2 × Z2

√
(v) GalQ (Q( 3 2, ω)) ∼
= S3
√
Proof. Let L = Q( 3 2, ω). Again, there are exactly 6 homomorphisms
√ ϕ :
3
L → C in Example 3.2. Also, for each of these homomorphisms ϕ( 2) and
ϕ(ω) belong to L. Hence by Lemma 3.7,

Galk (L) = {σi : 1 ≤ i ≤ 6}

It is a group of order 6, so Galk (L) ∼

= Z6 or S3 . However,
√
3
√
3
σ3 σ2 ( 2) 6= σ2 σ3 ( 2)

Hence, Galk (L) is non-abelian, which forces it to be S3

17
3.10. Lemma: Let k ⊂ L be a finite normal extension. If F is an intermediate field

k⊂F ⊂L

Then F ⊂ L is a finite normal extension.

Proof. (i) Clearly, [L : F ] ≤ [L : k] < ∞
(ii) If ϕ : L → C is an F -homomorphism, then ϕ is also a k-homomorphism.
Since k ⊂ L is normal, ϕ(L) ⊂ L.

3.11. Definition: Let k ⊂ L be a field extension and G := Galk (L)

(i) If k ⊂ F ⊂ L is an intermediate field, then

GalF (L) < Galk (L)

(ii) If H < G, then

LH := {x ∈ L : ϕ(x) = x ∀ϕ ∈ H} ⊂ L

is called the fixed field of H

Note: LH is a subfield of L containing k.
(iii) We set

F := {intermediate fields k ⊂ F ⊂ L} G := {subgroups H < G}

Φ:F →G Ψ:G→F
F 7→ GalF (L) H 7→ LH

3.12. Examples:
(i) If k ⊂ L is any field extension, and G = Galk (L)
(a) If H = {e} < G, then LH = L
However, LG may not be equal to k (See below)
(b) If F = L, then GalF (L) = {e}
If F = k, then Galk (L) = G
√ √
(ii) If k = Q, L = Q( 2, 3), then
(a) If H = hσ3 i where
√ √ √ √
σ3 : 2 7→ 2 and 3 7→ − 3
√
Then LH = Q( 2) and GalQ(√2) (L) = H

18
 (b) If H = hσ4 i where
√ √ √ √
σ4 : 2 7→ − 2 and 3 7→ − 3
√
Then LH = Q( 6)
√
(iii) If k = Q, L = Q( 3 2, ω), then
(a) If H = hσ2 i where
√ √
2 7→ 2 and ω 7→ ω 2
3 3
σ2 :
√ √
Then LH = Q( 3 2). Also, GalQ( √
3
3
2) (L) = H. Note that Q ⊂ Q( 2) is
not normal.
3.13. (Fundamental Theorem of Galois Theory - I): Let k ⊂ L be a Galois extension of
subfields of C with Galois group G. Then
(i) Ψ and Φ are mutual inverses. So there is a one-to-one correspondence

F ↔G

(ii) If F ∈ F is an intermediate field, then

[F : k] = [Galk (L) : GalF (L)]

We visualize this by tower diagrams

L Galk (L)
=

F GalF (L)
=

k {e}

3.14. (Fundamental Theorem of Galois Theory - II): Let k ⊂ L be a Galois extension of

subfields of C with Galois group G. Then
(i) If F ∈ F, k ⊂ F is normal iff GalF (L) C Galk (L).
(ii) In that case, the restriction map

π : Galk (L) → Galk (F )

is a well-defined, surjective, group homomorphism.

(iii) ker(π) = GalF (L)
(iv) Hence,
Galk (L)/ GalF (L) ∼
= Galk (F )

19
Exercises
3.1. If k ⊂ L is a Galois extension with Galois group G.
(i) If F is an intermediate field (k ⊂ F ⊂ L), then prove that GalF (L) < G
(ii) If H < G, then prove that LH is a field such that k ⊂ LH ⊂ L.
3.2. Let k ⊂ L be a Galois extension with Galois group G
(i) If F1 and F2 are intermediate fields such that F1 ⊂ F2 , then prove that
GalF2 (L) ⊂ GalF1 (L)
(ii) If H1 and H2 are subgroups of G such that H1 ⊂ H2 , then prove that LH2 ⊂
LH 1
√ √
3.3. Let L = Q( 3 2, ω), and G = GalQ (L). Let F = Q( 3 2) and H = GalF (L)
(i) List all the elements of H from Example 3.2(vi)
(ii) Prove that H is not normal in G.
√ √
3.4. Let L = Q( 3, 2) and let G = GalQ (L)
(i) Determine all the subgroups of G
(ii) For each subgroup H from part (i), determine LH (Use Theorem 3.13(ii))
√
3.5. Let L = Q( 3 2, ω) and G = GalQ (L).
(i) List all the subgroups of G from Example 3.2(vi)
(ii) For each subgroup H from part (i), determine LH (Use Theorem 3.13(ii))
√
3.6. Let L = Q( 4 2, i).
(i) Prove that Q ⊂ L is a normal extension of degree 8.
(ii) Prove that GalQ (L) has a subgroup H which is not normal.
Note: There is only one group of order 8 which satisfies (ii). This is the
dihedral group of order 8, usually denoted by D4 or D8 .
3.7. Let n ∈ N, ζn = e2πi/n and L = Q(ζn ).
(i) Prove that Q ⊂ L is a finite extension of degree ≤ n − 1.
(ii) If ϕ : L → C is a homomorphism, then prove that ∃1 ≤ i ≤ n − 1 such that
ϕ(ζn ) = ζni

Conclude that Q ⊂ L is a normal extension

(iii) Use part (ii) to prove that GalQ (L) is an abelian group.
3.8. Read Theorem 3.14 before attempting this problem: Let k ⊂ L be a finite normal
extension whose Galois group is abelian.
(i) If F is an intermediate field, k ⊂ F ⊂ L, then prove that k ⊂ F is a normal
extension. (Compare this with Exercise 3.1)
(ii) Prove that Galk (F ) is an abelian group.

20
4. Finite Fields
Note: All fields in this section will be finite of characteristic p > 0.

4.1. Remark: If L is a finite field of characteristic p > 0, then

(i) Fp ,→ L. Hence, |L| = pn for some n ≥ 1
(ii) We may construct an algebraically closed field, Fp which will play the role
that C played in the earlier discission.
(iii) Note that the only homomorphism ϕ : Fp → Fp is the inclusion map, so Fp
will play the role that Q played in the earlier discussion.
(iv) As before, if k is a finite field and α ∈ Fp , then there is a 1-1 correspondence

{k-homomorphisms ϕ : k(α) → Fp } ↔ {roots of mα in Fp }

(Exercise 2.5 and Corollary 1.10 hold for finite fields without any change in
the proof)
4.2. Theorem: If k is a finite field and f ∈ k[x] is irreducible, then it is separable. (ie.
It has no multiple roots in Fp )
4.3. Corollary: If k is a finite field and α ∈ Fp , then

|{k-homomorphisms ϕ : k(α) → Fp }| = [k(α) : k]

4.4. Theorem: If L is a finite field, then L∗ is cyclic (as a multiplicative group)

Proof. By the Fundamental theorem of finite abelian groups, we can write

L∗ ∼
= Zd1 × Zd2 × . . . × Zdm

such that d1 | d2 | . . . | dm . Let n = dm , then for any α ∈ L∗ , we have

αn = 1

Hence, α is a root of the polynomial xn − 1 ∈ L[x] which has atmost n roots.

Hence,
d1 d2 . . . dm = |L∗ | ≤ n = dm
Hence, m = 1 and L∗ ∼= Zn is cyclic.
4.5. Corollary: If k ⊂ L is a finite field extension of finite fields, then k ⊂ L is simple.
Proof. The primitive element may be chosen to be any generator of L∗
4.6. Corollary: If k ⊂ L is a finite extension of finite fields, then

the number of k-homomorphisms ϕ : L → Fp = [L : k]

21
 Proof. Write L = k(α) by Corollary 4.3. Now, as in 3.1,

|{k-homomorphisms ϕ : L → Fp }| = |{ roots of mα in Fp }|

But mα is separable, so the RHS is deg(mα ) = [L : k]

4.7. Theorem: If L is a finite field of characteristic p > 0 and cardinality pn , then
n
L = {α ∈ Fp : αp = α}

Proof. Let S be the set in the RHS, then S is the set of roots of the polynomial
n
xp − x ∈ Fp [x]. Hence,
|S| ≤ pn
Since |L| = pn , it suffices to prove that L ⊂ S. Now if α ∈ L, then we may assume
α 6= 0. Now α ∈ L∗ , which is a group of order pn − 1. So
n −1 n
αp = 1 ⇒ αp = α

Hence, α ∈ S, which proves the L ⊂ S and completes the proof.

4.8. Corollary: For any prime p ∈ N and any n ∈ N, there is a unique field of cardinality
pn . We denote this field by
Fp n

4.9. Definition: As before, we make the following definitions :

(i) A field extension k ⊂ L of finite fields is called normal if, for every k-
homomorphism ϕ : L → Fp , we have ϕ(L) ⊂ L.
(ii) If k ⊂ L is such an extension, we write

Galk (L) = {k-homomorphisms ϕ : L → Fp }

Note: As before, If k ⊂ L is a finite normal extension, Galk (L) is a group.

4.10. Theorem: If k ⊂ L is a finite extension of finite fields, then it is normal.
Proof. Suppose ϕ : L → Fp is a k-homomorphism, and β ∈ L, then we want to
show that ϕ(β) ∈ L. By Theorem 4.7,
n n
β p = β ⇒ ϕ(β)p = ϕ(β)
n
Hence, ϕ(β) ∈ {α ∈ Fp : αp = α} = L
4.11. Lemma: If L is a field of characteristic p > 0, then the map F : L → L defined by
x 7→ xp is a homomorphism, called the Frobenius map.
Proof. (i) F is a map from L to Fp , but Fp ⊂ L is a normal extension by 4.10,
so F (L) = L.
(ii) F is clearly multiplicative

22
 (iii) We only need to check that if α, β ∈ L, then (α + β)p = αp + β p . However,
this follows by the binomial theorem and the fact that
 
p
p| ∀1 ≤ k ≤ p − 1
k

4.12. Theorem: Let L = Fpn , then GalFp (L) is a cyclic group of order n generated by F
Proof. We know that F : L → L is a homomorphism, and every homomorphism
fixes Fp , so F ∈ GalFp (L). Also,

| GalFp (L)| = [L : Fp ] = n

So it suffices to prove that o(F ) = n.

(i) For any α ∈ L,
n
F n (α) = αp = α ⇒ o(F ) ≤ n

(ii) Suppose o(F ) = s ≤ n, then F s = idL . Then for any α ∈ L,

s s
αp = α ⇒ α is a root of xp − x

Hence,
s
|L| ≤ |{roots of xp − x in Fp }| ≤ ps
But |L| = pn , so n ≤ s, which proves the theorem.

4.13. Corollary: If k ⊂ L is a finite extension of finite fields of characteristic p > 0, then

Galk (L) is a cyclic group genered by F j where j = [k : Fp ]
Proof. Let G = Galk (L), then G < GalFp (L), which is a cyclic group generated by
F . Hence, G = hF j i for some 1 ≤ j 6= n. Now

o(F j ) = |G| = [L : k]

Hence, by the Tower Law

| GalFp (L)| [L : Fp ]
j= = = [k : Fp ]
|G| [L : k]

23
Exercises
4.1. Let L = F3 [x]/(x2 + 1).
(i) Prove that L is a field of cardinality 9.
(ii) List down all the elements of L
4.2. Let f (x) = x3 − 2 ∈ F7 [x] and L = F7 [x]/(f )
(i) Prove that L is a field of cardinality 73
(ii) Determine which elements of L are the roots of f . [Hint: 23 = 1 in F7 ]
4.3. Construct a field with 4 elements.
4.4. Let k be a finite field and f ∈ k[x] be an irreducible polynomial.
n
(i) Prove that ∃n ∈ N such that f | (xp − x) in k[x]
[Hint: Use Theorem 4.7 on the field k[x]/(f )]
(ii) Conclude that f is separable. (This proves Theorem 4.2)
4.5. Fix n ∈ N and let
n
L = {α ∈ Fp : αp = α}
(i) If d | n, then prove that
d
k = {α ∈ Fp : αp = α}
is a subfield of L
(ii) Conversely, if k ⊂ L is a subfield of L, then prove that ∃d | n such that k is
given as in part (i).
[Hint: Consider the tower Fp ⊂ k ⊂ L]
(iii) If k is as in part (i), determine Galk (L)
4.6. Let L be a finite field. Prove that there is a 1-1 correspondence
{subfields of L} ↔ {subgroups of GalFp (L)}
Note: This proves Theorem 3.13 in the case of finite fields. Theorem 3.14 is not
needed since all finite extensions of finite fields are normal.
4.7. Let k ⊂ F1 and k ⊂ F2 be Galois extensions of subfields of C. Prove that k ⊂
F1 ∩ F2 is a Galois extension.
4.8. Let k ⊂ F1 and k ⊂ F2 be Galois extensions of subfields of C. Let L = F1 F2 , and
prove that
(i) k ⊂ L is a Galois extension (Use the primitive element theorem)
(ii) Define a function
µ : Galk (L) → Galk (F1 ) × Galk (F2 )
by
ϕ 7→ (ϕ |F1 , ϕ |F2 )
Prove that µ is well-defined and injective.

24
5. Cyclotomic Extensions
5.1. Definition: Fix n ∈ N
(i) ζn = e2πi/n
(ii) µn = {e2πik/n : 0 ≤ k ≤ n − 1} = hζn i.
Note: µn is a cyclic group of order n.
(iii) Elements of µn are called roots of unity. Generators of µn are called primitive
nth roots of unity.
(iv) Q(ζn ) = Q(µn ) is called the nth cyclotomic field.
Note: Q ⊂ Q(µn ) is a normal extension.
(v) If G is a group, then Aut(G) = {ϕ : G → G : ϕ is an isomorphism}.
5.2. Theorem: Let k ⊂ C be any field, then
(i) k ⊂ k(ζn ) is a finite normal extension.
(ii) The map
Γ : Galk (k(ζn )) → Aut(µn )
given by
ϕ 7→ ϕ|µn
is a well-defined injective homomorphism.
Proof. (a) If ϕ ∈ Galk (k(ζn )), we want to show that ϕ|µn ∈ Aut(µn ).
A. Firstly, if ζ ∈ µn , then ϕ(ζ)n = 1, so ϕ(ζ) ∈ µn
B. Now, ϕ|µn is clearly a homomorphism from µn to itself.
C. Since ϕ is injective, ϕ|µn is injective. Since µn is finite, it is also
surjective.
(b) So Γ is a well-defined function. It is clearly a homomorphism since the
operation on both groups is composition.
(c) If Γ(ϕ) = idµn , then ϕ(ζn ) = ζn . Since ϕ|Q = idQ , it would follow that
ϕ = idQ(ζn )

5.3. Recall:
(i) If R is a ring, R∗ = {u ∈ R : ∃v ∈ R such that uv = 1}.
(ii) R∗ is a group under multiplication, called the group of units of R.
(iii) If R = Zn , then
R∗ = {a ∈ Zn : (a, n) = 1}

Note: Z∗n is an abelian group.

25
5.4. Theorem: Aut(µn ) ∼
= Z∗n
Proof. If a ∈ Z such that (a, n) = 1, then define
σa : µn → µn by ζn 7→ ζna
(i) σa is clearly a homomorphism of µn
(ii) Since (a, n) = 1, ∃s, t ∈ Z such that sa + tn = 1. Hence,
σa (ζ s ) = ζ as = ζ as+tn = ζ
Hence, σa is surjective.
(iii) Since µn is finite, this means that σa is injective as well. So σa ∈ Aut(µn ).
(iv) If a ≡ b in Z∗n , then n | (b − a), so
σa (ζ) = ζ a = ζ b = σb (ζ)
Hence, σa = σb . So we get a well-defined map
Θ : Z∗n → Aut(µn ) given by a 7→ σa

(v) Θ is a homomorphism since σab = σa ◦ σb

(vi) Θ is injective: If σa = idµn , then
ζa = ζ
⇒ ζ a−1 = 1
⇒ n | (a − 1)
⇒ a = 1 in Z∗n

(vii) Θ is surjective: If ϕ : µn → µn is an automorphism, then ∃1 ≤ j ≤ n such

that ϕ(ζ) = ζ j . Since ϕ is surjective, ∃t ∈ N such that
ϕ(ζ t ) = ζ ⇒ ζ tj = ζ 1
Hence, ∃s ∈ Z such that tj + sn = 1. So (j, n) = 1 whence j ∈ Z∗n and
ϕ = σj = Θ(j)

5.5. Lemma: Let n ∈ N and ζ ∈ µn be a primitive nth root of unity. ζ a is a primitive

nth root of unity if and only if (a, n) = 1
Proof. Suppose (a, n) = 1, then ∃s, t ∈ Z such that sa + tn = 1. Hence
ζn = ζnsa+tn = (ζna )s ∈ hζna i
Hence, hζna i = µn . Conversely, if ζnj is a primitive root of unity, then ∃k ∈ N such
that ζnjk = ζn . Hence, n | (jk − 1) and so (j, n) = 1.

26
 5.6. Definition: nth Cyclotomic polynomial is defined as
Y
Φn (x) = (x − ζ)
where the product is taken over all the primitive roots of unity. By Lemma 5.5,
deg(Φn ) = |Z∗n |

5.7. Theorem: Φn ∈ Q[x]

Proof. Write
Φn (x) = a0 + a1 x + . . . + al xl
If ψ ∈ GalQ (Q(ζn )) and a ∈ Z such that (a, n) = 1, then ∃b ∈ Z such that
(b, n) = 1 such that
ψ(ζna ) = ζnab
Since (ab, n) = 1, we see that ψ permutes the roots of Φn . Hence,
Y Y
ψ(Φn (x)) = (x − ψ(ζ)) = (x − ζ) = Φn (x)
Hence, ψ(ai ) = ai for all i. Hence, if G = GalQ (Q(ζn )), we have
ai ∈ Q(ζn )G
By the Fundamental Theorem of Galois Theory - I, this implies that ai ∈ Q.
5.8. Theorem: Φn is irreducible in Q[x]
5.9. Corollary: GalQ (Q(ζn )) ∼
= Z∗ n

Proof. By 5.2 and 5.4,

GalQ (Q(ζn )) ∼
= a subgroup of Z∗n
However, Φn is in Q[x], it satisfies ζn , and it is both irreducible and monic. Hence,
Φn is the minimal polynomial of ζn over Q, whence
[Q(ζn ) : Q] = deg(Φn ) = |Z∗n |
Hence, the map Γ from Theorem 5.2 must be surjective as well.
5.10. Corollary: If p ∈ N is a prime, then GalQ (Q(ζp )) is a cyclic group of order p − 1.
5.11. Corollary: If F is an intermediate field Q ⊂ F ⊂ Q(µn ) such that Q ⊂ F is Galois,
then GalQ (F ) is an abelian group.
Proof. Since Q ⊂ F is Galois, by Theorem 3.14, GalF (Q(µn )) C GalQ (Q(µn )) and
GalQ (Q(µn ))/ GalF (Q(µn )) ∼
= GalQ (F )
Hence, GalQ (F ) is a quotient of the abelian group Z∗n , and so it must be abelian.
5.12. (Kronecker-Weber Theorem) If Q ⊂ F is any finite normal extension such that
GalQ (F ) is abelian, then ∃n ∈ N such that F ⊂ Q(µn ).
The proof of this statement will take up the rest of the course.

27
Exercises
Read all the previous sections carefully and ask any questions you may have
5.1. If m, n ∈ N has lcm l, then
(i) Prove that Q(ζn ) ⊂ Q(ζl ) and Q(ζm ) ⊂ Q(ζl )
s t
(ii) Prove that ∃s, t ∈ Z such that ζl = ζm ζn
(iii) Conclude that
Q(ζn )Q(ζm ) = Q(ζl )
5.2. (i) For any n ∈ N, prove that
Y
xn − 1 = Φd (x)
d|n

(ii) If p ∈ N is prime, then prove that

Φ2p (x) = Φp (−x) (Compare with Exercise 1.5)
p−1
X
Φp2 (x) = xpk
k=0

(iii) Determine Φ8 (x) (Compare with Exercise 1.3)

Recall the following facts:
(i) Chinese Remainder theorem: If n = pe11 pe22 . . . pekk is the prime decomposition
of n with {pi } all distinct primes, then
Z∗n ∼
= (Zpe11 )∗ × (Zpe22 )∗ × . . . × (Zpekk )∗

(ii) If G = G1 × G2 is the direct product of two finite groups. Then G is cyclic

iff both G1 and G2 are cyclic and (|G1 |, |G2 |) = 1.
5.3. If p ∈ N is prime, then prove that |Z∗pe | = pe − pe−1
5.4. If n ∈ N is such that Z∗n is cyclic, then prove that n can be divisible by atmost one
odd prime.
5.5. Let n ≥ 3
(i) List down the element of Z∗8 , and prove that it is not cyclic.
(ii) Prove that there is a surjective map
Z∗2n → Z∗2n−1

(iii) Use induction and part (i) to prove that Z∗2n is not cyclic
5.6. If n ∈ N such that Z∗n is cyclic, then prove that either n = 4 or n = 2i pj for some
odd prime p and i ∈ {0, 1} and j ∈ N ∪ {0}
Note: The converse to the above statement is also true.

28
Bibliography
[Stewart] Ian Stewart, Galois Theory (3rd Ed.)

[Rotman] J. Rotman, Galois Theory (2nd Ed.)

[Garling] DJH Garling, A Course in Galois Theory

[E. Artin] Emil Artin and A.N. Milgram, Galois Theory

[Greenberg] M.J. Greenberg, An Elementary Proof of the Kronecker-Weber Theorem,

The American Mathematical Monthly, Vol. 81, No. 6 (Jun.-Jul. 1974), pp. 601-607.

[Lang] S. Lang, Algebraic Number Theory, Addison-Wesley, Reading, Mass. (1970)

[Neukrich] J. Neukrich, Algebraic Number Theory, Springer (1999)

29